ShamaDaVinci

app.py

@@ -1,29 +1,755 @@
-import importlib.util
-import sys
+"""
+DVNC.AI — app.py
+Refactored for functional "Use as main insight" logic with academic rigor.
+"""
+
+# ── Standard library ────────────────────────────────────────────────────────
+import html
+import json
+import math
+import os
+import random
+import re
+import urllib.parse
+import xml.etree.ElementTree as ET
 from pathlib import Path
+from typing import Dict, List, Optional
+from urllib.parse import quote
+
+# ── Third-party ──────────────────────────────────────────────────────────────
+import gradio as gr
+import requests
+
+try:
+    import fitz  # PyMuPDF
+except Exception:
+    fitz = None
+
+try:
+    from bs4 import BeautifulSoup
+except Exception:
+    BeautifulSoup = None
+
+# ── Internal modules ─────────────────────────────────────────────────────────
+from dvnc_ai_v2_hf.agent_route_cards import build_agent_route_cards_html
+from dvnc_ai_v2_hf.discovery_app_bridge import (
+    get_default_route_state,
+    get_discovery_css,
+    get_initial_discovery_timeline_html,
+)
+from dvnc_ai_v2_hf.dvnc_ui_layout import get_dvnc_layout_css
+from dvnc_ai_v2_hf.self_learning_graph import (
+    DEFAULT_SOURCES,
+    SEARCH_MODES,
+    SOURCE_OPTIONS,
+    build_learning_graph_html,
+    build_journal_html,
+    ingest_selected_papers,
+    parse_uploaded_pdf,
+    render_parse_result,
+    run_paper_discovery,
+    safe_text,
+)
+
+# ── Constants ────────────────────────────────────────────────────────────────
+MODELS = [
+    {"name": "DVNC Sovereign", "tag": "flagship", "desc": "Maximum depth orchestration for frontier discovery"},
+    {"name": "DVNC Atlas",     "tag": "research",  "desc": "Balanced reasoning, graph traversal, and synthesis"},
+    {"name": "DVNC Curie",     "tag": "lab",        "desc": "Experimental hypothesis generation for anomalous signals"},
+]
+
+AGENTS = [
+    "Query Interpreter",
+    "Graph Divergence Mapper",
+    "Evidence Harvester",
+    "Analogy Engine",
+    "Hypothesis Composer",
+    "Adversarial Critic",
+    "Experimental Program Designer",
+]
+
+NODES = [
+    {"id": "seed",    "label": "Seed Query",          "group": "core",      "x": 10,  "y":   0, "z":   0},
+    {"id": "bio",     "label": "Biomaterials",         "group": "domain",    "x": 24,  "y":  12, "z":  -8},
+    {"id": "card",    "label": "Cardiac Repair",       "group": "domain",    "x": 38,  "y":   3, "z":  14},
+    {"id": "nano",    "label": "Nanostructure",        "group": "bridge",    "x": 24,  "y": -18, "z":  16},
+    {"id": "selfasm", "label": "Self-Assembly",        "group": "bridge",    "x": 40,  "y": -16, "z": -16},
+    {"id": "electro", "label": "Electro-signalling",   "group": "mechanism", "x": 58,  "y":  10, "z": -10},
+    {"id": "immune",  "label": "Immune Modulation",    "group": "mechanism", "x": 64,  "y":  -8, "z":  10},
+    {"id": "trial",   "label": "Validation Path",      "group": "outcome",   "x": 80,  "y":   0, "z":   0},
+    {"id": "alt1",    "label": "Piezoelectric Scaffold","group": "candidate", "x": 56,  "y":  26, "z":  14},
+    {"id": "alt2",    "label": "Peptide Mesh",          "group": "candidate", "x": 54,  "y": -27, "z": -14},
+]
+
+EDGES = [
+    ("seed",    "bio"),    ("seed",   "nano"),
+    ("bio",     "card"),   ("nano",   "selfasm"),
+    ("selfasm", "electro"),("card",   "immune"),
+    ("electro", "trial"),  ("immune", "trial"),
+    ("card",    "alt1"),   ("selfasm","alt2"),
+    ("alt1",    "trial"),  ("alt2",   "trial"),
+]
+
+DEFAULT_PATH = ["seed", "nano", "selfasm", "electro", "trial"]
+
+CANDIDATES = [
+    {
+        "title":   "Piezoelectric Scaffold Cascade",
+        "front":   "Use mechano-electric scaffolds to convert cardiac strain into micro-current signalling.",
+        "back":    "Discovery path: anomalous healing signal -> piezoelectric analog -> ion-channel entrainment -> tissue regeneration. Risk: power density and fibrosis coupling.",
+        "score":   92,
+        "novelty": "High",
+        "agent":   "Hypothesis Composer",
+    },
+    {
+        "title":   "Peptide Self-Assembly Mesh",
+        "front":   "Deploy dynamic peptide meshes that self-assemble around damaged myocardium and guide repair.",
+        "back":    "Discovery path: self-assembly -> local immune choreography -> regenerative substrate formation. Risk: degradation timing and targeting specificity.",
+        "score":   88,
+        "novelty": "High",
+        "agent":   "Analogy Engine",
+    },
+    {
+        "title":   "Immune-Tuned Conductive Hydrogel",
+        "front":   "Blend conductivity with macrophage-state modulation to reduce scarring and restore conduction.",
+        "back":    "Discovery path: inflammation mismatch -> conductive medium -> macrophage polarization -> synchronized healing. Risk: persistence and biocompatibility.",
+        "score":   85,
+        "novelty": "Medium-High",
+        "agent":   "Adversarial Critic",
+    },
+]
+
+ACADEMIC_INSIGHTS = [
+    {
+        "hypothesis": "Implementation of mechano-electric scaffolds to transduce cardiac strain into localized micro-current signalling for myocardial regeneration.",
+        "metrics": {"Novelty": 92, "Mechanistic clarity": 85, "Experimental tractability": 78, "Cross-domain distance": 94},
+        "outline": "1. Synthesize candidate piezoelectric biomaterial scaffolds with tunable strain-electric coupling.\n2. Evaluate *in vitro* electromechanical transduction and subsequent ion-channel entrainment.\n3. Conduct *in vivo* comparative models to assess regenerative efficacy against gold-standard substrates.\n4. Rigorously validate to exclude pathological fibrosis and power-density toxicity.",
+        "path": ["seed", "bio", "card", "alt1", "trial"]
+    },
+    {
+        "hypothesis": "Deployment of dynamic peptide networks that self-assemble post-infarction to orchestrate local immunological responses and guide substrate regeneration.",
+        "metrics": {"Novelty": 88, "Mechanistic clarity": 82, "Experimental tractability": 86, "Cross-domain distance": 85},
+        "outline": "1. Formulate peptide sequences programmed for triggered *in situ* self-assembly within the myocardial infarct zone.\n2. Quantify macrophage polarization and local immune choreography post-deployment.\n3. Map the temporospatial degradation profile against *de novo* tissue formation.\n4. Falsify against off-target aggregation and delayed clearance risks.",
+        "path": ["seed", "nano", "selfasm", "alt2", "trial"]
+    },
+    {
+        "hypothesis": "Integration of conductive hydrogels with immunomodulatory properties to simultaneously bridge electrical uncoupling and mitigate adverse fibrotic scarring.",
+        "metrics": {"Novelty": 85, "Mechanistic clarity": 90, "Experimental tractability": 88, "Cross-domain distance": 79},
+        "outline": "1. Fabricate biocompatible hydrogels featuring precisely tuned electrical conductivity and immunomodulatory motifs.\n2. Monitor electrophysiological synchronization across the scaffold-tissue interface.\n3. Assess macrophage state transitions and suppression of adverse fibrotic remodelling.\n4. Validate long-term persistence, hemocompatibility, and mechanical integration.",
+        "path": ["seed", "bio", "card", "immune", "trial"]
+    }
+]
+
+JOURNALS = [
+    {"name": "Nature",     "url": "https://www.nature.com/search",                              "desc": "Flagship multidisciplinary research journal."},
+    {"name": "Science",    "url": "https://www.science.org/search",                             "desc": "High-impact science journal and family."},
+    {"name": "Cell",       "url": "https://www.cell.com/search",                                "desc": "Life sciences and translational biology."},
+    {"name": "The Lancet", "url": "https://www.thelancet.com/search",                           "desc": "Clinical and medical research."},
+    {"name": "IEEE Xplore","url": "https://ieeexplore.ieee.org/search/searchresult.jsp",        "desc": "Engineering, AI, signal processing, and systems."},
+]
+
+SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
+GROBID_URL               = os.getenv("GROBID_URL", "").strip()
+REQUEST_TIMEOUT          = 25
+
+
+# ── Utility helpers ──────────────────────────────────────────────────────────
+def safe_text(x, default: str = "") -> str:
+    return html.escape(str(x if x is not None else default))
+
+
+def norm_text(x: Optional[str]) -> str:
+    return re.sub(r"\s+", " ", (x or "")).strip()
+
+
+def detect_query_type(query: str) -> str:
+    q = (query or "").strip()
+    if re.match(r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$", q, flags=re.I):
+        return "doi"
+    if q.startswith("http://") or q.startswith("https://"):
+        return "link"
+    return "topic"
+
+
+def ensure_list(x):
+    return x if isinstance(x, list) else []
+
+
+# ── HTML builders ─────────────────────────────────────────────────────────────
+def build_connectome_html(path_ids: List[str]) -> str:
+    active    = set(path_ids)
+    node_map  = {n["id"]: n for n in NODES}
+    path_pairs = {
+        pair
+        for i in range(len(path_ids) - 1)
+        for pair in [(path_ids[i], path_ids[i + 1]), (path_ids[i + 1], path_ids[i])]
+    }
+
+    base_lines, active_lines, circles, labels = [], [], [], []
+
+    for a, b in EDGES:
+        na, nb = node_map[a], node_map[b]
+        x1, y1 = na["x"] * 8 + 80, na["y"] * 6 + 280
+        x2, y2 = nb["x"] * 8 + 80, nb["y"] * 6 + 280
+        base_lines.append(f'<line class="edge" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />')
+        if (a, b) in path_pairs:
+            active_lines.append(f'<line class="edge active" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />')
+
+    for n in NODES:
+        cx, cy   = n["x"] * 8 + 80, n["y"] * 6 + 280
+        is_active = n["id"] in active
+        state    = "chosen" if is_active else "idle"
+        halo_cls = "halo active" if is_active else "halo"
+        lbl_cls  = "label active" if is_active else "label"
+        radius   = 18 if is_active else 13
+        halo_r   = 30 if is_active else 0
+        circles.append(
+            f'<g class="node-wrap">'
+            f'<circle class="{halo_cls}" cx="{cx:.1f}" cy="{cy:.1f}" r="{halo_r}" />'
+            f'<circle class="node {n["group"]} {state}" cx="{cx:.1f}" cy="{cy:.1f}" r="{radius}" />'
+            f'</g>'
+        )
+        labels.append(f'<text class="{lbl_cls}" x="{cx + 18:.1f}" y="{cy - 16:.1f}">{safe_text(n["label"])}</text>')
+
+    return f"""
+    <div class="panel brain-shell">
+      <div class="brain-header">
+        <div>
+          <p class="eyebrow">Connectome</p>
+          <h3>3D Connectome</h3>
+        </div>
+        <div class="brain-legend">
+          <span><i class="dot dot-live"></i> lit path</span>
+          <span><i class="dot dot-chosen"></i> chosen node</span>
+          <span><i class="dot dot-idle"></i> available node</span>
+        </div>
+      </div>
+      <div class="brain-stage">
+        <svg viewBox="0 0 780 560" class="brain-svg" role="img" aria-label="DVNC 3D connectome visualisation">
+          {"".join(base_lines)}
+          {"".join(active_lines)}
+          {"".join(circles)}
+          {"".join(labels)}
+        </svg>
+      </div>
+    </div>
+    """
+
+
+def build_cards_html(cards: List[Dict]) -> str:
+    items = []
+    for i, c in enumerate(cards):
+        items.append(f"""
+        <article class="candidate-card" tabindex="0">
+          <div class="candidate-card-inner">
+            <div class="candidate-face candidate-front">
+              <div class="candidate-top">
+                <span class="chip">{safe_text(c["agent"])}</span>
+                <span class="score">{safe_text(c["score"])}</span>
+              </div>
+              <h4>{safe_text(c["title"])}</h4>
+              <p>{safe_text(c["front"])}</p>
+              <div class="meta-row"><span>Novelty</span><strong>{safe_text(c["novelty"])}</strong></div>
+              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
+            </div>
+            <div class="candidate-face candidate-back">
+              <div class="candidate-top">
+                <span class="chip alt">Alternative path</span>
+                <span class="score">{safe_text(c["score"])}</span>
+              </div>
+              <h4>{safe_text(c["title"])}</h4>
+              <p>{safe_text(c["back"])}</p>
+              <div class="meta-row"><span>Swap into route</span><strong>Enabled</strong></div>
+              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
+            </div>
+          </div>
+        </article>""")
+    return '<div class="panel" style="padding:20px;"><div class="candidate-grid">' + "".join(items) + "</div></div>"
+
+
+def build_agent_timeline(reasoning: List[Dict]) -> str:
+    rows = []
+    for r in reasoning:
+        rows.append(f"""
+        <details class="agent-step" {"open" if r["step"] == 1 else ""}>
+          <summary class="agent-summary">
+            <div class="agent-index">{safe_text(r["step"])}</div>
+            <div class="agent-head">
+              <h4>{safe_text(r["agent"])}</h4>
+              <span>{safe_text(r["tag"])}</span>
+            </div>
+          </summary>
+          <div class="agent-copy">
+            <p>{safe_text(r["summary"])}</p>
+          </div>
+        </details>""")
+    return '<div class="panel" style="padding:18px;"><div class="timeline">' + "".join(rows) + "</div></div>"
+
+
+def build_chat_html(query: str, result: Dict) -> str:
+    return f"""
+    <div class="panel chat-panel">
+      <div class="chat-thread">
+        <div class="bubble bubble-user">
+          <span class="role">You</span>
+          <p>{safe_text(query)}</p>
+        </div>
+        <div class="bubble bubble-ai">
+          <span class="role">DVNC Sovereign</span>
+          <p>{safe_text(result["summary"])}</p>
+        </div>
+        <div class="bubble bubble-system">
+          <span class="role">Discovery Signal</span>
+          <p><strong>Primary hypothesis:</strong> {safe_text(result["primary_hypothesis"])}</p>
+        </div>
+      </div>
+    </div>
+    """
+
+
+def build_models_html(selected: str) -> str:
+    items = []
+    for m in MODELS:
+        active = "active" if m["name"] == selected else ""
+        items.append(f"""
+            <div class="model-pill {active}">
+              <span class="model-name">{safe_text(m["name"])}</span>
+              <span class="model-tag">{safe_text(m["tag"])}</span>
+              <small>{safe_text(m["desc"])}</small>
+            </div>""")
+    return '<div class="panel" style="padding:18px;"><div class="model-switcher">' + "".join(items) + "</div></div>"
+
+
+# ── Discovery logic ───────────────────────────────────────────────────────────
+def run_discovery(query: str, model_name: str):
+    """
+    Runs the 7-agent discovery pipeline.
+    """
+    random.seed(len(query) + len(model_name))
+
+    if "curie" in query.lower() or "einstein" in query.lower():
+        primary = "Map the anomaly first, then force a distant analogy before composing the experimental programme."
+        path    = ["seed", "bio", "card", "immune", "trial"]
+    else:
+        primary = "Utilization of a self-assembling conductive scaffold to transduce mechanical strain into localized regenerative signalling pathways."
+        path    = DEFAULT_PATH
+
+    summaries = [
+        "Normalises the user prompt into a graph-searchable seed and isolates the tension inside the question.",
+        "Finds remote conceptual bridges instead of staying near the starting domain cluster.",
+        "Pulls evidence packets and conflict signals required for grounded hypothesis formation.",
+        "Generates cross-domain analogies with a bias toward mechanism transfer rather than keyword similarity.",
+        "Composes the lead hypothesis and two structurally different variants.",
+        "Attacks weak assumptions, hidden confounders, and feasibility gaps.",
+        "Produces a staged validation plan with measurable falsification criteria.",
+    ]
+
+    tags = ["input", "graph", "evidence", "analogy", "compose", "critique", "experiment"]
+
+    reasoning = [
+        {"step": i + 1, "agent": AGENTS[i], "tag": tags[i], "summary": summaries[i]}
+        for i in range(7)
+    ]
+
+    result = {
+        "summary":            "A deeper route was chosen through the connectome, with live alternatives preserved as swappable cards so the reasoning path can be inspected rather than hidden.",
+        "primary_hypothesis": primary,
+        "reasoning":          reasoning,
+        "cards":              CANDIDATES,
+        "path":               path,
+        "metrics": {
+            "Novelty":                   93,
+            "Mechanistic clarity":       89,
+            "Experimental tractability": 82,
+            "Cross-domain distance":     91,
+        },
+    }
+
+    chat_html       = build_chat_html(query, result)
+    connectome_html = build_connectome_html(path)
+    timeline_html   = build_agent_route_cards_html(reasoning)
+
+    metrics_md      = "\n".join(f"- {k}: {v}/100" for k, v in result["metrics"].items())
+    hypothesis_md   = (
+        "# Discovery Output\n\n"
+        f"**Model:** {model_name}\n\n"
+        f"**Primary hypothesis:** {result['primary_hypothesis']}\n\n"
+        "## Scoring\n"
+        f"{metrics_md}\n\n"
+        "## Experimental outline\n"
+        "1. Construct the candidate material or protocol.\n"
+        "2. Test mechanistic signal expression under controlled conditions.\n"
+        "3. Compare against baseline and nearest-neighbour alternatives.\n"
+        "4. Falsify using the adversarial risk criteria surfaced in the reasoning path.\n"
+    )
+
+    cards_html  = build_cards_html(CANDIDATES)
+    route_state = get_default_route_state()
+
+    return chat_html, connectome_html, timeline_html, cards_html, hypothesis_md, build_models_html(model_name), route_state
+
+
+def apply_route_swap(query: str, model_name: str, route_swap_payload: str, route_state):
+    """
+    Called when a user clicks 'Use as main insight' on a candidate card.
+    Sanitizes the output, adopts academic rigor, updates the connectome and discovery output.
+    """
+    try:
+        idx = int(route_swap_payload)
+    except ValueError:
+        idx = 0
+
+    if not (0 <= idx < len(ACADEMIC_INSIGHTS)):
+        idx = 0
+
+    academic = ACADEMIC_INSIGHTS[idx]
+    
+    # Update Connectome
+    connectome_html = build_connectome_html(academic["path"])
+    
+    # Update Chat Feedback
+    result = {
+        "summary": "Main insight formally adopted. The connectome pathway and validation protocol have been realigned to the selected candidate methodology.",
+        "primary_hypothesis": academic["hypothesis"]
+    }
+    chat_html = build_chat_html(query, result)
+
+    # Format Oxford-tier markdown output
+    metrics_md = "\n".join(f"- {k}: {v}/100" for k, v in academic["metrics"].items())
+    
+    hypothesis_md = (
+        "# Discovery Output\n\n"
+        f"**Model:** {model_name}\n\n"
+        f"**Primary hypothesis:** {academic['hypothesis']}\n\n"
+        "## Scoring\n"
+        f"{metrics_md}\n\n"
+        "## Experimental outline\n"
+        f"{academic['outline']}\n"
+    )
+
+    return chat_html, connectome_html, gr.update(), hypothesis_md, route_state
+
+
+# ── Example loaders ─────────────────���─────────────────────────────────────────
+def load_example() -> str:
+    return "How could a self-assembling conductive biomaterial improve cardiac tissue regeneration by converting mechanical strain into repair signalling?"
+
+def load_paper_topic() -> str:
+    return "self-assembling conductive biomaterials for cardiac repair"
+
+# ── CSS / HEAD ────────────────────────────────────────────────────────────────
+BASE_CSS = r"""
+:root {
+  --bg: #ffffff; --panel: #ffffff; --line: rgba(0,0,0,.12);
+  --text: #111111; --muted: #5b5b5b; --soft: rgba(0,0,0,.62);
+  --gold: #ff6600; --teal: #17b8a6; --blue: #628dff;
+  --chosen: #ff7a1a; --idle: #b8d8ff; --idle-stroke: #5e8fe6;
+  --query-node: #ffd8b3; --paper-node: #d7f6f2; --upload-node: #e7defe;
+  --shadow: 0 16px 40px rgba(0,0,0,.12);
+}
+html,body,.gradio-container { background:#ffffff !important; font-family:Inter,ui-sans-serif,system-ui,sans-serif; }
+.gradio-container { max-width:1640px !important; padding:20px !important; }
+#dvnc-shell { border:1px solid var(--line); border-radius:28px; overflow:hidden; background:#ffffff; box-shadow:var(--shadow); padding:20px 22px 22px; }
+.hero-bar { display:flex; justify-content:space-between; align-items:center; gap:16px; padding-bottom:12px; border-bottom:1px solid rgba(0,0,0,.06); margin-bottom:16px; }
+.brand { display:flex; align-items:center; gap:14px; }
+.logo { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; color:var(--gold); background:linear-gradient(135deg,rgba(255,122,26,.12),rgba(23,184,166,.10)); border:1px solid rgba(0,0,0,.08); }
+.logo svg { width:24px; height:24px; }
+.brand h1 { font-size:1.05rem; margin:0; font-weight:700; letter-spacing:.12em; text-transform:uppercase; }
+.brand p { margin:3px 0 0; color:var(--muted); font-size:.84rem; }
+.status { display:flex; gap:10px; align-items:center; color:var(--soft); font-size:.85rem; }
+.status-dot { width:10px; height:10px; border-radius:50%; background:var(--teal); box-shadow:0 0 0 6px rgba(23,184,166,.10),0 0 14px rgba(23,184,166,.25); }
+.panel { background:#ffffff; border:1px solid var(--line); border-radius:22px; box-shadow:inset 0 1px 0 rgba(255,255,255,.8); }
+.querybox textarea,.querybox input { background:transparent !important; color:var(--text) !important; }
+.querybox,.querybox>div { background:#ffffff !important; border-radius:18px !important; border-color:var(--line) !important; }
+.chat-panel { padding:18px; min-height:280px; }
+.chat-thread { display:flex; flex-direction:column; gap:14px; }
+.bubble { max-width:88%; padding:16px 18px; border-radius:22px; border:1px solid var(--line); }
+.bubble p { margin:8px 0 0; line-height:1.6; font-size:.96rem; color:var(--text); }
+.bubble .role { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
+.bubble-user { align-self:flex-end; background:linear-gradient(135deg,rgba(98,141,255,.16),rgba(98,141,255,.08)); }
+.bubble-ai { align-self:flex-start; background:#ffffff; }
+.bubble-system { align-self:flex-start; background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,122,26,.04)); }
+.model-switcher { display:grid; grid-template-columns:repeat(3,1fr); gap:12px; }
+.model-pill { padding:14px; border:1px solid var(--line); border-radius:18px; display:flex; flex-direction:column; gap:4px; min-height:98px; background:#ffffff; }
+.model-pill.active { border-color:rgba(255,122,26,.40); background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,255,255,.96)); }
+.model-name { font-weight:650; color:var(--text); }
+.model-tag { font-size:.76rem; text-transform:uppercase; letter-spacing:.12em; color:var(--gold); }
+.model-pill small { color:var(--muted); line-height:1.45; }
+.brain-shell { padding:18px; }
+.brain-header { display:flex; justify-content:space-between; align-items:flex-end; gap:16px; margin-bottom:10px; }
+.eyebrow { font-size:.72rem; letter-spacing:.16em; text-transform:uppercase; color:var(--gold); margin:0 0 4px; }
+.brain-header h3 { margin:0; font-size:1.12rem; color:var(--text); }
+.brain-legend { display:flex; gap:14px; color:var(--muted); font-size:.8rem; flex-wrap:wrap; }
+.dot { width:10px; height:10px; display:inline-block; border-radius:50%; margin-right:6px; }
+.dot-live { background:var(--chosen); box-shadow:0 0 10px rgba(255,122,26,.35); }
+.dot-chosen { background:var(--chosen); }
+.dot-idle { background:var(--idle); border:1px solid var(--idle-stroke); }
+.dot-query { background:var(--query-node); border:1px solid #de9e58; }
+.dot-paper { background:var(--paper-node); border:1px solid #4fb3a5; }
+.dot-upload { background:var(--upload-node); border:1px solid #8f73d9; }
+.brain-stage { position:relative; min-height:420px; overflow:hidden; background:linear-gradient(180deg,rgba(250,250,250,1),rgba(255,255,255,1)); border:1px solid rgba(0,0,0,.05); border-radius:20px; }
+.brain-svg { width:100%; height:520px; display:block; }
+.edge { stroke:rgba(0,0,0,.12); stroke-width:2.4; }
+.edge.active { stroke:var(--chosen); stroke-width:4.2; stroke-linecap:round; filter:drop-shadow(0 0 6px rgba(255,122,26,.45)); stroke-dasharray:8 12; animation:pulseEdge 1.5s linear infinite; }
+.node { stroke-width:2.2; transition:all .25s ease; }
+.node.idle { fill:var(--idle); stroke:var(--idle-stroke); }
+.node.chosen { fill:var(--chosen); stroke:#ffb16d; }
+.halo { fill:none; }
+.halo.active { stroke:rgba(255,122,26,.18); stroke-width:12; }
+.label { fill:#2c2c2c; font-size:13px; font-weight:500; letter-spacing:.01em; }
+.label.active { fill:#111111; font-weight:700; }
+.learn-edge { stroke:rgba(0,0,0,.18); stroke-width:2.2; stroke-linecap:round; }
+.learn-node { stroke-width:2.2; }
+.learn-node.query { fill:var(--query-node); stroke:#de9e58; }
+.learn-node.paper { fill:var(--paper-node); stroke:#36a091; }
+.learn-node.upload { fill:var(--upload-node); stroke:#7e63cb; }
+.learn-label { fill:#1e1e1e; font-size:12px; font-weight:600; }
+.learning-empty { display:grid; place-items:center; }
+.empty-graph-copy { text-align:center; max-width:440px; padding:40px 20px; }
+.empty-graph-copy h4 { margin:0 0 10px; font-size:1.05rem; }
+.empty-graph-copy p { margin:0; color:var(--muted); line-height:1.6; }
+.timeline { display:flex; flex-direction:column; gap:10px; }
+.agent-step { border:1px solid var(--line); border-radius:18px; background:#ffffff; overflow:hidden; }
+.agent-summary { list-style:none; display:grid; grid-template-columns:42px 1fr; gap:12px; align-items:center; padding:12px; cursor:pointer; }
+.agent-summary::-webkit-details-marker { display:none; }
+.agent-index { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; font-weight:700; color:var(--gold); background:rgba(255,122,26,.08); border:1px solid rgba(255,122,26,.18); }
+.agent-head { display:flex; justify-content:space-between; gap:12px; align-items:center; }
+.agent-head h4 { margin:0; font-size:.98rem; color:var(--text); }
+.agent-head span { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
+.agent-copy { padding:0 14px 16px 66px; }
+.agent-copy p { margin:0; color:#2d2d2d; font-size:.93rem; line-height:1.6; }
+.candidate-grid { display:grid; grid-template-columns:repeat(3,minmax(0,1fr)); gap:18px; }
+.candidate-card { background:none; perspective:1400px; min-height:330px; }
+.candidate-card-inner { position:relative; width:100%; min-height:330px; transition:transform .8s cubic-bezier(.2,.7,.1,1); transform-style:preserve-3d; }
+.candidate-card:hover .candidate-card-inner,.candidate-card:focus .candidate-card-inner,.candidate-card:focus-within .candidate-card-inner { transform:rotateY(180deg); }
+.candidate-face { position:absolute; inset:0; padding:20px; border-radius:22px; border:1px solid var(--line); background:#ffffff; color:var(--text); backface-visibility:hidden; box-shadow:0 12px 24px rgba(0,0,0,.06); display:flex; flex-direction:column; gap:14px; }
+.candidate-back { transform:rotateY(180deg); }
+.candidate-top { display:flex; justify-content:space-between; align-items:center; gap:8px; }
+.chip { font-size:.72rem; text-transform:uppercase; letter-spacing:.12em; color:#0b6f66; padding:7px 10px; border-radius:999px; background:rgba(23,184,166,.08); border:1px solid rgba(23,184,166,.18); }
+.chip.alt { color:var(--gold); background:rgba(255,122,26,.08); border-color:rgba(255,122,26,.18); }
+.score { font-weight:700; color:var(--gold); }
+.candidate-face h4 { margin:0; font-size:1.08rem; line-height:1.35; }
+.candidate-face p { margin:0; color:#1e1e1e; line-height:1.65; font-size:.96rem; overflow-wrap:anywhere; }
+.meta-row { margin-top:auto; display:flex; justify-content:space-between; color:var(--muted); font-size:.88rem; gap:14px; }
+.mini { cursor:pointer; margin-top:8px; align-self:flex-start; color:var(--text); padding:10px 12px; border-radius:14px; border:1px solid var(--line); background:#ffffff; transition:all 0.2s; }
+.mini:hover { background: #f5f5f5; border-color: var(--chosen); }
+.papers-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
+.paper-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
+.paper-topline { display:flex; gap:8px; flex-wrap:wrap; margin-bottom:10px; }
+.paper-badge { font-size:.72rem; padding:6px 10px; border-radius:999px; background:rgba(98,141,255,.08); color:#3456b5; border:1px solid rgba(98,141,255,.18); }
+.paper-badge.alt { background:rgba(0,0,0,.04); color:#444; border-color:rgba(0,0,0,.08); }
+.doi-badge { background:rgba(255,122,26,.08); color:#8a4105; border-color:rgba(255,122,26,.18); }
+.paper-card h4 { margin:0 0 10px; line-height:1.35; font-size:1rem; }
+.paper-card p { margin:0 0 12px; line-height:1.6; color:#222; }
+.paper-links { display:flex; gap:12px; flex-wrap:wrap; }
+.paper-meta-stack { display:flex; flex-direction:column; gap:6px; color:#444; margin-bottom:12px; font-size:.9rem; }
+.paper-links a,.journal-card,.upload-note a { color:#0b63ce; text-decoration:none; }
+.journal-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
+.journal-card { border:1px solid var(--line); border-radius:18px; padding:16px; display:flex; justify-content:space-between; gap:14px; align-items:center; background:#ffffff; }
+.journal-card h4 { margin:0 0 6px; }
+.journal-card p { margin:0; color:var(--muted); line-height:1.5; }
+.upload-note { border:1px dashed rgba(0,0,0,.16); border-radius:18px; padding:16px; background:rgba(0,0,0,.015); color:#1f1f1f; line-height:1.6; }
+.prosebox { padding:18px; white-space:pre-wrap; font-family:ui-monospace,SFMono-Regular,Menlo,monospace; line-height:1.55; color:#1b1b1b; }
+.gr-button-primary { background:linear-gradient(135deg,rgba(255,122,26,.92),rgba(240,108,22,.92)) !important; color:#ffffff !important; border:none !important; }
+.gr-button-secondary { background:#ffffff !important; color:var(--text) !important; border:1px solid var(--line) !important; }
+.ref-list { margin:0; padding-left:18px; }
+.ref-list li { margin-bottom:8px; line-height:1.5; }
+.parse-grid { display:grid; grid-template-columns:1.2fr 1fr; gap:14px; }
+.parse-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
+.selection-panel { padding:18px; }
+footer { display:none !important; }
+@keyframes pulseEdge { to { stroke-dashoffset:-40; } }
+@media (max-width:1180px) {
+  .model-switcher,.candidate-grid,.papers-grid,.journal-grid,.parse-grid { grid-template-columns:1fr; }
+  .brain-svg { height:460px; }
+}
+"""
+
+CSS  = BASE_CSS + "\n" + get_dvnc_layout_css()
+HEAD = """
+<link rel="preconnect" href="https://fonts.googleapis.com">
+<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
+<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap" rel="stylesheet">
+<script>
+function triggerRouteSwap(idx) {
+    const container = document.getElementById('route_swap_payload');
+    if(!container) return;
+    const input = container.querySelector('textarea') || container.querySelector('input');
+    if(input) {
+        input.value = idx.toString();
+        input.dispatchEvent(new Event('input', { bubbles: true }));
+        setTimeout(() => {
+            const btn = document.getElementById('route_swap_apply');
+            if(btn) btn.click();
+        }, 150);
+    }
+}
+</script>
+"""
+
+# ── Gradio layout ─────────────────────────────────────────────────────────────
+with gr.Blocks(css=CSS, head=HEAD, theme=gr.themes.Base(), fill_height=True) as demo:
+
+    # ── Shared state ──────────────────────────────────────────────────────────
+    papers_state        = gr.State([])
+    parsed_pdf_state    = gr.State({})
+    ingest_payload_state = gr.State({})
+    route_state         = gr.State(get_default_route_state())
+
+    # ── Header ────────────────────────────────────────────────────────────────
+    gr.HTML("""
+        <div id="dvnc-shell">
+          <div class="hero-bar">
+            <div class="brand">
+              <div class="logo" aria-hidden="true">
+                <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7">
+                  <path d="M5 17L12 4l7 13"/>
+                  <path d="M8.5 12.5h7"/>
+                  <circle cx="12" cy="12" r="1.8" fill="currentColor" stroke="none"/>
+                </svg>
+              </div>
+              <div>
+                <h1>DVNC.AI</h1>
+                <p>Sovereign discovery instrument · connectome-native reasoning</p>
+              </div>
+            </div>
+            <div class="status"><span class="status-dot"></span><span>Live orchestration</span></div>
+          </div>
+        </div>
+    """)
+
+    with gr.Tabs():
+
+        # ── Tab 1 · Discovery Engine ──────────────────────────────────────────
+        with gr.Tab("Discovery Engine"):
+            model_html = gr.HTML(build_models_html("DVNC Sovereign"))
+
+            with gr.Row():
+                with gr.Column(scale=2):
+                    model = gr.Dropdown(
+                        choices=[m["name"] for m in MODELS],
+                        value="DVNC Sovereign",
+                        label="Model tier",
+                    )
+                    query = gr.Textbox(
+                        label="Discovery query",
+                        elem_classes=["querybox"],
+                        placeholder="Enter a scientific question, anomaly, or breakthrough direction…",
+                        lines=4,
+                    )
+                    with gr.Row():
+                        run_btn     = gr.Button("Run discovery",  variant="primary")
+                        example_btn = gr.Button("Load example",   variant="secondary")
+                    chat = gr.HTML("""
+                        <div class="panel chat-panel">
+                          <div class="chat-thread">
+                            <div class="bubble bubble-ai">
+                              <span class="role">DVNC</span>
+                              <p>Enter a query to activate the 7-agent discovery stack and illuminate the chosen path through the 3D connectome.</p>
+                            </div>
+                          </div>
+                        </div>
+                    """)
+
+                with gr.Column(scale=3):
+                    connectome = gr.HTML(build_connectome_html(DEFAULT_PATH))
+                    cards      = gr.HTML("")
+
+            output   = gr.Markdown("# Discovery Output\n\nAwaiting query.")
+            timeline = gr.HTML(get_initial_discovery_timeline_html())
+
+            route_swap_payload = gr.Textbox(value="", visible=False, elem_id="route_swap_payload")
+            route_swap_apply   = gr.Button("Apply route swap", visible=False, elem_id="route_swap_apply")
+
+        # ── Tab 2 · Self-Learning Graph ───────────────────────────────────────
+        with gr.Tab("Self-Learning Graph"):
+            with gr.Row():
+                with gr.Column(scale=2):
+                    paper_query = gr.Textbox(
+                        label="Research topic / title / DOI / link",
+                        elem_classes=["querybox"],
+                        placeholder="e.g. self-assembling conductive biomaterials for cardiac repair",
+                        lines=3,
+                    )
+                    search_mode = gr.Dropdown(
+                        choices=SEARCH_MODES,
+                        value="topic",
+                        label="Search mode",
+                    )
+                    source_selector = gr.CheckboxGroup(
+                        choices=SOURCE_OPTIONS,
+                        value=DEFAULT_SOURCES,
+                        label="Sources",
+                    )
+                    pdf_upload = gr.File(label="Upload PDF papers", file_types=[".pdf"], file_count="single")
+
+                    with gr.Row():
+                        learn_btn      = gr.Button("Discover papers",      variant="primary")
+                        load_topic_btn = gr.Button("Load example topic",   variant="secondary")
+
+                    upload_status    = gr.Markdown("No PDF uploaded yet.")
+                    discovery_status = gr.Markdown("### No discovery results yet.")
+                    journal_panel    = gr.HTML(build_journal_html("biomaterials cardiac repair"))
+
+                    gr.HTML('<div class="panel selection-panel"><h3 style="margin:0 0 12px;">Select papers to ingest</h3></div>')
+                    selection_box = gr.CheckboxGroup(choices=[], value=[], label="Candidate papers")
+
+                    parser_order = gr.CheckboxGroup(
+                        choices=["grobid", "docling", "pymupdf"],
+                        value=["grobid", "docling", "pymupdf"],
+                        label="Parser routing order",
+                    )
+
+                    with gr.Row():
+                        parse_btn  = gr.Button("Parse uploaded PDF",        variant="secondary")
+                        ingest_btn = gr.Button("Ingest selected into graph", variant="primary")
+
+                with gr.Column(scale=3):
+                    learning_graph = gr.HTML(build_learning_graph_html([], []))
+                    papers_panel   = gr.HTML('<div class="panel papers-panel" style="padding:18px"><p>Search by topic, title, DOI, or link, then select papers before graph ingestion.</p></div>')
+                    parse_summary  = gr.Markdown("### PDF parse status\n\nAwaiting upload.")
+                    parse_panel    = gr.HTML('<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>')
+                    ingest_summary = gr.Markdown("### Graph ingest status\n\nAwaiting paper selection.")
+                    ingest_payload = gr.JSON(label="Graph ingest payload", value={"status": "empty", "nodes": [], "edges": []})
 
-ORIGINAL_APP_PATH = Path("dvnc_ai_v2_hf/app.py")
-PATCH_PATH = Path("dvnc_flip_insight_patch.py")
+    # ── Event wiring ──────────────────────────────────────────────────────────
+    example_btn.click(fn=load_example, outputs=query)
 
+    run_btn.click(
+        fn=run_discovery,
+        inputs=[query, model],
+        outputs=[chat, connectome, timeline, cards, output, model_html, route_state],
+    )
 
-def load_module(module_name: str, file_path: Path):
-    spec = importlib.util.spec_from_file_location(module_name, file_path)
-    if spec is None or spec.loader is None:
-        raise RuntimeError(f"Could not load module {module_name} from {file_path}")
-    module = importlib.util.module_from_spec(spec)
-    sys.modules[module_name] = module
-    spec.loader.exec_module(module)
-    return module
+    route_swap_apply.click(
+        fn=apply_route_swap,
+        inputs=[query, model, route_swap_payload, route_state],
+        outputs=[chat, connectome, timeline, output, route_state],
+    )
 
+    load_topic_btn.click(fn=load_paper_topic, outputs=paper_query)
 
-original_app = load_module("dvnc_original_app", ORIGINAL_APP_PATH)
-patch_module = load_module("dvnc_flip_insight_patch", PATCH_PATH)
+    learn_btn.click(
+        fn=run_paper_discovery,
+        inputs=[paper_query, search_mode, source_selector, pdf_upload],
+        outputs=[learning_graph, papers_panel, journal_panel, upload_status, selection_box, papers_state, discovery_status],
+    )
 
-patch_module.apply_patch(original_app)
+    parse_btn.click(
+        fn=parse_uploaded_pdf,
+        inputs=[pdf_upload, parser_order],
+        outputs=[parse_summary, parsed_pdf_state],
+    ).then(
+        fn=render_parse_result,
+        inputs=[parsed_pdf_state],
+        outputs=[parse_panel],
+    )
 
-demo = getattr(original_app, "demo", None)
-app = getattr(original_app, "app", None)
+    ingest_btn.click(
+        fn=ingest_selected_papers,
+        inputs=[paper_query, selection_box, papers_state, pdf_upload, parsed_pdf_state],
+        outputs=[learning_graph, ingest_summary, ingest_payload],
+    )
 
 if __name__ == "__main__":
-    if demo is not None:
-        demo.launch()
+    demo.launch()

app_old2.py

@@ -1,770 +0,0 @@
-"""
-DVNC.AI — app.py
-Refactored for functional "Use as main insight" logic with academic rigor.
-"""
-from dvnc_ai_v2_hf.graph_canvas_patch import render_graph_canvas_html
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    return render_graph_canvas_html(
-        {
-            "status": "ok" if (nodes or edges) else "empty",
-            "nodes": nodes or [],
-            "edges": edges or [],
-        },
-        title=title,
-        height=780,
-    )
-
-# ── Standard library ────────────────────────────────────────────────────────
-import html
-import json
-import math
-import os
-import random
-import re
-import urllib.parse
-import xml.etree.ElementTree as ET
-from pathlib import Path
-from typing import Dict, List, Optional
-from urllib.parse import quote
-from dvnc_ai_v2_hf.graph_canvas_patch import render_graph_canvas_html
-
-# ── Third-party ──────────────────────────────────────────────────────────────
-import gradio as gr
-import requests
-
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-
-learning_graph = gr.HTML(build_learning_graph_html([], []))
-
-# ── Internal modules ─────────────────────────────────────────────────────────
-from dvnc_ai_v2_hf.agent_route_cards import build_agent_route_cards_html
-from dvnc_ai_v2_hf.discovery_app_bridge import (
-    get_default_route_state,
-    get_discovery_css,
-    get_initial_discovery_timeline_html,
-)
-from dvnc_ai_v2_hf.dvnc_ui_layout import get_dvnc_layout_css
-from dvnc_ai_v2_hf.self_learning_graph import (
-    DEFAULT_SOURCES,
-    SEARCH_MODES,
-    SOURCE_OPTIONS,
-    build_journal_html,
-    ingest_selected_papers,
-    parse_uploaded_pdf,
-    render_parse_result,
-    run_paper_discovery,
-    safe_text,
-)
-
-from dvnc_ai_v2_hf.graph_canvas_patch import render_graph_canvas_html
-
-# ── Constants ────────────────────────────────────────────────────────────────
-MODELS = [
-    {"name": "DVNC Sovereign", "tag": "flagship", "desc": "Maximum depth orchestration for frontier discovery"},
-    {"name": "DVNC Atlas",     "tag": "research",  "desc": "Balanced reasoning, graph traversal, and synthesis"},
-    {"name": "DVNC Curie",     "tag": "lab",        "desc": "Experimental hypothesis generation for anomalous signals"},
-]
-
-AGENTS = [
-    "Query Interpreter",
-    "Graph Divergence Mapper",
-    "Evidence Harvester",
-    "Analogy Engine",
-    "Hypothesis Composer",
-    "Adversarial Critic",
-    "Experimental Program Designer",
-]
-
-NODES = [
-    {"id": "seed",    "label": "Seed Query",          "group": "core",      "x": 10,  "y":   0, "z":   0},
-    {"id": "bio",     "label": "Biomaterials",         "group": "domain",    "x": 24,  "y":  12, "z":  -8},
-    {"id": "card",    "label": "Cardiac Repair",       "group": "domain",    "x": 38,  "y":   3, "z":  14},
-    {"id": "nano",    "label": "Nanostructure",        "group": "bridge",    "x": 24,  "y": -18, "z":  16},
-    {"id": "selfasm", "label": "Self-Assembly",        "group": "bridge",    "x": 40,  "y": -16, "z": -16},
-    {"id": "electro", "label": "Electro-signalling",   "group": "mechanism", "x": 58,  "y":  10, "z": -10},
-    {"id": "immune",  "label": "Immune Modulation",    "group": "mechanism", "x": 64,  "y":  -8, "z":  10},
-    {"id": "trial",   "label": "Validation Path",      "group": "outcome",   "x": 80,  "y":   0, "z":   0},
-    {"id": "alt1",    "label": "Piezoelectric Scaffold","group": "candidate", "x": 56,  "y":  26, "z":  14},
-    {"id": "alt2",    "label": "Peptide Mesh",          "group": "candidate", "x": 54,  "y": -27, "z": -14},
-]
-
-EDGES = [
-    ("seed",    "bio"),    ("seed",   "nano"),
-    ("bio",     "card"),   ("nano",   "selfasm"),
-    ("selfasm", "electro"),("card",   "immune"),
-    ("electro", "trial"),  ("immune", "trial"),
-    ("card",    "alt1"),   ("selfasm","alt2"),
-    ("alt1",    "trial"),  ("alt2",   "trial"),
-]
-
-DEFAULT_PATH = ["seed", "nano", "selfasm", "electro", "trial"]
-
-CANDIDATES = [
-    {
-        "title":   "Piezoelectric Scaffold Cascade",
-        "front":   "Use mechano-electric scaffolds to convert cardiac strain into micro-current signalling.",
-        "back":    "Discovery path: anomalous healing signal -> piezoelectric analog -> ion-channel entrainment -> tissue regeneration. Risk: power density and fibrosis coupling.",
-        "score":   92,
-        "novelty": "High",
-        "agent":   "Hypothesis Composer",
-    },
-    {
-        "title":   "Peptide Self-Assembly Mesh",
-        "front":   "Deploy dynamic peptide meshes that self-assemble around damaged myocardium and guide repair.",
-        "back":    "Discovery path: self-assembly -> local immune choreography -> regenerative substrate formation. Risk: degradation timing and targeting specificity.",
-        "score":   88,
-        "novelty": "High",
-        "agent":   "Analogy Engine",
-    },
-    {
-        "title":   "Immune-Tuned Conductive Hydrogel",
-        "front":   "Blend conductivity with macrophage-state modulation to reduce scarring and restore conduction.",
-        "back":    "Discovery path: inflammation mismatch -> conductive medium -> macrophage polarization -> synchronized healing. Risk: persistence and biocompatibility.",
-        "score":   85,
-        "novelty": "Medium-High",
-        "agent":   "Adversarial Critic",
-    },
-]
-
-ACADEMIC_INSIGHTS = [
-    {
-        "hypothesis": "Implementation of mechano-electric scaffolds to transduce cardiac strain into localized micro-current signalling for myocardial regeneration.",
-        "metrics": {"Novelty": 92, "Mechanistic clarity": 85, "Experimental tractability": 78, "Cross-domain distance": 94},
-        "outline": "1. Synthesize candidate piezoelectric biomaterial scaffolds with tunable strain-electric coupling.\n2. Evaluate *in vitro* electromechanical transduction and subsequent ion-channel entrainment.\n3. Conduct *in vivo* comparative models to assess regenerative efficacy against gold-standard substrates.\n4. Rigorously validate to exclude pathological fibrosis and power-density toxicity.",
-        "path": ["seed", "bio", "card", "alt1", "trial"]
-    },
-    {
-        "hypothesis": "Deployment of dynamic peptide networks that self-assemble post-infarction to orchestrate local immunological responses and guide substrate regeneration.",
-        "metrics": {"Novelty": 88, "Mechanistic clarity": 82, "Experimental tractability": 86, "Cross-domain distance": 85},
-        "outline": "1. Formulate peptide sequences programmed for triggered *in situ* self-assembly within the myocardial infarct zone.\n2. Quantify macrophage polarization and local immune choreography post-deployment.\n3. Map the temporospatial degradation profile against *de novo* tissue formation.\n4. Falsify against off-target aggregation and delayed clearance risks.",
-        "path": ["seed", "nano", "selfasm", "alt2", "trial"]
-    },
-    {
-        "hypothesis": "Integration of conductive hydrogels with immunomodulatory properties to simultaneously bridge electrical uncoupling and mitigate adverse fibrotic scarring.",
-        "metrics": {"Novelty": 85, "Mechanistic clarity": 90, "Experimental tractability": 88, "Cross-domain distance": 79},
-        "outline": "1. Fabricate biocompatible hydrogels featuring precisely tuned electrical conductivity and immunomodulatory motifs.\n2. Monitor electrophysiological synchronization across the scaffold-tissue interface.\n3. Assess macrophage state transitions and suppression of adverse fibrotic remodelling.\n4. Validate long-term persistence, hemocompatibility, and mechanical integration.",
-        "path": ["seed", "bio", "card", "immune", "trial"]
-    }
-]
-
-JOURNALS = [
-    {"name": "Nature",     "url": "https://www.nature.com/search",                              "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science",    "url": "https://www.science.org/search",                             "desc": "High-impact science journal and family."},
-    {"name": "Cell",       "url": "https://www.cell.com/search",                                "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search",                           "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore","url": "https://ieeexplore.ieee.org/search/searchresult.jsp",        "desc": "Engineering, AI, signal processing, and systems."},
-]
-
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
-GROBID_URL               = os.getenv("GROBID_URL", "").strip()
-REQUEST_TIMEOUT          = 25
-
-
-# ── Utility helpers ──────────────────────────────────────────────────────────
-def safe_text(x, default: str = "") -> str:
-    return html.escape(str(x if x is not None else default))
-
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    if re.match(r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$", q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-
-
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-# ── HTML builders ─────────────────────────────────────────────────────────────
-def build_connectome_html(path_ids: List[str]) -> str:
-    active    = set(path_ids)
-    node_map  = {n["id"]: n for n in NODES}
-    path_pairs = {
-        pair
-        for i in range(len(path_ids) - 1)
-        for pair in [(path_ids[i], path_ids[i + 1]), (path_ids[i + 1], path_ids[i])]
-    }
-
-    base_lines, active_lines, circles, labels = [], [], [], []
-
-    for a, b in EDGES:
-        na, nb = node_map[a], node_map[b]
-        x1, y1 = na["x"] * 8 + 80, na["y"] * 6 + 280
-        x2, y2 = nb["x"] * 8 + 80, nb["y"] * 6 + 280
-        base_lines.append(f'<line class="edge" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />')
-        if (a, b) in path_pairs:
-            active_lines.append(f'<line class="edge active" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />')
-
-    for n in NODES:
-        cx, cy   = n["x"] * 8 + 80, n["y"] * 6 + 280
-        is_active = n["id"] in active
-        state    = "chosen" if is_active else "idle"
-        halo_cls = "halo active" if is_active else "halo"
-        lbl_cls  = "label active" if is_active else "label"
-        radius   = 18 if is_active else 13
-        halo_r   = 30 if is_active else 0
-        circles.append(
-            f'<g class="node-wrap">'
-            f'<circle class="{halo_cls}" cx="{cx:.1f}" cy="{cy:.1f}" r="{halo_r}" />'
-            f'<circle class="node {n["group"]} {state}" cx="{cx:.1f}" cy="{cy:.1f}" r="{radius}" />'
-            f'</g>'
-        )
-        labels.append(f'<text class="{lbl_cls}" x="{cx + 18:.1f}" y="{cy - 16:.1f}">{safe_text(n["label"])}</text>')
-
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Connectome</p>
-          <h3>3D Connectome</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-live"></i> lit path</span>
-          <span><i class="dot dot-chosen"></i> chosen node</span>
-          <span><i class="dot dot-idle"></i> available node</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 780 560" class="brain-svg" role="img" aria-label="DVNC 3D connectome visualisation">
-          {"".join(base_lines)}
-          {"".join(active_lines)}
-          {"".join(circles)}
-          {"".join(labels)}
-        </svg>
-      </div>
-    </div>
-    """
-
-
-def build_cards_html(cards: List[Dict]) -> str:
-    items = []
-    for i, c in enumerate(cards):
-        items.append(f"""
-        <article class="candidate-card" tabindex="0">
-          <div class="candidate-card-inner">
-            <div class="candidate-face candidate-front">
-              <div class="candidate-top">
-                <span class="chip">{safe_text(c["agent"])}</span>
-                <span class="score">{safe_text(c["score"])}</span>
-              </div>
-              <h4>{safe_text(c["title"])}</h4>
-              <p>{safe_text(c["front"])}</p>
-              <div class="meta-row"><span>Novelty</span><strong>{safe_text(c["novelty"])}</strong></div>
-              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-            </div>
-            <div class="candidate-face candidate-back">
-              <div class="candidate-top">
-                <span class="chip alt">Alternative path</span>
-                <span class="score">{safe_text(c["score"])}</span>
-              </div>
-              <h4>{safe_text(c["title"])}</h4>
-              <p>{safe_text(c["back"])}</p>
-              <div class="meta-row"><span>Swap into route</span><strong>Enabled</strong></div>
-              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-            </div>
-          </div>
-        </article>""")
-    return '<div class="panel" style="padding:20px;"><div class="candidate-grid">' + "".join(items) + "</div></div>"
-
-
-def build_agent_timeline(reasoning: List[Dict]) -> str:
-    rows = []
-    for r in reasoning:
-        rows.append(f"""
-        <details class="agent-step" {"open" if r["step"] == 1 else ""}>
-          <summary class="agent-summary">
-            <div class="agent-index">{safe_text(r["step"])}</div>
-            <div class="agent-head">
-              <h4>{safe_text(r["agent"])}</h4>
-              <span>{safe_text(r["tag"])}</span>
-            </div>
-          </summary>
-          <div class="agent-copy">
-            <p>{safe_text(r["summary"])}</p>
-          </div>
-        </details>""")
-    return '<div class="panel" style="padding:18px;"><div class="timeline">' + "".join(rows) + "</div></div>"
-
-
-def build_chat_html(query: str, result: Dict) -> str:
-    return f"""
-    <div class="panel chat-panel">
-      <div class="chat-thread">
-        <div class="bubble bubble-user">
-          <span class="role">You</span>
-          <p>{safe_text(query)}</p>
-        </div>
-        <div class="bubble bubble-ai">
-          <span class="role">DVNC Sovereign</span>
-          <p>{safe_text(result["summary"])}</p>
-        </div>
-        <div class="bubble bubble-system">
-          <span class="role">Discovery Signal</span>
-          <p><strong>Primary hypothesis:</strong> {safe_text(result["primary_hypothesis"])}</p>
-        </div>
-      </div>
-    </div>
-    """
-
-
-def build_models_html(selected: str) -> str:
-    items = []
-    for m in MODELS:
-        active = "active" if m["name"] == selected else ""
-        items.append(f"""
-            <div class="model-pill {active}">
-              <span class="model-name">{safe_text(m["name"])}</span>
-              <span class="model-tag">{safe_text(m["tag"])}</span>
-              <small>{safe_text(m["desc"])}</small>
-            </div>""")
-    return '<div class="panel" style="padding:18px;"><div class="model-switcher">' + "".join(items) + "</div></div>"
-
-
-# ── Discovery logic ───────────────────────────────────────────────────────────
-def run_discovery(query: str, model_name: str):
-    """
-    Runs the 7-agent discovery pipeline.
-    """
-    random.seed(len(query) + len(model_name))
-
-    if "curie" in query.lower() or "einstein" in query.lower():
-        primary = "Map the anomaly first, then force a distant analogy before composing the experimental programme."
-        path    = ["seed", "bio", "card", "immune", "trial"]
-    else:
-        primary = "Utilization of a self-assembling conductive scaffold to transduce mechanical strain into localized regenerative signalling pathways."
-        path    = DEFAULT_PATH
-
-    summaries = [
-        "Normalises the user prompt into a graph-searchable seed and isolates the tension inside the question.",
-        "Finds remote conceptual bridges instead of staying near the starting domain cluster.",
-        "Pulls evidence packets and conflict signals required for grounded hypothesis formation.",
-        "Generates cross-domain analogies with a bias toward mechanism transfer rather than keyword similarity.",
-        "Composes the lead hypothesis and two structurally different variants.",
-        "Attacks weak assumptions, hidden confounders, and feasibility gaps.",
-        "Produces a staged validation plan with measurable falsification criteria.",
-    ]
-
-    tags = ["input", "graph", "evidence", "analogy", "compose", "critique", "experiment"]
-
-    reasoning = [
-        {"step": i + 1, "agent": AGENTS[i], "tag": tags[i], "summary": summaries[i]}
-        for i in range(7)
-    ]
-
-    result = {
-        "summary":            "A deeper route was chosen through the connectome, with live alternatives preserved as swappable cards so the reasoning path can be inspected rather than hidden.",
-        "primary_hypothesis": primary,
-        "reasoning":          reasoning,
-        "cards":              CANDIDATES,
-        "path":               path,
-        "metrics": {
-            "Novelty":                   93,
-            "Mechanistic clarity":       89,
-            "Experimental tractability": 82,
-            "Cross-domain distance":     91,
-        },
-    }
-
-    chat_html       = build_chat_html(query, result)
-    connectome_html = build_connectome_html(path)
-    timeline_html   = build_agent_route_cards_html(reasoning)
-
-    metrics_md      = "\n".join(f"- {k}: {v}/100" for k, v in result["metrics"].items())
-    hypothesis_md   = (
-        "# Discovery Output\n\n"
-        f"**Model:** {model_name}\n\n"
-        f"**Primary hypothesis:** {result['primary_hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metrics_md}\n\n"
-        "## Experimental outline\n"
-        "1. Construct the candidate material or protocol.\n"
-        "2. Test mechanistic signal expression under controlled conditions.\n"
-        "3. Compare against baseline and nearest-neighbour alternatives.\n"
-        "4. Falsify using the adversarial risk criteria surfaced in the reasoning path.\n"
-    )
-
-    cards_html  = build_cards_html(CANDIDATES)
-    route_state = get_default_route_state()
-
-    return chat_html, connectome_html, timeline_html, cards_html, hypothesis_md, build_models_html(model_name), route_state
-
-
-def apply_route_swap(query: str, model_name: str, route_swap_payload: str, route_state):
-    """
-    Called when a user clicks 'Use as main insight' on a candidate card.
-    Sanitizes the output, adopts academic rigor, updates the connectome and discovery output.
-    """
-    try:
-        idx = int(route_swap_payload)
-    except ValueError:
-        idx = 0
-
-    if not (0 <= idx < len(ACADEMIC_INSIGHTS)):
-        idx = 0
-
-    academic = ACADEMIC_INSIGHTS[idx]
-    
-    # Update Connectome
-    connectome_html = build_connectome_html(academic["path"])
-    
-    # Update Chat Feedback
-    result = {
-        "summary": "Main insight formally adopted. The connectome pathway and validation protocol have been realigned to the selected candidate methodology.",
-        "primary_hypothesis": academic["hypothesis"]
-    }
-    chat_html = build_chat_html(query, result)
-
-    # Format Oxford-tier markdown output
-    metrics_md = "\n".join(f"- {k}: {v}/100" for k, v in academic["metrics"].items())
-    
-    hypothesis_md = (
-        "# Discovery Output\n\n"
-        f"**Model:** {model_name}\n\n"
-        f"**Primary hypothesis:** {academic['hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metrics_md}\n\n"
-        "## Experimental outline\n"
-        f"{academic['outline']}\n"
-    )
-
-    return chat_html, connectome_html, gr.update(), hypothesis_md, route_state
-
-
-# ── Example loaders ───────────────────────────────────────────────────────────
-def load_example() -> str:
-    return "How could a self-assembling conductive biomaterial improve cardiac tissue regeneration by converting mechanical strain into repair signalling?"
-
-def load_paper_topic() -> str:
-    return "self-assembling conductive biomaterials for cardiac repair"
-
-# ── CSS / HEAD ────────────────────────────────────────────────────────────────
-BASE_CSS = r"""
-:root {
-  --bg: #ffffff; --panel: #ffffff; --line: rgba(0,0,0,.12);
-  --text: #111111; --muted: #5b5b5b; --soft: rgba(0,0,0,.62);
-  --gold: #ff6600; --teal: #17b8a6; --blue: #628dff;
-  --chosen: #ff7a1a; --idle: #b8d8ff; --idle-stroke: #5e8fe6;
-  --query-node: #ffd8b3; --paper-node: #d7f6f2; --upload-node: #e7defe;
-  --shadow: 0 16px 40px rgba(0,0,0,.12);
-}
-html,body,.gradio-container { background:#ffffff !important; font-family:Inter,ui-sans-serif,system-ui,sans-serif; }
-.gradio-container { max-width:1640px !important; padding:20px !important; }
-#dvnc-shell { border:1px solid var(--line); border-radius:28px; overflow:hidden; background:#ffffff; box-shadow:var(--shadow); padding:20px 22px 22px; }
-.hero-bar { display:flex; justify-content:space-between; align-items:center; gap:16px; padding-bottom:12px; border-bottom:1px solid rgba(0,0,0,.06); margin-bottom:16px; }
-.brand { display:flex; align-items:center; gap:14px; }
-.logo { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; color:var(--gold); background:linear-gradient(135deg,rgba(255,122,26,.12),rgba(23,184,166,.10)); border:1px solid rgba(0,0,0,.08); }
-.logo svg { width:24px; height:24px; }
-.brand h1 { font-size:1.05rem; margin:0; font-weight:700; letter-spacing:.12em; text-transform:uppercase; }
-.brand p { margin:3px 0 0; color:var(--muted); font-size:.84rem; }
-.status { display:flex; gap:10px; align-items:center; color:var(--soft); font-size:.85rem; }
-.status-dot { width:10px; height:10px; border-radius:50%; background:var(--teal); box-shadow:0 0 0 6px rgba(23,184,166,.10),0 0 14px rgba(23,184,166,.25); }
-.panel { background:#ffffff; border:1px solid var(--line); border-radius:22px; box-shadow:inset 0 1px 0 rgba(255,255,255,.8); }
-.querybox textarea,.querybox input { background:transparent !important; color:var(--text) !important; }
-.querybox,.querybox>div { background:#ffffff !important; border-radius:18px !important; border-color:var(--line) !important; }
-.chat-panel { padding:18px; min-height:280px; }
-.chat-thread { display:flex; flex-direction:column; gap:14px; }
-.bubble { max-width:88%; padding:16px 18px; border-radius:22px; border:1px solid var(--line); }
-.bubble p { margin:8px 0 0; line-height:1.6; font-size:.96rem; color:var(--text); }
-.bubble .role { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
-.bubble-user { align-self:flex-end; background:linear-gradient(135deg,rgba(98,141,255,.16),rgba(98,141,255,.08)); }
-.bubble-ai { align-self:flex-start; background:#ffffff; }
-.bubble-system { align-self:flex-start; background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,122,26,.04)); }
-.model-switcher { display:grid; grid-template-columns:repeat(3,1fr); gap:12px; }
-.model-pill { padding:14px; border:1px solid var(--line); border-radius:18px; display:flex; flex-direction:column; gap:4px; min-height:98px; background:#ffffff; }
-.model-pill.active { border-color:rgba(255,122,26,.40); background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,255,255,.96)); }
-.model-name { font-weight:650; color:var(--text); }
-.model-tag { font-size:.76rem; text-transform:uppercase; letter-spacing:.12em; color:var(--gold); }
-.model-pill small { color:var(--muted); line-height:1.45; }
-.brain-shell { padding:18px; }
-.brain-header { display:flex; justify-content:space-between; align-items:flex-end; gap:16px; margin-bottom:10px; }
-.eyebrow { font-size:.72rem; letter-spacing:.16em; text-transform:uppercase; color:var(--gold); margin:0 0 4px; }
-.brain-header h3 { margin:0; font-size:1.12rem; color:var(--text); }
-.brain-legend { display:flex; gap:14px; color:var(--muted); font-size:.8rem; flex-wrap:wrap; }
-.dot { width:10px; height:10px; display:inline-block; border-radius:50%; margin-right:6px; }
-.dot-live { background:var(--chosen); box-shadow:0 0 10px rgba(255,122,26,.35); }
-.dot-chosen { background:var(--chosen); }
-.dot-idle { background:var(--idle); border:1px solid var(--idle-stroke); }
-.dot-query { background:var(--query-node); border:1px solid #de9e58; }
-.dot-paper { background:var(--paper-node); border:1px solid #4fb3a5; }
-.dot-upload { background:var(--upload-node); border:1px solid #8f73d9; }
-.brain-stage { position:relative; min-height:420px; overflow:hidden; background:linear-gradient(180deg,rgba(250,250,250,1),rgba(255,255,255,1)); border:1px solid rgba(0,0,0,.05); border-radius:20px; }
-.brain-svg { width:100%; height:520px; display:block; }
-.edge { stroke:rgba(0,0,0,.12); stroke-width:2.4; }
-.edge.active { stroke:var(--chosen); stroke-width:4.2; stroke-linecap:round; filter:drop-shadow(0 0 6px rgba(255,122,26,.45)); stroke-dasharray:8 12; animation:pulseEdge 1.5s linear infinite; }
-.node { stroke-width:2.2; transition:all .25s ease; }
-.node.idle { fill:var(--idle); stroke:var(--idle-stroke); }
-.node.chosen { fill:var(--chosen); stroke:#ffb16d; }
-.halo { fill:none; }
-.halo.active { stroke:rgba(255,122,26,.18); stroke-width:12; }
-.label { fill:#2c2c2c; font-size:13px; font-weight:500; letter-spacing:.01em; }
-.label.active { fill:#111111; font-weight:700; }
-.learn-edge { stroke:rgba(0,0,0,.18); stroke-width:2.2; stroke-linecap:round; }
-.learn-node { stroke-width:2.2; }
-.learn-node.query { fill:var(--query-node); stroke:#de9e58; }
-.learn-node.paper { fill:var(--paper-node); stroke:#36a091; }
-.learn-node.upload { fill:var(--upload-node); stroke:#7e63cb; }
-.learn-label { fill:#1e1e1e; font-size:12px; font-weight:600; }
-.learning-empty { display:grid; place-items:center; }
-.empty-graph-copy { text-align:center; max-width:440px; padding:40px 20px; }
-.empty-graph-copy h4 { margin:0 0 10px; font-size:1.05rem; }
-.empty-graph-copy p { margin:0; color:var(--muted); line-height:1.6; }
-.timeline { display:flex; flex-direction:column; gap:10px; }
-.agent-step { border:1px solid var(--line); border-radius:18px; background:#ffffff; overflow:hidden; }
-.agent-summary { list-style:none; display:grid; grid-template-columns:42px 1fr; gap:12px; align-items:center; padding:12px; cursor:pointer; }
-.agent-summary::-webkit-details-marker { display:none; }
-.agent-index { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; font-weight:700; color:var(--gold); background:rgba(255,122,26,.08); border:1px solid rgba(255,122,26,.18); }
-.agent-head { display:flex; justify-content:space-between; gap:12px; align-items:center; }
-.agent-head h4 { margin:0; font-size:.98rem; color:var(--text); }
-.agent-head span { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
-.agent-copy { padding:0 14px 16px 66px; }
-.agent-copy p { margin:0; color:#2d2d2d; font-size:.93rem; line-height:1.6; }
-.candidate-grid { display:grid; grid-template-columns:repeat(3,minmax(0,1fr)); gap:18px; }
-.candidate-card { background:none; perspective:1400px; min-height:330px; }
-.candidate-card-inner { position:relative; width:100%; min-height:330px; transition:transform .8s cubic-bezier(.2,.7,.1,1); transform-style:preserve-3d; }
-.candidate-card:hover .candidate-card-inner,.candidate-card:focus .candidate-card-inner,.candidate-card:focus-within .candidate-card-inner { transform:rotateY(180deg); }
-.candidate-face { position:absolute; inset:0; padding:20px; border-radius:22px; border:1px solid var(--line); background:#ffffff; color:var(--text); backface-visibility:hidden; box-shadow:0 12px 24px rgba(0,0,0,.06); display:flex; flex-direction:column; gap:14px; }
-.candidate-back { transform:rotateY(180deg); }
-.candidate-top { display:flex; justify-content:space-between; align-items:center; gap:8px; }
-.chip { font-size:.72rem; text-transform:uppercase; letter-spacing:.12em; color:#0b6f66; padding:7px 10px; border-radius:999px; background:rgba(23,184,166,.08); border:1px solid rgba(23,184,166,.18); }
-.chip.alt { color:var(--gold); background:rgba(255,122,26,.08); border-color:rgba(255,122,26,.18); }
-.score { font-weight:700; color:var(--gold); }
-.candidate-face h4 { margin:0; font-size:1.08rem; line-height:1.35; }
-.candidate-face p { margin:0; color:#1e1e1e; line-height:1.65; font-size:.96rem; overflow-wrap:anywhere; }
-.meta-row { margin-top:auto; display:flex; justify-content:space-between; color:var(--muted); font-size:.88rem; gap:14px; }
-.mini { cursor:pointer; margin-top:8px; align-self:flex-start; color:var(--text); padding:10px 12px; border-radius:14px; border:1px solid var(--line); background:#ffffff; transition:all 0.2s; }
-.mini:hover { background: #f5f5f5; border-color: var(--chosen); }
-.papers-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
-.paper-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
-.paper-topline { display:flex; gap:8px; flex-wrap:wrap; margin-bottom:10px; }
-.paper-badge { font-size:.72rem; padding:6px 10px; border-radius:999px; background:rgba(98,141,255,.08); color:#3456b5; border:1px solid rgba(98,141,255,.18); }
-.paper-badge.alt { background:rgba(0,0,0,.04); color:#444; border-color:rgba(0,0,0,.08); }
-.doi-badge { background:rgba(255,122,26,.08); color:#8a4105; border-color:rgba(255,122,26,.18); }
-.paper-card h4 { margin:0 0 10px; line-height:1.35; font-size:1rem; }
-.paper-card p { margin:0 0 12px; line-height:1.6; color:#222; }
-.paper-links { display:flex; gap:12px; flex-wrap:wrap; }
-.paper-meta-stack { display:flex; flex-direction:column; gap:6px; color:#444; margin-bottom:12px; font-size:.9rem; }
-.paper-links a,.journal-card,.upload-note a { color:#0b63ce; text-decoration:none; }
-.journal-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
-.journal-card { border:1px solid var(--line); border-radius:18px; padding:16px; display:flex; justify-content:space-between; gap:14px; align-items:center; background:#ffffff; }
-.journal-card h4 { margin:0 0 6px; }
-.journal-card p { margin:0; color:var(--muted); line-height:1.5; }
-.upload-note { border:1px dashed rgba(0,0,0,.16); border-radius:18px; padding:16px; background:rgba(0,0,0,.015); color:#1f1f1f; line-height:1.6; }
-.prosebox { padding:18px; white-space:pre-wrap; font-family:ui-monospace,SFMono-Regular,Menlo,monospace; line-height:1.55; color:#1b1b1b; }
-.gr-button-primary { background:linear-gradient(135deg,rgba(255,122,26,.92),rgba(240,108,22,.92)) !important; color:#ffffff !important; border:none !important; }
-.gr-button-secondary { background:#ffffff !important; color:var(--text) !important; border:1px solid var(--line) !important; }
-.ref-list { margin:0; padding-left:18px; }
-.ref-list li { margin-bottom:8px; line-height:1.5; }
-.parse-grid { display:grid; grid-template-columns:1.2fr 1fr; gap:14px; }
-.parse-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
-.selection-panel { padding:18px; }
-footer { display:none !important; }
-@keyframes pulseEdge { to { stroke-dashoffset:-40; } }
-@media (max-width:1180px) {
-  .model-switcher,.candidate-grid,.papers-grid,.journal-grid,.parse-grid { grid-template-columns:1fr; }
-  .brain-svg { height:460px; }
-}
-"""
-
-CSS  = BASE_CSS + "\n" + get_dvnc_layout_css()
-HEAD = """
-<link rel="preconnect" href="https://fonts.googleapis.com">
-<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
-<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap" rel="stylesheet">
-<script>
-function triggerRouteSwap(idx) {
-    const container = document.getElementById('route_swap_payload');
-    if(!container) return;
-    const input = container.querySelector('textarea') || container.querySelector('input');
-    if(input) {
-        input.value = idx.toString();
-        input.dispatchEvent(new Event('input', { bubbles: true }));
-        setTimeout(() => {
-            const btn = document.getElementById('route_swap_apply');
-            if(btn) btn.click();
-        }, 150);
-    }
-}
-</script>
-"""
-
-# ── Gradio layout ─────────────────────────────────────────────────────────────
-with gr.Blocks(fill_height=True) as demo:
-
-    # ── Shared state ──────────────────────────────────────────────────────────
-    papers_state        = gr.State([])
-    parsed_pdf_state    = gr.State({})
-    ingest_payload_state = gr.State({})
-    route_state         = gr.State(get_default_route_state())
-
-    # ── Header ────────────────────────────────────────────────────────────────
-    gr.HTML("""
-        <div id="dvnc-shell">
-          <div class="hero-bar">
-            <div class="brand">
-              <div class="logo" aria-hidden="true">
-                <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7">
-                  <path d="M5 17L12 4l7 13"/>
-                  <path d="M8.5 12.5h7"/>
-                  <circle cx="12" cy="12" r="1.8" fill="currentColor" stroke="none"/>
-                </svg>
-              </div>
-              <div>
-                <h1>DVNC.AI</h1>
-                <p>Sovereign discovery instrument · connectome-native reasoning</p>
-              </div>
-            </div>
-            <div class="status"><span class="status-dot"></span><span>Live orchestration</span></div>
-          </div>
-        </div>
-    """)
-
-    with gr.Tabs():
-
-        # ── Tab 1 · Discovery Engine ──────────────────────────────────────────
-        with gr.Tab("Discovery Engine"):
-            model_html = gr.HTML(build_models_html("DVNC Sovereign"))
-
-            with gr.Row():
-                with gr.Column(scale=2):
-                    model = gr.Dropdown(
-                        choices=[m["name"] for m in MODELS],
-                        value="DVNC Sovereign",
-                        label="Model tier",
-                    )
-                    query = gr.Textbox(
-                        label="Discovery query",
-                        elem_classes=["querybox"],
-                        placeholder="Enter a scientific question, anomaly, or breakthrough direction…",
-                        lines=4,
-                    )
-                    with gr.Row():
-                        run_btn     = gr.Button("Run discovery",  variant="primary")
-                        example_btn = gr.Button("Load example",   variant="secondary")
-                    chat = gr.HTML("""
-                        <div class="panel chat-panel">
-                          <div class="chat-thread">
-                            <div class="bubble bubble-ai">
-                              <span class="role">DVNC</span>
-                              <p>Enter a query to activate the 7-agent discovery stack and illuminate the chosen path through the 3D connectome.</p>
-                            </div>
-                          </div>
-                        </div>
-                    """)
-
-                with gr.Column(scale=3):
-                    connectome = gr.HTML(build_connectome_html(DEFAULT_PATH))
-                    cards      = gr.HTML("")
-
-            output   = gr.Markdown("# Discovery Output\n\nAwaiting query.")
-            timeline = gr.HTML(get_initial_discovery_timeline_html())
-
-            route_swap_payload = gr.Textbox(value="", visible=False, elem_id="route_swap_payload")
-            route_swap_apply   = gr.Button("Apply route swap", visible=False, elem_id="route_swap_apply")
-
-        # ── Tab 2 · Self-Learning Graph ───────────────────────────────────────
-        with gr.Tab("Self-Learning Graph"):
-            with gr.Row():
-                with gr.Column(scale=2):
-                    paper_query = gr.Textbox(
-                        label="Research topic / title / DOI / link",
-                        elem_classes=["querybox"],
-                        placeholder="e.g. self-assembling conductive biomaterials for cardiac repair",
-                        lines=3,
-                    )
-                    search_mode = gr.Dropdown(
-                        choices=SEARCH_MODES,
-                        value="topic",
-                        label="Search mode",
-                    )
-                    source_selector = gr.CheckboxGroup(
-                        choices=SOURCE_OPTIONS,
-                        value=DEFAULT_SOURCES,
-                        label="Sources",
-                    )
-                    pdf_upload = gr.File(label="Upload PDF papers", file_types=[".pdf"], file_count="single")
-
-                    with gr.Row():
-                        learn_btn      = gr.Button("Discover papers",      variant="primary")
-                        load_topic_btn = gr.Button("Load example topic",   variant="secondary")
-
-                    upload_status    = gr.Markdown("No PDF uploaded yet.")
-                    discovery_status = gr.Markdown("### No discovery results yet.")
-                    journal_panel    = gr.HTML(build_journal_html("biomaterials cardiac repair"))
-
-                    gr.HTML('<div class="panel selection-panel"><h3 style="margin:0 0 12px;">Select papers to ingest</h3></div>')
-                    selection_box = gr.CheckboxGroup(choices=[], value=[], label="Candidate papers")
-
-                    parser_order = gr.CheckboxGroup(
-                        choices=["grobid", "docling", "pymupdf"],
-                        value=["grobid", "docling", "pymupdf"],
-                        label="Parser routing order",
-                    )
-
-                    with gr.Row():
-                        parse_btn  = gr.Button("Parse uploaded PDF",        variant="secondary")
-                        ingest_btn = gr.Button("Ingest selected into graph", variant="primary")
-
-                with gr.Column(scale=3):
-                    learning_graph = gr.HTML(build_learning_graph_html([], []))
-                    papers_panel   = gr.HTML('<div class="panel papers-panel" style="padding:18px"><p>Search by topic, title, DOI, or link, then select papers before graph ingestion.</p></div>')
-                    parse_summary  = gr.Markdown("### PDF parse status\n\nAwaiting upload.")
-                    parse_panel    = gr.HTML('<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>')
-                    ingest_summary = gr.Markdown("### Graph ingest status\n\nAwaiting paper selection.")
-                    ingest_payload = gr.JSON(label="Graph ingest payload", value={"status": "empty", "nodes": [], "edges": []})
-                    graph_html = render_graph_canvas_html({"status": "ok", "nodes": NODES, "edges": EDGES},title="Some title",height=780,)
-    # ── Event wiring ────────────────────────────────────────────────────���─────
-    example_btn.click(fn=load_example, outputs=query)
-
-    run_btn.click(
-        fn=run_discovery,
-        inputs=[query, model],
-        outputs=[chat, connectome, timeline, cards, output, model_html, route_state],
-    )
-
-    route_swap_apply.click(
-        fn=apply_route_swap,
-        inputs=[query, model, route_swap_payload, route_state],
-        outputs=[chat, connectome, timeline, output, route_state],
-    )
-
-    load_topic_btn.click(fn=load_paper_topic, outputs=paper_query)
-
-    learn_btn.click(
-        fn=run_paper_discovery,
-        inputs=[paper_query, search_mode, source_selector, pdf_upload],
-        outputs=[learning_graph, papers_panel, journal_panel, upload_status, selection_box, papers_state, discovery_status],
-    )
-
-    parse_btn.click(
-        fn=parse_uploaded_pdf,
-        inputs=[pdf_upload, parser_order],
-        outputs=[parse_summary, parsed_pdf_state],
-    ).then(
-        fn=render_parse_result,
-        inputs=[parsed_pdf_state],
-        outputs=[parse_panel],
-    )
-
-    ingest_btn.click(
-        fn=ingest_selected_papers,
-        inputs=[paper_query, selection_box, papers_state, pdf_upload, parsed_pdf_state],
-        outputs=[learning_graph, ingest_summary, ingest_payload],
-    )
-
-if __name__ == "__main__":
-    demo.launch(css=CSS, head=HEAD, theme=gr.themes.Base())

app_old3.py

@@ -1,782 +0,0 @@
-"""
-DVNC.AI — root app.py
-Refactored to fix Gradio runtime startup and preserve current
-dvncaiv2hf self-learning graph integration.
-"""
-
-# ── Standard library ────────────────────────────────────────────────────────
-import html
-import random
-import re
-from typing import Dict, List, Optional
-
-# ── Third-party ─────────────────────────────────────────────────────────────
-import gradio as gr
-
-# ── Internal modules ────────────────────────────────────────────────────────
-from dvncaiv2hf.agentroutecards import buildagentroutecardshtml
-from dvncaiv2hf.discoveryappbridge import (
-    getdefaultroutestate,
-    getdiscoverycss,
-    getinitialdiscoverytimelinehtml,
-)
-from dvncaiv2hf.dvncuilayout import getdvnclayoutcss
-from dvncaiv2hf.graphcanvaspatch import rendergraphcanvashtml
-from dvncaiv2hf.selflearninggraph import (
-    DEFAULTSOURCES,
-    SEARCHMODES,
-    SOURCEOPTIONS,
-    buildjournalhtml,
-    ingestselectedpapers,
-    parseuploadedpdf,
-    renderparseresult,
-    runpaperdiscovery,
-    safetext,
-)
-
-# ── Constants ───────────────────────────────────────────────────────────────
-MODELS = [
-    {
-        "name": "DVNC Sovereign",
-        "tag": "flagship",
-        "desc": "Maximum depth orchestration for frontier discovery",
-    },
-    {
-        "name": "DVNC Atlas",
-        "tag": "research",
-        "desc": "Balanced reasoning, graph traversal, and synthesis",
-    },
-    {
-        "name": "DVNC Curie",
-        "tag": "lab",
-        "desc": "Experimental hypothesis generation for anomalous signals",
-    },
-]
-
-AGENTS = [
-    "Query Interpreter",
-    "Graph Divergence Mapper",
-    "Evidence Harvester",
-    "Analogy Engine",
-    "Hypothesis Composer",
-    "Adversarial Critic",
-    "Experimental Program Designer",
-]
-
-NODES = [
-    {"id": "seed", "label": "Seed Query", "group": "core", "x": 10, "y": 0, "z": 0},
-    {"id": "bio", "label": "Biomaterials", "group": "domain", "x": 24, "y": 12, "z": -8},
-    {"id": "card", "label": "Cardiac Repair", "group": "domain", "x": 38, "y": 3, "z": 14},
-    {"id": "nano", "label": "Nanostructure", "group": "bridge", "x": 24, "y": -18, "z": 16},
-    {"id": "selfasm", "label": "Self-Assembly", "group": "bridge", "x": 40, "y": -16, "z": -16},
-    {"id": "electro", "label": "Electro-signalling", "group": "mechanism", "x": 58, "y": 10, "z": -10},
-    {"id": "immune", "label": "Immune Modulation", "group": "mechanism", "x": 64, "y": -8, "z": 10},
-    {"id": "trial", "label": "Validation Path", "group": "outcome", "x": 80, "y": 0, "z": 0},
-    {"id": "alt1", "label": "Piezoelectric Scaffold", "group": "candidate", "x": 56, "y": 26, "z": 14},
-    {"id": "alt2", "label": "Peptide Mesh", "group": "candidate", "x": 54, "y": -27, "z": -14},
-]
-
-EDGES = [
-    ("seed", "bio"),
-    ("seed", "nano"),
-    ("bio", "card"),
-    ("nano", "selfasm"),
-    ("selfasm", "electro"),
-    ("card", "immune"),
-    ("electro", "trial"),
-    ("immune", "trial"),
-    ("card", "alt1"),
-    ("selfasm", "alt2"),
-    ("alt1", "trial"),
-    ("alt2", "trial"),
-]
-
-DEFAULTPATH = ["seed", "nano", "selfasm", "electro", "trial"]
-
-CANDIDATES = [
-    {
-        "title": "Piezoelectric Scaffold Cascade",
-        "front": "Use mechano-electric scaffolds to convert cardiac strain into micro-current signalling.",
-        "back": "Discovery path: anomalous healing signal -> piezoelectric analog -> ion-channel entrainment -> tissue regeneration. Risk: power density and fibrosis coupling.",
-        "score": 92,
-        "novelty": "High",
-        "agent": "Hypothesis Composer",
-    },
-    {
-        "title": "Peptide Self-Assembly Mesh",
-        "front": "Deploy dynamic peptide meshes that self-assemble around damaged myocardium and guide repair.",
-        "back": "Discovery path: self-assembly -> local immune choreography -> regenerative substrate formation. Risk: degradation timing and targeting specificity.",
-        "score": 88,
-        "novelty": "High",
-        "agent": "Analogy Engine",
-    },
-    {
-        "title": "Immune-Tuned Conductive Hydrogel",
-        "front": "Blend conductivity with macrophage-state modulation to reduce scarring and restore conduction.",
-        "back": "Discovery path: inflammation mismatch -> conductive medium -> macrophage polarization -> synchronized healing. Risk: persistence and biocompatibility.",
-        "score": 85,
-        "novelty": "Medium-High",
-        "agent": "Adversarial Critic",
-    },
-]
-
-ACADEMICINSIGHTS = [
-    {
-        "hypothesis": "Implementation of mechano-electric scaffolds to transduce cardiac strain into localized micro-current signalling for myocardial regeneration.",
-        "metrics": {
-            "Novelty": 92,
-            "Mechanistic clarity": 85,
-            "Experimental tractability": 78,
-            "Cross-domain distance": 94,
-        },
-        "outline": (
-            "1. Synthesize candidate piezoelectric biomaterial scaffolds with tunable strain-electric coupling.\n"
-            "2. Evaluate in vitro electromechanical transduction and subsequent ion-channel entrainment.\n"
-            "3. Conduct in vivo comparative models to assess regenerative efficacy against gold-standard substrates.\n"
-            "4. Rigorously validate to exclude pathological fibrosis and power-density toxicity."
-        ),
-        "path": ["seed", "bio", "card", "alt1", "trial"],
-    },
-    {
-        "hypothesis": "Deployment of dynamic peptide networks that self-assemble post-infarction to orchestrate local immunological responses and guide substrate regeneration.",
-        "metrics": {
-            "Novelty": 88,
-            "Mechanistic clarity": 82,
-            "Experimental tractability": 86,
-            "Cross-domain distance": 85,
-        },
-        "outline": (
-            "1. Formulate peptide sequences programmed for triggered in situ self-assembly within the myocardial infarct zone.\n"
-            "2. Quantify macrophage polarization and local immune choreography post-deployment.\n"
-            "3. Map the temporospatial degradation profile against de novo tissue formation.\n"
-            "4. Falsify against off-target aggregation and delayed clearance risks."
-        ),
-        "path": ["seed", "nano", "selfasm", "alt2", "trial"],
-    },
-    {
-        "hypothesis": "Integration of conductive hydrogels with immunomodulatory properties to simultaneously bridge electrical uncoupling and mitigate adverse fibrotic scarring.",
-        "metrics": {
-            "Novelty": 85,
-            "Mechanistic clarity": 90,
-            "Experimental tractability": 88,
-            "Cross-domain distance": 79,
-        },
-        "outline": (
-            "1. Fabricate biocompatible hydrogels featuring precisely tuned electrical conductivity and immunomodulatory motifs.\n"
-            "2. Monitor electrophysiological synchronization across the scaffold-tissue interface.\n"
-            "3. Assess macrophage state transitions and suppression of adverse fibrotic remodelling.\n"
-            "4. Validate long-term persistence, hemocompatibility, and mechanical integration."
-        ),
-        "path": ["seed", "bio", "card", "immune", "trial"],
-    },
-]
-
-
-# ── Utility helpers ─────────────────────────────────────────────────────────
-def normtext(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def buildlearninggraphhtml(nodes, edges, title="Self-Learning Knowledge Graph"):
-    return rendergraphcanvashtml(
-        {
-            "status": "ok" if nodes or edges else "empty",
-            "nodes": nodes or [],
-            "edges": edges or [],
-        },
-        title=title,
-        height=780,
-    )
-
-
-# ── HTML builders ───────────────────────────────────────────────────────────
-def buildconnectomehtml(pathids: List[str]) -> str:
-    active = set(pathids)
-    nodemap = {n["id"]: n for n in NODES}
-    pathpairs = {
-        pair
-        for i in range(len(pathids) - 1)
-        for pair in [(pathids[i], pathids[i + 1]), (pathids[i + 1], pathids[i])]
-    }
-
-    baselines, activelines, circles, labels = [], [], [], []
-
-    for a, b in EDGES:
-        na, nb = nodemap[a], nodemap[b]
-        x1, y1 = na["x"] * 8 + 80, na["y"] * 6 + 280
-        x2, y2 = nb["x"] * 8 + 80, nb["y"] * 6 + 280
-        baselines.append(
-            f'<line class="edge" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />'
-        )
-        if (a, b) in pathpairs:
-            activelines.append(
-                f'<line class="edge active" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />'
-            )
-
-    for n in NODES:
-        cx, cy = n["x"] * 8 + 80, n["y"] * 6 + 280
-        isactive = n["id"] in active
-        state = "chosen" if isactive else "idle"
-        halocls = "halo active" if isactive else "halo"
-        lblcls = "label active" if isactive else "label"
-        radius = 18 if isactive else 13
-        halor = 30 if isactive else 0
-
-        circles.append(
-            f'<g class="node-wrap">'
-            f'<circle class="{halocls}" cx="{cx:.1f}" cy="{cy:.1f}" r="{halor}" />'
-            f'<circle class="node {n["group"]} {state}" cx="{cx:.1f}" cy="{cy:.1f}" r="{radius}" />'
-            f"</g>"
-        )
-        labels.append(
-            f'<text class="{lblcls}" x="{cx + 18:.1f}" y="{cy - 16:.1f}">{safe_text(n["label"])}</text>'
-        )
-
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Connectome</p>
-          <h3>3D Connectome</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-live"></i> lit path</span>
-          <span><i class="dot dot-chosen"></i> chosen node</span>
-          <span><i class="dot dot-idle"></i> available node</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 780 560" class="brain-svg" role="img" aria-label="DVNC 3D connectome visualisation">
-          {''.join(baselines)}
-          {''.join(activelines)}
-          {''.join(circles)}
-          {''.join(labels)}
-        </svg>
-      </div>
-    </div>
-    """
-
-
-def buildcardshtml(cards: List[Dict]) -> str:
-    items = []
-    for i, c in enumerate(cards):
-        items.append(
-            f"""
-            <article class="candidate-card" tabindex="0">
-              <div class="candidate-card-inner">
-                <div class="candidate-face candidate-front">
-                  <div class="candidate-top">
-                    <span class="chip">{safe_text(c["agent"])}</span>
-                    <span class="score">{safe_text(c["score"])}</span>
-                  </div>
-                  <h4>{safe_text(c["title"])}</h4>
-                  <p>{safe_text(c["front"])}</p>
-                  <div class="meta-row"><span>Novelty</span><strong>{safe_text(c["novelty"])}</strong></div>
-                  <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-                </div>
-                <div class="candidate-face candidate-back">
-                  <div class="candidate-top">
-                    <span class="chip alt">Alternative path</span>
-                    <span class="score">{safe_text(c["score"])}</span>
-                  </div>
-                  <h4>{safe_text(c["title"])}</h4>
-                  <p>{safe_text(c["back"])}</p>
-                  <div class="meta-row"><span>Swap into route</span><strong>Enabled</strong></div>
-                  <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-                </div>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="panel" style="padding:20px;"><div class="candidate-grid">' + "".join(items) + "</div></div>"
-
-
-def buildchathtml(query: str, result: Dict) -> str:
-    return f"""
-    <div class="panel chat-panel">
-      <div class="chat-thread">
-        <div class="bubble bubble-user">
-          <span class="role">You</span>
-          <p>{safe_text(query)}</p>
-        </div>
-        <div class="bubble bubble-ai">
-          <span class="role">DVNC Sovereign</span>
-          <p>{safe_text(result["summary"])}</p>
-        </div>
-        <div class="bubble bubble-system">
-          <span class="role">Discovery Signal</span>
-          <p><strong>Primary hypothesis:</strong> {safe_text(result["primary_hypothesis"])}</p>
-        </div>
-      </div>
-    </div>
-    """
-
-
-def buildmodelshtml(selected: str) -> str:
-    items = []
-    for m in MODELS:
-        active = "active" if m["name"] == selected else ""
-        items.append(
-            f"""
-            <div class="model-pill {active}">
-              <span class="model-name">{safe_text(m["name"])}</span>
-              <span class="model-tag">{safe_text(m["tag"])}</span>
-              <small>{safe_text(m["desc"])}</small>
-            </div>
-            """
-        )
-    return '<div class="panel" style="padding:18px;"><div class="model-switcher">' + "".join(items) + "</div></div>"
-
-
-# ── Discovery logic ─────────────────────────────────────────────────────────
-def rundiscovery(query: str, modelname: str):
-    random.seed(len(query or "") + len(modelname or ""))
-
-    if "curie" in (query or "").lower() or "einstein" in (query or "").lower():
-        primary = "Map the anomaly first, then force a distant analogy before composing the experimental programme."
-        path = ["seed", "bio", "card", "immune", "trial"]
-    else:
-        primary = (
-            "Utilization of a self-assembling conductive scaffold to transduce mechanical "
-            "strain into localized regenerative signalling pathways."
-        )
-        path = DEFAULTPATH
-
-    summaries = [
-        "Normalises the user prompt into a graph-searchable seed and isolates the tension inside the question.",
-        "Finds remote conceptual bridges instead of staying near the starting domain cluster.",
-        "Pulls evidence packets and conflict signals required for grounded hypothesis formation.",
-        "Generates cross-domain analogies with a bias toward mechanism transfer rather than keyword similarity.",
-        "Composes the lead hypothesis and two structurally different variants.",
-        "Attacks weak assumptions, hidden confounders, and feasibility gaps.",
-        "Produces a staged validation plan with measurable falsification criteria.",
-    ]
-    tags = ["input", "graph", "evidence", "analogy", "compose", "critique", "experiment"]
-    reasoning = [
-        {"step": i + 1, "agent": AGENTS[i], "tag": tags[i], "summary": summaries[i]}
-        for i in range(7)
-    ]
-
-    result = {
-        "summary": (
-            "A deeper route was chosen through the connectome, with live alternatives preserved "
-            "as swappable cards so the reasoning path can be inspected rather than hidden."
-        ),
-        "primary_hypothesis": primary,
-        "reasoning": reasoning,
-        "cards": CANDIDATES,
-        "path": path,
-        "metrics": {
-            "Novelty": 93,
-            "Mechanistic clarity": 89,
-            "Experimental tractability": 82,
-            "Cross-domain distance": 91,
-        },
-    }
-
-    chathtml = buildchathtml(query, result)
-    connectomehtml = buildconnectomehtml(path)
-    timelinehtml = buildagentroutecardshtml(reasoning)
-    metricsmd = "\n".join(f"- {k}: {v}/100" for k, v in result["metrics"].items())
-
-    hypothesismd = (
-        "# Discovery Output\n\n"
-        f"**Model:** {modelname}\n\n"
-        f"**Primary hypothesis:** {result['primary_hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metricsmd}\n\n"
-        "## Experimental outline\n"
-        "1. Construct the candidate material or protocol.\n"
-        "2. Test mechanistic signal expression under controlled conditions.\n"
-        "3. Compare against baseline and nearest-neighbour alternatives.\n"
-        "4. Falsify using the adversarial risk criteria surfaced in the reasoning path.\n"
-    )
-
-    cardshtml = buildcardshtml(CANDIDATES)
-    routestate = getdefaultroutestate()
-    return (
-        chathtml,
-        connectomehtml,
-        timelinehtml,
-        cardshtml,
-        hypothesismd,
-        buildmodelshtml(modelname),
-        routestate,
-    )
-
-
-def applyrouteswap(query: str, modelname: str, routeswappayload: str, routestate):
-    try:
-        idx = int(routeswappayload)
-    except Exception:
-        idx = 0
-
-    if not (0 <= idx < len(ACADEMICINSIGHTS)):
-        idx = 0
-
-    academic = ACADEMICINSIGHTS[idx]
-    connectomehtml = buildconnectomehtml(academic["path"])
-
-    result = {
-        "summary": (
-            "Main insight formally adopted. The connectome pathway and validation protocol "
-            "have been realigned to the selected candidate methodology."
-        ),
-        "primary_hypothesis": academic["hypothesis"],
-    }
-    chathtml = buildchathtml(query, result)
-
-    metricsmd = "\n".join(f"- {k}: {v}/100" for k, v in academic["metrics"].items())
-    hypothesismd = (
-        "# Discovery Output\n\n"
-        f"**Model:** {modelname}\n\n"
-        f"**Primary hypothesis:** {academic['hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metricsmd}\n\n"
-        "## Experimental outline\n"
-        f"{academic['outline']}\n"
-    )
-
-    return chathtml, connectomehtml, gr.update(), hypothesismd, routestate
-
-
-# ── Example loaders ────────────────────────────────────────────────────────
-def loadexample() -> str:
-    return "How could a self-assembling conductive biomaterial improve cardiac tissue regeneration by converting mechanical strain into repair signalling?"
-
-
-def loadpapertopic() -> str:
-    return "self-assembling conductive biomaterials for cardiac repair"
-
-
-# ── CSS / HEAD ──────────────────────────────────────────────────────────────
-BASECSS = r"""
-:root {
-  --bg: #ffffff; --panel: #ffffff; --line: rgba(0,0,0,.12);
-  --text: #111111; --muted: #5b5b5b; --soft: rgba(0,0,0,.62);
-  --gold: #ff6600; --teal: #17b8a6; --blue: #628dff;
-  --chosen: #ff7a1a; --idle: #b8d8ff; --idle-stroke: #5e8fe6;
-  --query-node: #ffd8b3; --paper-node: #d7f6f2; --upload-node: #e7defe;
-  --shadow: 0 16px 40px rgba(0,0,0,.12);
-}
-html,body,.gradio-container { background:#ffffff !important; font-family:Inter,ui-sans-serif,system-ui,sans-serif; }
-.gradio-container { max-width:1640px !important; padding:20px !important; }
-#dvnc-shell { border:1px solid var(--line); border-radius:28px; overflow:hidden; background:#ffffff; box-shadow:var(--shadow); padding:20px 22px 22px; }
-.hero-bar { display:flex; justify-content:space-between; align-items:center; gap:16px; padding-bottom:12px; border-bottom:1px solid rgba(0,0,0,.06); margin-bottom:16px; }
-.brand { display:flex; align-items:center; gap:14px; }
-.logo { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; color:var(--gold); background:linear-gradient(135deg,rgba(255,122,26,.12),rgba(23,184,166,.10)); border:1px solid rgba(0,0,0,.08); }
-.logo svg { width:24px; height:24px; }
-.brand h1 { font-size:1.05rem; margin:0; font-weight:700; letter-spacing:.12em; text-transform:uppercase; }
-.brand p { margin:3px 0 0; color:var(--muted); font-size:.84rem; }
-.status { display:flex; gap:10px; align-items:center; color:var(--soft); font-size:.85rem; }
-.status-dot { width:10px; height:10px; border-radius:50%; background:var(--teal); box-shadow:0 0 0 6px rgba(23,184,166,.10),0 0 14px rgba(23,184,166,.25); }
-.panel { background:#ffffff; border:1px solid var(--line); border-radius:22px; box-shadow:inset 0 1px 0 rgba(255,255,255,.8); }
-.querybox textarea,.querybox input { background:transparent !important; color:var(--text) !important; }
-.querybox,.querybox>div { background:#ffffff !important; border-radius:18px !important; border-color:var(--line) !important; }
-.chat-panel { padding:18px; min-height:280px; }
-.chat-thread { display:flex; flex-direction:column; gap:14px; }
-.bubble { max-width:88%; padding:16px 18px; border-radius:22px; border:1px solid var(--line); }
-.bubble p { margin:8px 0 0; line-height:1.6; font-size:.96rem; color:var(--text); }
-.bubble .role { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
-.bubble-user { align-self:flex-end; background:linear-gradient(135deg,rgba(98,141,255,.16),rgba(98,141,255,.08)); }
-.bubble-ai { align-self:flex-start; background:#ffffff; }
-.bubble-system { align-self:flex-start; background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,122,26,.04)); }
-.model-switcher { display:grid; grid-template-columns:repeat(3,1fr); gap:12px; }
-.model-pill { padding:14px; border:1px solid var(--line); border-radius:18px; display:flex; flex-direction:column; gap:4px; min-height:98px; background:#ffffff; }
-.model-pill.active { border-color:rgba(255,122,26,.40); background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,255,255,.96)); }
-.model-name { font-weight:650; color:var(--text); }
-.model-tag { font-size:.76rem; text-transform:uppercase; letter-spacing:.12em; color:var(--gold); }
-.model-pill small { color:var(--muted); line-height:1.45; }
-.brain-shell { padding:18px; }
-.brain-header { display:flex; justify-content:space-between; align-items:flex-end; gap:16px; margin-bottom:10px; }
-.eyebrow { font-size:.72rem; letter-spacing:.16em; text-transform:uppercase; color:var(--gold); margin:0 0 4px; }
-.brain-header h3 { margin:0; font-size:1.12rem; color:var(--text); }
-.brain-legend { display:flex; gap:14px; color:var(--muted); font-size:.8rem; flex-wrap:wrap; }
-.dot { width:10px; height:10px; display:inline-block; border-radius:50%; margin-right:6px; }
-.dot-live { background:var(--chosen); box-shadow:0 0 10px rgba(255,122,26,.35); }
-.dot-chosen { background:var(--chosen); }
-.dot-idle { background:var(--idle); border:1px solid var(--idle-stroke); }
-.dot-query { background:var(--query-node); border:1px solid #de9e58; }
-.dot-paper { background:var(--paper-node); border:1px solid #4fb3a5; }
-.dot-upload { background:var(--upload-node); border:1px solid #8f73d9; }
-.brain-stage { position:relative; min-height:420px; overflow:hidden; background:linear-gradient(180deg,rgba(250,250,250,1),rgba(255,255,255,1)); border:1px solid rgba(0,0,0,.05); border-radius:20px; }
-.brain-svg { width:100%; height:520px; display:block; }
-.edge { stroke:rgba(0,0,0,.12); stroke-width:2.4; }
-.edge.active { stroke:var(--chosen); stroke-width:4.2; stroke-linecap:round; filter:drop-shadow(0 0 6px rgba(255,122,26,.45)); stroke-dasharray:8 12; animation:pulseEdge 1.5s linear infinite; }
-.node { stroke-width:2.2; transition:all .25s ease; }
-.node.idle { fill:var(--idle); stroke:var(--idle-stroke); }
-.node.chosen { fill:var(--chosen); stroke:#ffb16d; }
-.halo { fill:none; }
-.halo.active { stroke:rgba(255,122,26,.18); stroke-width:12; }
-.label { fill:#2c2c2c; font-size:13px; font-weight:500; letter-spacing:.01em; }
-.label.active { fill:#111111; font-weight:700; }
-.timeline { display:flex; flex-direction:column; gap:10px; }
-.agent-step { border:1px solid var(--line); border-radius:18px; background:#ffffff; overflow:hidden; }
-.agent-summary { list-style:none; display:grid; grid-template-columns:42px 1fr; gap:12px; align-items:center; padding:12px; cursor:pointer; }
-.agent-summary::-webkit-details-marker { display:none; }
-.agent-index { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; font-weight:700; color:var(--gold); background:rgba(255,122,26,.08); border:1px solid rgba(255,122,26,.18); }
-.agent-head { display:flex; justify-content:space-between; gap:12px; align-items:center; }
-.agent-head h4 { margin:0; font-size:.98rem; color:var(--text); }
-.agent-head span { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
-.agent-copy { padding:0 14px 16px 66px; }
-.agent-copy p { margin:0; color:#2d2d2d; font-size:.93rem; line-height:1.6; }
-.candidate-grid { display:grid; grid-template-columns:repeat(3,minmax(0,1fr)); gap:18px; }
-.candidate-card { background:none; perspective:1400px; min-height:330px; }
-.candidate-card-inner { position:relative; width:100%; min-height:330px; transition:transform .8s cubic-bezier(.2,.7,.1,1); transform-style:preserve-3d; }
-.candidate-card:hover .candidate-card-inner,.candidate-card:focus .candidate-card-inner,.candidate-card:focus-within .candidate-card-inner { transform:rotateY(180deg); }
-.candidate-face { position:absolute; inset:0; padding:20px; border-radius:22px; border:1px solid var(--line); background:#ffffff; color:var(--text); backface-visibility:hidden; box-shadow:0 12px 24px rgba(0,0,0,.06); display:flex; flex-direction:column; gap:14px; }
-.candidate-back { transform:rotateY(180deg); }
-.candidate-top { display:flex; justify-content:space-between; align-items:center; gap:8px; }
-.chip { font-size:.72rem; text-transform:uppercase; letter-spacing:.12em; color:#0b6f66; padding:7px 10px; border-radius:999px; background:rgba(23,184,166,.08); border:1px solid rgba(23,184,166,.18); }
-.chip.alt { color:var(--gold); background:rgba(255,122,26,.08); border-color:rgba(255,122,26,.18); }
-.score { font-weight:700; color:var(--gold); }
-.candidate-face h4 { margin:0; font-size:1.08rem; line-height:1.35; }
-.candidate-face p { margin:0; color:#1e1e1e; line-height:1.65; font-size:.96rem; overflow-wrap:anywhere; }
-.meta-row { margin-top:auto; display:flex; justify-content:space-between; color:var(--muted); font-size:.88rem; gap:14px; }
-.mini { cursor:pointer; margin-top:8px; align-self:flex-start; color:var(--text); padding:10px 12px; border-radius:14px; border:1px solid var(--line); background:#ffffff; transition:all 0.2s; }
-.mini:hover { background:#f5f5f5; border-color:var(--chosen); }
-.papers-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
-.paper-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
-.paper-topline { display:flex; gap:8px; flex-wrap:wrap; margin-bottom:10px; }
-.paper-badge { font-size:.72rem; padding:6px 10px; border-radius:999px; background:rgba(98,141,255,.08); color:#3456b5; border:1px solid rgba(98,141,255,.18); }
-.paper-badge.alt { background:rgba(0,0,0,.04); color:#444; border-color:rgba(0,0,0,.08); }
-.doi-badge { background:rgba(255,122,26,.08); color:#8a4105; border-color:rgba(255,122,26,.18); }
-.paper-card h4 { margin:0 0 10px; line-height:1.35; font-size:1rem; }
-.paper-card p { margin:0 0 12px; line-height:1.6; color:#222; }
-.paper-links { display:flex; gap:12px; flex-wrap:wrap; }
-.paper-meta-stack { display:flex; flex-direction:column; gap:6px; color:#444; margin-bottom:12px; font-size:.9rem; }
-.paper-links a,.journal-card,.upload-note a { color:#0b63ce; text-decoration:none; }
-.journal-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
-.journal-card { border:1px solid var(--line); border-radius:18px; padding:16px; display:flex; justify-content:space-between; gap:14px; align-items:center; background:#ffffff; }
-.journal-card h4 { margin:0 0 6px; }
-.journal-card p { margin:0; color:var(--muted); line-height:1.5; }
-.selection-panel { padding:18px; }
-.parse-grid { display:grid; grid-template-columns:1.2fr 1fr; gap:14px; }
-.parse-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
-.ref-list { margin:0; padding-left:18px; }
-.ref-list li { margin-bottom:8px; line-height:1.5; }
-.gr-button-primary { background:linear-gradient(135deg,rgba(255,122,26,.92),rgba(240,108,22,.92)) !important; color:#ffffff !important; border:none !important; }
-.gr-button-secondary { background:#ffffff !important; color:var(--text) !important; border:1px solid var(--line) !important; }
-footer { display:none !important; }
-@keyframes pulseEdge { to { stroke-dashoffset:-40; } }
-@media (max-width:1180px) {
-  .model-switcher,.candidate-grid,.papers-grid,.journal-grid,.parse-grid { grid-template-columns:1fr; }
-  .brain-svg { height:460px; }
-}
-"""
-
-CSS = BASECSS + "\n" + getdvnclayoutcss() + "\n" + getdiscoverycss()
-
-HEAD = """
-<link rel="preconnect" href="https://fonts.googleapis.com">
-<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
-<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap" rel="stylesheet">
-<script>
-function triggerRouteSwap(idx) {
-    const container = document.getElementById('route_swap_payload');
-    if (!container) return;
-    const input = container.querySelector('textarea') || container.querySelector('input');
-    if (input) {
-        input.value = idx.toString();
-        input.dispatchEvent(new Event('input', { bubbles: true }));
-        setTimeout(() => {
-            const btn = document.getElementById('route_swap_apply');
-            if (btn) btn.click();
-        }, 150);
-    }
-}
-</script>
-"""
-
-
-# ── Gradio layout ───────────────────────────────────────────────────────────
-with gr.Blocks(css=CSS, head=HEAD, theme=gr.themes.Base(), fill_height=True) as demo:
-    papersstate = gr.State([])
-    parsedpdfstate = gr.State({})
-    ingestpayloadstate = gr.State({})
-    routestate = gr.State(getdefaultroutestate())
-
-    gr.HTML(
-        """
-        <div id="dvnc-shell">
-          <div class="hero-bar">
-            <div class="brand">
-              <div class="logo" aria-hidden="true">
-                <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7">
-                  <path d="M5 17L12 4l7 13"/>
-                  <path d="M8.5 12.5h7"/>
-                  <circle cx="12" cy="12" r="1.8" fill="currentColor" stroke="none"/>
-                </svg>
-              </div>
-              <div>
-                <h1>DVNC.AI</h1>
-                <p>Sovereign discovery instrument · connectome-native reasoning</p>
-              </div>
-            </div>
-            <div class="status"><span class="status-dot"></span><span>Live orchestration</span></div>
-          </div>
-        </div>
-        """
-    )
-
-    with gr.Tabs():
-        with gr.Tab("Discovery Engine"):
-            modelhtml = gr.HTML(buildmodelshtml("DVNC Sovereign"))
-
-            with gr.Row():
-                with gr.Column(scale=2):
-                    model = gr.Dropdown(
-                        choices=[m["name"] for m in MODELS],
-                        value="DVNC Sovereign",
-                        label="Model tier",
-                    )
-                    query = gr.Textbox(
-                        label="Discovery query",
-                        elem_classes=["querybox"],
-                        placeholder="Enter a scientific question, anomaly, or breakthrough direction…",
-                        lines=4,
-                    )
-                    with gr.Row():
-                        runbtn = gr.Button("Run discovery", variant="primary")
-                        examplebtn = gr.Button("Load example", variant="secondary")
-
-                    chat = gr.HTML(
-                        """
-                        <div class="panel chat-panel">
-                          <div class="chat-thread">
-                            <div class="bubble bubble-ai">
-                              <span class="role">DVNC</span>
-                              <p>Enter a query to activate the 7-agent discovery stack and illuminate the chosen path through the 3D connectome.</p>
-                            </div>
-                          </div>
-                        </div>
-                        """
-                    )
-
-                with gr.Column(scale=3):
-                    connectome = gr.HTML(buildconnectomehtml(DEFAULTPATH))
-                    cards = gr.HTML("")
-
-            output = gr.Markdown("# Discovery Output\n\nAwaiting query.")
-            timeline = gr.HTML(getinitialdiscoverytimelinehtml())
-            routeswappayload = gr.Textbox(value="", visible=False, elem_id="route_swap_payload")
-            routeswapapply = gr.Button("Apply route swap", visible=False, elem_id="route_swap_apply")
-
-        with gr.Tab("Self-Learning Graph"):
-            with gr.Row():
-                with gr.Column(scale=2):
-                    paperquery = gr.Textbox(
-                        label="Research topic / title / DOI / link",
-                        elem_classes=["querybox"],
-                        placeholder="e.g. self-assembling conductive biomaterials for cardiac repair",
-                        lines=3,
-                    )
-                    searchmode = gr.Dropdown(
-                        choices=SEARCHMODES,
-                        value="topic",
-                        label="Search mode",
-                    )
-                    sourceselector = gr.CheckboxGroup(
-                        choices=SOURCEOPTIONS,
-                        value=DEFAULTSOURCES,
-                        label="Sources",
-                    )
-                    pdfupload = gr.File(
-                        label="Upload PDF papers",
-                        file_types=[".pdf"],
-                        file_count="single",
-                    )
-
-                    with gr.Row():
-                        learnbtn = gr.Button("Discover papers", variant="primary")
-                        loadtopicbtn = gr.Button("Load example topic", variant="secondary")
-
-                    uploadstatus = gr.Markdown("No PDF uploaded yet.")
-                    discoverystatus = gr.Markdown("### No discovery results yet.")
-                    journalpanel = gr.HTML(buildjournalhtml("biomaterials cardiac repair"))
-
-                    gr.HTML(
-                        '<div class="panel selection-panel"><h3 style="margin:0 0 12px;">Select papers to ingest</h3></div>'
-                    )
-
-                    selectionbox = gr.CheckboxGroup(
-                        choices=[],
-                        value=[],
-                        label="Candidate papers",
-                    )
-
-                    parserorder = gr.CheckboxGroup(
-                        choices=["grobid", "docling", "pymupdf"],
-                        value=["grobid", "docling", "pymupdf"],
-                        label="Parser routing order",
-                    )
-
-                    with gr.Row():
-                        parsebtn = gr.Button("Parse uploaded PDF", variant="secondary")
-                        ingestbtn = gr.Button("Ingest selected into graph", variant="primary")
-
-                with gr.Column(scale=3):
-                    learninggraph = gr.HTML(buildlearninggraphhtml([], []))
-                    paperspanel = gr.HTML(
-                        '<div class="panel papers-panel" style="padding:18px"><p>Search by topic, title, DOI, or link, then select papers before graph ingestion.</p></div>'
-                    )
-                    parsesummary = gr.Markdown("### PDF parse status\n\nAwaiting upload.")
-                    parsepanel = gr.HTML(
-                        '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-                    )
-                    ingestsummary = gr.Markdown("### Graph ingest status\n\nAwaiting paper selection.")
-                    ingestpayload = gr.JSON(
-                        label="Graph ingest payload",
-                        value={"status": "empty", "nodes": [], "edges": []},
-                    )
-
-    # ── Event wiring ────────────────────────────────────────────────────────
-    examplebtn.click(fn=loadexample, outputs=query)
-
-    runbtn.click(
-        fn=rundiscovery,
-        inputs=[query, model],
-        outputs=[chat, connectome, timeline, cards, output, modelhtml, routestate],
-    )
-
-    routeswapapply.click(
-        fn=applyrouteswap,
-        inputs=[query, model, routeswappayload, routestate],
-        outputs=[chat, connectome, timeline, output, routestate],
-    )
-
-    loadtopicbtn.click(fn=loadpapertopic, outputs=paperquery)
-
-    learnbtn.click(
-        fn=runpaperdiscovery,
-        inputs=[paperquery, searchmode, sourceselector, pdfupload],
-        outputs=[
-            learninggraph,
-            paperspanel,
-            journalpanel,
-            uploadstatus,
-            selectionbox,
-            papersstate,
-            discoverystatus,
-        ],
-    )
-
-    parsebtn.click(
-        fn=parseuploadedpdf,
-        inputs=[pdfupload, parserorder],
-        outputs=[parsesummary, parsedpdfstate],
-    ).then(
-        fn=renderparseresult,
-        inputs=[parsedpdfstate],
-        outputs=[parsepanel],
-    )
-
-    ingestbtn.click(
-        fn=ingestselectedpapers,
-        inputs=[paperquery, selectionbox, papersstate, pdfupload, parsedpdfstate],
-        outputs=[learninggraph, ingestsummary, ingestpayload],
-    )
-
-
-if __name__ == "__main__":
-    demo.launch()

app_old4.py

@@ -1,412 +0,0 @@
-"""
-DVNC.AI — root app.py
-Fixed: all internal module imports now use correct underscore names
-matching the actual function/constant names in dvnc_ai_v2_hf/.
-"""
-
-# ── Standard library ──────────────────────────────────────────────────────
-import html
-import random
-import re
-import sys
-from pathlib import Path
-from typing import Dict, List, Optional
-
-# Ensure the repository root is on sys.path so the package is importable
-ROOT = Path(__file__).resolve().parent
-if str(ROOT) not in sys.path:
-    sys.path.insert(0, str(ROOT))
-
-# ── Third-party ───────────────────────────────────────────────────────────
-import gradio as gr
-
-# ── Internal modules ──────────────────────────────────────────────────────
-from dvnc_ai_v2_hf.agent_route_cards import build_agent_route_cards_html
-from dvnc_ai_v2_hf.discovery_app_bridge import (
-    get_default_route_state,
-    get_discovery_css,
-    get_initial_discovery_timeline_html,
-)
-from dvnc_ai_v2_hf.dvnc_ui_layout import get_dvnc_layout_css
-from dvnc_ai_v2_hf.graph_canvas_patch import render_graph_canvas_html
-from dvnc_ai_v2_hf.self_learning_graph import (
-    DEFAULT_SOURCES,
-    SEARCH_MODES,
-    SOURCE_OPTIONS,
-    build_journal_html,
-    ingest_selected_papers,
-    parse_uploaded_pdf,
-    render_parse_result,
-    run_paper_discovery,
-    safe_text,
-)
-
-# ── Constants ─────────────────────────────────────────────────────────────
-MODELS = [
-    {
-        "name": "DVNC Sovereign",
-        "tag": "flagship",
-        "desc": "Maximum depth orchestration for frontier discovery",
-    },
-    {
-        "name": "DVNC Atlas",
-        "tag": "research",
-        "desc": "Balanced reasoning, graph traversal, and synthesis",
-    },
-    {
-        "name": "DVNC Curie",
-        "tag": "lab",
-        "desc": "Experimental hypothesis generation for anomalous signals",
-    },
-]
-
-AGENTS = [
-    "Query Interpreter",
-    "Graph Divergence Mapper",
-    "Evidence Harvester",
-    "Analogy Engine",
-    "Hypothesis Composer",
-    "Adversarial Critic",
-    "Experimental Program Designer",
-]
-
-NODES = [
-    {"id": "seed", "label": "Seed Query", "group": "core", "x": 10, "y": 0, "z": 0},
-    {"id": "bio", "label": "Biomaterials", "group": "domain", "x": 24, "y": 12, "z": -8},
-    {"id": "card", "label": "Cardiac Repair", "group": "domain", "x": 38, "y": 3, "z": 14},
-    {"id": "nano", "label": "Nanostructure", "group": "bridge", "x": 24, "y": -18, "z": 16},
-    {"id": "selfasm", "label": "Self-Assembly", "group": "bridge", "x": 40, "y": -16, "z": -16},
-    {"id": "electro", "label": "Electro-signalling", "group": "mechanism", "x": 58, "y": 10, "z": -10},
-    {"id": "immune", "label": "Immune Modulation", "group": "mechanism", "x": 64, "y": -8, "z": 10},
-    {"id": "trial", "label": "Validation Path", "group": "outcome", "x": 80, "y": 0, "z": 0},
-    {"id": "alt1", "label": "Piezoelectric Scaffold", "group": "candidate", "x": 56, "y": 26, "z": 14},
-    {"id": "alt2", "label": "Peptide Mesh", "group": "candidate", "x": 54, "y": -27, "z": -14},
-]
-
-EDGES = [
-    ("seed", "bio"), ("seed", "nano"), ("bio", "card"), ("nano", "selfasm"),
-    ("selfasm", "electro"), ("card", "immune"), ("electro", "trial"),
-    ("immune", "trial"), ("card", "alt1"), ("selfasm", "alt2"),
-    ("alt1", "trial"), ("alt2", "trial"),
-]
-
-DEFAULT_PATH = ["seed", "nano", "selfasm", "electro", "trial"]
-
-CANDIDATES = [
-    {
-        "title": "Piezoelectric Scaffold Cascade",
-        "front": "Use mechano-electric scaffolds to convert cardiac strain into micro-current signalling.",
-        "back": "Discovery path: anomalous healing signal -> piezoelectric analog -> ion-channel entrainment -> tissue regeneration. Risk: power density and fibrosis coupling.",
-        "score": 92,
-        "novelty": "High",
-        "agent": "Hypothesis Composer",
-    },
-    {
-        "title": "Peptide Self-Assembly Mesh",
-        "front": "Deploy dynamic peptide meshes that self-assemble around damaged myocardium and guide repair.",
-        "back": "Discovery path: self-assembly -> local immune choreography -> regenerative substrate formation. Risk: degradation timing and targeting specificity.",
-        "score": 88,
-        "novelty": "High",
-        "agent": "Analogy Engine",
-    },
-    {
-        "title": "Immune-Tuned Conductive Hydrogel",
-        "front": "Blend conductivity with macrophage-state modulation to reduce scarring and restore conduction.",
-        "back": "Discovery path: inflammation mismatch -> conductive medium -> macrophage polarization -> synchronized healing. Risk: persistence and biocompatibility.",
-        "score": 85,
-        "novelty": "Medium-High",
-        "agent": "Adversarial Critic",
-    },
-]
-
-ACADEMIC_INSIGHTS = [
-    {
-        "hypothesis": "Implementation of mechano-electric scaffolds to transduce cardiac strain into localized micro-current signalling for myocardial regeneration.",
-        "metrics": {
-            "Novelty": 92,
-            "Mechanistic clarity": 85,
-            "Experimental tractability": 78,
-            "Cross-domain distance": 94,
-        },
-        "outline": (
-            "1. Synthesize candidate piezoelectric biomaterial scaffolds with tunable strain-electric coupling.\n"
-            "2. Evaluate in vitro electromechanical transduction and subsequent ion-channel entrainment.\n"
-            "3. Conduct in vivo comparative models to assess regenerative efficacy against gold-standard substrates.\n"
-            "4. Rigorously validate to exclude pathological fibrosis and power-density toxicity."
-        ),
-        "path": ["seed", "bio", "card", "alt1", "trial"],
-    },
-    {
-        "hypothesis": "Deployment of dynamic peptide networks that self-assemble post-infarction to orchestrate local immunological responses and guide substrate regeneration.",
-        "metrics": {
-            "Novelty": 88,
-            "Mechanistic clarity": 82,
-            "Experimental tractability": 86,
-            "Cross-domain distance": 85,
-        },
-        "outline": (
-            "1. Formulate peptide sequences programmed for triggered in situ self-assembly within the myocardial infarct zone.\n"
-            "2. Quantify macrophage polarization and local immune choreography post-deployment.\n"
-            "3. Map the temporospatial degradation profile against de novo tissue formation.\n"
-            "4. Falsify against off-target aggregation and delayed clearance risks."
-        ),
-        "path": ["seed", "nano", "selfasm", "alt2", "trial"],
-    },
-    {
-        "hypothesis": "Integration of conductive hydrogels with immunomodulatory properties to simultaneously bridge electrical uncoupling and mitigate adverse fibrotic scarring.",
-        "metrics": {
-            "Novelty": 85,
-            "Mechanistic clarity": 90,
-            "Experimental tractability": 88,
-            "Cross-domain distance": 79,
-        },
-        "outline": (
-            "1. Fabricate biocompatible hydrogels featuring precisely tuned electrical conductivity and immunomodulatory motifs.\n"
-            "2. Monitor electrophysiological synchronization across the scaffold-tissue interface.\n"
-            "3. Assess macrophage state transitions and suppression of adverse fibrotic remodelling.\n"
-            "4. Validate long-term persistence, hemocompatibility, and mechanical integration."
-        ),
-        "path": ["seed", "bio", "card", "immune", "trial"],
-    },
-]
-
-# ── Utility helpers ───────────────────────────────────────────────────────
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    return render_graph_canvas_html(
-        {
-            "status": "ok" if (nodes or edges) else "empty",
-            "nodes": nodes or [],
-            "edges": edges or [],
-        },
-        title=title,
-        height=780,
-    )
-
-# ── HTML builders ─────────────────────────────────────────────────────────
-
-def build_connectome_html(path_ids: List[str]) -> str:
-    active = set(path_ids)
-    node_map = {n["id"]: n for n in NODES}
-    path_pairs = {
-        pair
-        for i in range(len(path_ids) - 1)
-        for pair in [(path_ids[i], path_ids[i + 1]), (path_ids[i + 1], path_ids[i])]
-    }
-    baselines, activelines, circles, labels = [], [], [], []
-    for a, b in EDGES:
-        na, nb = node_map[a], node_map[b]
-        x1, y1 = na["x"] * 8 + 80, na["y"] * 6 + 280
-        x2, y2 = nb["x"] * 8 + 80, nb["y"] * 6 + 280
-        baselines.append(f'e class="edge" x1="{x1}" y1="{y1}" x2="{x2}" y2="{y2}"/>')
-        if (a, b) in path_pairs:
-            activelines.append(f'e class="edge active" x1="{x1}" y1="{y1}" x2="{x2}" y2="{y2}"/>')
-    for n in NODES:
-        cx, cy = n["x"] * 8 + 80, n["y"] * 6 + 280
-        is_active = n["id"] in active
-        state = "chosen" if is_active else "idle"
-        halo_cls = "halo active" if is_active else "halo"
-        lbl_cls = "label active" if is_active else "label"
-        radius = 18 if is_active else 13
-        halo_r = 30 if is_active else 0
-        circles.append(
-            f'ircle class="{halo_cls}" cx="{cx}" cy="{cy}" r="{halo_r}"/>'
-            f'ircle class="node {state}" cx="{cx}" cy="{cy}" r="{radius}"/>'
-            f'<title>{safe_text(n["label"])}</title>'
-            f'<text class="{lbl_cls}" x="{cx}" y="{cy + radius + 14}" text-anchor="middle">{safe_text(n["label"][:18])}</text>'
-        )
-    return f"""<div class="brain-shell panel">
-<div class="brain-header">
-<div><p class="eyebrow">Connectome</p><h3>3D Connectome</h3></div>
-<div class="brain-legend">
-<span><span class="dot dot-live"></span>lit path</span>
-<span><span class="dot dot-chosen"></span>chosen node</span>
-<span><span class="dot dot-idle"></span>available node</span>
-</div></div>
-<div class="brain-stage">
-<svg class="brain-svg" viewBox="0 0 880 560">
-{''.join(baselines)} {''.join(activelines)} {''.join(circles)}
-</svg></div></div>"""
-
-
-def build_cards_html(cards: List[Dict]) -> str:
-    items = []
-    for i, c in enumerate(cards):
-        items.append(
-            f"""<div class="candidate-card" tabindex="0">
-<div class="candidate-card-inner">
-<div class="candidate-face">
-<div class="candidate-top"><span class="chip">{safe_text(c["agent"])}</span><span class="score">{safe_text(c["score"])}</span></div>
-<h4>{safe_text(c["title"])}</h4>
-<p>{safe_text(c["front"])}</p>
-<div class="meta-row"><span>Novelty <strong>{safe_text(c["novelty"])}</strong></span></div>
-<button class="mini" onclick="triggerRouteSwap({i})">Use as main insight</button>
-</div>
-<div class="candidate-face candidate-back">
-<div class="candidate-top"><span class="chip alt">Alternative path</span><span class="score">{safe_text(c["score"])}</span></div>
-<h4>{safe_text(c["title"])}</h4>
-<p>{safe_text(c["back"])}</p>
-<div class="meta-row"><span>Swap into route <strong>Enabled</strong></span></div>
-<button class="mini" onclick="triggerRouteSwap({i})">Use as main insight</button>
-</div>
-</div></div>"""
-        )
-    return '<div class="candidate-grid">' + "".join(items) + "</div>"
-
-
-def build_chat_html(query: str, result: Dict) -> str:
-    return f"""<div class="chat-panel panel"><div class="chat-thread">
-<div class="bubble bubble-user"><span class="role">You</span><p>{safe_text(query)}</p></div>
-<div class="bubble bubble-ai"><span class="role">DVNC Sovereign</span><p>{safe_text(result["summary"])}</p></div>
-<div class="bubble bubble-system"><span class="role">Discovery Signal</span>
-<p><strong>Primary hypothesis:</strong> {safe_text(result["primary_hypothesis"])}</p>
-</div></div></div>"""
-
-
-def build_models_html(selected: str) -> str:
-    items = []
-    for m in MODELS:
-        active = "active" if m["name"] == selected else ""
-        items.append(
-            f'<div class="model-pill {active}"><span class="model-name">{safe_text(m["name"])}</span>'
-            f'<span class="model-tag">{safe_text(m["tag"])}</span>'
-            f'<small>{safe_text(m["desc"])}</small></div>'
-        )
-    return '<div class="model-switcher">' + "".join(items) + "</div>"
-
-# ── Discovery logic ───────────────────────────────────────────────────────
-
-def run_discovery(query: str, model_name: str):
-    random.seed(len(query or "") + len(model_name or ""))
-    if "curie" in (query or "").lower() or "einstein" in (query or "").lower():
-        primary = "Map the anomaly first, then force a distant analogy before composing the experimental programme."
-        path = ["seed", "bio", "card", "immune", "trial"]
-    else:
-        primary = (
-            "Utilization of a self-assembling conductive scaffold to transduce mechanical "
-            "strain into localized regenerative signalling pathways."
-        )
-        path = DEFAULT_PATH
-
-    summaries = [
-        "Normalises the user prompt into a graph-searchable seed and isolates the tension inside the question.",
-        "Finds remote conceptual bridges instead of staying near the starting domain cluster.",
-        "Pulls evidence packets and conflict signals required for grounded hypothesis formation.",
-        "Generates cross-domain analogies with a bias toward mechanism transfer rather than keyword similarity.",
-        "Composes the lead hypothesis and two structurally different variants.",
-        "Attacks weak assumptions, hidden confounders, and feasibility gaps.",
-        "Produces a staged validation plan with measurable falsification criteria.",
-    ]
-    tags = ["input", "graph", "evidence", "analogy", "compose", "critique", "experiment"]
-    reasoning = [
-        {"step": i + 1, "agent": AGENTS[i], "tag": tags[i], "summary": summaries[i]}
-        for i in range(7)
-    ]
-    result = {
-        "summary": (
-            "A deeper route was chosen through the connectome, with live alternatives preserved "
-            "as swappable cards so the reasoning path can be inspected rather than hidden."
-        ),
-        "primary_hypothesis": primary,
-        "reasoning": reasoning,
-        "cards": CANDIDATES,
-        "path": path,
-        "metrics": {
-            "Novelty": 93,
-            "Mechanistic clarity": 89,
-            "Experimental tractability": 82,
-            "Cross-domain distance": 91,
-        },
-    }
-    chat_html = build_chat_html(query, result)
-    connectome_html = build_connectome_html(path)
-    timeline_html = build_agent_route_cards_html(reasoning)
-    metrics_md = "\n".join(f"- {k}: {v}/100" for k, v in result["metrics"].items())
-    hypothesis_md = (
-        "# Discovery Output\n\n"
-        f"**Model:** {model_name}\n\n"
-        f"**Primary hypothesis:** {result['primary_hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metrics_md}\n\n"
-        "## Experimental outline\n"
-        "1. Construct the candidate material or protocol.\n"
-        "2. Test mechanistic signal expression under controlled conditions.\n"
-        "3. Compare against baseline and nearest-neighbour alternatives.\n"
-        "4. Falsify using the adversarial risk criteria surfaced in the reasoning path.\n"
-    )
-    cards_html = build_cards_html(CANDIDATES)
-    route_state = get_default_route_state()
-    return (
-        chat_html,
-        connectome_html,
-        timeline_html,
-        cards_html,
-        hypothesis_md,
-        build_models_html(model_name),
-        route_state,
-    )
-
-
-def apply_route_swap(query: str, model_name: str, route_swap_payload: str, route_state):
-    try:
-        idx = int(route_swap_payload)
-    except Exception:
-        idx = 0
-    if not (0 <= idx < len(ACADEMIC_INSIGHTS)):
-        idx = 0
-    academic = ACADEMIC_INSIGHTS[idx]
-    connectome_html = build_connectome_html(academic["path"])
-    result = {
-        "summary": (
-            "Main insight formally adopted. The connectome pathway and validation protocol "
-            "have been realigned to the selected candidate methodology."
-        ),
-        "primary_hypothesis": academic["hypothesis"],
-    }
-    chat_html = build_chat_html(query, result)
-    metrics_md = "\n".join(f"- {k}: {v}/100" for k, v in academic["metrics"].items())
-    hypothesis_md = (
-        "# Discovery Output\n\n"
-        f"**Model:** {model_name}\n\n"
-        f"**Primary hypothesis:** {academic['hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metrics_md}\n\n"
-        "## Experimental outline\n"
-        f"{academic['outline']}\n"
-    )
-    return chat_html, connectome_html, gr.update(), hypothesis_md, route_state
-
-# ── Example loaders ──────────────────────────────────────────────────────
-
-def load_example() -> str:
-    return "How could a self-assembling conductive biomaterial improve cardiac tissue regeneration by converting mechanical strain into repair signalling?"
-
-
-def load_paper_topic() -> str:
-    return "self-assembling conductive biomaterials for cardiac repair"
-
-# ── CSS / HEAD ────────────────────────────────────────────────────────────
-
-BASE_CSS = r"""
-:root {
-  --bg:#ffffff; --panel:#ffffff; --line:rgba(0,0,0,.12); --text:#111111;
-  --muted:#5b5b5b; --soft:rgba(0,0,0,.62); --gold:#ff6600; --teal:#17b8a6;
-  --blue:#628dff; --chosen:#ff7a1a; --idle:#b8d8ff; --idle-stroke:#5e8fe6;
-  --query-node:#ffd8b3; --paper-node:#d7f6f2; --upload-node:#e7defe;
-  --shadow:0 16px 40px rgba(0,0,0,.12);
-}
-html,body,.gradio-container{background:#ffffff !important;font-family:Inter,ui-sans-serif,system-ui,sans-serif;}
-.gradio-container{max-width:1640px !important;padding:20px !important;}
-#dvnc-shell{border:1px solid var(--line);border-radius:28px;overflow:hidden;background:#ffffff;box-shadow:var(--shadow);padding:20px 22px 22px;}
-.hero-bar{display:flex;justify-content:space-between;align-items:center;gap:16px;padding-bottom:12px;border-bottom:1px solid rgba(0,0,0,.06);margin-bottom:16px;}
-.brand{display:flex;align-items:center;gap:14px;}
-.logo{width:42px;height:42px;border-radius:14px;display:grid;place-items:center;color:var(--gold);background:linear-gradient(135deg,rgba(255,122,26,.12),rgba(23,184,166,.10));border:1px solid rgba(0,0,0,.08);}
-.brand h1{font-size:1.05rem;margin:0;font-weight:700;letter-spacing:.12em;text-transform:uppercase;}
-.brand p{margin:3px 0 0;color:var(--muted);font-size:.84rem;}
-.status{display:flex;gap:10px;align-items:center;color:var(--soft);font-size:.85rem;}
-.status-dot{width:10px;height:10px;border-radius:50%;background:var(--teal);box-shadow:0 0 0 6px rgba(23,184,166,.10),0 0 14px rgba(23,184,166,.25);}
-.panel{background:#ffffff;border:1px solid var(--line);border-radius:22px;box-shadow:inset 0 1px 0 rgba(255,255,255,.

app_old5.py

@@ -1,52 +0,0 @@
-"""
-DVNC.AI — root app.py
-Fixed: corrected all imports to use dvnc_ai_v2_hf package with proper underscore function names.
-"""
-
-# ── Standard library ──────────────────────────────────────────────────────
-import html
-import random
-import re
-import sys
-from pathlib import Path
-from typing import Dict, List, Optional
-
-# Ensure repository root is on sys.path
-ROOT = Path(__file__).resolve().parent
-if str(ROOT) not in sys.path:
-    sys.path.insert(0, str(ROOT))
-
-# ── Third-party ───────────────────────────────────────────────────────────
-import gradio as gr
-
-# ── Internal modules ──────────────────────────────────────────────────────
-from dvnc_ai_v2_hf.agent_route_cards import build_agent_route_cards_html
-from dvnc_ai_v2_hf.discovery_app_bridge import (
-    get_default_route_state,
-    get_discovery_css,
-    get_initial_discovery_timeline_html,
-)
-from dvnc_ai_v2_hf.dvnc_ui_layout import get_dvnc_layout_css
-from dvnc_ai_v2_hf.graph_canvas_patch import render_graph_canvas_html
-from dvnc_ai_v2_hf.self_learning_graph import (
-    DEFAULT_SOURCES,
-    SEARCH_MODES,
-    SOURCE_OPTIONS,
-    build_journal_html,
-    ingest_selected_papers,
-    parse_uploaded_pdf,
-    render_parse_result,
-    run_paper_discovery,
-    safe_text,
-)
-
-# ── Simplified launcher: use existing app from dvnc_ai_v2_hf ─────────────
-if __name__ == "__main__":
-    try:
-        from dvnc_ai_v2_hf import app as internal_app
-        internal_app.demo.launch()
-    except (ImportError, AttributeError):
-        print("Could not import dvnc_ai_v2_hf.app — falling back to minimal Gradio demo")
-        with gr.Blocks() as demo:
-            gr.Markdown("# DVNC.AI\n\nSpace configuration in progress. Check Files tab for dvnc_ai_v2_hf/app.py")
-        demo.launch()

dvnc_ai_v2_hf/app.py

@@ -1,18 +1,31 @@
-"""
+app_py = '''"""
 DVNC.AI — app.py
-Production Gradio entrypoint for the DVNC discovery workspace.
-Aligned to the current dvnc_ai_v2_hf.self_learning_graph interface.
+Refactored for functional "Use as main insight" logic with academic rigor.
 """
-
 # ── Standard library ────────────────────────────────────────────────────────
+import html
+import json
+import math
+import os
 import random
 import re
+import urllib.parse
+import xml.etree.ElementTree as ET
+from pathlib import Path
 from typing import Dict, List, Optional
-
-# ── Third-party ─────────────────────────────────────────────────────────────
+from urllib.parse import quote
+# ── Third-party ──────────────────────────────────────────────────────────────
 import gradio as gr
-
-# ── Internal modules ────────────────────────────────────────────────────────
+import requests
+try:
+    import fitz  # PyMuPDF
+except Exception:
+    fitz = None
+try:
+    from bs4 import BeautifulSoup
+except Exception:
+    BeautifulSoup = None
+# ── Internal modules ─────────────────────────────────────────────────────────
 from dvnc_ai_v2_hf.agent_route_cards import build_agent_route_cards_html
 from dvnc_ai_v2_hf.discovery_app_bridge import (
     get_default_route_state,
@@ -20,39 +33,24 @@ from dvnc_ai_v2_hf.discovery_app_bridge import (
     get_initial_discovery_timeline_html,
 )
 from dvnc_ai_v2_hf.dvnc_ui_layout import get_dvnc_layout_css
-from dvnc_ai_v2_hf import self_learning_graph as slg
 from dvnc_ai_v2_hf.self_learning_graph import (
     DEFAULT_SOURCES,
     SEARCH_MODES,
     SOURCE_OPTIONS,
-    build_journal_html,
     build_learning_graph_html,
+    build_journal_html,
     ingest_selected_papers,
     parse_uploaded_pdf,
     render_parse_result,
     run_paper_discovery,
     safe_text,
 )
-
-# ── Constants ───────────────────────────────────────────────────────────────
+# ── Constants ────────────────────────────────────────────────────────────────
 MODELS = [
-    {
-        "name": "DVNC Sovereign",
-        "tag": "flagship",
-        "desc": "Maximum depth orchestration for frontier discovery",
-    },
-    {
-        "name": "DVNC Atlas",
-        "tag": "research",
-        "desc": "Balanced reasoning, graph traversal, and synthesis",
-    },
-    {
-        "name": "DVNC Curie",
-        "tag": "lab",
-        "desc": "Experimental hypothesis generation for anomalous signals",
-    },
+    {"name": "DVNC Sovereign", "tag": "flagship", "desc": "Maximum depth orchestration for frontier discovery"},
+    {"name": "DVNC Atlas",     "tag": "research",  "desc": "Balanced reasoning, graph traversal, and synthesis"},
+    {"name": "DVNC Curie",     "tag": "lab",        "desc": "Experimental hypothesis generation for anomalous signals"},
 ]
-
 AGENTS = [
     "Query Interpreter",
     "Graph Divergence Mapper",
@@ -62,171 +60,129 @@ AGENTS = [
     "Adversarial Critic",
     "Experimental Program Designer",
 ]
-
 NODES = [
-    {"id": "seed", "label": "Seed Query", "group": "core", "x": 10, "y": 0, "z": 0},
-    {"id": "bio", "label": "Biomaterials", "group": "domain", "x": 24, "y": 12, "z": -8},
-    {"id": "card", "label": "Cardiac Repair", "group": "domain", "x": 38, "y": 3, "z": 14},
-    {"id": "nano", "label": "Nanostructure", "group": "bridge", "x": 24, "y": -18, "z": 16},
-    {"id": "selfasm", "label": "Self-Assembly", "group": "bridge", "x": 40, "y": -16, "z": -16},
-    {"id": "electro", "label": "Electro-signalling", "group": "mechanism", "x": 58, "y": 10, "z": -10},
-    {"id": "immune", "label": "Immune Modulation", "group": "mechanism", "x": 64, "y": -8, "z": 10},
-    {"id": "trial", "label": "Validation Path", "group": "outcome", "x": 80, "y": 0, "z": 0},
-    {"id": "alt1", "label": "Piezoelectric Scaffold", "group": "candidate", "x": 56, "y": 26, "z": 14},
-    {"id": "alt2", "label": "Peptide Mesh", "group": "candidate", "x": 54, "y": -27, "z": -14},
+    {"id": "seed",    "label": "Seed Query",          "group": "core",      "x": 10,  "y":   0, "z":   0},
+    {"id": "bio",     "label": "Biomaterials",         "group": "domain",    "x": 24,  "y":  12, "z":  -8},
+    {"id": "card",    "label": "Cardiac Repair",       "group": "domain",    "x": 38,  "y":   3, "z":  14},
+    {"id": "nano",    "label": "Nanostructure",        "group": "bridge",    "x": 24,  "y": -18, "z":  16},
+    {"id": "selfasm", "label": "Self-Assembly",        "group": "bridge",    "x": 40,  "y": -16, "z": -16},
+    {"id": "electro", "label": "Electro-signalling",   "group": "mechanism", "x": 58,  "y":  10, "z": -10},
+    {"id": "immune",  "label": "Immune Modulation",    "group": "mechanism", "x": 64,  "y":  -8, "z":  10},
+    {"id": "trial",   "label": "Validation Path",      "group": "outcome",   "x": 80,  "y":   0, "z":   0},
+    {"id": "alt1",    "label": "Piezoelectric Scaffold","group": "candidate", "x": 56,  "y":  26, "z":  14},
+    {"id": "alt2",    "label": "Peptide Mesh",          "group": "candidate", "x": 54,  "y": -27, "z": -14},
 ]
-
 EDGES = [
-    ("seed", "bio"),
-    ("seed", "nano"),
-    ("bio", "card"),
-    ("nano", "selfasm"),
-    ("selfasm", "electro"),
-    ("card", "immune"),
-    ("electro", "trial"),
-    ("immune", "trial"),
-    ("card", "alt1"),
-    ("selfasm", "alt2"),
-    ("alt1", "trial"),
-    ("alt2", "trial"),
+    ("seed",    "bio"),    ("seed",   "nano"),
+    ("bio",     "card"),   ("nano",   "selfasm"),
+    ("selfasm", "electro"),("card",   "immune"),
+    ("electro", "trial"),  ("immune", "trial"),
+    ("card",    "alt1"),   ("selfasm","alt2"),
+    ("alt1",    "trial"),  ("alt2",   "trial"),
 ]
-
 DEFAULT_PATH = ["seed", "nano", "selfasm", "electro", "trial"]
-
 CANDIDATES = [
     {
-        "title": "Piezoelectric Scaffold Cascade",
-        "front": "Use mechano-electric scaffolds to convert cardiac strain into micro-current signalling.",
-        "back": "Discovery path: anomalous healing signal -> piezoelectric analog -> ion-channel entrainment -> tissue regeneration. Risk: power density and fibrosis coupling.",
-        "score": 92,
+        "title":   "Piezoelectric Scaffold Cascade",
+        "front":   "Use mechano-electric scaffolds to convert cardiac strain into micro-current signalling.",
+        "back":    "Discovery path: anomalous healing signal -> piezoelectric analog -> ion-channel entrainment -> tissue regeneration. Risk: power density and fibrosis coupling.",
+        "score":   92,
         "novelty": "High",
-        "agent": "Hypothesis Composer",
+        "agent":   "Hypothesis Composer",
     },
     {
-        "title": "Peptide Self-Assembly Mesh",
-        "front": "Deploy dynamic peptide meshes that self-assemble around damaged myocardium and guide repair.",
-        "back": "Discovery path: self-assembly -> local immune choreography -> regenerative substrate formation. Risk: degradation timing and targeting specificity.",
-        "score": 88,
+        "title":   "Peptide Self-Assembly Mesh",
+        "front":   "Deploy dynamic peptide meshes that self-assemble around damaged myocardium and guide repair.",
+        "back":    "Discovery path: self-assembly -> local immune choreography -> regenerative substrate formation. Risk: degradation timing and targeting specificity.",
+        "score":   88,
         "novelty": "High",
-        "agent": "Analogy Engine",
+        "agent":   "Analogy Engine",
     },
     {
-        "title": "Immune-Tuned Conductive Hydrogel",
-        "front": "Blend conductivity with macrophage-state modulation to reduce scarring and restore conduction.",
-        "back": "Discovery path: inflammation mismatch -> conductive medium -> macrophage polarization -> synchronized healing. Risk: persistence and biocompatibility.",
-        "score": 85,
+        "title":   "Immune-Tuned Conductive Hydrogel",
+        "front":   "Blend conductivity with macrophage-state modulation to reduce scarring and restore conduction.",
+        "back":    "Discovery path: inflammation mismatch -> conductive medium -> macrophage polarization -> synchronized healing. Risk: persistence and biocompatibility.",
+        "score":   85,
         "novelty": "Medium-High",
-        "agent": "Adversarial Critic",
+        "agent":   "Adversarial Critic",
     },
 ]
-
 ACADEMIC_INSIGHTS = [
     {
         "hypothesis": "Implementation of mechano-electric scaffolds to transduce cardiac strain into localized micro-current signalling for myocardial regeneration.",
-        "metrics": {
-            "Novelty": 92,
-            "Mechanistic clarity": 85,
-            "Experimental tractability": 78,
-            "Cross-domain distance": 94,
-        },
-        "outline": (
-            "1. Synthesize candidate piezoelectric biomaterial scaffolds with tunable strain-electric coupling.\n"
-            "2. Evaluate in vitro electromechanical transduction and subsequent ion-channel entrainment.\n"
-            "3. Conduct in vivo comparative models to assess regenerative efficacy against gold-standard substrates.\n"
-            "4. Rigorously validate to exclude pathological fibrosis and power-density toxicity."
-        ),
-        "path": ["seed", "bio", "card", "alt1", "trial"],
+        "metrics": {"Novelty": 92, "Mechanistic clarity": 85, "Experimental tractability": 78, "Cross-domain distance": 94},
+        "outline": "1. Synthesize candidate piezoelectric biomaterial scaffolds with tunable strain-electric coupling.\\n2. Evaluate *in vitro* electromechanical transduction and subsequent ion-channel entrainment.\\n3. Conduct *in vivo* comparative models to assess regenerative efficacy against gold-standard substrates.\\n4. Rigorously validate to exclude pathological fibrosis and power-density toxicity.",
+        "path": ["seed", "bio", "card", "alt1", "trial"]
     },
     {
         "hypothesis": "Deployment of dynamic peptide networks that self-assemble post-infarction to orchestrate local immunological responses and guide substrate regeneration.",
-        "metrics": {
-            "Novelty": 88,
-            "Mechanistic clarity": 82,
-            "Experimental tractability": 86,
-            "Cross-domain distance": 85,
-        },
-        "outline": (
-            "1. Formulate peptide sequences programmed for triggered in situ self-assembly within the myocardial infarct zone.\n"
-            "2. Quantify macrophage polarization and local immune choreography post-deployment.\n"
-            "3. Map the temporospatial degradation profile against de novo tissue formation.\n"
-            "4. Falsify against off-target aggregation and delayed clearance risks."
-        ),
-        "path": ["seed", "nano", "selfasm", "alt2", "trial"],
+        "metrics": {"Novelty": 88, "Mechanistic clarity": 82, "Experimental tractability": 86, "Cross-domain distance": 85},
+        "outline": "1. Formulate peptide sequences programmed for triggered *in situ* self-assembly within the myocardial infarct zone.\\n2. Quantify macrophage polarization and local immune choreography post-deployment.\\n3. Map the temporospatial degradation profile against *de novo* tissue formation.\\n4. Falsify against off-target aggregation and delayed clearance risks.",
+        "path": ["seed", "nano", "selfasm", "alt2", "trial"]
     },
     {
         "hypothesis": "Integration of conductive hydrogels with immunomodulatory properties to simultaneously bridge electrical uncoupling and mitigate adverse fibrotic scarring.",
-        "metrics": {
-            "Novelty": 85,
-            "Mechanistic clarity": 90,
-            "Experimental tractability": 88,
-            "Cross-domain distance": 79,
-        },
-        "outline": (
-            "1. Fabricate biocompatible hydrogels featuring precisely tuned electrical conductivity and immunomodulatory motifs.\n"
-            "2. Monitor electrophysiological synchronization across the scaffold-tissue interface.\n"
-            "3. Assess macrophage state transitions and suppression of adverse fibrotic remodelling.\n"
-            "4. Validate long-term persistence, hemocompatibility, and mechanical integration."
-        ),
-        "path": ["seed", "bio", "card", "immune", "trial"],
-    },
+        "metrics": {"Novelty": 85, "Mechanistic clarity": 90, "Experimental tractability": 88, "Cross-domain distance": 79},
+        "outline": "1. Fabricate biocompatible hydrogels featuring precisely tuned electrical conductivity and immunomodulatory motifs.\\n2. Monitor electrophysiological synchronization across the scaffold-tissue interface.\\n3. Assess macrophage state transitions and suppression of adverse fibrotic remodelling.\\n4. Validate long-term persistence, hemocompatibility, and mechanical integration.",
+        "path": ["seed", "bio", "card", "immune", "trial"]
+    }
 ]
-
-PDF_PARSER_OPTIONS = getattr(slg, "PDF_PARSERS", ["pymupdf"])
-
-
-# ── Utility helpers ─────────────────────────────────────────────────────────
+JOURNALS = [
+    {"name": "Nature",     "url": "https://www.nature.com/search",                              "desc": "Flagship multidisciplinary research journal."},
+    {"name": "Science",    "url": "https://www.science.org/search",                             "desc": "High-impact science journal and family."},
+    {"name": "Cell",       "url": "https://www.cell.com/search",                                "desc": "Life sciences and translational biology."},
+    {"name": "The Lancet", "url": "https://www.thelancet.com/search",                           "desc": "Clinical and medical research."},
+    {"name": "IEEE Xplore","url": "https://ieeexplore.ieee.org/search/searchresult.jsp",        "desc": "Engineering, AI, signal processing, and systems."},
+]
+SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
+GROBID_URL               = os.getenv("GROBID_URL", "").strip()
+REQUEST_TIMEOUT          = 25
+# ── Utility helpers ──────────────────────────────────────────────────────────
+def safe_text(x, default: str = "") -> str:
+    return html.escape(str(x if x is not None else default))
 def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def build_metrics_markdown(metrics: Dict[str, int]) -> str:
-    return "\n".join(f"- {key}: {value}/100" for key, value in metrics.items())
-
-
-# ── HTML builders ────────────────────────────────────────────────────────────
+    return re.sub(r"\\s+", " ", (x or "")).strip()
+def detect_query_type(query: str) -> str:
+    q = (query or "").strip()
+    if re.match(r"^10\\.\\d{4,9}/[-._;()/:A-Z0-9]+$", q, flags=re.I):
+        return "doi"
+    if q.startswith("http://") or q.startswith("https://"):
+        return "link"
+    return "topic"
+def ensure_list(x):
+    return x if isinstance(x, list) else []
+# ── HTML builders ─────────────────────────────────────────────────────────────
 def build_connectome_html(path_ids: List[str]) -> str:
-    active = set(path_ids)
-    node_map = {n["id"]: n for n in NODES}
-
+    active    = set(path_ids)
+    node_map  = {n["id"]: n for n in NODES}
     path_pairs = {
         pair
         for i in range(len(path_ids) - 1)
         for pair in [(path_ids[i], path_ids[i + 1]), (path_ids[i + 1], path_ids[i])]
     }
-
     base_lines, active_lines, circles, labels = [], [], [], []
-
     for a, b in EDGES:
         na, nb = node_map[a], node_map[b]
         x1, y1 = na["x"] * 8 + 80, na["y"] * 6 + 280
         x2, y2 = nb["x"] * 8 + 80, nb["y"] * 6 + 280
-        base_lines.append(
-            f'<line class="edge" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />'
-        )
+        base_lines.append(f\'<line class="edge" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />\')
         if (a, b) in path_pairs:
-            active_lines.append(
-                f'<line class="edge active" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />'
-            )
-
+            active_lines.append(f\'<line class="edge active" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />\')
     for n in NODES:
-        cx, cy = n["x"] * 8 + 80, n["y"] * 6 + 280
+        cx, cy   = n["x"] * 8 + 80, n["y"] * 6 + 280
         is_active = n["id"] in active
-        state = "chosen" if is_active else "idle"
+        state    = "chosen" if is_active else "idle"
         halo_cls = "halo active" if is_active else "halo"
-        lbl_cls = "label active" if is_active else "label"
-        radius = 18 if is_active else 13
-        halo_r = 30 if is_active else 0
-
+        lbl_cls  = "label active" if is_active else "label"
+        radius   = 18 if is_active else 13
+        halo_r   = 30 if is_active else 0
         circles.append(
-            f'<g class="node-wrap">'
-            f'<circle class="{halo_cls}" cx="{cx:.1f}" cy="{cy:.1f}" r="{halo_r}" />'
-            f'<circle class="node {n["group"]} {state}" cx="{cx:.1f}" cy="{cy:.1f}" r="{radius}" />'
-            f"</g>"
-        )
-        labels.append(
-            f'<text class="{lbl_cls}" x="{cx + 18:.1f}" y="{cy - 16:.1f}">{safe_text(n["label"])}</text>'
+            f\'<g class="node-wrap">\'
+            f\'<circle class="{halo_cls}" cx="{cx:.1f}" cy="{cy:.1f}" r="{halo_r}" />\'
+            f\'<circle class="node {n["group"]} {state}" cx="{cx:.1f}" cy="{cy:.1f}" r="{radius}" />\'
+            f\'</g>\'
         )
-
+        labels.append(f\'<text class="{lbl_cls}" x="{cx + 18:.1f}" y="{cy - 16:.1f}">{safe_text(n["label"])}</text>\')
     return f"""
     <div class="panel brain-shell">
       <div class="brain-header">
@@ -242,50 +198,60 @@ def build_connectome_html(path_ids: List[str]) -> str:
       </div>
       <div class="brain-stage">
         <svg viewBox="0 0 780 560" class="brain-svg" role="img" aria-label="DVNC 3D connectome visualisation">
-          {''.join(base_lines)}
-          {''.join(active_lines)}
-          {''.join(circles)}
-          {''.join(labels)}
+          {"".join(base_lines)}
+          {"".join(active_lines)}
+          {"".join(circles)}
+          {"".join(labels)}
         </svg>
       </div>
     </div>
     """
-
-
 def build_cards_html(cards: List[Dict]) -> str:
     items = []
     for i, c in enumerate(cards):
-        items.append(
-            f"""
-            <article class="candidate-card" tabindex="0">
-              <div class="candidate-card-inner">
-                <div class="candidate-face candidate-front">
-                  <div class="candidate-top">
-                    <span class="chip">{safe_text(c["agent"])}</span>
-                    <span class="score">{safe_text(c["score"])}</span>
-                  </div>
-                  <h4>{safe_text(c["title"])}</h4>
-                  <p>{safe_text(c["front"])}</p>
-                  <div class="meta-row"><span>Novelty</span><strong>{safe_text(c["novelty"])}</strong></div>
-                  <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-                </div>
-                <div class="candidate-face candidate-back">
-                  <div class="candidate-top">
-                    <span class="chip alt">Alternative path</span>
-                    <span class="score">{safe_text(c["score"])}</span>
-                  </div>
-                  <h4>{safe_text(c["title"])}</h4>
-                  <p>{safe_text(c["back"])}</p>
-                  <div class="meta-row"><span>Swap into route</span><strong>Enabled</strong></div>
-                  <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-                </div>
+        items.append(f"""
+        <article class="candidate-card" tabindex="0">
+          <div class="candidate-card-inner">
+            <div class="candidate-face candidate-front">
+              <div class="candidate-top">
+                <span class="chip">{safe_text(c["agent"])}</span>
+                <span class="score">{safe_text(c["score"])}</span>
               </div>
-            </article>
-            """
-        )
-    return '<div class="panel" style="padding:20px;"><div class="candidate-grid">' + "".join(items) + "</div></div>"
-
-
+              <h4>{safe_text(c["title"])}</h4>
+              <p>{safe_text(c["front"])}</p>
+              <div class="meta-row"><span>Novelty</span><strong>{safe_text(c["novelty"])}</strong></div>
+              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
+            </div>
+            <div class="candidate-face candidate-back">
+              <div class="candidate-top">
+                <span class="chip alt">Alternative path</span>
+                <span class="score">{safe_text(c["score"])}</span>
+              </div>
+              <h4>{safe_text(c["title"])}</h4>
+              <p>{safe_text(c["back"])}</p>
+              <div class="meta-row"><span>Swap into route</span><strong>Enabled</strong></div>
+              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
+            </div>
+          </div>
+        </article>""")
+    return \'<div class="panel" style="padding:20px;"><div class="candidate-grid">\' + "".join(items) + "</div></div>"
+def build_agent_timeline(reasoning: List[Dict]) -> str:
+    rows = []
+    for r in reasoning:
+        rows.append(f"""
+        <details class="agent-step" {"open" if r["step"] == 1 else ""}>
+          <summary class="agent-summary">
+            <div class="agent-index">{safe_text(r["step"])}</div>
+            <div class="agent-head">
+              <h4>{safe_text(r["agent"])}</h4>
+              <span>{safe_text(r["tag"])}</span>
+            </div>
+          </summary>
+          <div class="agent-copy">
+            <p>{safe_text(r["summary"])}</p>
+          </div>
+        </details>""")
+    return \'<div class="panel" style="padding:18px;"><div class="timeline">\' + "".join(rows) + "</div></div>"
 def build_chat_html(query: str, result: Dict) -> str:
     return f"""
     <div class="panel chat-panel">
@@ -305,41 +271,29 @@ def build_chat_html(query: str, result: Dict) -> str:
       </div>
     </div>
     """
-
-
 def build_models_html(selected: str) -> str:
     items = []
     for m in MODELS:
         active = "active" if m["name"] == selected else ""
-        items.append(
-            f"""
+        items.append(f"""
             <div class="model-pill {active}">
               <span class="model-name">{safe_text(m["name"])}</span>
               <span class="model-tag">{safe_text(m["tag"])}</span>
               <small>{safe_text(m["desc"])}</small>
-            </div>
-            """
-        )
-    return '<div class="panel" style="padding:18px;"><div class="model-switcher">' + "".join(items) + "</div></div>"
-
-
-# ── Discovery logic ──────────────────────────────────────────────────────────
+            </div>""")
+    return \'<div class="panel" style="padding:18px;"><div class="model-switcher">\' + "".join(items) + "</div></div>"
+# ── Discovery logic ───────────────────────────────────────────────────────────
 def run_discovery(query: str, model_name: str):
-    query_text = norm_text(query)
-    model_text = norm_text(model_name) or "DVNC Sovereign"
-
-    random.seed(len(query_text) + len(model_text))
-
-    if "curie" in query_text.lower() or "einstein" in query_text.lower():
+    """
+    Runs the 7-agent discovery pipeline.
+    """
+    random.seed(len(query) + len(model_name))
+    if "curie" in query.lower() or "einstein" in query.lower():
         primary = "Map the anomaly first, then force a distant analogy before composing the experimental programme."
-        path = ["seed", "bio", "card", "immune", "trial"]
+        path    = ["seed", "bio", "card", "immune", "trial"]
     else:
-        primary = (
-            "Utilization of a self-assembling conductive scaffold to transduce mechanical strain "
-            "into localized regenerative signalling pathways."
-        )
-        path = DEFAULT_PATH
-
+        primary = "Utilization of a self-assembling conductive scaffold to transduce mechanical strain into localized regenerative signalling pathways."
+        path    = DEFAULT_PATH
     summaries = [
         "Normalises the user prompt into a graph-searchable seed and isolates the tension inside the question.",
         "Finds remote conceptual bridges instead of staying near the starting domain cluster.",
@@ -350,912 +304,240 @@ def run_discovery(query: str, model_name: str):
         "Produces a staged validation plan with measurable falsification criteria.",
     ]
     tags = ["input", "graph", "evidence", "analogy", "compose", "critique", "experiment"]
-
     reasoning = [
         {"step": i + 1, "agent": AGENTS[i], "tag": tags[i], "summary": summaries[i]}
         for i in range(7)
     ]
-
     result = {
-        "summary": (
-            "A deeper route was chosen through the connectome, with live alternatives preserved "
-            "as swappable cards so the reasoning path can be inspected rather than hidden."
-        ),
+        "summary":            "A deeper route was chosen through the connectome, with live alternatives preserved as swappable cards so the reasoning path can be inspected rather than hidden.",
         "primary_hypothesis": primary,
-        "reasoning": reasoning,
-        "cards": CANDIDATES,
-        "path": path,
+        "reasoning":          reasoning,
+        "cards":              CANDIDATES,
+        "path":               path,
         "metrics": {
-            "Novelty": 93,
-            "Mechanistic clarity": 89,
+            "Novelty":                   93,
+            "Mechanistic clarity":       89,
             "Experimental tractability": 82,
-            "Cross-domain distance": 91,
+            "Cross-domain distance":     91,
         },
     }
-
-    chat_html = build_chat_html(query_text, result)
+    chat_html       = build_chat_html(query, result)
     connectome_html = build_connectome_html(path)
-    timeline_html = build_agent_route_cards_html(reasoning)
-    metrics_md = build_metrics_markdown(result["metrics"])
-    cards_html = build_cards_html(CANDIDATES)
-    route_state = get_default_route_state()
-
-    hypothesis_md = (
-        "# Discovery Output\n\n"
-        f"**Model:** {model_text}\n\n"
-        f"**Primary hypothesis:** {result['primary_hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metrics_md}\n\n"
-        "## Experimental outline\n"
-        "1. Construct the candidate material or protocol.\n"
-        "2. Test mechanistic signal expression under controlled conditions.\n"
-        "3. Compare against baseline and nearest-neighbour alternatives.\n"
-        "4. Falsify using the adversarial risk criteria surfaced in the reasoning path.\n"
-    )
-
-    return (
-        chat_html,
-        connectome_html,
-        timeline_html,
-        cards_html,
-        hypothesis_md,
-        build_models_html(model_text),
-        route_state,
+    timeline_html   = build_agent_route_cards_html(reasoning)
+    metrics_md      = "\\n".join(f"- {k}: {v}/100" for k, v in result["metrics"].items())
+    hypothesis_md   = (
+        "# Discovery Output\\n\\n"
+        f"**Model:** {model_name}\\n\\n"
+        f"**Primary hypothesis:** {result['primary_hypothesis']}\\n\\n"
+        "## Scoring\\n"
+        f"{metrics_md}\\n\\n"
+        "## Experimental outline\\n"
+        "1. Construct the candidate material or protocol.\\n"
+        "2. Test mechanistic signal expression under controlled conditions.\\n"
+        "3. Compare against baseline and nearest-neighbour alternatives.\\n"
+        "4. Falsify using the adversarial risk criteria surfaced in the reasoning path.\\n"
     )
-
-
+    cards_html  = build_cards_html(CANDIDATES)
+    route_state = get_default_route_state()
+    return chat_html, connectome_html, timeline_html, cards_html, hypothesis_md, build_models_html(model_name), route_state
 def apply_route_swap(query: str, model_name: str, route_swap_payload: str, route_state):
+    """
+    Called when a user clicks 'Use as main insight' on a candidate card.
+    Sanitizes the output, adopts academic rigor, updates the connectome and discovery output.
+    """
     try:
-        idx = int((route_swap_payload or "").strip())
-    except Exception:
+        idx = int(route_swap_payload)
+    except ValueError:
         idx = 0
-
     if not (0 <= idx < len(ACADEMIC_INSIGHTS)):
         idx = 0
-
     academic = ACADEMIC_INSIGHTS[idx]
+    
+    # Update Connectome
     connectome_html = build_connectome_html(academic["path"])
-
+    
+    # Update Chat Feedback
     result = {
-        "summary": (
-            "Main insight formally adopted. The connectome pathway and validation protocol "
-            "have been realigned to the selected candidate methodology."
-        ),
-        "primary_hypothesis": academic["hypothesis"],
+        "summary": "Main insight formally adopted. The connectome pathway and validation protocol have been realigned to the selected candidate methodology.",
+        "primary_hypothesis": academic["hypothesis"]
     }
-    chat_html = build_chat_html(norm_text(query), result)
-    metrics_md = build_metrics_markdown(academic["metrics"])
-
+    chat_html = build_chat_html(query, result)
+    # Format Oxford-tier markdown output
+    metrics_md = "\\n".join(f"- {k}: {v}/100" for k, v in academic["metrics"].items())
+    
     hypothesis_md = (
-        "# Discovery Output\n\n"
-        f"**Model:** {norm_text(model_name) or 'DVNC Sovereign'}\n\n"
-        f"**Primary hypothesis:** {academic['hypothesis']}\n\n"
-        "## Scoring\n"
-        f"{metrics_md}\n\n"
-        "## Experimental outline\n"
-        f"{academic['outline']}\n"
+        "# Discovery Output\\n\\n"
+        f"**Model:** {model_name}\\n\\n"
+        f"**Primary hypothesis:** {academic['hypothesis']}\\n\\n"
+        "## Scoring\\n"
+        f"{metrics_md}\\n\\n"
+        "## Experimental outline\\n"
+        f"{academic['outline']}\\n"
     )
-
+    # We return the new chat, new connectome, leave timeline alone (gr.update()), new output, new state
     return chat_html, connectome_html, gr.update(), hypothesis_md, route_state
-
-
-# ── Example loaders ──────────────────────────────────────────────────────────
+# ── Example loaders ───────────────────────────────────────────────────────────
 def load_example() -> str:
-    return (
-        "How could a self-assembling conductive biomaterial improve cardiac tissue "
-        "regeneration by converting mechanical strain into repair signalling?"
-    )
-
-
+    return "How could a self-assembling conductive biomaterial improve cardiac tissue regeneration by converting mechanical strain into repair signalling?"
 def load_paper_topic() -> str:
     return "self-assembling conductive biomaterials for cardiac repair"
-
-
-# ── CSS / HEAD ───────────────────────────────────────────────────────────────
+# ── CSS / HEAD ────────────────────────────────────────────────────────────────
 BASE_CSS = r"""
 :root {
-  --bg: #ffffff;
-  --panel: #ffffff;
-  --line: rgba(0,0,0,.12);
-  --text: #111111;
-  --muted: #5b5b5b;
-  --soft: rgba(0,0,0,.62);
-  --gold: #ff6600;
-  --teal: #17b8a6;
-  --blue: #628dff;
-  --chosen: #ff7a1a;
-  --idle: #b8d8ff;
-  --idle-stroke: #5e8fe6;
-  --query-node: #ffd8b3;
-  --paper-node: #d7f6f2;
-  --upload-node: #e7defe;
+  --bg: #ffffff; --panel: #ffffff; --line: rgba(0,0,0,.12);
+  --text: #111111; --muted: #5b5b5b; --soft: rgba(0,0,0,.62);
+  --gold: #ff6600; --teal: #17b8a6; --blue: #628dff;
+  --chosen: #ff7a1a; --idle: #b8d8ff; --idle-stroke: #5e8fe6;
+  --query-node: #ffd8b3; --paper-node: #d7f6f2; --upload-node: #e7defe;
   --shadow: 0 16px 40px rgba(0,0,0,.12);
 }
-
-html, body, .gradio-container {
-  background: #ffffff !important;
-  font-family: Inter, ui-sans-serif, system-ui, sans-serif;
-}
-
-.gradio-container {
-  max-width: 1640px !important;
-  padding: 20px !important;
-}
-
-#dvnc-shell {
-  border: 1px solid var(--line);
-  border-radius: 28px;
-  overflow: hidden;
-  background: #ffffff;
-  box-shadow: var(--shadow);
-  padding: 20px 22px 22px;
-}
-
-.hero-bar {
-  display: flex;
-  justify-content: space-between;
-  align-items: center;
-  gap: 16px;
-  padding-bottom: 12px;
-  border-bottom: 1px solid rgba(0,0,0,.06);
-  margin-bottom: 16px;
-}
-
-.brand {
-  display: flex;
-  align-items: center;
-  gap: 14px;
-}
-
-.logo {
-  width: 42px;
-  height: 42px;
-  border-radius: 14px;
-  display: grid;
-  place-items: center;
-  color: var(--gold);
-  background: linear-gradient(135deg, rgba(255,122,26,.12), rgba(23,184,166,.10));
-  border: 1px solid rgba(0,0,0,.08);
-}
-
-.logo svg {
-  width: 24px;
-  height: 24px;
-}
-
-.brand h1 {
-  font-size: 1.05rem;
-  margin: 0;
-  font-weight: 700;
-  letter-spacing: .12em;
-  text-transform: uppercase;
-}
-
-.brand p {
-  margin: 3px 0 0;
-  color: var(--muted);
-  font-size: .84rem;
-}
-
-.status {
-  display: flex;
-  gap: 10px;
-  align-items: center;
-  color: var(--soft);
-  font-size: .85rem;
-}
-
-.status-dot {
-  width: 10px;
-  height: 10px;
-  border-radius: 50%;
-  background: var(--teal);
-  box-shadow: 0 0 0 6px rgba(23,184,166,.10), 0 0 14px rgba(23,184,166,.25);
-}
-
-.panel {
-  background: #ffffff;
-  border: 1px solid var(--line);
-  border-radius: 22px;
-  box-shadow: inset 0 1px 0 rgba(255,255,255,.8);
-}
-
-.querybox textarea,
-.querybox input {
-  background: transparent !important;
-  color: var(--text) !important;
-}
-
-.querybox,
-.querybox > div {
-  background: #ffffff !important;
-  border-radius: 18px !important;
-  border-color: var(--line) !important;
-}
-
-.chat-panel {
-  padding: 18px;
-  min-height: 280px;
-}
-
-.chat-thread {
-  display: flex;
-  flex-direction: column;
-  gap: 14px;
-}
-
-.bubble {
-  max-width: 88%;
-  padding: 16px 18px;
-  border-radius: 22px;
-  border: 1px solid var(--line);
-}
-
-.bubble p {
-  margin: 8px 0 0;
-  line-height: 1.6;
-  font-size: .96rem;
-  color: var(--text);
-}
-
-.bubble .role {
-  font-size: .72rem;
-  letter-spacing: .12em;
-  text-transform: uppercase;
-  color: var(--muted);
-}
-
-.bubble-user {
-  align-self: flex-end;
-  background: linear-gradient(135deg, rgba(98,141,255,.16), rgba(98,141,255,.08));
-}
-
-.bubble-ai {
-  align-self: flex-start;
-  background: #ffffff;
-}
-
-.bubble-system {
-  align-self: flex-start;
-  background: linear-gradient(135deg, rgba(255,122,26,.10), rgba(255,122,26,.04));
-}
-
-.model-switcher {
-  display: grid;
-  grid-template-columns: repeat(3, 1fr);
-  gap: 12px;
-}
-
-.model-pill {
-  padding: 14px;
-  border: 1px solid var(--line);
-  border-radius: 18px;
-  display: flex;
-  flex-direction: column;
-  gap: 4px;
-  min-height: 98px;
-  background: #ffffff;
-}
-
-.model-pill.active {
-  border-color: rgba(255,122,26,.40);
-  background: linear-gradient(135deg, rgba(255,122,26,.10), rgba(255,255,255,.96));
-}
-
-.model-name {
-  font-weight: 650;
-  color: var(--text);
-}
-
-.model-tag {
-  font-size: .76rem;
-  text-transform: uppercase;
-  letter-spacing: .12em;
-  color: var(--gold);
-}
-
-.model-pill small {
-  color: var(--muted);
-  line-height: 1.45;
-}
-
-.brain-shell {
-  padding: 18px;
-}
-
-.brain-header {
-  display: flex;
-  justify-content: space-between;
-  align-items: flex-end;
-  gap: 16px;
-  margin-bottom: 10px;
-}
-
-.eyebrow {
-  font-size: .72rem;
-  letter-spacing: .16em;
-  text-transform: uppercase;
-  color: var(--gold);
-  margin: 0 0 4px;
-}
-
-.brain-header h3 {
-  margin: 0;
-  font-size: 1.12rem;
-  color: var(--text);
-}
-
-.brain-legend {
-  display: flex;
-  gap: 14px;
-  color: var(--muted);
-  font-size: .8rem;
-  flex-wrap: wrap;
-}
-
-.dot {
-  width: 10px;
-  height: 10px;
-  display: inline-block;
-  border-radius: 50%;
-  margin-right: 6px;
-}
-
-.dot-live {
-  background: var(--chosen);
-  box-shadow: 0 0 10px rgba(255,122,26,.35);
-}
-
-.dot-chosen {
-  background: var(--chosen);
-}
-
-.dot-idle {
-  background: var(--idle);
-  border: 1px solid var(--idle-stroke);
-}
-
-.dot-query {
-  background: var(--query-node);
-  border: 1px solid #de9e58;
-}
-
-.dot-paper {
-  background: var(--paper-node);
-  border: 1px solid #4fb3a5;
-}
-
-.dot-upload {
-  background: var(--upload-node);
-  border: 1px solid #8f73d9;
-}
-
-.brain-stage {
-  position: relative;
-  min-height: 420px;
-  overflow: hidden;
-  background: linear-gradient(180deg, rgba(250,250,250,1), rgba(255,255,255,1));
-  border: 1px solid rgba(0,0,0,.05);
-  border-radius: 20px;
-}
-
-.brain-svg {
-  width: 100%;
-  height: 520px;
-  display: block;
-}
-
-.edge {
-  stroke: rgba(0,0,0,.12);
-  stroke-width: 2.4;
-}
-
-.edge.active {
-  stroke: var(--chosen);
-  stroke-width: 4.2;
-  stroke-linecap: round;
-  filter: drop-shadow(0 0 6px rgba(255,122,26,.45));
-  stroke-dasharray: 8 12;
-  animation: pulseEdge 1.5s linear infinite;
-}
-
-.node {
-  stroke-width: 2.2;
-  transition: all .25s ease;
-}
-
-.node.idle {
-  fill: var(--idle);
-  stroke: var(--idle-stroke);
-}
-
-.node.chosen {
-  fill: var(--chosen);
-  stroke: #ffb16d;
-}
-
-.halo {
-  fill: none;
-}
-
-.halo.active {
-  stroke: rgba(255,122,26,.18);
-  stroke-width: 12;
-}
-
-.label {
-  fill: #2c2c2c;
-  font-size: 13px;
-  font-weight: 500;
-  letter-spacing: .01em;
-}
-
-.label.active {
-  fill: #111111;
-  font-weight: 700;
-}
-
-.learn-edge {
-  stroke: rgba(0,0,0,.18);
-  stroke-width: 2.2;
-  stroke-linecap: round;
-}
-
-.learn-node {
-  stroke-width: 2.2;
-}
-
-.learn-node.query {
-  fill: var(--query-node);
-  stroke: #de9e58;
-}
-
-.learn-node.paper {
-  fill: var(--paper-node);
-  stroke: #36a091;
-}
-
-.learn-node.upload {
-  fill: var(--upload-node);
-  stroke: #7e63cb;
-}
-
-.learn-label {
-  fill: #1e1e1e;
-  font-size: 12px;
-  font-weight: 600;
-}
-
-.learning-empty {
-  display: grid;
-  place-items: center;
-}
-
-.empty-graph-copy {
-  text-align: center;
-  max-width: 440px;
-  padding: 40px 20px;
-}
-
-.empty-graph-copy h4 {
-  margin: 0 0 10px;
-  font-size: 1.05rem;
-}
-
-.empty-graph-copy p {
-  margin: 0;
-  color: var(--muted);
-  line-height: 1.6;
-}
-
-.timeline {
-  display: flex;
-  flex-direction: column;
-  gap: 10px;
-}
-
-.agent-step {
-  border: 1px solid var(--line);
-  border-radius: 18px;
-  background: #ffffff;
-  overflow: hidden;
-}
-
-.agent-summary {
-  list-style: none;
-  display: grid;
-  grid-template-columns: 42px 1fr;
-  gap: 12px;
-  align-items: center;
-  padding: 12px;
-  cursor: pointer;
-}
-
-.agent-summary::-webkit-details-marker {
-  display: none;
-}
-
-.agent-index {
-  width: 42px;
-  height: 42px;
-  border-radius: 14px;
-  display: grid;
-  place-items: center;
-  font-weight: 700;
-  color: var(--gold);
-  background: rgba(255,122,26,.08);
-  border: 1px solid rgba(255,122,26,.18);
-}
-
-.agent-head {
-  display: flex;
-  justify-content: space-between;
-  gap: 12px;
-  align-items: center;
-}
-
-.agent-head h4 {
-  margin: 0;
-  font-size: .98rem;
-  color: var(--text);
-}
-
-.agent-head span {
-  font-size: .72rem;
-  letter-spacing: .12em;
-  text-transform: uppercase;
-  color: var(--muted);
-}
-
-.agent-copy {
-  padding: 0 14px 16px 66px;
-}
-
-.agent-copy p {
-  margin: 0;
-  color: #2d2d2d;
-  font-size: .93rem;
-  line-height: 1.6;
-}
-
-.candidate-grid {
-  display: grid;
-  grid-template-columns: repeat(3, minmax(0,1fr));
-  gap: 18px;
-}
-
-.candidate-card {
-  background: none;
-  perspective: 1400px;
-  min-height: 330px;
-}
-
-.candidate-card-inner {
-  position: relative;
-  width: 100%;
-  min-height: 330px;
-  transition: transform .8s cubic-bezier(.2,.7,.1,1);
-  transform-style: preserve-3d;
-}
-
-.candidate-card:hover .candidate-card-inner,
-.candidate-card:focus .candidate-card-inner,
-.candidate-card:focus-within .candidate-card-inner {
-  transform: rotateY(180deg);
-}
-
-.candidate-face {
-  position: absolute;
-  inset: 0;
-  padding: 20px;
-  border-radius: 22px;
-  border: 1px solid var(--line);
-  background: #ffffff;
-  color: var(--text);
-  backface-visibility: hidden;
-  box-shadow: 0 12px 24px rgba(0,0,0,.06);
-  display: flex;
-  flex-direction: column;
-  gap: 14px;
-}
-
-.candidate-back {
-  transform: rotateY(180deg);
-}
-
-.candidate-top {
-  display: flex;
-  justify-content: space-between;
-  align-items: center;
-  gap: 8px;
-}
-
-.chip {
-  font-size: .72rem;
-  text-transform: uppercase;
-  letter-spacing: .12em;
-  color: #0b6f66;
-  padding: 7px 10px;
-  border-radius: 999px;
-  background: rgba(23,184,166,.08);
-  border: 1px solid rgba(23,184,166,.18);
-}
-
-.chip.alt {
-  color: var(--gold);
-  background: rgba(255,122,26,.08);
-  border-color: rgba(255,122,26,.18);
-}
-
-.score {
-  font-weight: 700;
-  color: var(--gold);
-}
-
-.candidate-face h4 {
-  margin: 0;
-  font-size: 1.08rem;
-  line-height: 1.35;
-}
-
-.candidate-face p {
-  margin: 0;
-  color: #1e1e1e;
-  line-height: 1.65;
-  font-size: .96rem;
-  overflow-wrap: anywhere;
-}
-
-.meta-row {
-  margin-top: auto;
-  display: flex;
-  justify-content: space-between;
-  color: var(--muted);
-  font-size: .88rem;
-  gap: 14px;
-}
-
-.mini {
-  cursor: pointer;
-  margin-top: 8px;
-  align-self: flex-start;
-  color: var(--text);
-  padding: 10px 12px;
-  border-radius: 14px;
-  border: 1px solid var(--line);
-  background: #ffffff;
-  transition: all 0.2s;
-}
-
-.mini:hover {
-  background: #f5f5f5;
-  border-color: var(--chosen);
-}
-
-.papers-grid {
-  display: grid;
-  grid-template-columns: repeat(2, minmax(0,1fr));
-  gap: 14px;
-}
-
-.paper-card {
-  border: 1px solid var(--line);
-  border-radius: 18px;
-  padding: 16px;
-  background: #ffffff;
-}
-
-.paper-topline {
-  display: flex;
-  gap: 8px;
-  flex-wrap: wrap;
-  margin-bottom: 10px;
-}
-
-.paper-badge {
-  font-size: .72rem;
-  padding: 6px 10px;
-  border-radius: 999px;
-  background: rgba(98,141,255,.08);
-  color: #3456b5;
-  border: 1px solid rgba(98,141,255,.18);
-}
-
-.paper-badge.alt {
-  background: rgba(0,0,0,.04);
-  color: #444;
-  border-color: rgba(0,0,0,.08);
-}
-
-.doi-badge {
-  background: rgba(255,122,26,.08);
-  color: #8a4105;
-  border-color: rgba(255,122,26,.18);
-}
-
-.paper-card h4 {
-  margin: 0 0 10px;
-  line-height: 1.35;
-  font-size: 1rem;
-}
-
-.paper-card p {
-  margin: 0 0 12px;
-  line-height: 1.6;
-  color: #222;
-}
-
-.paper-links {
-  display: flex;
-  gap: 12px;
-  flex-wrap: wrap;
-}
-
-.paper-meta-stack {
-  display: flex;
-  flex-direction: column;
-  gap: 6px;
-  color: #444;
-  margin-bottom: 12px;
-  font-size: .9rem;
-}
-
-.paper-links a,
-.journal-card,
-.upload-note a {
-  color: #0b63ce;
-  text-decoration: none;
-}
-
-.journal-grid {
-  display: grid;
-  grid-template-columns: repeat(2, minmax(0,1fr));
-  gap: 14px;
-}
-
-.journal-card {
-  border: 1px solid var(--line);
-  border-radius: 18px;
-  padding: 16px;
-  display: flex;
-  justify-content: space-between;
-  gap: 14px;
-  align-items: center;
-  background: #ffffff;
-}
-
-.journal-card h4 {
-  margin: 0 0 6px;
-}
-
-.journal-card p {
-  margin: 0;
-  color: var(--muted);
-  line-height: 1.5;
-}
-
-.upload-note {
-  border: 1px dashed rgba(0,0,0,.16);
-  border-radius: 18px;
-  padding: 16px;
-  background: rgba(0,0,0,.015);
-  color: #1f1f1f;
-  line-height: 1.6;
-}
-
-.selection-panel {
-  padding: 18px;
-}
-
-.parse-grid {
-  display: grid;
-  grid-template-columns: 1.2fr 1fr;
-  gap: 14px;
-}
-
-.parse-card {
-  border: 1px solid var(--line);
-  border-radius: 18px;
-  padding: 16px;
-  background: #ffffff;
-}
-
-.ref-list {
-  margin: 0;
-  padding-left: 18px;
-}
-
-.ref-list li {
-  margin-bottom: 8px;
-  line-height: 1.5;
-}
-
-.prosebox {
-  padding: 18px;
-  white-space: pre-wrap;
-  font-family: ui-monospace, SFMono-Regular, Menlo, monospace;
-  line-height: 1.55;
-  color: #1b1b1b;
-}
-
-.gr-button-primary {
-  background: linear-gradient(135deg, rgba(255,122,26,.92), rgba(240,108,22,.92)) !important;
-  color: #ffffff !important;
-  border: none !important;
-}
-
-.gr-button-secondary {
-  background: #ffffff !important;
-  color: var(--text) !important;
-  border: 1px solid var(--line) !important;
-}
-
-footer {
-  display: none !important;
-}
-
-@keyframes pulseEdge {
-  to { stroke-dashoffset: -40; }
-}
-
-@media (max-width: 1180px) {
-  .model-switcher,
-  .candidate-grid,
-  .papers-grid,
-  .journal-grid,
-  .parse-grid {
-    grid-template-columns: 1fr;
-  }
-
-  .brain-svg {
-    height: 460px;
-  }
+html,body,.gradio-container { background:#ffffff !important; font-family:Inter,ui-sans-serif,system-ui,sans-serif; }
+.gradio-container { max-width:1640px !important; padding:20px !important; }
+#dvnc-shell { border:1px solid var(--line); border-radius:28px; overflow:hidden; background:#ffffff; box-shadow:var(--shadow); padding:20px 22px 22px; }
+.hero-bar { display:flex; justify-content:space-between; align-items:center; gap:16px; padding-bottom:12px; border-bottom:1px solid rgba(0,0,0,.06); margin-bottom:16px; }
+.brand { display:flex; align-items:center; gap:14px; }
+.logo { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; color:var(--gold); background:linear-gradient(135deg,rgba(255,122,26,.12),rgba(23,184,166,.10)); border:1px solid rgba(0,0,0,.08); }
+.logo svg { width:24px; height:24px; }
+.brand h1 { font-size:1.05rem; margin:0; font-weight:700; letter-spacing:.12em; text-transform:uppercase; }
+.brand p { margin:3px 0 0; color:var(--muted); font-size:.84rem; }
+.status { display:flex; gap:10px; align-items:center; color:var(--soft); font-size:.85rem; }
+.status-dot { width:10px; height:10px; border-radius:50%; background:var(--teal); box-shadow:0 0 0 6px rgba(23,184,166,.10),0 0 14px rgba(23,184,166,.25); }
+.panel { background:#ffffff; border:1px solid var(--line); border-radius:22px; box-shadow:inset 0 1px 0 rgba(255,255,255,.8); }
+.querybox textarea,.querybox input { background:transparent !important; color:var(--text) !important; }
+.querybox,.querybox>div { background:#ffffff !important; border-radius:18px !important; border-color:var(--line) !important; }
+.chat-panel { padding:18px; min-height:280px; }
+.chat-thread { display:flex; flex-direction:column; gap:14px; }
+.bubble { max-width:88%; padding:16px 18px; border-radius:22px; border:1px solid var(--line); }
+.bubble p { margin:8px 0 0; line-height:1.6; font-size:.96rem; color:var(--text); }
+.bubble .role { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
+.bubble-user { align-self:flex-end; background:linear-gradient(135deg,rgba(98,141,255,.16),rgba(98,141,255,.08)); }
+.bubble-ai { align-self:flex-start; background:#ffffff; }
+.bubble-system { align-self:flex-start; background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,122,26,.04)); }
+.model-switcher { display:grid; grid-template-columns:repeat(3,1fr); gap:12px; }
+.model-pill { padding:14px; border:1px solid var(--line); border-radius:18px; display:flex; flex-direction:column; gap:4px; min-height:98px; background:#ffffff; }
+.model-pill.active { border-color:rgba(255,122,26,.40); background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,255,255,.96)); }
+.model-name { font-weight:650; color:var(--text); }
+.model-tag { font-size:.76rem; text-transform:uppercase; letter-spacing:.12em; color:var(--gold); }
+.model-pill small { color:var(--muted); line-height:1.45; }
+.brain-shell { padding:18px; }
+.brain-header { display:flex; justify-content:space-between; align-items:flex-end; gap:16px; margin-bottom:10px; }
+.eyebrow { font-size:.72rem; letter-spacing:.16em; text-transform:uppercase; color:var(--gold); margin:0 0 4px; }
+.brain-header h3 { margin:0; font-size:1.12rem; color:var(--text); }
+.brain-legend { display:flex; gap:14px; color:var(--muted); font-size:.8rem; flex-wrap:wrap; }
+.dot { width:10px; height:10px; display:inline-block; border-radius:50%; margin-right:6px; }
+.dot-live { background:var(--chosen); box-shadow:0 0 10px rgba(255,122,26,.35); }
+.dot-chosen { background:var(--chosen); }
+.dot-idle { background:var(--idle); border:1px solid var(--idle-stroke); }
+.dot-query { background:var(--query-node); border:1px solid #de9e58; }
+.dot-paper { background:var(--paper-node); border:1px solid #4fb3a5; }
+.dot-upload { background:var(--upload-node); border:1px solid #8f73d9; }
+.brain-stage { position:relative; min-height:420px; overflow:hidden; background:linear-gradient(180deg,rgba(250,250,250,1),rgba(255,255,255,1)); border:1px solid rgba(0,0,0,.05); border-radius:20px; }
+.brain-svg { width:100%; height:520px; display:block; }
+.edge { stroke:rgba(0,0,0,.12); stroke-width:2.4; }
+.edge.active { stroke:var(--chosen); stroke-width:4.2; stroke-linecap:round; filter:drop-shadow(0 0 6px rgba(255,122,26,.45)); stroke-dasharray:8 12; animation:pulseEdge 1.5s linear infinite; }
+.node { stroke-width:2.2; transition:all .25s ease; }
+.node.idle { fill:var(--idle); stroke:var(--idle-stroke); }
+.node.chosen { fill:var(--chosen); stroke:#ffb16d; }
+.halo { fill:none; }
+.halo.active { stroke:rgba(255,122,26,.18); stroke-width:12; }
+.label { fill:#2c2c2c; font-size:13px; font-weight:500; letter-spacing:.01em; }
+.label.active { fill:#111111; font-weight:700; }
+.learn-edge { stroke:rgba(0,0,0,.18); stroke-width:2.2; stroke-linecap:round; }
+.learn-node { stroke-width:2.2; }
+.learn-node.query { fill:var(--query-node); stroke:#de9e58; }
+.learn-node.paper { fill:var(--paper-node); stroke:#36a091; }
+.learn-node.upload { fill:var(--upload-node); stroke:#7e63cb; }
+.learn-label { fill:#1e1e1e; font-size:12px; font-weight:600; }
+.learning-empty { display:grid; place-items:center; }
+.empty-graph-copy { text-align:center; max-width:440px; padding:40px 20px; }
+.empty-graph-copy h4 { margin:0 0 10px; font-size:1.05rem; }
+.empty-graph-copy p { margin:0; color:var(--muted); line-height:1.6; }
+.timeline { display:flex; flex-direction:column; gap:10px; }
+.agent-step { border:1px solid var(--line); border-radius:18px; background:#ffffff; overflow:hidden; }
+.agent-summary { list-style:none; display:grid; grid-template-columns:42px 1fr; gap:12px; align-items:center; padding:12px; cursor:pointer; }
+.agent-summary::-webkit-details-marker { display:none; }
+.agent-index { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; font-weight:700; color:var(--gold); background:rgba(255,122,26,.08); border:1px solid rgba(255,122,26,.18); }
+.agent-head { display:flex; justify-content:space-between; gap:12px; align-items:center; }
+.agent-head h4 { margin:0; font-size:.98rem; color:var(--text); }
+.agent-head span { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
+.agent-copy { padding:0 14px 16px 66px; }
+.agent-copy p { margin:0; color:#2d2d2d; font-size:.93rem; line-height:1.6; }
+.candidate-grid { display:grid; grid-template-columns:repeat(3,minmax(0,1fr)); gap:18px; }
+.candidate-card { background:none; perspective:1400px; min-height:330px; }
+.candidate-card-inner { position:relative; width:100%; min-height:330px; transition:transform .8s cubic-bezier(.2,.7,.1,1); transform-style:preserve-3d; }
+.candidate-card:hover .candidate-card-inner,.candidate-card:focus .candidate-card-inner,.candidate-card:focus-within .candidate-card-inner { transform:rotateY(180deg); }
+.candidate-face { position:absolute; inset:0; padding:20px; border-radius:22px; border:1px solid var(--line); background:#ffffff; color:var(--text); backface-visibility:hidden; box-shadow:0 12px 24px rgba(0,0,0,.06); display:flex; flex-direction:column; gap:14px; }
+.candidate-back { transform:rotateY(180deg); }
+.candidate-top { display:flex; justify-content:space-between; align-items:center; gap:8px; }
+.chip { font-size:.72rem; text-transform:uppercase; letter-spacing:.12em; color:#0b6f66; padding:7px 10px; border-radius:999px; background:rgba(23,184,166,.08); border:1px solid rgba(23,184,166,.18); }
+.chip.alt { color:var(--gold); background:rgba(255,122,26,.08); border-color:rgba(255,122,26,.18); }
+.score { font-weight:700; color:var(--gold); }
+.candidate-face h4 { margin:0; font-size:1.08rem; line-height:1.35; }
+.candidate-face p { margin:0; color:#1e1e1e; line-height:1.65; font-size:.96rem; overflow-wrap:anywhere; }
+.meta-row { margin-top:auto; display:flex; justify-content:space-between; color:var(--muted); font-size:.88rem; gap:14px; }
+.mini { cursor:pointer; margin-top:8px; align-self:flex-start; color:var(--text); padding:10px 12px; border-radius:14px; border:1px solid var(--line); background:#ffffff; transition:all 0.2s; }
+.mini:hover { background: #f5f5f5; border-color: var(--chosen); }
+.papers-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
+.paper-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
+.paper-topline { display:flex; gap:8px; flex-wrap:wrap; margin-bottom:10px; }
+.paper-badge { font-size:.72rem; padding:6px 10px; border-radius:999px; background:rgba(98,141,255,.08); color:#3456b5; border:1px solid rgba(98,141,255,.18); }
+.paper-badge.alt { background:rgba(0,0,0,.04); color:#444; border-color:rgba(0,0,0,.08); }
+.doi-badge { background:rgba(255,122,26,.08); color:#8a4105; border-color:rgba(255,122,26,.18); }
+.paper-card h4 { margin:0 0 10px; line-height:1.35; font-size:1rem; }
+.paper-card p { margin:0 0 12px; line-height:1.6; color:#222; }
+.paper-links { display:flex; gap:12px; flex-wrap:wrap; }
+.paper-meta-stack { display:flex; flex-direction:column; gap:6px; color:#444; margin-bottom:12px; font-size:.9rem; }
+.paper-links a,.journal-card,.upload-note a { color:#0b63ce; text-decoration:none; }
+.journal-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
+.journal-card { border:1px solid var(--line); border-radius:18px; padding:16px; display:flex; justify-content:space-between; gap:14px; align-items:center; background:#ffffff; }
+.journal-card h4 { margin:0 0 6px; }
+.journal-card p { margin:0; color:var(--muted); line-height:1.5; }
+.upload-note { border:1px dashed rgba(0,0,0,.16); border-radius:18px; padding:16px; background:rgba(0,0,0,.015); color:#1f1f1f; line-height:1.6; }
+.prosebox { padding:18px; white-space:pre-wrap; font-family:ui-monospace,SFMono-Regular,Menlo,monospace; line-height:1.55; color:#1b1b1b; }
+.gr-button-primary { background:linear-gradient(135deg,rgba(255,122,26,.92),rgba(240,108,22,.92)) !important; color:#ffffff !important; border:none !important; }
+.gr-button-secondary { background:#ffffff !important; color:var(--text) !important; border:1px solid var(--line) !important; }
+.ref-list { margin:0; padding-left:18px; }
+.ref-list li { margin-bottom:8px; line-height:1.5; }
+.parse-grid { display:grid; grid-template-columns:1.2fr 1fr; gap:14px; }
+.parse-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
+.selection-panel { padding:18px; }
+footer { display:none !important; }
+@keyframes pulseEdge { to { stroke-dashoffset:-40; } }
+@media (max-width:1180px) {
+  .model-switcher,.candidate-grid,.papers-grid,.journal-grid,.parse-grid { grid-template-columns:1fr; }
+  .brain-svg { height:460px; }
 }
 """
-
-CSS = BASE_CSS + "\n" + get_dvnc_layout_css() + "\n" + get_discovery_css()
-
+CSS  = BASE_CSS + "\\n" + get_dvnc_layout_css()
 HEAD = """
 <link rel="preconnect" href="https://fonts.googleapis.com">
 <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
 <link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap" rel="stylesheet">
 <script>
 function triggerRouteSwap(idx) {
-  const container = document.getElementById("route_swap_payload");
-  if (!container) return;
-  const input = container.querySelector("textarea, input");
-  if (!input) return;
-
-  input.value = String(idx);
-  input.dispatchEvent(new Event("input", { bubbles: true }));
-  input.dispatchEvent(new Event("change", { bubbles: true }));
-
-  setTimeout(() => {
-    const btn = document.getElementById("route_swap_apply");
-    if (btn) btn.click();
-  }, 120);
+    const container = document.getElementById('route_swap_payload');
+    if(!container) return;
+    const input = container.querySelector('textarea') || container.querySelector('input');
+    if(input) {
+        input.value = idx.toString();
+        input.dispatchEvent(new Event('input', { bubbles: true }));
+        setTimeout(() => {
+            const btn = document.getElementById('route_swap_apply');
+            if(btn) btn.click();
+        }, 150);
+    }
 }
 </script>
 """
-
-
-# ── Gradio layout ────────────────────────────────────────────────────────────
-with gr.Blocks(
-    css=CSS,
-    head=HEAD,
-    theme=gr.themes.Base(),
-    fill_height=True,
-    title="DVNC.AI",
-) as demo:
-    # Shared state
-    papers_state = gr.State([])
-    parsed_pdf_state = gr.State({})
+# ── Gradio layout ─────────────────────────────────────────────────────────────
+with gr.Blocks(css=CSS, head=HEAD, theme=gr.themes.Base(), fill_height=True) as demo:
+    # ── Shared state ──────────────────────────────────────────────────────────
+    papers_state        = gr.State([])
+    parsed_pdf_state    = gr.State({})
     ingest_payload_state = gr.State({})
-    route_state = gr.State(get_default_route_state())
-
-    # Header
-    gr.HTML(
-        """
+    route_state         = gr.State(get_default_route_state())
+    # ── Header ────────────────────────────────────────────────────────────────
+    gr.HTML("""
         <div id="dvnc-shell">
           <div class="hero-bar">
             <div class="brand">
@@ -1274,14 +556,11 @@ with gr.Blocks(
             <div class="status"><span class="status-dot"></span><span>Live orchestration</span></div>
           </div>
         </div>
-        """
-    )
-
+    """)
     with gr.Tabs():
-        # ── Tab 1: Discovery Engine ──────────────────────────────────────────
+        # ── Tab 1 · Discovery Engine ──────────────────────────────────────────
         with gr.Tab("Discovery Engine"):
             model_html = gr.HTML(build_models_html("DVNC Sovereign"))
-
             with gr.Row():
                 with gr.Column(scale=2):
                     model = gr.Dropdown(
@@ -1296,11 +575,9 @@ with gr.Blocks(
                         lines=4,
                     )
                     with gr.Row():
-                        run_btn = gr.Button("Run discovery", variant="primary")
-                        example_btn = gr.Button("Load example", variant="secondary")
-
-                    chat = gr.HTML(
-                        """
+                        run_btn     = gr.Button("Run discovery",  variant="primary")
+                        example_btn = gr.Button("Load example",   variant="secondary")
+                    chat = gr.HTML("""
                         <div class="panel chat-panel">
                           <div class="chat-thread">
                             <div class="bubble bubble-ai">
@@ -1309,20 +586,15 @@ with gr.Blocks(
                             </div>
                           </div>
                         </div>
-                        """
-                    )
-
+                    """)
                 with gr.Column(scale=3):
                     connectome = gr.HTML(build_connectome_html(DEFAULT_PATH))
-                    cards = gr.HTML("")
-
-            output = gr.Markdown("# Discovery Output\n\nAwaiting query.")
+                    cards      = gr.HTML("")
+            output   = gr.Markdown("# Discovery Output\\n\\nAwaiting query.")
             timeline = gr.HTML(get_initial_discovery_timeline_html())
-
             route_swap_payload = gr.Textbox(value="", visible=False, elem_id="route_swap_payload")
-            route_swap_apply = gr.Button("Apply route swap", visible=False, elem_id="route_swap_apply")
-
-        # ── Tab 2: Self-Learning Graph ───────────────────────────────────────
+            route_swap_apply   = gr.Button("Apply route swap", visible=False, elem_id="route_swap_apply")
+        # ── Tab 2 · Self-Learning Graph ───────────────────────────────────────
         with gr.Tab("Self-Learning Graph"):
             with gr.Row():
                 with gr.Column(scale=2):
@@ -1342,86 +614,48 @@ with gr.Blocks(
                         value=DEFAULT_SOURCES,
                         label="Sources",
                     )
-                    pdf_upload = gr.File(
-                        label="Upload PDF papers",
-                        file_types=[".pdf"],
-                        file_count="single",
-                    )
-
+                    pdf_upload = gr.File(label="Upload PDF papers", file_types=[".pdf"], file_count="single")
                     with gr.Row():
-                        learn_btn = gr.Button("Discover papers", variant="primary")
-                        load_topic_btn = gr.Button("Load example topic", variant="secondary")
-
-                    upload_status = gr.Markdown("No PDF uploaded yet.")
+                        learn_btn      = gr.Button("Discover papers",      variant="primary")
+                        load_topic_btn = gr.Button("Load example topic",   variant="secondary")
+                    upload_status    = gr.Markdown("No PDF uploaded yet.")
                     discovery_status = gr.Markdown("### No discovery results yet.")
-                    journal_panel = gr.HTML(build_journal_html("biomaterials cardiac repair"))
-
-                    gr.HTML(
-                        '<div class="panel selection-panel"><h3 style="margin:0 0 12px;">Select papers to ingest</h3></div>'
-                    )
-
-                    selection_box = gr.CheckboxGroup(
-                        choices=[],
-                        value=[],
-                        label="Candidate papers",
-                    )
-
+                    journal_panel    = gr.HTML(build_journal_html("biomaterials cardiac repair"))
+                    gr.HTML(\'<div class="panel selection-panel"><h3 style="margin:0 0 12px;">Select papers to ingest</h3></div>\')
+                    selection_box = gr.CheckboxGroup(choices=[], value=[], label="Candidate papers")
                     parser_order = gr.CheckboxGroup(
-                        choices=PDF_PARSER_OPTIONS,
-                        value=PDF_PARSER_OPTIONS,
+                        choices=["grobid", "docling", "pymupdf"],
+                        value=["grobid", "docling", "pymupdf"],
                         label="Parser routing order",
                     )
-
                     with gr.Row():
-                        parse_btn = gr.Button("Parse uploaded PDF", variant="secondary")
+                        parse_btn  = gr.Button("Parse uploaded PDF",        variant="secondary")
                         ingest_btn = gr.Button("Ingest selected into graph", variant="primary")
-
                 with gr.Column(scale=3):
                     learning_graph = gr.HTML(build_learning_graph_html([], []))
-                    papers_panel = gr.HTML(
-                        '<div class="panel papers-panel" style="padding:18px"><p>Search by topic, title, DOI, or link, then select papers before graph ingestion.</p></div>'
-                    )
-                    parse_summary = gr.Markdown("### PDF parse status\n\nAwaiting upload.")
-                    parse_panel = gr.HTML(
-                        '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-                    )
-                    ingest_summary = gr.Markdown("### Graph ingest status\n\nAwaiting paper selection.")
-                    ingest_payload = gr.JSON(
-                        label="Graph ingest payload",
-                        value={"status": "empty", "nodes": [], "edges": []},
-                    )
-
-    # ── Event wiring ─────────────────────────────────────────────────────────
+                    papers_panel   = gr.HTML(\'<div class="panel papers-panel" style="padding:18px"><p>Search by topic, title, DOI, or link, then select papers before graph ingestion.</p></div>\')
+                    parse_summary  = gr.Markdown("### PDF parse status\\n\\nAwaiting upload.")
+                    parse_panel    = gr.HTML(\'<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>\')
+                    ingest_summary = gr.Markdown("### Graph ingest status\\n\\nAwaiting paper selection.")
+                    ingest_payload = gr.JSON(label="Graph ingest payload", value={"status": "empty", "nodes": [], "edges": []})
+    # ── Event wiring ──────────────────────────────────────────────────────────
     example_btn.click(fn=load_example, outputs=query)
-
     run_btn.click(
         fn=run_discovery,
         inputs=[query, model],
         outputs=[chat, connectome, timeline, cards, output, model_html, route_state],
     )
-
     route_swap_apply.click(
         fn=apply_route_swap,
         inputs=[query, model, route_swap_payload, route_state],
         outputs=[chat, connectome, timeline, output, route_state],
     )
-
     load_topic_btn.click(fn=load_paper_topic, outputs=paper_query)
-
     learn_btn.click(
         fn=run_paper_discovery,
         inputs=[paper_query, search_mode, source_selector, pdf_upload],
-        outputs=[
-            learning_graph,
-            papers_panel,
-            journal_panel,
-            upload_status,
-            selection_box,
-            papers_state,
-            discovery_status,
-        ],
+        outputs=[learning_graph, papers_panel, journal_panel, upload_status, selection_box, papers_state, discovery_status],
     )
-
     parse_btn.click(
         fn=parse_uploaded_pdf,
         inputs=[pdf_upload, parser_order],
@@ -1431,17 +665,14 @@ with gr.Blocks(
         inputs=[parsed_pdf_state],
         outputs=[parse_panel],
     )
-
     ingest_btn.click(
         fn=ingest_selected_papers,
         inputs=[paper_query, selection_box, papers_state, pdf_upload, parsed_pdf_state],
         outputs=[learning_graph, ingest_summary, ingest_payload],
-    ).then(
-        fn=lambda payload: payload,
-        inputs=[ingest_payload],
-        outputs=[ingest_payload_state],
     )
-
-
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()
+'''
+
+with open("app.py", "w") as f:
+    f.write(app_py)

dvnc_ai_v2_hf/deprecated/app_old10.py

@@ -1,677 +0,0 @@
-app_py = """
-DVNC.AI — app.py
-Refactored for functional "Use as main insight" logic with academic rigor.
-"""
-# ── Standard library ────────────────────────────────────────────────────────
-import html
-import json
-import math
-import os
-import random
-import re
-import urllib.parse
-import xml.etree.ElementTree as ET
-from pathlib import Path
-from typing import Dict, List, Optional
-from urllib.parse import quote
-# ── Third-party ──────────────────────────────────────────────────────────────
-import gradio as gr
-import requests
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-# ── Internal modules ─────────────────────────────────────────────────────────
-from dvnc_ai_v2_hf.agent_route_cards import build_agent_route_cards_html
-from dvnc_ai_v2_hf.discovery_app_bridge import (
-    get_default_route_state,
-    get_discovery_css,
-    get_initial_discovery_timeline_html,
-)
-from dvnc_ai_v2_hf.dvnc_ui_layout import get_dvnc_layout_css
-from dvnc_ai_v2_hf.self_learning_graph import (
-    DEFAULT_SOURCES,
-    SEARCH_MODES,
-    SOURCE_OPTIONS,
-    build_learning_graph_html,
-    build_journal_html,
-    ingest_selected_papers,
-    parse_uploaded_pdf,
-    render_parse_result,
-    run_paper_discovery,
-    safe_text,
-)
-# ── Constants ────────────────────────────────────────────────────────────────
-MODELS = [
-    {"name": "DVNC Sovereign", "tag": "flagship", "desc": "Maximum depth orchestration for frontier discovery"},
-    {"name": "DVNC Atlas",     "tag": "research",  "desc": "Balanced reasoning, graph traversal, and synthesis"},
-    {"name": "DVNC Curie",     "tag": "lab",        "desc": "Experimental hypothesis generation for anomalous signals"},
-]
-AGENTS = [
-    "Query Interpreter",
-    "Graph Divergence Mapper",
-    "Evidence Harvester",
-    "Analogy Engine",
-    "Hypothesis Composer",
-    "Adversarial Critic",
-    "Experimental Program Designer",
-]
-NODES = [
-    {"id": "seed",    "label": "Seed Query",          "group": "core",      "x": 10,  "y":   0, "z":   0},
-    {"id": "bio",     "label": "Biomaterials",         "group": "domain",    "x": 24,  "y":  12, "z":  -8},
-    {"id": "card",    "label": "Cardiac Repair",       "group": "domain",    "x": 38,  "y":   3, "z":  14},
-    {"id": "nano",    "label": "Nanostructure",        "group": "bridge",    "x": 24,  "y": -18, "z":  16},
-    {"id": "selfasm", "label": "Self-Assembly",        "group": "bridge",    "x": 40,  "y": -16, "z": -16},
-    {"id": "electro", "label": "Electro-signalling",   "group": "mechanism", "x": 58,  "y":  10, "z": -10},
-    {"id": "immune",  "label": "Immune Modulation",    "group": "mechanism", "x": 64,  "y":  -8, "z":  10},
-    {"id": "trial",   "label": "Validation Path",      "group": "outcome",   "x": 80,  "y":   0, "z":   0},
-    {"id": "alt1",    "label": "Piezoelectric Scaffold","group": "candidate", "x": 56,  "y":  26, "z":  14},
-    {"id": "alt2",    "label": "Peptide Mesh",          "group": "candidate", "x": 54,  "y": -27, "z": -14},
-]
-EDGES = [
-    ("seed",    "bio"),    ("seed",   "nano"),
-    ("bio",     "card"),   ("nano",   "selfasm"),
-    ("selfasm", "electro"),("card",   "immune"),
-    ("electro", "trial"),  ("immune", "trial"),
-    ("card",    "alt1"),   ("selfasm","alt2"),
-    ("alt1",    "trial"),  ("alt2",   "trial"),
-]
-DEFAULT_PATH = ["seed", "nano", "selfasm", "electro", "trial"]
-CANDIDATES = [
-    {
-        "title":   "Piezoelectric Scaffold Cascade",
-        "front":   "Use mechano-electric scaffolds to convert cardiac strain into micro-current signalling.",
-        "back":    "Discovery path: anomalous healing signal -> piezoelectric analog -> ion-channel entrainment -> tissue regeneration. Risk: power density and fibrosis coupling.",
-        "score":   92,
-        "novelty": "High",
-        "agent":   "Hypothesis Composer",
-    },
-    {
-        "title":   "Peptide Self-Assembly Mesh",
-        "front":   "Deploy dynamic peptide meshes that self-assemble around damaged myocardium and guide repair.",
-        "back":    "Discovery path: self-assembly -> local immune choreography -> regenerative substrate formation. Risk: degradation timing and targeting specificity.",
-        "score":   88,
-        "novelty": "High",
-        "agent":   "Analogy Engine",
-    },
-    {
-        "title":   "Immune-Tuned Conductive Hydrogel",
-        "front":   "Blend conductivity with macrophage-state modulation to reduce scarring and restore conduction.",
-        "back":    "Discovery path: inflammation mismatch -> conductive medium -> macrophage polarization -> synchronized healing. Risk: persistence and biocompatibility.",
-        "score":   85,
-        "novelty": "Medium-High",
-        "agent":   "Adversarial Critic",
-    },
-]
-ACADEMIC_INSIGHTS = [
-    {
-        "hypothesis": "Implementation of mechano-electric scaffolds to transduce cardiac strain into localized micro-current signalling for myocardial regeneration.",
-        "metrics": {"Novelty": 92, "Mechanistic clarity": 85, "Experimental tractability": 78, "Cross-domain distance": 94},
-        "outline": "1. Synthesize candidate piezoelectric biomaterial scaffolds with tunable strain-electric coupling.\\n2. Evaluate *in vitro* electromechanical transduction and subsequent ion-channel entrainment.\\n3. Conduct *in vivo* comparative models to assess regenerative efficacy against gold-standard substrates.\\n4. Rigorously validate to exclude pathological fibrosis and power-density toxicity.",
-        "path": ["seed", "bio", "card", "alt1", "trial"]
-    },
-    {
-        "hypothesis": "Deployment of dynamic peptide networks that self-assemble post-infarction to orchestrate local immunological responses and guide substrate regeneration.",
-        "metrics": {"Novelty": 88, "Mechanistic clarity": 82, "Experimental tractability": 86, "Cross-domain distance": 85},
-        "outline": "1. Formulate peptide sequences programmed for triggered *in situ* self-assembly within the myocardial infarct zone.\\n2. Quantify macrophage polarization and local immune choreography post-deployment.\\n3. Map the temporospatial degradation profile against *de novo* tissue formation.\\n4. Falsify against off-target aggregation and delayed clearance risks.",
-        "path": ["seed", "nano", "selfasm", "alt2", "trial"]
-    },
-    {
-        "hypothesis": "Integration of conductive hydrogels with immunomodulatory properties to simultaneously bridge electrical uncoupling and mitigate adverse fibrotic scarring.",
-        "metrics": {"Novelty": 85, "Mechanistic clarity": 90, "Experimental tractability": 88, "Cross-domain distance": 79},
-        "outline": "1. Fabricate biocompatible hydrogels featuring precisely tuned electrical conductivity and immunomodulatory motifs.\\n2. Monitor electrophysiological synchronization across the scaffold-tissue interface.\\n3. Assess macrophage state transitions and suppression of adverse fibrotic remodelling.\\n4. Validate long-term persistence, hemocompatibility, and mechanical integration.",
-        "path": ["seed", "bio", "card", "immune", "trial"]
-    }
-]
-JOURNALS = [
-    {"name": "Nature",     "url": "https://www.nature.com/search",                              "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science",    "url": "https://www.science.org/search",                             "desc": "High-impact science journal and family."},
-    {"name": "Cell",       "url": "https://www.cell.com/search",                                "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search",                           "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore","url": "https://ieeexplore.ieee.org/search/searchresult.jsp",        "desc": "Engineering, AI, signal processing, and systems."},
-]
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
-GROBID_URL               = os.getenv("GROBID_URL", "").strip()
-REQUEST_TIMEOUT          = 25
-# ── Utility helpers ──────────────────────────────────────────────────────────
-def safe_text(x, default: str = "") -> str:
-    return html.escape(str(x if x is not None else default))
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\\s+", " ", (x or "")).strip()
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    if re.match(r"^10\\.\\d{4,9}/[-._;()/:A-Z0-9]+$", q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-# ── HTML builders ─────────────────────────────────────────────────────────────
-def build_connectome_html(path_ids: List[str]) -> str:
-    active    = set(path_ids)
-    node_map  = {n["id"]: n for n in NODES}
-    path_pairs = {
-        pair
-        for i in range(len(path_ids) - 1)
-        for pair in [(path_ids[i], path_ids[i + 1]), (path_ids[i + 1], path_ids[i])]
-    }
-    base_lines, active_lines, circles, labels = [], [], [], []
-    for a, b in EDGES:
-        na, nb = node_map[a], node_map[b]
-        x1, y1 = na["x"] * 8 + 80, na["y"] * 6 + 280
-        x2, y2 = nb["x"] * 8 + 80, nb["y"] * 6 + 280
-        base_lines.append(f\'<line class="edge" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />\')
-        if (a, b) in path_pairs:
-            active_lines.append(f\'<line class="edge active" x1="{x1:.1f}" y1="{y1:.1f}" x2="{x2:.1f}" y2="{y2:.1f}" />\')
-    for n in NODES:
-        cx, cy   = n["x"] * 8 + 80, n["y"] * 6 + 280
-        is_active = n["id"] in active
-        state    = "chosen" if is_active else "idle"
-        halo_cls = "halo active" if is_active else "halo"
-        lbl_cls  = "label active" if is_active else "label"
-        radius   = 18 if is_active else 13
-        halo_r   = 30 if is_active else 0
-        circles.append(
-            f\'<g class="node-wrap">\'
-            f\'<circle class="{halo_cls}" cx="{cx:.1f}" cy="{cy:.1f}" r="{halo_r}" />\'
-            f\'<circle class="node {n["group"]} {state}" cx="{cx:.1f}" cy="{cy:.1f}" r="{radius}" />\'
-            f\'</g>\'
-        )
-        labels.append(f\'<text class="{lbl_cls}" x="{cx + 18:.1f}" y="{cy - 16:.1f}">{safe_text(n["label"])}</text>\')
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Connectome</p>
-          <h3>3D Connectome</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-live"></i> lit path</span>
-          <span><i class="dot dot-chosen"></i> chosen node</span>
-          <span><i class="dot dot-idle"></i> available node</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 780 560" class="brain-svg" role="img" aria-label="DVNC 3D connectome visualisation">
-          {"".join(base_lines)}
-          {"".join(active_lines)}
-          {"".join(circles)}
-          {"".join(labels)}
-        </svg>
-      </div>
-    </div>
-    """
-def build_cards_html(cards: List[Dict]) -> str:
-    items = []
-    for i, c in enumerate(cards):
-        items.append(f"""
-        <article class="candidate-card" tabindex="0">
-          <div class="candidate-card-inner">
-            <div class="candidate-face candidate-front">
-              <div class="candidate-top">
-                <span class="chip">{safe_text(c["agent"])}</span>
-                <span class="score">{safe_text(c["score"])}</span>
-              </div>
-              <h4>{safe_text(c["title"])}</h4>
-              <p>{safe_text(c["front"])}</p>
-              <div class="meta-row"><span>Novelty</span><strong>{safe_text(c["novelty"])}</strong></div>
-              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-            </div>
-            <div class="candidate-face candidate-back">
-              <div class="candidate-top">
-                <span class="chip alt">Alternative path</span>
-                <span class="score">{safe_text(c["score"])}</span>
-              </div>
-              <h4>{safe_text(c["title"])}</h4>
-              <p>{safe_text(c["back"])}</p>
-              <div class="meta-row"><span>Swap into route</span><strong>Enabled</strong></div>
-              <button class="mini" type="button" onclick="triggerRouteSwap('{i}')">Use as main insight</button>
-            </div>
-          </div>
-        </article>""")
-    return \'<div class="panel" style="padding:20px;"><div class="candidate-grid">\' + "".join(items) + "</div></div>"
-def build_agent_timeline(reasoning: List[Dict]) -> str:
-    rows = []
-    for r in reasoning:
-        rows.append(f"""
-        <details class="agent-step" {"open" if r["step"] == 1 else ""}>
-          <summary class="agent-summary">
-            <div class="agent-index">{safe_text(r["step"])}</div>
-            <div class="agent-head">
-              <h4>{safe_text(r["agent"])}</h4>
-              <span>{safe_text(r["tag"])}</span>
-            </div>
-          </summary>
-          <div class="agent-copy">
-            <p>{safe_text(r["summary"])}</p>
-          </div>
-        </details>""")
-    return \'<div class="panel" style="padding:18px;"><div class="timeline">\' + "".join(rows) + "</div></div>"
-def build_chat_html(query: str, result: Dict) -> str:
-    return f"""
-    <div class="panel chat-panel">
-      <div class="chat-thread">
-        <div class="bubble bubble-user">
-          <span class="role">You</span>
-          <p>{safe_text(query)}</p>
-        </div>
-        <div class="bubble bubble-ai">
-          <span class="role">DVNC Sovereign</span>
-          <p>{safe_text(result["summary"])}</p>
-        </div>
-        <div class="bubble bubble-system">
-          <span class="role">Discovery Signal</span>
-          <p><strong>Primary hypothesis:</strong> {safe_text(result["primary_hypothesis"])}</p>
-        </div>
-      </div>
-    </div>
-    """
-def build_models_html(selected: str) -> str:
-    items = []
-    for m in MODELS:
-        active = "active" if m["name"] == selected else ""
-        items.append(f"""
-            <div class="model-pill {active}">
-              <span class="model-name">{safe_text(m["name"])}</span>
-              <span class="model-tag">{safe_text(m["tag"])}</span>
-              <small>{safe_text(m["desc"])}</small>
-            </div>""")
-    return \'<div class="panel" style="padding:18px;"><div class="model-switcher">\' + "".join(items) + "</div></div>"
-# ── Discovery logic ───────────────────────────────────────────────────────────
-def run_discovery(query: str, model_name: str):
-    """
-    Runs the 7-agent discovery pipeline.
-    """
-    random.seed(len(query) + len(model_name))
-    if "curie" in query.lower() or "einstein" in query.lower():
-        primary = "Map the anomaly first, then force a distant analogy before composing the experimental programme."
-        path    = ["seed", "bio", "card", "immune", "trial"]
-    else:
-        primary = "Utilization of a self-assembling conductive scaffold to transduce mechanical strain into localized regenerative signalling pathways."
-        path    = DEFAULT_PATH
-    summaries = [
-        "Normalises the user prompt into a graph-searchable seed and isolates the tension inside the question.",
-        "Finds remote conceptual bridges instead of staying near the starting domain cluster.",
-        "Pulls evidence packets and conflict signals required for grounded hypothesis formation.",
-        "Generates cross-domain analogies with a bias toward mechanism transfer rather than keyword similarity.",
-        "Composes the lead hypothesis and two structurally different variants.",
-        "Attacks weak assumptions, hidden confounders, and feasibility gaps.",
-        "Produces a staged validation plan with measurable falsification criteria.",
-    ]
-    tags = ["input", "graph", "evidence", "analogy", "compose", "critique", "experiment"]
-    reasoning = [
-        {"step": i + 1, "agent": AGENTS[i], "tag": tags[i], "summary": summaries[i]}
-        for i in range(7)
-    ]
-    result = {
-        "summary":            "A deeper route was chosen through the connectome, with live alternatives preserved as swappable cards so the reasoning path can be inspected rather than hidden.",
-        "primary_hypothesis": primary,
-        "reasoning":          reasoning,
-        "cards":              CANDIDATES,
-        "path":               path,
-        "metrics": {
-            "Novelty":                   93,
-            "Mechanistic clarity":       89,
-            "Experimental tractability": 82,
-            "Cross-domain distance":     91,
-        },
-    }
-    chat_html       = build_chat_html(query, result)
-    connectome_html = build_connectome_html(path)
-    timeline_html   = build_agent_route_cards_html(reasoning)
-    metrics_md      = "\\n".join(f"- {k}: {v}/100" for k, v in result["metrics"].items())
-    hypothesis_md   = (
-        "# Discovery Output\\n\\n"
-        f"**Model:** {model_name}\\n\\n"
-        f"**Primary hypothesis:** {result['primary_hypothesis']}\\n\\n"
-        "## Scoring\\n"
-        f"{metrics_md}\\n\\n"
-        "## Experimental outline\\n"
-        "1. Construct the candidate material or protocol.\\n"
-        "2. Test mechanistic signal expression under controlled conditions.\\n"
-        "3. Compare against baseline and nearest-neighbour alternatives.\\n"
-        "4. Falsify using the adversarial risk criteria surfaced in the reasoning path.\\n"
-    )
-    cards_html  = build_cards_html(CANDIDATES)
-    route_state = get_default_route_state()
-    return chat_html, connectome_html, timeline_html, cards_html, hypothesis_md, build_models_html(model_name), route_state
-def apply_route_swap(query: str, model_name: str, route_swap_payload: str, route_state):
-    """
-    Called when a user clicks 'Use as main insight' on a candidate card.
-    Sanitizes the output, adopts academic rigor, updates the connectome and discovery output.
-    """
-    try:
-        idx = int(route_swap_payload)
-    except ValueError:
-        idx = 0
-    if not (0 <= idx < len(ACADEMIC_INSIGHTS)):
-        idx = 0
-    academic = ACADEMIC_INSIGHTS[idx]
-    
-    # Update Connectome
-    connectome_html = build_connectome_html(academic["path"])
-    
-    # Update Chat Feedback
-    result = {
-        "summary": "Main insight formally adopted. The connectome pathway and validation protocol have been realigned to the selected candidate methodology.",
-        "primary_hypothesis": academic["hypothesis"]
-    }
-    chat_html = build_chat_html(query, result)
-    # Format Oxford-tier markdown output
-    metrics_md = "\\n".join(f"- {k}: {v}/100" for k, v in academic["metrics"].items())
-    
-    hypothesis_md = (
-        "# Discovery Output\\n\\n"
-        f"**Model:** {model_name}\\n\\n"
-        f"**Primary hypothesis:** {academic['hypothesis']}\\n\\n"
-        "## Scoring\\n"
-        f"{metrics_md}\\n\\n"
-        "## Experimental outline\\n"
-        f"{academic['outline']}\\n"
-    )
-    # We return the new chat, new connectome, leave timeline alone (gr.update()), new output, new state
-    return chat_html, connectome_html, gr.update(), hypothesis_md, route_state
-# ── Example loaders ────────���──────────────────────────────────────────────────
-def load_example() -> str:
-    return "How could a self-assembling conductive biomaterial improve cardiac tissue regeneration by converting mechanical strain into repair signalling?"
-def load_paper_topic() -> str:
-    return "self-assembling conductive biomaterials for cardiac repair"
-# ── CSS / HEAD ────────────────────────────────────────────────────────────────
-BASE_CSS = r"""
-:root {
-  --bg: #ffffff; --panel: #ffffff; --line: rgba(0,0,0,.12);
-  --text: #111111; --muted: #5b5b5b; --soft: rgba(0,0,0,.62);
-  --gold: #ff6600; --teal: #17b8a6; --blue: #628dff;
-  --chosen: #ff7a1a; --idle: #b8d8ff; --idle-stroke: #5e8fe6;
-  --query-node: #ffd8b3; --paper-node: #d7f6f2; --upload-node: #e7defe;
-  --shadow: 0 16px 40px rgba(0,0,0,.12);
-}
-html,body,.gradio-container { background:#ffffff !important; font-family:Inter,ui-sans-serif,system-ui,sans-serif; }
-.gradio-container { max-width:1640px !important; padding:20px !important; }
-#dvnc-shell { border:1px solid var(--line); border-radius:28px; overflow:hidden; background:#ffffff; box-shadow:var(--shadow); padding:20px 22px 22px; }
-.hero-bar { display:flex; justify-content:space-between; align-items:center; gap:16px; padding-bottom:12px; border-bottom:1px solid rgba(0,0,0,.06); margin-bottom:16px; }
-.brand { display:flex; align-items:center; gap:14px; }
-.logo { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; color:var(--gold); background:linear-gradient(135deg,rgba(255,122,26,.12),rgba(23,184,166,.10)); border:1px solid rgba(0,0,0,.08); }
-.logo svg { width:24px; height:24px; }
-.brand h1 { font-size:1.05rem; margin:0; font-weight:700; letter-spacing:.12em; text-transform:uppercase; }
-.brand p { margin:3px 0 0; color:var(--muted); font-size:.84rem; }
-.status { display:flex; gap:10px; align-items:center; color:var(--soft); font-size:.85rem; }
-.status-dot { width:10px; height:10px; border-radius:50%; background:var(--teal); box-shadow:0 0 0 6px rgba(23,184,166,.10),0 0 14px rgba(23,184,166,.25); }
-.panel { background:#ffffff; border:1px solid var(--line); border-radius:22px; box-shadow:inset 0 1px 0 rgba(255,255,255,.8); }
-.querybox textarea,.querybox input { background:transparent !important; color:var(--text) !important; }
-.querybox,.querybox>div { background:#ffffff !important; border-radius:18px !important; border-color:var(--line) !important; }
-.chat-panel { padding:18px; min-height:280px; }
-.chat-thread { display:flex; flex-direction:column; gap:14px; }
-.bubble { max-width:88%; padding:16px 18px; border-radius:22px; border:1px solid var(--line); }
-.bubble p { margin:8px 0 0; line-height:1.6; font-size:.96rem; color:var(--text); }
-.bubble .role { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
-.bubble-user { align-self:flex-end; background:linear-gradient(135deg,rgba(98,141,255,.16),rgba(98,141,255,.08)); }
-.bubble-ai { align-self:flex-start; background:#ffffff; }
-.bubble-system { align-self:flex-start; background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,122,26,.04)); }
-.model-switcher { display:grid; grid-template-columns:repeat(3,1fr); gap:12px; }
-.model-pill { padding:14px; border:1px solid var(--line); border-radius:18px; display:flex; flex-direction:column; gap:4px; min-height:98px; background:#ffffff; }
-.model-pill.active { border-color:rgba(255,122,26,.40); background:linear-gradient(135deg,rgba(255,122,26,.10),rgba(255,255,255,.96)); }
-.model-name { font-weight:650; color:var(--text); }
-.model-tag { font-size:.76rem; text-transform:uppercase; letter-spacing:.12em; color:var(--gold); }
-.model-pill small { color:var(--muted); line-height:1.45; }
-.brain-shell { padding:18px; }
-.brain-header { display:flex; justify-content:space-between; align-items:flex-end; gap:16px; margin-bottom:10px; }
-.eyebrow { font-size:.72rem; letter-spacing:.16em; text-transform:uppercase; color:var(--gold); margin:0 0 4px; }
-.brain-header h3 { margin:0; font-size:1.12rem; color:var(--text); }
-.brain-legend { display:flex; gap:14px; color:var(--muted); font-size:.8rem; flex-wrap:wrap; }
-.dot { width:10px; height:10px; display:inline-block; border-radius:50%; margin-right:6px; }
-.dot-live { background:var(--chosen); box-shadow:0 0 10px rgba(255,122,26,.35); }
-.dot-chosen { background:var(--chosen); }
-.dot-idle { background:var(--idle); border:1px solid var(--idle-stroke); }
-.dot-query { background:var(--query-node); border:1px solid #de9e58; }
-.dot-paper { background:var(--paper-node); border:1px solid #4fb3a5; }
-.dot-upload { background:var(--upload-node); border:1px solid #8f73d9; }
-.brain-stage { position:relative; min-height:420px; overflow:hidden; background:linear-gradient(180deg,rgba(250,250,250,1),rgba(255,255,255,1)); border:1px solid rgba(0,0,0,.05); border-radius:20px; }
-.brain-svg { width:100%; height:520px; display:block; }
-.edge { stroke:rgba(0,0,0,.12); stroke-width:2.4; }
-.edge.active { stroke:var(--chosen); stroke-width:4.2; stroke-linecap:round; filter:drop-shadow(0 0 6px rgba(255,122,26,.45)); stroke-dasharray:8 12; animation:pulseEdge 1.5s linear infinite; }
-.node { stroke-width:2.2; transition:all .25s ease; }
-.node.idle { fill:var(--idle); stroke:var(--idle-stroke); }
-.node.chosen { fill:var(--chosen); stroke:#ffb16d; }
-.halo { fill:none; }
-.halo.active { stroke:rgba(255,122,26,.18); stroke-width:12; }
-.label { fill:#2c2c2c; font-size:13px; font-weight:500; letter-spacing:.01em; }
-.label.active { fill:#111111; font-weight:700; }
-.learn-edge { stroke:rgba(0,0,0,.18); stroke-width:2.2; stroke-linecap:round; }
-.learn-node { stroke-width:2.2; }
-.learn-node.query { fill:var(--query-node); stroke:#de9e58; }
-.learn-node.paper { fill:var(--paper-node); stroke:#36a091; }
-.learn-node.upload { fill:var(--upload-node); stroke:#7e63cb; }
-.learn-label { fill:#1e1e1e; font-size:12px; font-weight:600; }
-.learning-empty { display:grid; place-items:center; }
-.empty-graph-copy { text-align:center; max-width:440px; padding:40px 20px; }
-.empty-graph-copy h4 { margin:0 0 10px; font-size:1.05rem; }
-.empty-graph-copy p { margin:0; color:var(--muted); line-height:1.6; }
-.timeline { display:flex; flex-direction:column; gap:10px; }
-.agent-step { border:1px solid var(--line); border-radius:18px; background:#ffffff; overflow:hidden; }
-.agent-summary { list-style:none; display:grid; grid-template-columns:42px 1fr; gap:12px; align-items:center; padding:12px; cursor:pointer; }
-.agent-summary::-webkit-details-marker { display:none; }
-.agent-index { width:42px; height:42px; border-radius:14px; display:grid; place-items:center; font-weight:700; color:var(--gold); background:rgba(255,122,26,.08); border:1px solid rgba(255,122,26,.18); }
-.agent-head { display:flex; justify-content:space-between; gap:12px; align-items:center; }
-.agent-head h4 { margin:0; font-size:.98rem; color:var(--text); }
-.agent-head span { font-size:.72rem; letter-spacing:.12em; text-transform:uppercase; color:var(--muted); }
-.agent-copy { padding:0 14px 16px 66px; }
-.agent-copy p { margin:0; color:#2d2d2d; font-size:.93rem; line-height:1.6; }
-.candidate-grid { display:grid; grid-template-columns:repeat(3,minmax(0,1fr)); gap:18px; }
-.candidate-card { background:none; perspective:1400px; min-height:330px; }
-.candidate-card-inner { position:relative; width:100%; min-height:330px; transition:transform .8s cubic-bezier(.2,.7,.1,1); transform-style:preserve-3d; }
-.candidate-card:hover .candidate-card-inner,.candidate-card:focus .candidate-card-inner,.candidate-card:focus-within .candidate-card-inner { transform:rotateY(180deg); }
-.candidate-face { position:absolute; inset:0; padding:20px; border-radius:22px; border:1px solid var(--line); background:#ffffff; color:var(--text); backface-visibility:hidden; box-shadow:0 12px 24px rgba(0,0,0,.06); display:flex; flex-direction:column; gap:14px; }
-.candidate-back { transform:rotateY(180deg); }
-.candidate-top { display:flex; justify-content:space-between; align-items:center; gap:8px; }
-.chip { font-size:.72rem; text-transform:uppercase; letter-spacing:.12em; color:#0b6f66; padding:7px 10px; border-radius:999px; background:rgba(23,184,166,.08); border:1px solid rgba(23,184,166,.18); }
-.chip.alt { color:var(--gold); background:rgba(255,122,26,.08); border-color:rgba(255,122,26,.18); }
-.score { font-weight:700; color:var(--gold); }
-.candidate-face h4 { margin:0; font-size:1.08rem; line-height:1.35; }
-.candidate-face p { margin:0; color:#1e1e1e; line-height:1.65; font-size:.96rem; overflow-wrap:anywhere; }
-.meta-row { margin-top:auto; display:flex; justify-content:space-between; color:var(--muted); font-size:.88rem; gap:14px; }
-.mini { cursor:pointer; margin-top:8px; align-self:flex-start; color:var(--text); padding:10px 12px; border-radius:14px; border:1px solid var(--line); background:#ffffff; transition:all 0.2s; }
-.mini:hover { background: #f5f5f5; border-color: var(--chosen); }
-.papers-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
-.paper-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
-.paper-topline { display:flex; gap:8px; flex-wrap:wrap; margin-bottom:10px; }
-.paper-badge { font-size:.72rem; padding:6px 10px; border-radius:999px; background:rgba(98,141,255,.08); color:#3456b5; border:1px solid rgba(98,141,255,.18); }
-.paper-badge.alt { background:rgba(0,0,0,.04); color:#444; border-color:rgba(0,0,0,.08); }
-.doi-badge { background:rgba(255,122,26,.08); color:#8a4105; border-color:rgba(255,122,26,.18); }
-.paper-card h4 { margin:0 0 10px; line-height:1.35; font-size:1rem; }
-.paper-card p { margin:0 0 12px; line-height:1.6; color:#222; }
-.paper-links { display:flex; gap:12px; flex-wrap:wrap; }
-.paper-meta-stack { display:flex; flex-direction:column; gap:6px; color:#444; margin-bottom:12px; font-size:.9rem; }
-.paper-links a,.journal-card,.upload-note a { color:#0b63ce; text-decoration:none; }
-.journal-grid { display:grid; grid-template-columns:repeat(2,minmax(0,1fr)); gap:14px; }
-.journal-card { border:1px solid var(--line); border-radius:18px; padding:16px; display:flex; justify-content:space-between; gap:14px; align-items:center; background:#ffffff; }
-.journal-card h4 { margin:0 0 6px; }
-.journal-card p { margin:0; color:var(--muted); line-height:1.5; }
-.upload-note { border:1px dashed rgba(0,0,0,.16); border-radius:18px; padding:16px; background:rgba(0,0,0,.015); color:#1f1f1f; line-height:1.6; }
-.prosebox { padding:18px; white-space:pre-wrap; font-family:ui-monospace,SFMono-Regular,Menlo,monospace; line-height:1.55; color:#1b1b1b; }
-.gr-button-primary { background:linear-gradient(135deg,rgba(255,122,26,.92),rgba(240,108,22,.92)) !important; color:#ffffff !important; border:none !important; }
-.gr-button-secondary { background:#ffffff !important; color:var(--text) !important; border:1px solid var(--line) !important; }
-.ref-list { margin:0; padding-left:18px; }
-.ref-list li { margin-bottom:8px; line-height:1.5; }
-.parse-grid { display:grid; grid-template-columns:1.2fr 1fr; gap:14px; }
-.parse-card { border:1px solid var(--line); border-radius:18px; padding:16px; background:#ffffff; }
-.selection-panel { padding:18px; }
-footer { display:none !important; }
-@keyframes pulseEdge { to { stroke-dashoffset:-40; } }
-@media (max-width:1180px) {
-  .model-switcher,.candidate-grid,.papers-grid,.journal-grid,.parse-grid { grid-template-columns:1fr; }
-  .brain-svg { height:460px; }
-}
-"""
-CSS  = BASE_CSS + "\\n" + get_dvnc_layout_css()
-HEAD = """
-<link rel="preconnect" href="https://fonts.googleapis.com">
-<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
-<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700;800&display=swap" rel="stylesheet">
-<script>
-function triggerRouteSwap(idx) {
-    const container = document.getElementById('route_swap_payload');
-    if(!container) return;
-    const input = container.querySelector('textarea') || container.querySelector('input');
-    if(input) {
-        input.value = idx.toString();
-        input.dispatchEvent(new Event('input', { bubbles: true }));
-        setTimeout(() => {
-            const btn = document.getElementById('route_swap_apply');
-            if(btn) btn.click();
-        }, 150);
-    }
-}
-</script>
-"""
-# ── Gradio layout ─────────────────────────────────────────────────────────────
-with gr.Blocks(css=CSS, head=HEAD, theme=gr.themes.Base(), fill_height=True) as demo:
-    # ── Shared state ──────────────────────────────────────────────────────────
-    papers_state        = gr.State([])
-    parsed_pdf_state    = gr.State({})
-    ingest_payload_state = gr.State({})
-    route_state         = gr.State(get_default_route_state())
-    # ── Header ────────────────────────────────────────────────────────────────
-    gr.HTML("""
-        <div id="dvnc-shell">
-          <div class="hero-bar">
-            <div class="brand">
-              <div class="logo" aria-hidden="true">
-                <svg viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="1.7">
-                  <path d="M5 17L12 4l7 13"/>
-                  <path d="M8.5 12.5h7"/>
-                  <circle cx="12" cy="12" r="1.8" fill="currentColor" stroke="none"/>
-                </svg>
-              </div>
-              <div>
-                <h1>DVNC.AI</h1>
-                <p>Sovereign discovery instrument · connectome-native reasoning</p>
-              </div>
-            </div>
-            <div class="status"><span class="status-dot"></span><span>Live orchestration</span></div>
-          </div>
-        </div>
-    """)
-    with gr.Tabs():
-        # ── Tab 1 · Discovery Engine ──────────────────────────────────────────
-        with gr.Tab("Discovery Engine"):
-            model_html = gr.HTML(build_models_html("DVNC Sovereign"))
-            with gr.Row():
-                with gr.Column(scale=2):
-                    model = gr.Dropdown(
-                        choices=[m["name"] for m in MODELS],
-                        value="DVNC Sovereign",
-                        label="Model tier",
-                    )
-                    query = gr.Textbox(
-                        label="Discovery query",
-                        elem_classes=["querybox"],
-                        placeholder="Enter a scientific question, anomaly, or breakthrough direction…",
-                        lines=4,
-                    )
-                    with gr.Row():
-                        run_btn     = gr.Button("Run discovery",  variant="primary")
-                        example_btn = gr.Button("Load example",   variant="secondary")
-                    chat = gr.HTML("""
-                        <div class="panel chat-panel">
-                          <div class="chat-thread">
-                            <div class="bubble bubble-ai">
-                              <span class="role">DVNC</span>
-                              <p>Enter a query to activate the 7-agent discovery stack and illuminate the chosen path through the 3D connectome.</p>
-                            </div>
-                          </div>
-                        </div>
-                    """)
-                with gr.Column(scale=3):
-                    connectome = gr.HTML(build_connectome_html(DEFAULT_PATH))
-                    cards      = gr.HTML("")
-            output   = gr.Markdown("# Discovery Output\\n\\nAwaiting query.")
-            timeline = gr.HTML(get_initial_discovery_timeline_html())
-            route_swap_payload = gr.Textbox(value="", visible=False, elem_id="route_swap_payload")
-            route_swap_apply   = gr.Button("Apply route swap", visible=False, elem_id="route_swap_apply")
-        # ── Tab 2 · Self-Learning Graph ───────────────────────────────────────
-        with gr.Tab("Self-Learning Graph"):
-            with gr.Row():
-                with gr.Column(scale=2):
-                    paper_query = gr.Textbox(
-                        label="Research topic / title / DOI / link",
-                        elem_classes=["querybox"],
-                        placeholder="e.g. self-assembling conductive biomaterials for cardiac repair",
-                        lines=3,
-                    )
-                    search_mode = gr.Dropdown(
-                        choices=SEARCH_MODES,
-                        value="topic",
-                        label="Search mode",
-                    )
-                    source_selector = gr.CheckboxGroup(
-                        choices=SOURCE_OPTIONS,
-                        value=DEFAULT_SOURCES,
-                        label="Sources",
-                    )
-                    pdf_upload = gr.File(label="Upload PDF papers", file_types=[".pdf"], file_count="single")
-                    with gr.Row():
-                        learn_btn      = gr.Button("Discover papers",      variant="primary")
-                        load_topic_btn = gr.Button("Load example topic",   variant="secondary")
-                    upload_status    = gr.Markdown("No PDF uploaded yet.")
-                    discovery_status = gr.Markdown("### No discovery results yet.")
-                    journal_panel    = gr.HTML(build_journal_html("biomaterials cardiac repair"))
-                    gr.HTML(\'<div class="panel selection-panel"><h3 style="margin:0 0 12px;">Select papers to ingest</h3></div>\')
-                    selection_box = gr.CheckboxGroup(choices=[], value=[], label="Candidate papers")
-                    parser_order = gr.CheckboxGroup(
-                        choices=["grobid", "docling", "pymupdf"],
-                        value=["grobid", "docling", "pymupdf"],
-                        label="Parser routing order",
-                    )
-                    with gr.Row():
-                        parse_btn  = gr.Button("Parse uploaded PDF",        variant="secondary")
-                        ingest_btn = gr.Button("Ingest selected into graph", variant="primary")
-                with gr.Column(scale=3):
-                    learning_graph = gr.HTML(build_learning_graph_html([], []))
-                    papers_panel   = gr.HTML(\'<div class="panel papers-panel" style="padding:18px"><p>Search by topic, title, DOI, or link, then select papers before graph ingestion.</p></div>\')
-                    parse_summary  = gr.Markdown("### PDF parse status\\n\\nAwaiting upload.")
-                    parse_panel    = gr.HTML(\'<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>\')
-                    ingest_summary = gr.Markdown("### Graph ingest status\\n\\nAwaiting paper selection.")
-                    ingest_payload = gr.JSON(label="Graph ingest payload", value={"status": "empty", "nodes": [], "edges": []})
-    # ── Event wiring ──────────────────────────────────────────────────────────
-    example_btn.click(fn=load_example, outputs=query)
-    run_btn.click(
-        fn=run_discovery,
-        inputs=[query, model],
-        outputs=[chat, connectome, timeline, cards, output, model_html, route_state],
-    )
-    route_swap_apply.click(
-        fn=apply_route_swap,
-        inputs=[query, model, route_swap_payload, route_state],
-        outputs=[chat, connectome, timeline, output, route_state],
-    )
-    load_topic_btn.click(fn=load_paper_topic, outputs=paper_query)
-    learn_btn.click(
-        fn=run_paper_discovery,
-        inputs=[paper_query, search_mode, source_selector, pdf_upload],
-        outputs=[learning_graph, papers_panel, journal_panel, upload_status, selection_box, papers_state, discovery_status],
-    )
-    parse_btn.click(
-        fn=parse_uploaded_pdf,
-        inputs=[pdf_upload, parser_order],
-        outputs=[parse_summary, parsed_pdf_state],
-    ).then(
-        fn=render_parse_result,
-        inputs=[parsed_pdf_state],
-        outputs=[parse_panel],
-    )
-    ingest_btn.click(
-        fn=ingest_selected_papers,
-        inputs=[paper_query, selection_box, papers_state, pdf_upload, parsed_pdf_state],
-        outputs=[learning_graph, ingest_summary, ingest_payload],
-    )
-if __name__ == "__main__":
-    demo.launch()
-
-with open("app.py", "w") as f:
-    f.write(app_py)

dvnc_ai_v2_hf/deprecated/self_learning_graph_old.py

@@ -1,1001 +0,0 @@
-import html
-import os
-import re
-import urllib.parse
-import xml.etree.ElementTree as ET
-from pathlib import Path
-from typing import Dict, List, Optional
-from urllib.parse import quote
-
-import gradio as gr
-import requests
-
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-
-
-JOURNALS = [
-    {"name": "Nature", "url": "https://www.nature.com/search", "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science", "url": "https://www.science.org/search", "desc": "High-impact science journal and family."},
-    {"name": "Cell", "url": "https://www.cell.com/search", "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search", "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore", "url": "https://ieeexplore.ieee.org/search/searchresult.jsp", "desc": "Engineering, AI, signal processing, and systems."},
-]
-
-SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
-SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-DEFAULT_SOURCES = ["arxiv", "openalex", "crossref", "semantic_scholar", "europe_pmc"]
-
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
-GROBID_URL = os.getenv("GROBID_URL", "").strip()
-REQUEST_TIMEOUT = 25
-
-
-def safe_text(x, default=""):
-    return html.escape(str(x if x is not None else default))
-
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    doi_pattern = r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$"
-    if re.match(doi_pattern, q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-
-
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    if not nodes:
-        return """
-        <div class="panel brain-shell">
-          <div class="brain-header">
-            <div>
-              <p class="eyebrow">Learning Graph</p>
-              <h3>Self-Learning Knowledge Graph</h3>
-            </div>
-          </div>
-          <div class="brain-stage learning-empty">
-            <div class="empty-graph-copy">
-              <h4>No papers mapped yet</h4>
-              <p>Search papers, pick a topic, select candidates, or upload a PDF to grow the graph in real time.</p>
-            </div>
-          </div>
-        </div>
-        """
-
-    node_items = []
-    label_items = []
-    edge_items = []
-    coords = [(110, 110), (320, 80), (540, 130), (660, 270), (550, 410), (300, 450), (110, 340), (380, 260)]
-
-    graph_nodes = [dict(n) for n in nodes[:8]]
-    for i, node in enumerate(graph_nodes):
-        x, y = coords[i]
-        node["sx"] = x
-        node["sy"] = y
-
-    node_map = {n["id"]: n for n in graph_nodes}
-    for a, b in edges:
-        if a in node_map and b in node_map:
-            na = node_map[a]
-            nb = node_map[b]
-            edge_items.append(
-                f'<line class="learn-edge" x1="{na["sx"]}" y1="{na["sy"]}" x2="{nb["sx"]}" y2="{nb["sy"]}" />'
-            )
-
-    for node in graph_nodes:
-        kind = node.get("kind", "paper")
-        klass = f'learn-node {kind}'
-        radius = 24 if kind == "query" else 20
-        node_items.append(
-            f'<circle class="{klass}" cx="{node["sx"]}" cy="{node["sy"]}" r="{radius}" />'
-        )
-        label_items.append(
-            f'<text class="learn-label" x="{node["sx"] + 28}" y="{node["sy"] - 10}">{safe_text(node["label"][:44])}</text>'
-        )
-
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Learning Graph</p>
-          <h3>{safe_text(title)}</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-query"></i> query</span>
-          <span><i class="dot dot-paper"></i> paper</span>
-          <span><i class="dot dot-upload"></i> uploaded PDF</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 760 520" class="brain-svg" role="img" aria-label="Self-learning knowledge graph">
-          {''.join(edge_items)}
-          {''.join(node_items)}
-          {''.join(label_items)}
-        </svg>
-      </div>
-    </div>
-    """
-
-
-def build_journal_html(query):
-    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
-    rows = []
-    for journal in JOURNALS:
-        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
-        if "ieeexplore" in journal["url"]:
-            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
-        rows.append(
-            f"""
-            <a class="journal-card" href="{safe_text(url)}" target="_blank" rel="noopener noreferrer">
-              <div>
-                <h4>{safe_text(journal['name'])}</h4>
-                <p>{safe_text(journal['desc'])}</p>
-              </div>
-              <span>Open</span>
-            </a>
-            """
-        )
-    return '<div class="journal-grid">' + ''.join(rows) + '</div>'
-
-
-def search_arxiv(query, max_results=8):
-    encoded = urllib.parse.quote(query)
-    url = (
-        "http://export.arxiv.org/api/query?search_query=all:"
-        f"{encoded}&start=0&max_results={max_results}&sortBy=relevance&sortOrder=descending"
-    )
-    response = requests.get(url, timeout=REQUEST_TIMEOUT)
-    response.raise_for_status()
-    root = ET.fromstring(response.text)
-    ns = {"atom": "http://www.w3.org/2005/Atom"}
-    papers = []
-    for entry in root.findall("atom:entry", ns):
-        title = " ".join((entry.findtext("atom:title", default="", namespaces=ns) or "").split())
-        summary = " ".join((entry.findtext("atom:summary", default="", namespaces=ns) or "").split())
-        published = entry.findtext("atom:published", default="", namespaces=ns)
-        paper_id = entry.findtext("atom:id", default="", namespaces=ns)
-        authors = [a.findtext("atom:name", default="", namespaces=ns) for a in entry.findall("atom:author", ns)]
-        pdf_url = ""
-        for link in entry.findall("atom:link", ns):
-            if link.attrib.get("title") == "pdf":
-                pdf_url = link.attrib.get("href", "")
-                break
-        papers.append(
-            {
-                "id": paper_id or title,
-                "title": title,
-                "summary": summary,
-                "abstract": summary,
-                "published": published[:10],
-                "authors": [a for a in authors[:8] if a],
-                "authors_text": ", ".join([a for a in authors[:4] if a]),
-                "url": paper_id,
-                "pdf": pdf_url,
-                "doi": "",
-                "venue": "arXiv",
-                "year": published[:4] if published else "",
-                "source": "arxiv",
-                "score": 0.76,
-                "open_access": True,
-                "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
-            }
-        )
-    return papers
-
-
-def search_crossref(query, mode="topic", max_results=8):
-    headers = {"User-Agent": "dvnc-ai-space/0.1"}
-    if mode == "doi":
-        url = f"https://api.crossref.org/works/{quote(query)}"
-        response = requests.get(url, headers=headers, timeout=REQUEST_TIMEOUT)
-        if response.status_code != 200:
-            return []
-        items = [response.json().get("message", {})]
-    else:
-        params = {"rows": max_results}
-        if mode in ("title", "paper_name"):
-            params["query.title"] = query
-        else:
-            params["query.bibliographic"] = query
-        response = requests.get("https://api.crossref.org/works", params=params, headers=headers, timeout=REQUEST_TIMEOUT)
-        response.raise_for_status()
-        items = response.json().get("message", {}).get("items", [])
-
-    out = []
-    for item in items:
-        authors = []
-        for a in item.get("author", []) or []:
-            name = " ".join(filter(None, [a.get("given"), a.get("family")])).strip()
-            if name:
-                authors.append(name)
-        title = (item.get("title") or ["Untitled"])[0]
-        year = ""
-        for key in ["published-print", "published-online", "created"]:
-            if item.get(key, {}).get("date-parts"):
-                year = str(item[key]["date-parts"][0][0])
-                break
-        abstract = item.get("abstract") or ""
-        abstract = re.sub("<.*?>", "", abstract)
-        doi = item.get("DOI", "")
-        out.append({
-            "id": doi or title,
-            "title": norm_text(title),
-            "summary": norm_text(abstract)[:400],
-            "abstract": norm_text(abstract),
-            "published": year,
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("URL", ""),
-            "pdf": "",
-            "doi": doi,
-            "venue": (item.get("container-title") or [""])[0],
-            "year": year,
-            "source": "crossref",
-            "score": 0.72,
-            "open_access": None,
-            "external_ids": {"crossref": doi} if doi else {},
-        })
-    return out
-
-
-def search_openalex(query, mode="topic", max_results=8):
-    params = {"per-page": max_results}
-    if mode == "doi":
-        doi = query.lower().replace("https://doi.org/", "").replace("http://doi.org/", "")
-        params["filter"] = f"doi:https://doi.org/{doi}"
-    else:
-        params["search"] = query
-    response = requests.get("https://api.openalex.org/works", params=params, timeout=REQUEST_TIMEOUT)
-    response.raise_for_status()
-    items = response.json().get("results", [])
-    out = []
-    for item in items:
-        authors = []
-        for auth in item.get("authorships", [])[:8]:
-            author = auth.get("author") or {}
-            if author.get("display_name"):
-                authors.append(author["display_name"])
-        oa = item.get("open_access") or {}
-        doi = (item.get("doi") or "").replace("https://doi.org/", "")
-        out.append({
-            "id": item.get("id") or doi or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": "",
-            "abstract": "",
-            "published": str(item.get("publication_year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": (item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or "",
-            "pdf": oa.get("oa_url") or "",
-            "doi": doi,
-            "venue": ((item.get("primary_location") or {}).get("source") or {}).get("display_name") or "",
-            "year": str(item.get("publication_year") or ""),
-            "source": "openalex",
-            "score": 0.80,
-            "open_access": oa.get("is_oa"),
-            "external_ids": item.get("ids") or {},
-        })
-    return out
-
-
-def search_semantic_scholar(query, mode="topic", max_results=8):
-    headers = {}
-    if SEMANTIC_SCHOLAR_API_KEY:
-        headers["x-api-key"] = SEMANTIC_SCHOLAR_API_KEY
-    fields = "title,authors,year,abstract,venue,externalIds,url,openAccessPdf"
-    if mode == "doi":
-        doi = query.lower().replace("https://doi.org/", "").replace("http://doi.org/", "")
-        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{quote(doi)}"
-        response = requests.get(url, params={"fields": fields}, headers=headers, timeout=REQUEST_TIMEOUT)
-        if response.status_code != 200:
-            return []
-        items = [response.json()]
-    else:
-        response = requests.get(
-            "https://api.semanticscholar.org/graph/v1/paper/search",
-            params={"query": query, "limit": max_results, "fields": fields},
-            headers=headers,
-            timeout=REQUEST_TIMEOUT,
-        )
-        if response.status_code != 200:
-            return []
-        items = response.json().get("data", [])
-
-    out = []
-    for item in items:
-        external = item.get("externalIds") or {}
-        authors = [a.get("name") for a in item.get("authors", []) if a.get("name")]
-        out.append({
-            "id": external.get("CorpusId") or external.get("DOI") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(item.get("abstract", ""))[:400],
-            "abstract": norm_text(item.get("abstract", "")),
-            "published": str(item.get("year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("url") or "",
-            "pdf": (item.get("openAccessPdf") or {}).get("url") or "",
-            "doi": external.get("DOI", ""),
-            "venue": item.get("venue") or "",
-            "year": str(item.get("year") or ""),
-            "source": "semantic_scholar",
-            "score": 0.84,
-            "open_access": bool((item.get("openAccessPdf") or {}).get("url")),
-            "external_ids": external,
-        })
-    return out
-
-
-def search_europe_pmc(query, mode="topic", max_results=8):
-    epmc_query = f'DOI:"{query}"' if mode == "doi" else query
-    params = {
-        "query": epmc_query,
-        "format": "json",
-        "pageSize": max_results,
-        "resultType": "core",
-    }
-    response = requests.get("https://www.ebi.ac.uk/europepmc/webservices/rest/search", params=params, timeout=REQUEST_TIMEOUT)
-    if response.status_code != 200:
-        return []
-    items = response.json().get("resultList", {}).get("result", [])
-    out = []
-    for item in items:
-        author_string = item.get("authorString", "")
-        authors = [x.strip() for x in author_string.split(",")[:8] if x.strip()]
-        pmcid = item.get("pmcid", "")
-        pdf_url = f"https://europepmc.org/articles/{pmcid}?pdf=render" if pmcid else ""
-        landing_url = f"https://europepmc.org/article/PMC/{pmcid}" if pmcid else ""
-        out.append({
-            "id": item.get("id") or item.get("doi") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(item.get("abstractText", ""))[:400],
-            "abstract": norm_text(item.get("abstractText", "")),
-            "published": str(item.get("pubYear") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": landing_url,
-            "pdf": pdf_url,
-            "doi": item.get("doi", ""),
-            "venue": item.get("journalTitle", ""),
-            "year": str(item.get("pubYear") or ""),
-            "source": "europe_pmc",
-            "score": 0.78,
-            "open_access": bool(pmcid),
-            "external_ids": {"pmid": item.get("pmid"), "pmcid": pmcid},
-        })
-    return out
-
-
-def resolve_link(query):
-    url = (query or "").strip()
-    if not url:
-        return []
-    try:
-        response = requests.get(
-            url,
-            timeout=REQUEST_TIMEOUT,
-            allow_redirects=True,
-            headers={"User-Agent": "dvnc-ai-space/0.1"},
-        )
-        content_type = response.headers.get("content-type", "")
-        if "pdf" in content_type or url.lower().endswith(".pdf"):
-            name = Path(url.split("?")[0]).name or "linked-paper.pdf"
-            return [{
-                "id": url,
-                "title": name,
-                "summary": "Direct PDF link detected.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": url,
-                "doi": "",
-                "venue": "Direct PDF",
-                "year": "",
-                "source": "link",
-                "score": 0.66,
-                "open_access": True,
-                "external_ids": {},
-            }]
-
-        if BeautifulSoup is None:
-            return [{
-                "id": url,
-                "title": url,
-                "summary": "Web page resolved. Install beautifulsoup4 for DOI extraction.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": "",
-                "doi": "",
-                "venue": "Web Link",
-                "year": "",
-                "source": "link",
-                "score": 0.48,
-                "open_access": None,
-                "external_ids": {},
-            }]
-
-        soup = BeautifulSoup(response.text, "html.parser")
-        doi = ""
-        for meta_name in ["citation_doi", "dc.identifier", "dc.Identifier"]:
-            tag = soup.find("meta", attrs={"name": meta_name})
-            if tag and tag.get("content"):
-                doi = tag["content"].strip()
-                break
-
-        title = soup.title.text.strip() if soup.title else url
-        pdf_link = ""
-        for a in soup.find_all("a", href=True):
-            href = a["href"]
-            if ".pdf" in href.lower():
-                pdf_link = href if href.startswith("http") else ""
-                break
-
-        if doi:
-            results = search_crossref(doi, mode="doi", max_results=1)
-            if results:
-                if pdf_link and not results[0].get("pdf"):
-                    results[0]["pdf"] = pdf_link
-                if url and not results[0].get("url"):
-                    results[0]["url"] = url
-                return results
-
-        return [{
-            "id": url,
-            "title": title,
-            "summary": "Landing page resolved from direct link.",
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": pdf_link,
-            "doi": doi,
-            "venue": "Web Link",
-            "year": "",
-            "source": "link",
-            "score": 0.54,
-            "open_access": bool(pdf_link),
-            "external_ids": {},
-        }]
-    except Exception as e:
-        return [{
-            "id": url,
-            "title": "Link resolution error",
-            "summary": str(e),
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": "",
-            "doi": "",
-            "venue": "Link",
-            "year": "",
-            "source": "link",
-            "score": 0.20,
-            "open_access": None,
-            "external_ids": {},
-        }]
-
-
-def dedupe_papers(items: List[Dict]) -> List[Dict]:
-    seen = {}
-    for item in items:
-        key = (
-            (item.get("doi") or "").lower().strip()
-            or (item.get("external_ids") or {}).get("arxiv")
-            or norm_text(item.get("title", "")).lower()
-            or item.get("id")
-        )
-        if not key:
-            key = f"{item.get('source')}::{item.get('title')}"
-        if key not in seen or float(item.get("score", 0)) > float(seen[key].get("score", 0)):
-            seen[key] = item
-    return sorted(seen.values(), key=lambda x: float(x.get("score", 0)), reverse=True)
-
-
-def discover_papers(query, mode, sources, max_results=10):
-    query = (query or "").strip()
-    if not query:
-        return []
-
-    mode = detect_query_type(query) if mode == "autonomous_web" else mode
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    results = []
-
-    if mode == "link":
-        return dedupe_papers(resolve_link(query))
-
-    try:
-        if "arxiv" in selected_sources and mode != "doi":
-            results.extend(search_arxiv(query, max_results=max_results))
-    except Exception:
-        pass
-    try:
-        if "crossref" in selected_sources:
-            results.extend(search_crossref(query, mode=mode, max_results=max_results))
-    except Exception:
-        pass
-    try:
-        if "openalex" in selected_sources:
-            results.extend(search_openalex(query, mode=mode, max_results=max_results))
-    except Exception:
-        pass
-    try:
-        if "semantic_scholar" in selected_sources:
-            results.extend(search_semantic_scholar(query, mode=mode, max_results=max_results))
-    except Exception:
-        pass
-    try:
-        if "europe_pmc" in selected_sources:
-            results.extend(search_europe_pmc(query, mode=mode, max_results=max_results))
-    except Exception:
-        pass
-
-    return dedupe_papers(results)
-
-
-def format_papers_html(papers):
-    if not papers:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No papers found yet.</p></div>'
-
-    items = []
-    for i, paper in enumerate(papers, start=1):
-        summary = safe_text((paper.get("summary") or "")[:280])
-        doi_line = f'<span class="paper-badge doi-badge">{safe_text(paper.get("doi"))}</span>' if paper.get("doi") else ""
-        pdf_link = paper.get("pdf") or "#"
-        abs_link = paper.get("url") or "#"
-        items.append(
-            f"""
-            <article class="paper-card">
-              <div class="paper-topline">
-                <span class="paper-badge">{safe_text(paper.get('source', 'paper'))}</span>
-                <span class="paper-badge alt">{safe_text(paper.get('published', '') or 'Paper')}</span>
-                {doi_line}
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{summary or 'No abstract snippet available.'}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-                <div><strong>Venue:</strong> {safe_text(paper.get('venue', 'Unknown venue'))}</div>
-                <div><strong>Score:</strong> {safe_text(round(float(paper.get('score', 0)), 3))}</div>
-              </div>
-              <div class="paper-links">
-                <a href="{safe_text(abs_link)}" target="_blank" rel="noopener noreferrer">Abstract</a>
-                <a href="{safe_text(pdf_link)}" target="_blank" rel="noopener noreferrer">PDF</a>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(items) + '</div>'
-
-
-def format_selection_choices(papers):
-    choices = []
-    for i, paper in enumerate(papers):
-        label = f"[{paper.get('source', 'src')}] {paper.get('title', 'Untitled')} — {paper.get('authors_text', 'Unknown authors')[:80]}"
-        choices.append((label, str(i)))
-    return choices
-
-
-def uploaded_pdf_summary(file_obj):
-    if not file_obj:
-        return "No PDF uploaded yet."
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    p = Path(path)
-    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, and references."
-
-
-def build_learning_graph_state(query, papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-    for i, paper in enumerate(papers[:5], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper["title"], "kind": "paper"})
-        edges.append(("query", pid))
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append(("query", "upload"))
-        if len(nodes) > 2:
-            edges.append(("upload", "paper_1"))
-    return nodes, edges
-
-
-def graph_from_selected(query, selected_papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-    for i, paper in enumerate(selected_papers[:6], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper["title"], "kind": "paper"})
-        edges.append(("query", pid))
-        if paper.get("doi"):
-            doi_id = f"doi_{i}"
-            nodes.append({"id": doi_id, "label": f"DOI {paper['doi']}", "kind": "paper"})
-            edges.append((pid, doi_id))
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append(("query", "upload"))
-        if len(selected_papers) > 0:
-            edges.append(("upload", "paper_1"))
-    return nodes, edges
-
-
-def parse_pdf_with_grobid(pdf_path):
-    if not GROBID_URL:
-        raise RuntimeError("GROBID_URL is not set")
-
-    with open(pdf_path, "rb") as f:
-        files = {"input": (Path(pdf_path).name, f, "application/pdf")}
-        response = requests.post(
-            f"{GROBID_URL.rstrip('/')}/api/processFulltextDocument",
-            files=files,
-            data={"includeRawAffiliations": "1", "segmentSentences": "1"},
-            timeout=120,
-        )
-
-    response.raise_for_status()
-    tei_xml = response.text
-
-    if BeautifulSoup is None:
-        return {
-            "parser": "grobid",
-            "title": Path(pdf_path).name,
-            "abstract": "",
-            "authors": [],
-            "sections": [],
-            "references": [],
-            "raw_text": "",
-            "tei_xml": tei_xml[:30000],
-        }
-
-    soup = BeautifulSoup(tei_xml, "xml")
-    title_stmt = soup.find("titleStmt")
-    title_tag = title_stmt.find("title") if title_stmt else soup.find("title")
-    abstract_tag = soup.find("abstract")
-
-    authors = []
-    for author in soup.find_all("author"):
-        pers = author.find("persName")
-        if pers:
-            name = " ".join(
-                x.get_text(" ", strip=True)
-                for x in pers.find_all(["forename", "surname"])
-            ).strip()
-            if name:
-                authors.append(name)
-
-    sections = []
-    for div in soup.find_all("div"):
-        head = div.find("head")
-        paras = [p.get_text(" ", strip=True) for p in div.find_all("p")]
-        text = "\n".join([p for p in paras if p.strip()])
-        if head and text.strip():
-            sections.append({"heading": head.get_text(" ", strip=True), "text": text[:4000]})
-
-    references = []
-    for bibl in soup.find_all("biblStruct")[:30]:
-        ref_title = ""
-        ref_doi = ""
-        title_node = bibl.find("title")
-        if title_node:
-            ref_title = title_node.get_text(" ", strip=True)
-        doi_node = bibl.find("idno", attrs={"type": "DOI"})
-        if doi_node:
-            ref_doi = doi_node.get_text(" ", strip=True)
-        references.append({"title": ref_title, "doi": ref_doi})
-
-    return {
-        "parser": "grobid",
-        "title": title_tag.get_text(" ", strip=True) if title_tag else Path(pdf_path).name,
-        "abstract": abstract_tag.get_text(" ", strip=True) if abstract_tag else "",
-        "authors": authors,
-        "sections": sections,
-        "references": references,
-        "raw_text": "",
-        "tei_xml": tei_xml[:60000],
-    }
-
-
-def parse_pdf_with_pymupdf(pdf_path):
-    if fitz is None:
-        raise RuntimeError("PyMuPDF not installed")
-
-    doc = fitz.open(pdf_path)
-    pages = [page.get_text("text") for page in doc]
-    raw_text = "\n".join(pages).strip()
-    first_page = raw_text[:4000]
-    lines = [x.strip() for x in first_page.splitlines() if x.strip()]
-    title = lines[0][:300] if lines else Path(pdf_path).name
-
-    abstract = ""
-    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n1[\.\s]|introduction)", raw_text, re.I | re.S)
-    if match:
-        abstract = norm_text(match.group(1))[:2000]
-
-    return {
-        "parser": "pymupdf",
-        "title": title,
-        "abstract": abstract,
-        "authors": [],
-        "sections": [{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else [],
-        "references": [],
-        "raw_text": raw_text[:50000],
-        "tei_xml": "",
-    }
-
-
-def parse_pdf_with_docling(pdf_path):
-    try:
-        from docling.document_converter import DocumentConverter
-    except Exception as e:
-        raise RuntimeError(f"Docling import failed: {e}")
-
-    converter = DocumentConverter()
-    result = converter.convert(pdf_path)
-    doc = result.document
-    markdown = doc.export_to_markdown()
-
-    title = Path(pdf_path).name
-    first_nonempty = next((line.strip("# ").strip() for line in markdown.splitlines() if line.strip()), "")
-    if first_nonempty:
-        title = first_nonempty[:300]
-
-    return {
-        "parser": "docling",
-        "title": title,
-        "abstract": "",
-        "authors": [],
-        "sections": [{"heading": "Document", "text": markdown[:12000]}] if markdown else [],
-        "references": [],
-        "raw_text": markdown[:50000],
-        "tei_xml": "",
-    }
-
-
-def parse_uploaded_pdf(file_obj, parser_order):
-    if not file_obj:
-        return "No PDF uploaded yet.", {}
-
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    parser_order = ensure_list(parser_order) or ["grobid", "docling", "pymupdf"]
-    errors = []
-
-    for parser_name in parser_order:
-        try:
-            if parser_name == "grobid":
-                result = parse_pdf_with_grobid(path)
-            elif parser_name == "docling":
-                result = parse_pdf_with_docling(path)
-            elif parser_name == "pymupdf":
-                result = parse_pdf_with_pymupdf(path)
-            else:
-                continue
-
-            summary = (
-                f"### Parsed PDF\n\n"
-                f"- Parser used: {result['parser']}\n"
-                f"- Title: {result.get('title') or 'Unknown'}\n"
-                f"- Authors: {', '.join(result.get('authors')[:6]) if result.get('authors') else 'Unknown'}\n"
-                f"- Abstract found: {'Yes' if result.get('abstract') else 'No'}\n"
-                f"- Sections extracted: {len(result.get('sections') or [])}\n"
-                f"- References extracted: {len(result.get('references') or [])}\n"
-            )
-            return summary, result
-        except Exception as e:
-            errors.append(f"{parser_name}: {e}")
-
-    fail_summary = "### PDF parsing failed\n\n" + "\n".join([f"- {x}" for x in errors])
-    return fail_summary, {"parser": None, "errors": errors}
-
-
-def render_parse_result(parsed):
-    if not parsed:
-        return '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-
-    sections_html = []
-    for section in parsed.get("sections", [])[:6]:
-        sections_html.append(
-            f"""
-            <details class="agent-step">
-              <summary class="agent-summary">
-                <div class="agent-index">§</div>
-                <div class="agent-head">
-                  <h4>{safe_text(section.get('heading', 'Section'))}</h4>
-                  <span>section</span>
-                </div>
-              </summary>
-              <div class="agent-copy">
-                <p>{safe_text(section.get('text', '')[:1800])}</p>
-              </div>
-            </details>
-            """
-        )
-
-    refs = parsed.get("references", [])[:12]
-    refs_html = "".join(
-        f"<li>{safe_text(r.get('title') or 'Untitled')} {'· DOI ' + safe_text(r.get('doi')) if r.get('doi') else ''}</li>"
-        for r in refs
-    ) or "<li>No references extracted.</li>"
-
-    title = safe_text(parsed.get("title") or "Parsed document")
-    abstract = safe_text((parsed.get("abstract") or "")[:2400]) or "No abstract extracted."
-    parser_name = safe_text(parsed.get("parser") or "unknown")
-
-    return f"""
-    <div class="panel" style="padding:18px">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">PDF Parse</p>
-          <h3>{title}</h3>
-        </div>
-        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name}</span></div>
-      </div>
-      <div class="parse-grid">
-        <div class="parse-card">
-          <h4>Abstract</h4>
-          <p>{abstract}</p>
-        </div>
-        <div class="parse-card">
-          <h4>References</h4>
-          <ul class="ref-list">{refs_html}</ul>
-        </div>
-      </div>
-      <div class="timeline" style="margin-top:14px;">
-        {''.join(sections_html) if sections_html else '<div class="panel" style="padding:16px;"><p>No sections extracted.</p></div>'}
-      </div>
-    </div>
-    """
-
-
-def build_ingest_payload(query, selected_papers, parsed_pdf=None):
-    nodes = [{"id": "topic:query", "type": "Topic", "label": query or "Research topic"}]
-    edges = []
-
-    for i, p in enumerate(selected_papers, start=1):
-        paper_id = p.get("doi") or (p.get("external_ids") or {}).get("arxiv") or f"paper:{i}"
-        nodes.append({
-            "id": paper_id,
-            "type": "Paper",
-            "title": p.get("title"),
-            "year": p.get("year"),
-            "venue": p.get("venue"),
-            "doi": p.get("doi"),
-            "source": p.get("source"),
-            "url": p.get("url"),
-            "pdf": p.get("pdf"),
-        })
-        edges.append({"source": "topic:query", "target": paper_id, "type": "ABOUT", "weight": p.get("score", 0)})
-
-        for author in p.get("authors", [])[:8]:
-            author_id = f"author:{author.lower()}"
-            nodes.append({"id": author_id, "type": "Author", "label": author})
-            edges.append({"source": paper_id, "target": author_id, "type": "WRITTEN_BY"})
-
-    if parsed_pdf and parsed_pdf.get("title"):
-        doc_id = "upload:pdf"
-        nodes.append({
-            "id": doc_id,
-            "type": "UploadedPDF",
-            "title": parsed_pdf.get("title"),
-            "parser": parsed_pdf.get("parser"),
-        })
-        edges.append({"source": "topic:query", "target": doc_id, "type": "UPLOADED_SOURCE"})
-        for idx, section in enumerate(parsed_pdf.get("sections", [])[:8], start=1):
-            sec_id = f"{doc_id}:section:{idx}"
-            nodes.append({"id": sec_id, "type": "Section", "label": section.get("heading") or f"Section {idx}"})
-            edges.append({"source": doc_id, "target": sec_id, "type": "HAS_SECTION"})
-        for idx, ref in enumerate(parsed_pdf.get("references", [])[:12], start=1):
-            ref_id = f"{doc_id}:ref:{idx}"
-            nodes.append({"id": ref_id, "type": "Reference", "label": ref.get("title") or f"Reference {idx}", "doi": ref.get("doi")})
-            edges.append({"source": doc_id, "target": ref_id, "type": "CITES"})
-
-    return {"status": "ok", "nodes": nodes, "edges": edges}
-
-
-def run_paper_discovery(query, search_mode, sources, pdf_file):
-    query_text = (query or "").strip()
-
-    if not query_text and not pdf_file:
-        empty_graph = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            empty_graph,
-            '<div class="panel papers-panel" style="padding:18px"><p>Enter a topic, title, DOI, link, or upload a PDF to start learning.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            "No PDF uploaded yet.",
-            gr.update(choices=[], value=[]),
-            [],
-            "### No discovery results yet.",
-        )
-
-    papers = []
-    if query_text:
-        try:
-            papers = discover_papers(query_text, search_mode, sources, max_results=10)
-        except Exception as e:
-            graph_nodes, graph_edges = build_learning_graph_state(
-                query_text,
-                [],
-                Path(getattr(pdf_file, "name", "uploaded.pdf")).name if pdf_file else None,
-            )
-            error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
-            return (
-                build_learning_graph_html(graph_nodes, graph_edges),
-                error_html,
-                build_journal_html(query_text or "biomaterials cardiac repair"),
-                uploaded_pdf_summary(pdf_file),
-                gr.update(choices=[], value=[]),
-                [],
-                f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
-            )
-
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-    graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:5], uploaded_name)
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges)
-    papers_html = format_papers_html(papers)
-    journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
-    pdf_summary = uploaded_pdf_summary(pdf_file)
-    choices = format_selection_choices(papers)
-
-    status_md = (
-        f"### Discovery results\n\n"
-        f"- Search mode: {search_mode}\n"
-        f"- Sources: {', '.join(ensure_list(sources) or DEFAULT_SOURCES)}\n"
-        f"- Candidates found: {len(papers)}\n"
-        f"- Select papers below, then click **Ingest selected into graph**.\n"
-    )
-    return graph_html, papers_html, journals_html, pdf_summary, gr.update(choices=choices, value=[]), papers, status_md
-
-
-def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
-    papers = ensure_list(papers_state)
-    selected_indices = ensure_list(selected_indices)
-
-    selected = []
-    for idx in selected_indices:
-        try:
-            selected.append(papers[int(idx)])
-        except Exception:
-            pass
-
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    if not selected and not parsed_state:
-        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return graph_html, "### Nothing ingested yet.\n\nSelect papers or parse an uploaded PDF first.", {"status": "empty", "nodes": [], "edges": []}
-
-    graph_nodes, graph_edges = graph_from_selected(query, selected, uploaded_name)
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
-    payload = build_ingest_payload(query, selected, parsed_state if isinstance(parsed_state, dict) else None)
-
-    summary_lines = [
-        "### Graph ingest ready",
-        "",
-        f"- Topic: {query or 'Research topic'}",
-        f"- Selected papers: {len(selected)}",
-        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
-        f"- Nodes created: {len(payload['nodes'])}",
-        f"- Edges created: {len(payload['edges'])}",
-    ]
-    return graph_html, "\n".join(summary_lines), payload

dvnc_ai_v2_hf/deprecated/self_learning_graph_old3.py

@@ -1,1626 +0,0 @@
-import html
-import math
-import os
-import re
-import time
-import urllib.parse
-import xml.etree.ElementTree as ET
-from collections import Counter, defaultdict
-from pathlib import Path
-from typing import Dict, List, Optional, Tuple
-from urllib.parse import quote
-
-import gradio as gr
-import requests
-
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-
-
-JOURNALS = [
-    {"name": "Nature", "url": "https://www.nature.com/search", "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science", "url": "https://www.science.org/search", "desc": "High-impact science journal and family."},
-    {"name": "Cell", "url": "https://www.cell.com/search", "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search", "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore", "url": "https://ieeexplore.ieee.org/search/searchresult.jsp", "desc": "Engineering, AI, signal processing, and systems."},
-]
-
-SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
-SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-DEFAULT_SOURCES = ["arxiv", "openalex", "crossref", "semantic_scholar", "europe_pmc"]
-
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
-GROBID_URL = os.getenv("GROBID_URL", "").strip()
-REQUEST_TIMEOUT = int(os.getenv("REQUEST_TIMEOUT", "25"))
-GRAPH_MAX_ROUNDS = int(os.getenv("GRAPH_MAX_ROUNDS", "2"))
-GRAPH_MAX_RESULTS_PER_SOURCE = int(os.getenv("GRAPH_MAX_RESULTS_PER_SOURCE", "8"))
-GRAPH_MAX_FRONTIER = int(os.getenv("GRAPH_MAX_FRONTIER", "5"))
-GRAPH_MAX_CONCEPTS = int(os.getenv("GRAPH_MAX_CONCEPTS", "12"))
-GRAPH_MAX_CLAIMS = int(os.getenv("GRAPH_MAX_CLAIMS", "6"))
-GRAPH_MAX_GRAPH_NODES = int(os.getenv("GRAPH_MAX_GRAPH_NODES", "18"))
-
-STOPWORDS = {
-    "a", "an", "and", "are", "as", "at", "be", "been", "being", "by", "can", "could", "did", "do", "does",
-    "for", "from", "had", "has", "have", "if", "in", "into", "is", "it", "its", "may", "might", "of", "on",
-    "or", "our", "such", "that", "the", "their", "there", "these", "this", "those", "to", "using", "use",
-    "used", "via", "was", "were", "will", "with", "within", "without", "we", "they", "you", "your", "study",
-    "paper", "research", "results", "method", "methods", "analysis", "approach", "toward", "towards",
-    "based", "new", "novel", "effect", "effects", "model", "models", "system", "systems",
-}
-
-GRAPH_MEMORY = {
-    "queries": [],
-    "papers": {},
-    "nodes": {},
-    "edges": [],
-    "events": [],
-    "concept_counts": Counter(),
-    "claim_counts": Counter(),
-    "seen_queries": set(),
-    "seen_dois": set(),
-    "seen_titles": set(),
-}
-
-
-def safe_text(x, default=""):
-    return html.escape(str(x if x is not None else default))
-
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def slugify(text: str) -> str:
-    return re.sub(r"[^a-z0-9]+", "-", (text or "").lower()).strip("-")
-
-
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-def unique_keep_order(items):
-    seen = set()
-    out = []
-    for item in items:
-        key = norm_text(str(item)).lower()
-        if key and key not in seen:
-            seen.add(key)
-            out.append(item)
-    return out
-
-
-def normalize_doi(text: str) -> str:
-    text = (text or "").strip()
-    text = text.replace("https://doi.org/", "").replace("http://doi.org/", "")
-    return text.strip()
-
-
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    doi_pattern = r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$"
-    if re.match(doi_pattern, q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-
-
-def tokenize(text: str) -> List[str]:
-    return [t for t in re.findall(r"[a-zA-Z][a-zA-Z0-9\-]{2,}", (text or "").lower()) if t not in STOPWORDS]
-
-
-def text_overlap_score(a: str, b: str) -> float:
-    sa = set(tokenize(a))
-    sb = set(tokenize(b))
-    if not sa or not sb:
-        return 0.0
-    return len(sa & sb) / max(len(sa | sb), 1)
-
-
-def compute_recency_bonus(year: str) -> float:
-    try:
-        y = int(str(year)[:4])
-    except Exception:
-        return 0.0
-    current = time.gmtime().tm_year
-    age = max(current - y, 0)
-    return max(0.0, 0.12 - age * 0.015)
-
-
-def extract_candidate_phrases(text: str, max_terms: int = 30) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-
-    phrases = []
-    tokens = re.findall(r"[A-Za-z][A-Za-z0-9\-]{2,}", text)
-    for n in (3, 2, 1):
-        for i in range(len(tokens) - n + 1):
-            phrase = " ".join(tokens[i:i + n]).strip().lower()
-            if len(phrase) < 4:
-                continue
-            parts = phrase.split()
-            if any(p in STOPWORDS for p in parts):
-                continue
-            if all(len(p) <= 2 for p in parts):
-                continue
-            phrases.append(phrase)
-
-    counts = Counter(phrases)
-    ranked = [p for p, _ in counts.most_common(max_terms * 3)]
-    filtered = []
-    for phrase in ranked:
-        if phrase in filtered:
-            continue
-        if any(phrase != other and phrase in other for other in filtered):
-            continue
-        filtered.append(phrase)
-        if len(filtered) >= max_terms:
-            break
-    return filtered
-
-
-def extract_concepts_from_text(text: str, max_terms: int = GRAPH_MAX_CONCEPTS) -> List[str]:
-    return extract_candidate_phrases(text, max_terms=max_terms)
-
-
-def extract_claim_like_sentences(text: str, max_items: int = GRAPH_MAX_CLAIMS) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-    parts = re.split(r"(?<=[\.\!\?])\s+", text)
-    scored = []
-    for s in parts:
-        ss = norm_text(s)
-        if len(ss) < 40 or len(ss) > 280:
-            continue
-        score = 0
-        lower = ss.lower()
-        if any(k in lower for k in ["improves", "reduces", "increases", "suggests", "demonstrates", "shows", "reveals", "predicts", "achieves", "outperforms"]):
-            score += 2
-        if any(k in lower for k in ["significant", "associated", "correlated", "effective", "robust", "accurate", "validated"]):
-            score += 1
-        score += min(len(tokenize(ss)) / 15.0, 2.0)
-        scored.append((score, ss))
-    return [s for _, s in sorted(scored, key=lambda x: x[0], reverse=True)[:max_items]]
-
-
-def parse_openalex_abstract(inverted_index) -> str:
-    if not inverted_index or not isinstance(inverted_index, dict):
-        return ""
-    pos_to_word = {}
-    for word, positions in inverted_index.items():
-        for p in positions:
-            pos_to_word[p] = word
-    if not pos_to_word:
-        return ""
-    ordered = [pos_to_word[i] for i in sorted(pos_to_word)]
-    return " ".join(ordered)
-
-
-def journal_query_links(query: str):
-    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
-    rows = []
-    for journal in JOURNALS:
-        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
-        if "ieeexplore" in journal["url"]:
-            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
-        rows.append({
-            "name": journal["name"],
-            "desc": journal["desc"],
-            "url": url,
-        })
-    return rows
-
-
-def build_journal_html(query):
-    rows = []
-    for journal in journal_query_links(query):
-        rows.append(
-            f"""
-            <a class="journal-card" href="{safe_text(journal['url'])}" target="_blank" rel="noopener noreferrer">
-              <div>
-                <h4>{safe_text(journal['name'])}</h4>
-                <p>{safe_text(journal['desc'])}</p>
-              </div>
-              <span>Open</span>
-            </a>
-            """
-        )
-    return '<div class="journal-grid">' + ''.join(rows) + '</div>'
-
-
-def search_arxiv(query, max_results=8):
-    encoded = urllib.parse.quote(query)
-    url = (
-        "http://export.arxiv.org/api/query?search_query=all:"
-        f"{encoded}&start=0&max_results={max_results}&sortBy=relevance&sortOrder=descending"
-    )
-    response = requests.get(url, timeout=REQUEST_TIMEOUT)
-    response.raise_for_status()
-    root = ET.fromstring(response.text)
-    ns = {"atom": "http://www.w3.org/2005/Atom"}
-    papers = []
-    for entry in root.findall("atom:entry", ns):
-        title = " ".join((entry.findtext("atom:title", default="", namespaces=ns) or "").split())
-        summary = " ".join((entry.findtext("atom:summary", default="", namespaces=ns) or "").split())
-        published = entry.findtext("atom:published", default="", namespaces=ns)
-        paper_id = entry.findtext("atom:id", default="", namespaces=ns)
-        authors = [a.findtext("atom:name", default="", namespaces=ns) for a in entry.findall("atom:author", ns)]
-        pdf_url = ""
-        for link in entry.findall("atom:link", ns):
-            if link.attrib.get("title") == "pdf":
-                pdf_url = link.attrib.get("href", "")
-                break
-        papers.append(
-            {
-                "id": paper_id or title,
-                "title": title,
-                "summary": summary,
-                "abstract": summary,
-                "published": published[:10],
-                "authors": [a for a in authors[:8] if a],
-                "authors_text": ", ".join([a for a in authors[:4] if a]),
-                "url": paper_id,
-                "pdf": pdf_url,
-                "doi": "",
-                "venue": "arXiv",
-                "year": published[:4] if published else "",
-                "source": "arxiv",
-                "score": 0.76,
-                "open_access": True,
-                "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
-            }
-        )
-    return papers
-
-
-def search_crossref(query, mode="topic", max_results=8):
-    headers = {"User-Agent": "dvnc-ai-space/0.2"}
-    if mode == "doi":
-        url = f"https://api.crossref.org/works/{quote(query)}"
-        response = requests.get(url, headers=headers, timeout=REQUEST_TIMEOUT)
-        if response.status_code != 200:
-            return []
-        items = [response.json().get("message", {})]
-    else:
-        params = {"rows": max_results}
-        if mode in ("title", "paper_name"):
-            params["query.title"] = query
-        else:
-            params["query.bibliographic"] = query
-        response = requests.get("https://api.crossref.org/works", params=params, headers=headers, timeout=REQUEST_TIMEOUT)
-        response.raise_for_status()
-        items = response.json().get("message", {}).get("items", [])
-
-    out = []
-    for item in items:
-        authors = []
-        for a in item.get("author", []) or []:
-            name = " ".join(filter(None, [a.get("given"), a.get("family")])).strip()
-            if name:
-                authors.append(name)
-
-        title = (item.get("title") or ["Untitled"])[0]
-        year = ""
-        for key in ["published-print", "published-online", "created"]:
-            if item.get(key, {}).get("date-parts"):
-                year = str(item[key]["date-parts"][0][0])
-                break
-
-        abstract = item.get("abstract") or ""
-        abstract = re.sub("<.*?>", "", abstract)
-        doi = normalize_doi(item.get("DOI", ""))
-        out.append({
-            "id": doi or title,
-            "title": norm_text(title),
-            "summary": norm_text(abstract)[:500],
-            "abstract": norm_text(abstract),
-            "published": year,
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("URL", ""),
-            "pdf": "",
-            "doi": doi,
-            "venue": (item.get("container-title") or [""])[0],
-            "year": year,
-            "source": "crossref",
-            "score": 0.72,
-            "open_access": None,
-            "external_ids": {"crossref": doi} if doi else {},
-        })
-    return out
-
-
-def search_openalex(query, mode="topic", max_results=8):
-    params = {"per-page": max_results}
-    if mode == "doi":
-        doi = normalize_doi(query)
-        params["filter"] = f"doi:https://doi.org/{doi}"
-    else:
-        params["search"] = query
-
-    response = requests.get("https://api.openalex.org/works", params=params, timeout=REQUEST_TIMEOUT)
-    response.raise_for_status()
-    items = response.json().get("results", [])
-    out = []
-    for item in items:
-        authors = []
-        for auth in item.get("authorships", [])[:8]:
-            author = auth.get("author") or {}
-            if author.get("display_name"):
-                authors.append(author["display_name"])
-
-        oa = item.get("open_access") or {}
-        doi = normalize_doi(item.get("doi") or "")
-        abstract = parse_openalex_abstract(item.get("abstract_inverted_index"))
-        out.append({
-            "id": item.get("id") or doi or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(abstract)[:500],
-            "abstract": norm_text(abstract),
-            "published": str(item.get("publication_year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": (item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or "",
-            "pdf": oa.get("oa_url") or "",
-            "doi": doi,
-            "venue": ((item.get("primary_location") or {}).get("source") or {}).get("display_name") or "",
-            "year": str(item.get("publication_year") or ""),
-            "source": "openalex",
-            "score": 0.80,
-            "open_access": oa.get("is_oa"),
-            "external_ids": item.get("ids") or {},
-        })
-    return out
-
-
-def search_semantic_scholar(query, mode="topic", max_results=8):
-    headers = {}
-    if SEMANTIC_SCHOLAR_API_KEY:
-        headers["x-api-key"] = SEMANTIC_SCHOLAR_API_KEY
-
-    fields = "title,authors,year,abstract,venue,externalIds,url,openAccessPdf"
-    if mode == "doi":
-        doi = normalize_doi(query)
-        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{quote(doi)}"
-        response = requests.get(url, params={"fields": fields}, headers=headers, timeout=REQUEST_TIMEOUT)
-        if response.status_code != 200:
-            return []
-        items = [response.json()]
-    else:
-        response = requests.get(
-            "https://api.semanticscholar.org/graph/v1/paper/search",
-            params={"query": query, "limit": max_results, "fields": fields},
-            headers=headers,
-            timeout=REQUEST_TIMEOUT,
-        )
-        if response.status_code != 200:
-            return []
-        items = response.json().get("data", [])
-
-    out = []
-    for item in items:
-        external = item.get("externalIds") or {}
-        authors = [a.get("name") for a in item.get("authors", []) if a.get("name")]
-        out.append({
-            "id": external.get("CorpusId") or external.get("DOI") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(item.get("abstract", ""))[:500],
-            "abstract": norm_text(item.get("abstract", "")),
-            "published": str(item.get("year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("url") or "",
-            "pdf": (item.get("openAccessPdf") or {}).get("url") or "",
-            "doi": normalize_doi(external.get("DOI", "")),
-            "venue": item.get("venue") or "",
-            "year": str(item.get("year") or ""),
-            "source": "semantic_scholar",
-            "score": 0.84,
-            "open_access": bool((item.get("openAccessPdf") or {}).get("url")),
-            "external_ids": external,
-        })
-    return out
-
-
-def search_europe_pmc(query, mode="topic", max_results=8):
-    epmc_query = f'DOI:"{query}"' if mode == "doi" else query
-    params = {
-        "query": epmc_query,
-        "format": "json",
-        "pageSize": max_results,
-        "resultType": "core",
-    }
-    response = requests.get("https://www.ebi.ac.uk/europepmc/webservices/rest/search", params=params, timeout=REQUEST_TIMEOUT)
-    if response.status_code != 200:
-        return []
-    items = response.json().get("resultList", {}).get("result", [])
-    out = []
-    for item in items:
-        author_string = item.get("authorString", "")
-        authors = [x.strip() for x in author_string.split(",")[:8] if x.strip()]
-        pmcid = item.get("pmcid", "")
-        pdf_url = f"https://europepmc.org/articles/{pmcid}?pdf=render" if pmcid else ""
-        landing_url = f"https://europepmc.org/article/PMC/{pmcid}" if pmcid else ""
-        out.append({
-            "id": item.get("id") or item.get("doi") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(item.get("abstractText", ""))[:500],
-            "abstract": norm_text(item.get("abstractText", "")),
-            "published": str(item.get("pubYear") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": landing_url,
-            "pdf": pdf_url,
-            "doi": normalize_doi(item.get("doi", "")),
-            "venue": item.get("journalTitle", ""),
-            "year": str(item.get("pubYear") or ""),
-            "source": "europe_pmc",
-            "score": 0.78,
-            "open_access": bool(pmcid),
-            "external_ids": {"pmid": item.get("pmid"), "pmcid": pmcid},
-        })
-    return out
-
-
-def resolve_link(query):
-    url = (query or "").strip()
-    if not url:
-        return []
-    try:
-        response = requests.get(
-            url,
-            timeout=REQUEST_TIMEOUT,
-            allow_redirects=True,
-            headers={"User-Agent": "dvnc-ai-space/0.2"},
-        )
-        content_type = response.headers.get("content-type", "")
-        if "pdf" in content_type or url.lower().endswith(".pdf"):
-            name = Path(url.split("?")[0]).name or "linked-paper.pdf"
-            return [{
-                "id": url,
-                "title": name,
-                "summary": "Direct PDF link detected.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": url,
-                "doi": "",
-                "venue": "Direct PDF",
-                "year": "",
-                "source": "link",
-                "score": 0.66,
-                "open_access": True,
-                "external_ids": {},
-            }]
-
-        if BeautifulSoup is None:
-            return [{
-                "id": url,
-                "title": url,
-                "summary": "Web page resolved. Install beautifulsoup4 for DOI extraction.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": "",
-                "doi": "",
-                "venue": "Web Link",
-                "year": "",
-                "source": "link",
-                "score": 0.48,
-                "open_access": None,
-                "external_ids": {},
-            }]
-
-        soup = BeautifulSoup(response.text, "html.parser")
-        doi = ""
-        for meta_name in ["citation_doi", "dc.identifier", "dc.Identifier"]:
-            tag = soup.find("meta", attrs={"name": meta_name})
-            if tag and tag.get("content"):
-                doi = normalize_doi(tag["content"].strip())
-                break
-
-        title = soup.title.text.strip() if soup.title else url
-        pdf_link = ""
-        for a in soup.find_all("a", href=True):
-            href = a["href"]
-            if ".pdf" in href.lower():
-                pdf_link = href if href.startswith("http") else ""
-                break
-
-        if doi:
-            results = search_crossref(doi, mode="doi", max_results=1)
-            if results:
-                if pdf_link and not results[0].get("pdf"):
-                    results[0]["pdf"] = pdf_link
-                if url and not results[0].get("url"):
-                    results[0]["url"] = url
-                return results
-
-        return [{
-            "id": url,
-            "title": title,
-            "summary": "Landing page resolved from direct link.",
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": pdf_link,
-            "doi": doi,
-            "venue": "Web Link",
-            "year": "",
-            "source": "link",
-            "score": 0.54,
-            "open_access": bool(pdf_link),
-            "external_ids": {},
-        }]
-    except Exception as e:
-        return [{
-            "id": url,
-            "title": "Link resolution error",
-            "summary": str(e),
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": "",
-            "doi": "",
-            "venue": "Link",
-            "year": "",
-            "source": "link",
-            "score": 0.20,
-            "open_access": None,
-            "external_ids": {},
-        }]
-
-
-def dedupe_papers(items: List[Dict]) -> List[Dict]:
-    seen = {}
-    for item in items:
-        key = (
-            normalize_doi(item.get("doi") or "")
-            or (item.get("external_ids") or {}).get("arxiv")
-            or norm_text(item.get("title", "")).lower()
-            or str(item.get("id"))
-        )
-        if not key:
-            key = f"{item.get('source')}::{item.get('title')}"
-        if key not in seen or float(item.get("score", 0)) > float(seen[key].get("score", 0)):
-            seen[key] = item
-    return sorted(seen.values(), key=lambda x: float(x.get("score", 0)), reverse=True)
-
-
-def enrich_paper_semantics(query: str, paper: Dict) -> Dict:
-    paper = dict(paper)
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue]).strip()
-
-    concepts = extract_concepts_from_text(base_text, max_terms=GRAPH_MAX_CONCEPTS)
-    claims = extract_claim_like_sentences(abstract, max_items=GRAPH_MAX_CLAIMS)
-
-    rel = text_overlap_score(query, f"{title} {abstract}")
-    recency = compute_recency_bonus(paper.get("year"))
-    citation_hints = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    semantic_bonus = min(len(concepts), 8) * 0.01
-    learned_score = float(paper.get("score", 0)) + rel * 0.5 + recency + citation_hints + oa_bonus + semantic_bonus
-
-    paper["concepts"] = concepts[:GRAPH_MAX_CONCEPTS]
-    paper["claims"] = claims[:GRAPH_MAX_CLAIMS]
-    paper["relevance"] = round(rel, 4)
-    paper["learned_score"] = round(learned_score, 4)
-    paper["semantic_summary"] = "; ".join(concepts[:5]) if concepts else ""
-    return paper
-
-
-def discover_papers(query, mode, sources, max_results=10):
-    query = (query or "").strip()
-    if not query:
-        return []
-
-    mode = detect_query_type(query) if mode == "autonomous_web" else mode
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    results = []
-
-    if mode == "link":
-        return dedupe_papers(resolve_link(query))
-
-    try:
-        if "arxiv" in selected_sources and mode != "doi":
-            results.extend(search_arxiv(query, max_results=min(max_results, GRAPH_MAX_RESULTS_PER_SOURCE)))
-    except Exception:
-        pass
-    try:
-        if "crossref" in selected_sources:
-            results.extend(search_crossref(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS_PER_SOURCE)))
-    except Exception:
-        pass
-    try:
-        if "openalex" in selected_sources:
-            results.extend(search_openalex(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS_PER_SOURCE)))
-    except Exception:
-        pass
-    try:
-        if "semantic_scholar" in selected_sources:
-            results.extend(search_semantic_scholar(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS_PER_SOURCE)))
-    except Exception:
-        pass
-    try:
-        if "europe_pmc" in selected_sources:
-            results.extend(search_europe_pmc(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS_PER_SOURCE)))
-    except Exception:
-        pass
-
-    ranked = [enrich_paper_semantics(query, p) for p in dedupe_papers(results)]
-    ranked = sorted(ranked, key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-    return ranked
-
-
-def extract_reference_queries_from_parsed(parsed_pdf: Optional[Dict], max_items: int = 8) -> List[str]:
-    if not parsed_pdf or not isinstance(parsed_pdf, dict):
-        return []
-    out = []
-    for ref in (parsed_pdf.get("references") or [])[:max_items]:
-        doi = normalize_doi(ref.get("doi") or "")
-        title = norm_text(ref.get("title") or "")
-        if doi:
-            out.append(doi)
-        elif title:
-            out.append(title)
-    return unique_keep_order(out)
-
-
-def propose_expansion_queries(seed_query: str, papers: List[Dict], parsed_pdf: Optional[Dict] = None, max_items: int = 10) -> List[str]:
-    counter = Counter()
-    seed_tokens = set(tokenize(seed_query))
-
-    for paper in papers[:10]:
-        title = paper.get("title", "")
-        abstract = paper.get("abstract", "") or paper.get("summary", "")
-        concepts = paper.get("concepts") or extract_concepts_from_text(f"{title} {abstract}")
-        for concept in concepts[:8]:
-            score = 1.0
-            ctoks = set(tokenize(concept))
-            if seed_tokens & ctoks:
-                score += 1.0
-            if len(concept.split()) >= 2:
-                score += 0.5
-            counter[concept] += score
-
-        title_query = norm_text(title)
-        if title_query:
-            counter[title_query] += 0.4
-
-    for rq in extract_reference_queries_from_parsed(parsed_pdf, max_items=6):
-        counter[rq] += 1.8
-
-    expansions = []
-    for concept, _ in counter.most_common(max_items * 3):
-        c = norm_text(concept)
-        if not c:
-            continue
-        if c.lower() == seed_query.lower():
-            continue
-        if len(c) < 4:
-            continue
-        expansions.append(c)
-
-    expansions = unique_keep_order(expansions)
-    return expansions[:max_items]
-
-
-def autonomous_expand_graph(query: str, sources: List[str], parsed_pdf: Optional[Dict] = None,
-                            initial_mode: str = "autonomous_web", max_rounds: int = GRAPH_MAX_ROUNDS,
-                            frontier_limit: int = GRAPH_MAX_FRONTIER, max_results: int = 8) -> Dict:
-    visited_queries = []
-    query_frontier = [query]
-    all_papers = []
-    round_summaries = []
-
-    for round_idx in range(max_rounds):
-        this_round_queries = []
-        this_round_papers = []
-
-        for q in query_frontier[:frontier_limit]:
-            qn = norm_text(q)
-            if not qn or qn.lower() in GRAPH_MEMORY["seen_queries"]:
-                continue
-            GRAPH_MEMORY["seen_queries"].add(qn.lower())
-            visited_queries.append(qn)
-            this_round_queries.append(qn)
-
-            mode = detect_query_type(qn) if initial_mode == "autonomous_web" else initial_mode
-            discovered = discover_papers(qn, mode, sources, max_results=max_results)
-            this_round_papers.extend(discovered)
-
-        all_papers.extend(this_round_papers)
-        all_papers = dedupe_papers(all_papers)
-        all_papers = [enrich_paper_semantics(query, p) for p in all_papers]
-        all_papers = sorted(all_papers, key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-
-        top_context = all_papers[:12]
-        proposed = propose_expansion_queries(query, top_context, parsed_pdf=parsed_pdf, max_items=frontier_limit * 2)
-        next_frontier = [q for q in proposed if q.lower() not in GRAPH_MEMORY["seen_queries"]][:frontier_limit]
-
-        round_summaries.append({
-            "round": round_idx + 1,
-            "queries": this_round_queries,
-            "papers_found": len(this_round_papers),
-            "frontier_next": next_frontier,
-        })
-
-        if not next_frontier:
-            break
-        query_frontier = next_frontier
-
-    graph_snapshot = build_autonomous_payload(query, all_papers[:15], parsed_pdf, visited_queries, round_summaries)
-    return {
-        "query": query,
-        "papers": all_papers,
-        "visited_queries": visited_queries,
-        "rounds": round_summaries,
-        "payload": graph_snapshot,
-    }
-
-
-def format_papers_html(papers):
-    if not papers:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No papers found yet.</p></div>'
-
-    items = []
-    for i, paper in enumerate(papers, start=1):
-        summary = safe_text((paper.get("summary") or paper.get("abstract") or "")[:280])
-        doi_line = f'<span class="paper-badge doi-badge">{safe_text(paper.get("doi"))}</span>' if paper.get("doi") else ""
-        pdf_link = paper.get("pdf") or "#"
-        abs_link = paper.get("url") or "#"
-        concept_line = ", ".join((paper.get("concepts") or [])[:4])
-        items.append(
-            f"""
-            <article class="paper-card">
-              <div class="paper-topline">
-                <span class="paper-badge">{safe_text(paper.get('source', 'paper'))}</span>
-                <span class="paper-badge alt">{safe_text(paper.get('published', '') or 'Paper')}</span>
-                {doi_line}
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{summary or 'No abstract snippet available.'}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-                <div><strong>Venue:</strong> {safe_text(paper.get('venue', 'Unknown venue'))}</div>
-                <div><strong>Base score:</strong> {safe_text(round(float(paper.get('score', 0)), 3))}</div>
-                <div><strong>Learned score:</strong> {safe_text(round(float(paper.get('learned_score', paper.get('score', 0)), 3)))}</div>
-                <div><strong>Concepts:</strong> {safe_text(concept_line or 'None extracted')}</div>
-              </div>
-              <div class="paper-links">
-                <a href="{safe_text(abs_link)}" target="_blank" rel="noopener noreferrer">Abstract</a>
-                <a href="{safe_text(pdf_link)}" target="_blank" rel="noopener noreferrer">PDF</a>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(items) + '</div>'
-
-
-def format_selection_choices(papers):
-    choices = []
-    for i, paper in enumerate(papers):
-        learned = paper.get("learned_score", paper.get("score", 0))
-        label = f"[{paper.get('source', 'src')}] {paper.get('title', 'Untitled')} — {paper.get('authors_text', 'Unknown authors')[:80]} — score {round(float(learned), 3)}"
-        choices.append((label, str(i)))
-    return choices
-
-
-def uploaded_pdf_summary(file_obj):
-    if not file_obj:
-        return "No PDF uploaded yet."
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    p = Path(path)
-    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, references, concepts, and graph edges."
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    if not nodes:
-        return """
-        <div class="panel brain-shell">
-          <div class="brain-header">
-            <div>
-              <p class="eyebrow">Learning Graph</p>
-              <h3>Self-Learning Knowledge Graph</h3>
-            </div>
-          </div>
-          <div class="brain-stage learning-empty">
-            <div class="empty-graph-copy">
-              <h4>No papers mapped yet</h4>
-              <p>Search papers, pick a topic, select candidates, or upload a PDF to grow the graph in real time.</p>
-            </div>
-          </div>
-        </div>
-        """
-
-    node_items = []
-    label_items = []
-    edge_items = []
-
-    coords = [
-        (100, 90), (250, 60), (420, 75), (590, 115), (690, 250), (620, 395),
-        (455, 455), (280, 455), (110, 395), (60, 250), (215, 250), (365, 245),
-        (525, 250), (300, 145), (480, 340), (180, 340), (545, 175), (130, 170)
-    ]
-
-    graph_nodes = [dict(n) for n in nodes[:GRAPH_MAX_GRAPH_NODES]]
-    for i, node in enumerate(graph_nodes):
-        x, y = coords[i % len(coords)]
-        node["sx"] = x
-        node["sy"] = y
-
-    node_map = {n["id"]: n for n in graph_nodes}
-    for edge in edges[:60]:
-        if isinstance(edge, dict):
-            a = edge.get("source")
-            b = edge.get("target")
-            etype = edge.get("type", "")
-        else:
-            a, b = edge
-            etype = ""
-        if a in node_map and b in node_map:
-            na = node_map[a]
-            nb = node_map[b]
-            edge_items.append(
-                f'<line class="learn-edge edge-{safe_text(etype.lower())}" x1="{na["sx"]}" y1="{na["sy"]}" x2="{nb["sx"]}" y2="{nb["sy"]}" />'
-            )
-
-    for node in graph_nodes:
-        kind = node.get("kind", node.get("type", "paper")).lower()
-        if kind == "topic":
-            kind = "query"
-        if kind == "uploadedpdf":
-            kind = "upload"
-        radius = 25 if kind == "query" else 18 if kind in {"concept", "author", "claim", "reference"} else 20
-        klass = f'learn-node {kind}'
-        node_items.append(
-            f'<circle class="{klass}" cx="{node["sx"]}" cy="{node["sy"]}" r="{radius}" />'
-        )
-        label = node.get("label") or node.get("title") or node.get("id")
-        label_items.append(
-            f'<text class="learn-label" x="{node["sx"] + 26}" y="{node["sy"] - 8}">{safe_text(str(label)[:46])}</text>'
-        )
-
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Learning Graph</p>
-          <h3>{safe_text(title)}</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-query"></i> topic</span>
-          <span><i class="dot dot-paper"></i> paper</span>
-          <span><i class="dot dot-upload"></i> uploaded PDF</span>
-          <span><i class="dot dot-concept"></i> concept</span>
-          <span><i class="dot dot-author"></i> author</span>
-          <span><i class="dot dot-ref"></i> reference</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 760 520" class="brain-svg" role="img" aria-label="Self-learning knowledge graph">
-          {''.join(edge_items)}
-          {''.join(node_items)}
-          {''.join(label_items)}
-        </svg>
-      </div>
-    </div>
-    """
-
-
-def build_learning_graph_state(query, papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-
-    for i, paper in enumerate(papers[:5], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper["title"], "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{i}_{slugify(concept)[:20]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-        if len(nodes) > 2:
-            edges.append({"source": "upload", "target": "paper_1", "type": "RELATED_TO"})
-
-    return nodes, edges
-
-
-def graph_from_selected(query, selected_papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-
-    for i, paper in enumerate(selected_papers[:8], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper["title"], "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-
-        if paper.get("doi"):
-            doi_id = f"doi_{i}"
-            nodes.append({"id": doi_id, "label": f"DOI {paper['doi']}", "kind": "reference"})
-            edges.append({"source": pid, "target": doi_id, "type": "HAS_ID"})
-
-        for author in paper.get("authors", [])[:2]:
-            aid = f"author_{slugify(author)[:24]}_{i}"
-            nodes.append({"id": aid, "label": author, "kind": "author"})
-            edges.append({"source": pid, "target": aid, "type": "WRITTEN_BY"})
-
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{slugify(concept)[:20]}_{i}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-        if len(selected_papers) > 0:
-            edges.append({"source": "upload", "target": "paper_1", "type": "RELATED_TO"})
-
-    return nodes, edges
-
-
-def parse_pdf_with_grobid(pdf_path):
-    if not GROBID_URL:
-        raise RuntimeError("GROBID_URL is not set")
-
-    with open(pdf_path, "rb") as f:
-        files = {"input": (Path(pdf_path).name, f, "application/pdf")}
-        response = requests.post(
-            f"{GROBID_URL.rstrip('/')}/api/processFulltextDocument",
-            files=files,
-            data={"includeRawAffiliations": "1", "segmentSentences": "1"},
-            timeout=120,
-        )
-
-    response.raise_for_status()
-    tei_xml = response.text
-
-    if BeautifulSoup is None:
-        return {
-            "parser": "grobid",
-            "title": Path(pdf_path).name,
-            "abstract": "",
-            "authors": [],
-            "sections": [],
-            "references": [],
-            "claims": [],
-            "concepts": [],
-            "raw_text": "",
-            "tei_xml": tei_xml[:30000],
-        }
-
-    soup = BeautifulSoup(tei_xml, "xml")
-    title_stmt = soup.find("titleStmt")
-    title_tag = title_stmt.find("title") if title_stmt else soup.find("title")
-    abstract_tag = soup.find("abstract")
-
-    authors = []
-    for author in soup.find_all("author"):
-        pers = author.find("persName")
-        if pers:
-            name = " ".join(
-                x.get_text(" ", strip=True)
-                for x in pers.find_all(["forename", "surname"])
-            ).strip()
-            if name:
-                authors.append(name)
-
-    sections = []
-    section_texts = []
-    for div in soup.find_all("div"):
-        head = div.find("head")
-        paras = [p.get_text(" ", strip=True) for p in div.find_all("p")]
-        text = "\n".join([p for p in paras if p.strip()])
-        if head and text.strip():
-            sections.append({"heading": head.get_text(" ", strip=True), "text": text[:4000]})
-            section_texts.append(text)
-
-    references = []
-    for bibl in soup.find_all("biblStruct")[:40]:
-        ref_title = ""
-        ref_doi = ""
-        title_node = bibl.find("title")
-        if title_node:
-            ref_title = title_node.get_text(" ", strip=True)
-        doi_node = bibl.find("idno", attrs={"type": "DOI"})
-        if doi_node:
-            ref_doi = doi_node.get_text(" ", strip=True)
-        references.append({"title": ref_title, "doi": ref_doi})
-
-    abstract = abstract_tag.get_text(" ", strip=True) if abstract_tag else ""
-    full_text_for_semantics = " ".join([abstract] + section_texts[:4])
-    return {
-        "parser": "grobid",
-        "title": title_tag.get_text(" ", strip=True) if title_tag else Path(pdf_path).name,
-        "abstract": abstract,
-        "authors": authors,
-        "sections": sections,
-        "references": references,
-        "claims": extract_claim_like_sentences(full_text_for_semantics, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(full_text_for_semantics, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": "",
-        "tei_xml": tei_xml[:60000],
-    }
-
-
-def parse_pdf_with_pymupdf(pdf_path):
-    if fitz is None:
-        raise RuntimeError("PyMuPDF not installed")
-
-    doc = fitz.open(pdf_path)
-    pages = [page.get_text("text") for page in doc]
-    raw_text = "\n".join(pages).strip()
-    first_page = raw_text[:4000]
-    lines = [x.strip() for x in first_page.splitlines() if x.strip()]
-    title = lines[0][:300] if lines else Path(pdf_path).name
-
-    abstract = ""
-    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n1[\.\s]|introduction)", raw_text, re.I | re.S)
-    if match:
-        abstract = norm_text(match.group(1))[:2500]
-
-    return {
-        "parser": "pymupdf",
-        "title": title,
-        "abstract": abstract,
-        "authors": [],
-        "sections": [{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(raw_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(raw_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": raw_text[:50000],
-        "tei_xml": "",
-    }
-
-
-def parse_pdf_with_docling(pdf_path):
-    try:
-        from docling.document_converter import DocumentConverter
-    except Exception as e:
-        raise RuntimeError(f"Docling import failed: {e}")
-
-    converter = DocumentConverter()
-    result = converter.convert(pdf_path)
-    doc = result.document
-    markdown = doc.export_to_markdown()
-
-    title = Path(pdf_path).name
-    first_nonempty = next((line.strip("# ").strip() for line in markdown.splitlines() if line.strip()), "")
-    if first_nonempty:
-        title = first_nonempty[:300]
-
-    return {
-        "parser": "docling",
-        "title": title,
-        "abstract": "",
-        "authors": [],
-        "sections": [{"heading": "Document", "text": markdown[:12000]}] if markdown else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(markdown, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(markdown, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": markdown[:50000],
-        "tei_xml": "",
-    }
-
-
-def parse_uploaded_pdf(file_obj, parser_order):
-    if not file_obj:
-        return "No PDF uploaded yet.", {}
-
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    parser_order = ensure_list(parser_order) or ["grobid", "docling", "pymupdf"]
-    errors = []
-
-    for parser_name in parser_order:
-        try:
-            if parser_name == "grobid":
-                result = parse_pdf_with_grobid(path)
-            elif parser_name == "docling":
-                result = parse_pdf_with_docling(path)
-            elif parser_name == "pymupdf":
-                result = parse_pdf_with_pymupdf(path)
-            else:
-                continue
-
-            summary = (
-                f"### Parsed PDF\n\n"
-                f"- Parser used: {result['parser']}\n"
-                f"- Title: {result.get('title') or 'Unknown'}\n"
-                f"- Authors: {', '.join(result.get('authors')[:6]) if result.get('authors') else 'Unknown'}\n"
-                f"- Abstract found: {'Yes' if result.get('abstract') else 'No'}\n"
-                f"- Sections extracted: {len(result.get('sections') or [])}\n"
-                f"- References extracted: {len(result.get('references') or [])}\n"
-                f"- Concepts extracted: {len(result.get('concepts') or [])}\n"
-                f"- Claims extracted: {len(result.get('claims') or [])}\n"
-            )
-            return summary, result
-        except Exception as e:
-            errors.append(f"{parser_name}: {e}")
-
-    fail_summary = "### PDF parsing failed\n\n" + "\n".join([f"- {x}" for x in errors])
-    return fail_summary, {"parser": None, "errors": errors}
-
-
-def render_parse_result(parsed):
-    if not parsed:
-        return '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-
-    sections_html = []
-    for section in parsed.get("sections", [])[:6]:
-        sections_html.append(
-            f"""
-            <details class="agent-step">
-              <summary class="agent-summary">
-                <div class="agent-index">§</div>
-                <div class="agent-head">
-                  <h4>{safe_text(section.get('heading', 'Section'))}</h4>
-                  <span>section</span>
-                </div>
-              </summary>
-              <div class="agent-copy">
-                <p>{safe_text(section.get('text', '')[:1800])}</p>
-              </div>
-            </details>
-            """
-        )
-
-    refs = parsed.get("references", [])[:12]
-    refs_html = "".join(
-        f"<li>{safe_text(r.get('title') or 'Untitled')} {'· DOI ' + safe_text(r.get('doi')) if r.get('doi') else ''}</li>"
-        for r in refs
-    ) or "<li>No references extracted.</li>"
-
-    concepts = parsed.get("concepts", [])[:10]
-    claims = parsed.get("claims", [])[:6]
-    concepts_html = "".join(f"<li>{safe_text(x)}</li>" for x in concepts) or "<li>No concepts extracted.</li>"
-    claims_html = "".join(f"<li>{safe_text(x)}</li>" for x in claims) or "<li>No claims extracted.</li>"
-
-    title = safe_text(parsed.get("title") or "Parsed document")
-    abstract = safe_text((parsed.get("abstract") or "")[:2400]) or "No abstract extracted."
-    parser_name = safe_text(parsed.get("parser") or "unknown")
-
-    return f"""
-    <div class="panel" style="padding:18px">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">PDF Parse</p>
-          <h3>{title}</h3>
-        </div>
-        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name}</span></div>
-      </div>
-      <div class="parse-grid">
-        <div class="parse-card">
-          <h4>Abstract</h4>
-          <p>{abstract}</p>
-        </div>
-        <div class="parse-card">
-          <h4>References</h4>
-          <ul class="ref-list">{refs_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Concepts</h4>
-          <ul class="ref-list">{concepts_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Claims</h4>
-          <ul class="ref-list">{claims_html}</ul>
-        </div>
-      </div>
-      <div class="timeline" style="margin-top:14px;">
-        {''.join(sections_html) if sections_html else '<div class="panel" style="padding:16px;"><p>No sections extracted.</p></div>'}
-      </div>
-    </div>
-    """
-
-
-def add_node(nodes_by_id: Dict[str, Dict], node_id: str, node_type: str, label: str = "", **attrs):
-    if not node_id:
-        return
-    existing = nodes_by_id.get(node_id, {})
-    merged = {"id": node_id, "type": node_type, "label": label or existing.get("label", node_id)}
-    merged.update(existing)
-    merged.update({k: v for k, v in attrs.items() if v not in [None, ""]})
-    nodes_by_id[node_id] = merged
-
-
-def add_edge(edges: List[Dict], source: str, target: str, edge_type: str, **attrs):
-    if not source or not target or source == target:
-        return
-    edge = {"source": source, "target": target, "type": edge_type}
-    edge.update({k: v for k, v in attrs.items() if v not in [None, ""]})
-    edges.append(edge)
-
-
-def build_ingest_payload(query, selected_papers, parsed_pdf=None):
-    nodes_by_id = {}
-    edges = []
-
-    topic_id = "topic:query"
-    add_node(nodes_by_id, topic_id, "Topic", label=query or "Research topic", query=query or "")
-
-    for i, p in enumerate(selected_papers, start=1):
-        paper_id = normalize_doi(p.get("doi")) or (p.get("external_ids") or {}).get("arxiv") or f"paper:{i}:{slugify(p.get('title', 'paper'))[:32]}"
-        add_node(
-            nodes_by_id,
-            paper_id,
-            "Paper",
-            label=p.get("title") or f"Paper {i}",
-            title=p.get("title"),
-            year=p.get("year"),
-            venue=p.get("venue"),
-            doi=normalize_doi(p.get("doi")),
-            source=p.get("source"),
-            url=p.get("url"),
-            pdf=p.get("pdf"),
-            score=p.get("score"),
-            learned_score=p.get("learned_score", p.get("score")),
-            open_access=p.get("open_access"),
-        )
-        add_edge(edges, topic_id, paper_id, "ABOUT", weight=p.get("learned_score", p.get("score", 0)))
-
-        for author in p.get("authors", [])[:8]:
-            author_id = f"author:{slugify(author)[:64]}"
-            add_node(nodes_by_id, author_id, "Author", label=author, name=author)
-            add_edge(edges, paper_id, author_id, "WRITTEN_BY")
-
-        for concept in (p.get("concepts") or [])[:8]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, paper_id, concept_id, "MENTIONS")
-
-        for claim in (p.get("claims") or [])[:4]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, paper_id, claim_id, "ASSERTS")
-
-        if p.get("venue"):
-            venue_id = f"venue:{slugify(p['venue'])[:72]}"
-            add_node(nodes_by_id, venue_id, "Venue", label=p["venue"], name=p["venue"])
-            add_edge(edges, paper_id, venue_id, "PUBLISHED_IN")
-
-    if parsed_pdf and parsed_pdf.get("title"):
-        doc_id = "upload:pdf"
-        add_node(
-            nodes_by_id,
-            doc_id,
-            "UploadedPDF",
-            label=parsed_pdf.get("title"),
-            title=parsed_pdf.get("title"),
-            parser=parsed_pdf.get("parser"),
-        )
-        add_edge(edges, topic_id, doc_id, "UPLOADED_SOURCE")
-
-        for idx, author in enumerate(parsed_pdf.get("authors", [])[:8], start=1):
-            author_id = f"author:upload:{slugify(author)[:64]}"
-            add_node(nodes_by_id, author_id, "Author", label=author, name=author)
-            add_edge(edges, doc_id, author_id, "WRITTEN_BY")
-
-        for idx, section in enumerate(parsed_pdf.get("sections", [])[:8], start=1):
-            sec_id = f"{doc_id}:section:{idx}"
-            add_node(nodes_by_id, sec_id, "Section", label=section.get("heading") or f"Section {idx}", text=section.get("text", ""))
-            add_edge(edges, doc_id, sec_id, "HAS_SECTION")
-
-        for idx, ref in enumerate(parsed_pdf.get("references", [])[:12], start=1):
-            ref_label = ref.get("title") or f"Reference {idx}"
-            ref_doi = normalize_doi(ref.get("doi") or "")
-            ref_id = ref_doi or f"{doc_id}:ref:{idx}:{slugify(ref_label)[:32]}"
-            add_node(nodes_by_id, ref_id, "Reference", label=ref_label, title=ref_label, doi=ref_doi)
-            add_edge(edges, doc_id, ref_id, "CITES")
-
-        for concept in (parsed_pdf.get("concepts") or [])[:8]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, doc_id, concept_id, "MENTIONS")
-
-        for claim in (parsed_pdf.get("claims") or [])[:6]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, doc_id, claim_id, "ASSERTS")
-
-    payload = {"status": "ok", "nodes": list(nodes_by_id.values()), "edges": edges}
-    return payload
-
-
-def build_autonomous_payload(query: str, papers: List[Dict], parsed_pdf: Optional[Dict], visited_queries: List[str], rounds: List[Dict]) -> Dict:
-    base_payload = build_ingest_payload(query, papers, parsed_pdf=parsed_pdf)
-    nodes_by_id = {n["id"]: dict(n) for n in base_payload["nodes"]}
-    edges = list(base_payload["edges"])
-
-    for q in visited_queries[:12]:
-        qid = f"query:{slugify(q)[:72]}"
-        add_node(nodes_by_id, qid, "Query", label=q, query=q)
-        add_edge(edges, "topic:query", qid, "EXPANDS_TO")
-
-    for r in rounds:
-        rid = f"round:{r['round']}"
-        add_node(nodes_by_id, rid, "Round", label=f"Round {r['round']}", round=r["round"])
-        add_edge(edges, "topic:query", rid, "HAS_ROUND")
-        for q in r.get("queries", [])[:5]:
-            qid = f"query:{slugify(q)[:72]}"
-            add_node(nodes_by_id, qid, "Query", label=q, query=q)
-            add_edge(edges, rid, qid, "EXECUTED_QUERY")
-        for nq in r.get("frontier_next", [])[:5]:
-            qid = f"query:{slugify(nq)[:72]}"
-            add_node(nodes_by_id, qid, "Query", label=nq, query=nq)
-            add_edge(edges, rid, qid, "PROPOSED_QUERY")
-
-    return {"status": "ok", "nodes": list(nodes_by_id.values()), "edges": edges}
-
-
-def learn_from_payload(payload: Dict, query: str = "") -> Dict:
-    if not payload:
-        return GRAPH_MEMORY
-
-    GRAPH_MEMORY["queries"].append(query or "")
-    GRAPH_MEMORY["events"].append({
-        "ts": time.time(),
-        "query": query or "",
-        "nodes": len(payload.get("nodes", [])),
-        "edges": len(payload.get("edges", [])),
-    })
-
-    for node in payload.get("nodes", []):
-        nid = node.get("id")
-        if not nid:
-            continue
-        GRAPH_MEMORY["nodes"][nid] = node
-        ntype = (node.get("type") or "").lower()
-        if ntype == "paper":
-            key = normalize_doi(node.get("doi") or "") or nid
-            GRAPH_MEMORY["papers"][key] = node
-            if node.get("doi"):
-                GRAPH_MEMORY["seen_dois"].add(normalize_doi(node["doi"]))
-            if node.get("title"):
-                GRAPH_MEMORY["seen_titles"].add(norm_text(node["title"]).lower())
-        if ntype == "concept" and node.get("label"):
-            GRAPH_MEMORY["concept_counts"][node["label"].lower()] += 1
-        if ntype == "claim" and node.get("label"):
-            GRAPH_MEMORY["claim_counts"][node["label"].lower()] += 1
-
-    GRAPH_MEMORY["edges"].extend(payload.get("edges", []))
-    return GRAPH_MEMORY
-
-
-def summarize_learning_state(learning_result: Dict) -> str:
-    if not learning_result:
-        return "### No learning results yet."
-
-    papers = learning_result.get("papers") or []
-    visited_queries = learning_result.get("visited_queries") or []
-    rounds = learning_result.get("rounds") or []
-
-    top_concepts = []
-    for p in papers[:8]:
-        top_concepts.extend((p.get("concepts") or [])[:3])
-    concept_counts = Counter([c.lower() for c in top_concepts if c])
-
-    lines = [
-        "### Self-learning graph update",
-        "",
-        f"- Seed query: {learning_result.get('query') or 'Research topic'}",
-        f"- Frontier queries executed: {len(visited_queries)}",
-        f"- Expansion rounds: {len(rounds)}",
-        f"- Unique papers discovered: {len(papers)}",
-        f"- Top learned concepts: {', '.join([c for c, _ in concept_counts.most_common(6)]) if concept_counts else 'None'}",
-    ]
-
-    for r in rounds[:3]:
-        lines.append(f"- Round {r['round']}: {len(r.get('queries', []))} queries, {r.get('papers_found', 0)} papers, next frontier {len(r.get('frontier_next', []))}")
-
-    return "\n".join(lines)
-
-
-def run_paper_discovery(query, search_mode, sources, pdf_file):
-    query_text = (query or "").strip()
-
-    if not query_text and not pdf_file:
-        empty_graph = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            empty_graph,
-            '<div class="panel papers-panel" style="padding:18px"><p>Enter a topic, title, DOI, link, or upload a PDF to start learning.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            "No PDF uploaded yet.",
-            gr.update(choices=[], value=[]),
-            [],
-            "### No discovery results yet.",
-        )
-
-    parsed_pdf = None
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    papers = []
-    try:
-        if search_mode == "autonomous_web" and query_text:
-            learning = autonomous_expand_graph(
-                query=query_text,
-                sources=ensure_list(sources) or DEFAULT_SOURCES,
-                parsed_pdf=parsed_pdf,
-                initial_mode="autonomous_web",
-                max_rounds=GRAPH_MAX_ROUNDS,
-                frontier_limit=GRAPH_MAX_FRONTIER,
-                max_results=GRAPH_MAX_RESULTS_PER_SOURCE,
-            )
-            papers = learning["papers"][:15]
-            payload = learning["payload"]
-            graph_html = build_learning_graph_html(payload["nodes"], payload["edges"], "Autonomous Self-Learning Graph")
-            status_md = summarize_learning_state(learning)
-        else:
-            if query_text:
-                papers = discover_papers(query_text, search_mode, sources, max_results=10)
-            graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:6], uploaded_name)
-            graph_html = build_learning_graph_html(graph_nodes, graph_edges)
-            status_md = (
-                f"### Discovery results\n\n"
-                f"- Search mode: {search_mode}\n"
-                f"- Sources: {', '.join(ensure_list(sources) or DEFAULT_SOURCES)}\n"
-                f"- Candidates found: {len(papers)}\n"
-                f"- Select papers below, then click **Ingest selected into graph**.\n"
-            )
-
-    except Exception as e:
-        graph_nodes, graph_edges = build_learning_graph_state(
-            query_text,
-            [],
-            Path(getattr(pdf_file, "name", "uploaded.pdf")).name if pdf_file else None,
-        )
-        error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges),
-            error_html,
-            build_journal_html(query_text or "biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
-        )
-
-    papers_html = format_papers_html(papers)
-    journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
-    pdf_summary = uploaded_pdf_summary(pdf_file)
-    choices = format_selection_choices(papers)
-    return graph_html, papers_html, journals_html, pdf_summary, gr.update(choices=choices, value=[]), papers, status_md
-
-
-def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
-    papers = ensure_list(papers_state)
-    selected_indices = ensure_list(selected_indices)
-
-    selected = []
-    for idx in selected_indices:
-        try:
-            selected.append(papers[int(idx)])
-        except Exception:
-            pass
-
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    if not selected and parsed_state and isinstance(parsed_state, dict) and papers:
-        selected = papers[:3]
-
-    if not selected and not parsed_state:
-        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return graph_html, "### Nothing ingested yet.\n\nSelect papers or parse an uploaded PDF first.", {"status": "empty", "nodes": [], "edges": []}
-
-    query_text = query or (parsed_state.get("title") if isinstance(parsed_state, dict) else "") or "Research topic"
-
-    for i, p in enumerate(selected):
-        selected[i] = enrich_paper_semantics(query_text, p)
-
-    graph_nodes, graph_edges = graph_from_selected(query_text, selected, uploaded_name)
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
-
-    payload = build_ingest_payload(query_text, selected, parsed_state if isinstance(parsed_state, dict) else None)
-    learn_from_payload(payload, query=query_text)
-
-    top_concepts = []
-    for p in selected:
-        top_concepts.extend((p.get("concepts") or [])[:3])
-    if isinstance(parsed_state, dict):
-        top_concepts.extend((parsed_state.get("concepts") or [])[:3])
-
-    summary_lines = [
-        "### Graph ingest ready",
-        "",
-        f"- Topic: {query_text}",
-        f"- Selected papers: {len(selected)}",
-        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
-        f"- Nodes created: {len(payload['nodes'])}",
-        f"- Edges created: {len(payload['edges'])}",
-        f"- Learned concepts: {', '.join(unique_keep_order(top_concepts)[:8]) if top_concepts else 'None'}",
-        f"- Memory papers stored: {len(GRAPH_MEMORY['papers'])}",
-        f"- Memory concepts stored: {len(GRAPH_MEMORY['concept_counts'])}",
-    ]
-    return graph_html, "\n".join(summary_lines), payload
-
-
-def run_self_learning_cycle(query, search_mode, sources, pdf_file, parser_order, selected_indices, papers_state, parsed_state):
-    query_text = (query or "").strip()
-
-    if pdf_file and (not parsed_state or not isinstance(parsed_state, dict) or not parsed_state.get("title")):
-        _, parsed_state = parse_uploaded_pdf(pdf_file, parser_order)
-
-    learning = autonomous_expand_graph(
-        query=query_text or (parsed_state.get("title") if isinstance(parsed_state, dict) else "Research topic"),
-        sources=ensure_list(sources) or DEFAULT_SOURCES,
-        parsed_pdf=parsed_state if isinstance(parsed_state, dict) else None,
-        initial_mode="autonomous_web" if search_mode == "autonomous_web" else search_mode,
-        max_rounds=GRAPH_MAX_ROUNDS,
-        frontier_limit=GRAPH_MAX_FRONTIER,
-        max_results=GRAPH_MAX_RESULTS_PER_SOURCE,
-    )
-
-    papers = learning["papers"][:15]
-
-    selected = []
-    for idx in ensure_list(selected_indices):
-        try:
-            selected.append(papers[int(idx)])
-        except Exception:
-            pass
-    if not selected:
-        selected = papers[:5]
-
-    payload = build_autonomous_payload(
-        learning["query"],
-        selected,
-        parsed_pdf=parsed_state if isinstance(parsed_state, dict) else None,
-        visited_queries=learning.get("visited_queries", []),
-        rounds=learning.get("rounds", []),
-    )
-    learn_from_payload(payload, query=learning["query"])
-
-    graph_html = build_learning_graph_html(payload["nodes"], payload["edges"], "Self-Learning Graph Cycle")
-    papers_html = format_papers_html(papers)
-    status_md = summarize_learning_state(learning)
-
-    return graph_html, papers_html, gr.update(choices=format_selection_choices(papers), value=[]), papers, status_md, payload
-
-
-def get_graph_memory_snapshot():
-    return {
-        "queries": list(GRAPH_MEMORY["queries"]),
-        "papers": list(GRAPH_MEMORY["papers"].values()),
-        "nodes": list(GRAPH_MEMORY["nodes"].values()),
-        "edges": list(GRAPH_MEMORY["edges"]),
-        "concept_counts": dict(GRAPH_MEMORY["concept_counts"]),
-        "claim_counts": dict(GRAPH_MEMORY["claim_counts"]),
-        "events": list(GRAPH_MEMORY["events"]),
-    }
-
-
-def reset_graph_memory():
-    GRAPH_MEMORY["queries"] = []
-    GRAPH_MEMORY["papers"] = {}
-    GRAPH_MEMORY["nodes"] = {}
-    GRAPH_MEMORY["edges"] = []
-    GRAPH_MEMORY["events"] = []
-    GRAPH_MEMORY["concept_counts"] = Counter()
-    GRAPH_MEMORY["claim_counts"] = Counter()
-    GRAPH_MEMORY["seen_queries"] = set()
-    GRAPH_MEMORY["seen_dois"] = set()
-    GRAPH_MEMORY["seen_titles"] = set()
-    return "### Graph memory reset.", get_graph_memory_snapshot()

dvnc_ai_v2_hf/deprecated/self_learning_graph_old4.py

@@ -1,1430 +0,0 @@
-import html
-import os
-import re
-import time
-import urllib.parse
-import xml.etree.ElementTree as ET
-from collections import Counter
-from pathlib import Path
-from typing import Dict, List, Optional
-
-import gradio as gr
-import requests
-
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-
-
-JOURNALS = [
-    {"name": "Nature", "url": "https://www.nature.com/search", "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science", "url": "https://www.science.org/search", "desc": "High-impact science journal and family."},
-    {"name": "Cell", "url": "https://www.cell.com/search", "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search", "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore", "url": "https://ieeexplore.ieee.org/search/searchresult.jsp", "desc": "Engineering, AI, signal processing, and systems."},
-]
-
-SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
-SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-DEFAULT_SOURCES = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "").strip()
-GROBID_URL = os.getenv("GROBID_URL", "").strip()
-REQUEST_TIMEOUT = int(os.getenv("REQUEST_TIMEOUT", "25"))
-GRAPH_MAX_CONCEPTS = int(os.getenv("GRAPH_MAX_CONCEPTS", "10"))
-GRAPH_MAX_CLAIMS = int(os.getenv("GRAPH_MAX_CLAIMS", "6"))
-GRAPH_MAX_RESULTS = int(os.getenv("GRAPH_MAX_RESULTS", "10"))
-
-STOPWORDS = {
-    "a", "an", "and", "are", "as", "at", "be", "been", "being", "by", "can", "could", "did", "do", "does",
-    "for", "from", "had", "has", "have", "if", "in", "into", "is", "it", "its", "may", "might", "of", "on",
-    "or", "our", "such", "that", "the", "their", "there", "these", "this", "those", "to", "using", "use",
-    "used", "via", "was", "were", "will", "with", "within", "without", "we", "they", "you", "your", "study",
-    "paper", "research", "results", "method", "methods", "analysis", "approach", "toward", "towards",
-    "based", "new", "novel", "effect", "effects", "model", "models", "system", "systems", "show", "shows",
-}
-
-GRAPH_MEMORY = {
-    "papers": {},
-    "nodes": {},
-    "edges": [],
-    "concept_counts": Counter(),
-    "claim_counts": Counter(),
-    "queries": [],
-    "events": [],
-}
-
-
-def safe_text(x, default=""):
-    return html.escape(str(x if x is not None else default))
-
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def slugify(text: str) -> str:
-    return re.sub(r"[^a-z0-9]+", "-", (text or "").lower()).strip("-")
-
-
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-def normalize_doi(text: str) -> str:
-    text = (text or "").strip()
-    text = text.replace("https://doi.org/", "").replace("http://doi.org/", "")
-    return text.strip()
-
-
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    doi_pattern = r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$"
-    if re.match(doi_pattern, q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-
-
-def tokenize(text: str) -> List[str]:
-    return [t for t in re.findall(r"[a-zA-Z][a-zA-Z0-9\-]{2,}", (text or "").lower()) if t not in STOPWORDS]
-
-
-def unique_keep_order(items: List[str]) -> List[str]:
-    seen = set()
-    out = []
-    for item in items:
-        key = norm_text(item).lower()
-        if key and key not in seen:
-            seen.add(key)
-            out.append(norm_text(item))
-    return out
-
-
-def text_overlap_score(a: str, b: str) -> float:
-    sa = set(tokenize(a))
-    sb = set(tokenize(b))
-    if not sa or not sb:
-        return 0.0
-    return len(sa & sb) / max(1, len(sa | sb))
-
-
-def compute_recency_bonus(year: str) -> float:
-    try:
-        y = int(str(year)[:4])
-    except Exception:
-        return 0.0
-    current = time.gmtime().tm_year
-    age = max(current - y, 0)
-    return max(0.0, 0.12 - age * 0.015)
-
-
-def extract_candidate_phrases(text: str, max_terms: int = 20) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-    tokens = re.findall(r"[A-Za-z][A-Za-z0-9\-]{2,}", text)
-    phrases = []
-
-    for n in (3, 2, 1):
-        for i in range(len(tokens) - n + 1):
-            phrase = " ".join(tokens[i:i + n]).strip().lower()
-            if len(phrase) < 4:
-                continue
-            parts = phrase.split()
-            if any(p in STOPWORDS for p in parts):
-                continue
-            if all(len(p) <= 2 for p in parts):
-                continue
-            phrases.append(phrase)
-
-    counts = Counter(phrases)
-    ranked = [p for p, _ in counts.most_common(max_terms * 3)]
-
-    filtered = []
-    for phrase in ranked:
-        if phrase in filtered:
-            continue
-        if any(phrase != other and phrase in other for other in filtered):
-            continue
-        filtered.append(phrase)
-        if len(filtered) >= max_terms:
-            break
-
-    return filtered
-
-
-def extract_concepts_from_text(text: str, max_terms: int = GRAPH_MAX_CONCEPTS) -> List[str]:
-    return extract_candidate_phrases(text, max_terms=max_terms)
-
-
-def extract_claim_like_sentences(text: str, max_items: int = GRAPH_MAX_CLAIMS) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-
-    parts = re.split(r"(?<=[\.\!\?])\s+", text)
-    scored = []
-    for sentence in parts:
-        s = norm_text(sentence)
-        if len(s) < 40 or len(s) > 280:
-            continue
-        lower = s.lower()
-        score = 0.0
-        if any(k in lower for k in ["improves", "reduces", "increases", "suggests", "demonstrates", "shows", "reveals", "predicts", "achieves", "outperforms"]):
-            score += 2.0
-        if any(k in lower for k in ["significant", "associated", "correlated", "effective", "robust", "accurate", "validated"]):
-            score += 1.0
-        score += min(len(tokenize(s)) / 15.0, 2.0)
-        scored.append((score, s))
-    return [s for _, s in sorted(scored, key=lambda x: x[0], reverse=True)[:max_items]]
-
-
-def parse_openalex_abstract(inverted_index) -> str:
-    if not inverted_index or not isinstance(inverted_index, dict):
-        return ""
-    pos_to_word = {}
-    for word, positions in inverted_index.items():
-        for pos in positions:
-            pos_to_word[pos] = word
-    if not pos_to_word:
-        return ""
-    return " ".join(pos_to_word[i] for i in sorted(pos_to_word))
-
-
-def enrich_paper_semantics(query: str, paper: Dict) -> Dict:
-    paper = dict(paper)
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue]).strip()
-
-    concepts = extract_concepts_from_text(base_text, max_terms=GRAPH_MAX_CONCEPTS)
-    claims = extract_claim_like_sentences(abstract, max_items=GRAPH_MAX_CLAIMS)
-
-    rel = text_overlap_score(query, f"{title} {abstract}")
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    concept_bonus = min(len(concepts), 8) * 0.01
-
-    learned_score = float(paper.get("score", 0)) + rel * 0.5 + recency + doi_bonus + oa_bonus + concept_bonus
-
-    paper["concepts"] = concepts[:GRAPH_MAX_CONCEPTS]
-    paper["claims"] = claims[:GRAPH_MAX_CLAIMS]
-    paper["relevance"] = round(rel, 4)
-    paper["learned_score"] = round(learned_score, 4)
-    return paper
-
-
-def paper_identity_key(paper: Dict) -> str:
-    return (
-        normalize_doi(paper.get("doi") or "")
-        or (paper.get("external_ids") or {}).get("arxiv")
-        or norm_text(paper.get("title", "")).lower()
-        or str(paper.get("id"))
-    )
-
-
-def journal_query_links(query: str):
-    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
-    rows = []
-    for journal in JOURNALS:
-        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
-        if "ieeexplore" in journal["url"]:
-            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
-        rows.append({"name": journal["name"], "desc": journal["desc"], "url": url})
-    return rows
-
-
-def build_journal_html(query):
-    rows = []
-    for journal in journal_query_links(query):
-        rows.append(
-            f"""
-            <a class="journal-card" href="{safe_text(journal['url'])}" target="_blank" rel="noopener noreferrer">
-              <div>
-                <h4>{safe_text(journal['name'])}</h4>
-                <p>{safe_text(journal['desc'])}</p>
-              </div>
-              <span>Open</span>
-            </a>
-            """
-        )
-    return '<div class="journal-grid">' + ''.join(rows) + '</div>'
-
-
-def search_arxiv(query, max_results=8):
-    encoded = urllib.parse.quote(query)
-    url = (
-        "http://export.arxiv.org/api/query?search_query=all:"
-        f"{encoded}&start=0&max_results={max_results}&sortBy=relevance&sortOrder=descending"
-    )
-    response = requests.get(url, timeout=REQUEST_TIMEOUT)
-    response.raise_for_status()
-
-    root = ET.fromstring(response.text)
-    ns = {"atom": "http://www.w3.org/2005/Atom"}
-    papers = []
-
-    for entry in root.findall("atom:entry", ns):
-        title = " ".join((entry.findtext("atom:title", default="", namespaces=ns) or "").split())
-        summary = " ".join((entry.findtext("atom:summary", default="", namespaces=ns) or "").split())
-        published = entry.findtext("atom:published", default="", namespaces=ns)
-        paper_id = entry.findtext("atom:id", default="", namespaces=ns)
-        authors = [a.findtext("atom:name", default="", namespaces=ns) for a in entry.findall("atom:author", ns)]
-        pdf_url = ""
-
-        for link in entry.findall("atom:link", ns):
-            if link.attrib.get("title") == "pdf":
-                pdf_url = link.attrib.get("href", "")
-                break
-
-        papers.append({
-            "id": paper_id or title,
-            "title": title,
-            "summary": summary,
-            "abstract": summary,
-            "published": published[:10],
-            "authors": [a for a in authors[:8] if a],
-            "authors_text": ", ".join([a for a in authors[:4] if a]) or "Unknown authors",
-            "url": paper_id,
-            "pdf": pdf_url,
-            "doi": "",
-            "venue": "arXiv",
-            "year": published[:4] if published else "",
-            "source": "arxiv",
-            "score": 0.76,
-            "open_access": True,
-            "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
-        })
-
-    return papers
-
-
-def search_crossref(query, mode="topic", max_results=8):
-    headers = {"User-Agent": "dvnc-ai-space/0.3"}
-
-    if mode == "doi":
-        url = f"https://api.crossref.org/works/{urllib.parse.quote(query)}"
-        response = requests.get(url, headers=headers, timeout=REQUEST_TIMEOUT)
-        if response.status_code != 200:
-            return []
-        items = [response.json().get("message", {})]
-    else:
-        params = {"rows": max_results}
-        if mode in ("title", "paper_name"):
-            params["query.title"] = query
-        else:
-            params["query.bibliographic"] = query
-        response = requests.get("https://api.crossref.org/works", params=params, headers=headers, timeout=REQUEST_TIMEOUT)
-        response.raise_for_status()
-        items = response.json().get("message", {}).get("items", [])
-
-    out = []
-    for item in items:
-        authors = []
-        for a in item.get("author", []) or []:
-            name = " ".join(filter(None, [a.get("given"), a.get("family")])).strip()
-            if name:
-                authors.append(name)
-
-        title = (item.get("title") or ["Untitled"])[0]
-        year = ""
-        for key in ["published-print", "published-online", "created"]:
-            if item.get(key, {}).get("date-parts"):
-                year = str(item[key]["date-parts"][0][0])
-                break
-
-        abstract = re.sub("<.*?>", "", item.get("abstract") or "")
-        doi = normalize_doi(item.get("DOI", ""))
-
-        out.append({
-            "id": doi or title,
-            "title": norm_text(title),
-            "summary": norm_text(abstract)[:500],
-            "abstract": norm_text(abstract),
-            "published": year,
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("URL", ""),
-            "pdf": "",
-            "doi": doi,
-            "venue": (item.get("container-title") or [""])[0],
-            "year": year,
-            "source": "crossref",
-            "score": 0.72,
-            "open_access": None,
-            "external_ids": {"crossref": doi} if doi else {},
-        })
-
-    return out
-
-
-def search_openalex(query, mode="topic", max_results=8):
-    params = {"per-page": max_results}
-    if mode == "doi":
-        doi = normalize_doi(query)
-        params["filter"] = f"doi:https://doi.org/{doi}"
-    else:
-        params["search"] = query
-
-    response = requests.get("https://api.openalex.org/works", params=params, timeout=REQUEST_TIMEOUT)
-    response.raise_for_status()
-    items = response.json().get("results", [])
-
-    out = []
-    for item in items:
-        authors = []
-        for auth in item.get("authorships", [])[:8]:
-            author = auth.get("author") or {}
-            if author.get("display_name"):
-                authors.append(author["display_name"])
-
-        oa = item.get("open_access") or {}
-        doi = normalize_doi(item.get("doi") or "")
-        abstract = parse_openalex_abstract(item.get("abstract_inverted_index"))
-
-        out.append({
-            "id": item.get("id") or doi or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(abstract)[:500],
-            "abstract": norm_text(abstract),
-            "published": str(item.get("publication_year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": (item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or "",
-            "pdf": oa.get("oa_url") or "",
-            "doi": doi,
-            "venue": ((item.get("primary_location") or {}).get("source") or {}).get("display_name") or "",
-            "year": str(item.get("publication_year") or ""),
-            "source": "openalex",
-            "score": 0.80,
-            "open_access": oa.get("is_oa"),
-            "external_ids": item.get("ids") or {},
-        })
-
-    return out
-
-
-def search_semantic_scholar(query, mode="topic", max_results=8):
-    headers = {}
-    if SEMANTIC_SCHOLAR_API_KEY:
-        headers["x-api-key"] = SEMANTIC_SCHOLAR_API_KEY
-
-    fields = "title,authors,year,abstract,venue,externalIds,url,openAccessPdf"
-
-    if mode == "doi":
-        doi = normalize_doi(query)
-        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{urllib.parse.quote(doi)}"
-        response = requests.get(url, params={"fields": fields}, headers=headers, timeout=REQUEST_TIMEOUT)
-        if response.status_code != 200:
-            return []
-        items = [response.json()]
-    else:
-        response = requests.get(
-            "https://api.semanticscholar.org/graph/v1/paper/search",
-            params={"query": query, "limit": max_results, "fields": fields},
-            headers=headers,
-            timeout=REQUEST_TIMEOUT,
-        )
-        if response.status_code != 200:
-            return []
-        items = response.json().get("data", [])
-
-    out = []
-    for item in items:
-        external = item.get("externalIds") or {}
-        authors = [a.get("name") for a in item.get("authors", []) if a.get("name")]
-
-        out.append({
-            "id": external.get("CorpusId") or external.get("DOI") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(item.get("abstract", ""))[:500],
-            "abstract": norm_text(item.get("abstract", "")),
-            "published": str(item.get("year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("url") or "",
-            "pdf": (item.get("openAccessPdf") or {}).get("url") or "",
-            "doi": normalize_doi(external.get("DOI", "")),
-            "venue": item.get("venue") or "",
-            "year": str(item.get("year") or ""),
-            "source": "semantic_scholar",
-            "score": 0.84,
-            "open_access": bool((item.get("openAccessPdf") or {}).get("url")),
-            "external_ids": external,
-        })
-
-    return out
-
-
-def search_europe_pmc(query, mode="topic", max_results=8):
-    epmc_query = f'DOI:"{query}"' if mode == "doi" else query
-    params = {"query": epmc_query, "format": "json", "pageSize": max_results, "resultType": "core"}
-    response = requests.get("https://www.ebi.ac.uk/europepmc/webservices/rest/search", params=params, timeout=REQUEST_TIMEOUT)
-    if response.status_code != 200:
-        return []
-
-    items = response.json().get("resultList", {}).get("result", [])
-    out = []
-    for item in items:
-        author_string = item.get("authorString", "")
-        authors = [x.strip() for x in author_string.split(",")[:8] if x.strip()]
-        pmcid = item.get("pmcid", "")
-        pdf_url = f"https://europepmc.org/articles/{pmcid}?pdf=render" if pmcid else ""
-        landing_url = f"https://europepmc.org/article/PMC/{pmcid}" if pmcid else ""
-
-        out.append({
-            "id": item.get("id") or item.get("doi") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": norm_text(item.get("abstractText", ""))[:500],
-            "abstract": norm_text(item.get("abstractText", "")),
-            "published": str(item.get("pubYear") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": landing_url,
-            "pdf": pdf_url,
-            "doi": normalize_doi(item.get("doi", "")),
-            "venue": item.get("journalTitle", ""),
-            "year": str(item.get("pubYear") or ""),
-            "source": "europe_pmc",
-            "score": 0.78,
-            "open_access": bool(pmcid),
-            "external_ids": {"pmid": item.get("pmid"), "pmcid": pmcid},
-        })
-
-    return out
-
-
-def resolve_link(query):
-    url = (query or "").strip()
-    if not url:
-        return []
-
-    try:
-        response = requests.get(
-            url,
-            timeout=REQUEST_TIMEOUT,
-            allow_redirects=True,
-            headers={"User-Agent": "dvnc-ai-space/0.3"},
-        )
-
-        content_type = response.headers.get("content-type", "")
-        if "pdf" in content_type or url.lower().endswith(".pdf"):
-            name = Path(url.split("?")[0]).name or "linked-paper.pdf"
-            return [{
-                "id": url,
-                "title": name,
-                "summary": "Direct PDF link detected.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": url,
-                "doi": "",
-                "venue": "Direct PDF",
-                "year": "",
-                "source": "link",
-                "score": 0.66,
-                "open_access": True,
-                "external_ids": {},
-            }]
-
-        doi = ""
-        title = url
-        pdf_link = ""
-
-        if BeautifulSoup is not None:
-            soup = BeautifulSoup(response.text, "html.parser")
-            title = soup.title.text.strip() if soup.title else url
-
-            for meta_name in ["citation_doi", "dc.identifier", "dc.Identifier"]:
-                tag = soup.find("meta", attrs={"name": meta_name})
-                if tag and tag.get("content"):
-                    doi = normalize_doi(tag["content"].strip())
-                    break
-
-            for a in soup.find_all("a", href=True):
-                href = a["href"]
-                if ".pdf" in href.lower():
-                    pdf_link = href if href.startswith("http") else ""
-                    break
-
-        if doi:
-            results = search_crossref(doi, mode="doi", max_results=1)
-            if results:
-                if pdf_link and not results[0].get("pdf"):
-                    results[0]["pdf"] = pdf_link
-                if url and not results[0].get("url"):
-                    results[0]["url"] = url
-                return results
-
-        return [{
-            "id": url,
-            "title": title,
-            "summary": "Landing page resolved from direct link.",
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": pdf_link,
-            "doi": doi,
-            "venue": "Web Link",
-            "year": "",
-            "source": "link",
-            "score": 0.54,
-            "open_access": bool(pdf_link),
-            "external_ids": {},
-        }]
-    except Exception as e:
-        return [{
-            "id": url,
-            "title": "Link resolution error",
-            "summary": str(e),
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": "",
-            "doi": "",
-            "venue": "Link",
-            "year": "",
-            "source": "link",
-            "score": 0.20,
-            "open_access": None,
-            "external_ids": {},
-        }]
-
-
-def dedupe_papers(items: List[Dict]) -> List[Dict]:
-    seen = {}
-    for item in items:
-        key = paper_identity_key(item) or f"{item.get('source', 'src')}::{item.get('title', 'paper')}"
-        if key not in seen or float(item.get("score", 0)) > float(seen[key].get("score", 0)):
-            seen[key] = item
-    return sorted(seen.values(), key=lambda x: float(x.get("score", 0)), reverse=True)
-
-
-def discover_papers(query, mode, sources, max_results=10):
-    query = (query or "").strip()
-    if not query:
-        return []
-
-    mode = detect_query_type(query) if mode == "autonomous_web" else mode
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    results = []
-
-    if mode == "link":
-        return dedupe_papers(resolve_link(query))
-
-    if "arxiv" in selected_sources and mode != "doi":
-        try:
-            results.extend(search_arxiv(query, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-
-    if "crossref" in selected_sources:
-        try:
-            results.extend(search_crossref(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-
-    if "openalex" in selected_sources:
-        try:
-            results.extend(search_openalex(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-
-    if "semantic_scholar" in selected_sources:
-        try:
-            results.extend(search_semantic_scholar(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-
-    if "europe_pmc" in selected_sources:
-        try:
-            results.extend(search_europe_pmc(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-
-    papers = dedupe_papers(results)
-    papers = [enrich_paper_semantics(query, p) for p in papers]
-    papers = sorted(papers, key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-    return papers
-
-
-def paper_choice_value(index: int, paper: Dict) -> str:
-    doi = normalize_doi(paper.get("doi") or "")
-    title_slug = slugify(paper.get("title", ""))[:40]
-    return f"{index}|{doi}|{title_slug}"
-
-
-def paper_choice_label(index: int, paper: Dict) -> str:
-    score = round(float(paper.get("learned_score", paper.get("score", 0))), 3)
-    title = paper.get("title", "Untitled")
-    authors_text = paper.get("authors_text", "Unknown authors")[:80]
-    source = paper.get("source", "src")
-    return f"[{source}] {title} — {authors_text} — score {score}"
-
-
-def format_selection_choices(papers):
-    return [(paper_choice_label(i, paper), paper_choice_value(i, paper)) for i, paper in enumerate(papers)]
-
-
-def format_papers_html(papers):
-    if not papers:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No papers found yet.</p></div>'
-
-    items = []
-    for i, paper in enumerate(papers, start=1):
-        summary = safe_text((paper.get("summary") or paper.get("abstract") or "")[:280])
-        doi_line = f'<span class="paper-badge doi-badge">{safe_text(paper.get("doi"))}</span>' if paper.get("doi") else ""
-        pdf_link = paper.get("pdf") or "#"
-        abs_link = paper.get("url") or "#"
-        concepts_text = ", ".join((paper.get("concepts") or [])[:4])
-
-        items.append(
-            f"""
-            <article class="paper-card">
-              <div class="paper-topline">
-                <span class="paper-badge">{safe_text(paper.get('source', 'paper'))}</span>
-                <span class="paper-badge alt">{safe_text(paper.get('published', '') or 'Paper')}</span>
-                {doi_line}
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{summary or 'No abstract snippet available.'}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-                <div><strong>Venue:</strong> {safe_text(paper.get('venue', 'Unknown venue'))}</div>
-                <div><strong>Learned score:</strong> {safe_text(round(float(paper.get('learned_score', paper.get('score', 0))), 3))}</div>
-                <div><strong>Concepts:</strong> {safe_text(concepts_text or 'None extracted')}</div>
-              </div>
-              <div class="paper-links">
-                <a href="{safe_text(abs_link)}" target="_blank" rel="noopener noreferrer">Abstract</a>
-                <a href="{safe_text(pdf_link)}" target="_blank" rel="noopener noreferrer">PDF</a>
-              </div>
-            </article>
-            """
-        )
-
-    return '<div class="papers-grid">' + ''.join(items) + '</div>'
-
-
-def uploaded_pdf_summary(file_obj):
-    if not file_obj:
-        return "No PDF uploaded yet."
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    p = Path(path)
-    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, references, concepts, and claims."
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    if not nodes:
-        return """
-        <div class="panel brain-shell">
-          <div class="brain-header">
-            <div>
-              <p class="eyebrow">Learning Graph</p>
-              <h3>Self-Learning Knowledge Graph</h3>
-            </div>
-          </div>
-          <div class="brain-stage learning-empty">
-            <div class="empty-graph-copy">
-              <h4>No papers mapped yet</h4>
-              <p>Search papers, pick a topic, select candidates, or upload a PDF to grow the graph in real time.</p>
-            </div>
-          </div>
-        </div>
-        """
-
-    coords = [
-        (100, 90), (250, 60), (420, 75), (590, 115), (690, 250), (620, 395),
-        (455, 455), (280, 455), (110, 395), (60, 250), (215, 250), (365, 245),
-        (525, 250), (300, 145), (480, 340), (180, 340), (545, 175), (130, 170)
-    ]
-
-    graph_nodes = [dict(n) for n in nodes[:18]]
-    for i, node in enumerate(graph_nodes):
-        x, y = coords[i % len(coords)]
-        node["sx"] = x
-        node["sy"] = y
-
-    node_map = {n["id"]: n for n in graph_nodes}
-    edge_items = []
-    node_items = []
-    label_items = []
-
-    for edge in edges[:80]:
-        source = edge.get("source")
-        target = edge.get("target")
-        edge_type = edge.get("type", "")
-        if source in node_map and target in node_map:
-            a = node_map[source]
-            b = node_map[target]
-            edge_items.append(
-                f'<line class="learn-edge edge-{safe_text(edge_type.lower())}" x1="{a["sx"]}" y1="{a["sy"]}" x2="{b["sx"]}" y2="{b["sy"]}" />'
-            )
-
-    for node in graph_nodes:
-        kind = (node.get("kind") or node.get("type") or "paper").lower()
-        if kind == "topic":
-            kind = "query"
-        if kind == "uploadedpdf":
-            kind = "upload"
-
-        radius = 25 if kind == "query" else 18 if kind in {"concept", "author", "claim", "reference"} else 20
-        css_class = f"learn-node {kind}"
-
-        node_items.append(
-            f'<circle class="{css_class}" cx="{node["sx"]}" cy="{node["sy"]}" r="{radius}" />'
-        )
-
-        label = node.get("label") or node.get("title") or node.get("id")
-        label_items.append(
-            f'<text class="learn-label" x="{node["sx"] + 26}" y="{node["sy"] - 8}">{safe_text(str(label)[:46])}</text>'
-        )
-
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Learning Graph</p>
-          <h3>{safe_text(title)}</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-query"></i> topic</span>
-          <span><i class="dot dot-paper"></i> paper</span>
-          <span><i class="dot dot-upload"></i> uploaded PDF</span>
-          <span><i class="dot dot-concept"></i> concept</span>
-          <span><i class="dot dot-author"></i> author</span>
-          <span><i class="dot dot-ref"></i> reference</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 760 520" class="brain-svg" role="img" aria-label="Self-learning knowledge graph">
-          {''.join(edge_items)}
-          {''.join(node_items)}
-          {''.join(label_items)}
-        </svg>
-      </div>
-    </div>
-    """
-
-
-def build_learning_graph_state(query, papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-
-    for i, paper in enumerate(papers[:5], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper.get("title", f"Paper {i}"), "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{i}_{slugify(concept)[:20]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-
-    return nodes, edges
-
-
-def graph_from_selected(query, selected_papers, uploaded_name=None, parsed_state=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-
-    for i, paper in enumerate(selected_papers[:6], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper.get("title", f"Paper {i}"), "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-
-        for author in paper.get("authors", [])[:2]:
-            aid = f"author_{i}_{slugify(author)[:24]}"
-            nodes.append({"id": aid, "label": author, "kind": "author"})
-            edges.append({"source": pid, "target": aid, "type": "WRITTEN_BY"})
-
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{i}_{slugify(concept)[:24]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-
-        for claim in (paper.get("claims") or [])[:1]:
-            cid = f"claim_{i}_{slugify(claim)[:24]}"
-            nodes.append({"id": cid, "label": claim[:42], "kind": "claim"})
-            edges.append({"source": pid, "target": cid, "type": "ASSERTS"})
-
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-
-        if parsed_state and isinstance(parsed_state, dict):
-            for concept in (parsed_state.get("concepts") or [])[:3]:
-                cid = f"upload_concept_{slugify(concept)[:24]}"
-                nodes.append({"id": cid, "label": concept, "kind": "concept"})
-                edges.append({"source": "upload", "target": cid, "type": "MENTIONS"})
-
-    return nodes, edges
-
-
-def parse_pdf_with_grobid(pdf_path):
-    if not GROBID_URL:
-        raise RuntimeError("GROBID_URL is not set")
-
-    with open(pdf_path, "rb") as f:
-        files = {"input": (Path(pdf_path).name, f, "application/pdf")}
-        response = requests.post(
-            f"{GROBID_URL.rstrip('/')}/api/processFulltextDocument",
-            files=files,
-            data={"includeRawAffiliations": "1", "segmentSentences": "1"},
-            timeout=120,
-        )
-
-    response.raise_for_status()
-    tei_xml = response.text
-    root = ET.fromstring(tei_xml)
-    ns = {"tei": "http://www.tei-c.org/ns/1.0"}
-
-    title = root.findtext(".//tei:titleStmt/tei:title", default="", namespaces=ns)
-    if not title:
-        title = Path(pdf_path).name
-
-    abstract_parts = []
-    for p in root.findall(".//tei:profileDesc/tei:abstract//tei:p", ns):
-        abstract_parts.append(" ".join(list(p.itertext())))
-    abstract = norm_text(" ".join(abstract_parts))
-
-    authors = []
-    for author in root.findall(".//tei:sourceDesc//tei:author", ns):
-        parts = []
-        forename = author.findall(".//tei:forename", ns)
-        surname = author.findall(".//tei:surname", ns)
-        parts.extend([norm_text(" ".join(x.itertext())) for x in forename])
-        parts.extend([norm_text(" ".join(x.itertext())) for x in surname])
-        name = norm_text(" ".join(parts))
-        if name:
-            authors.append(name)
-
-    sections = []
-    text_pool = []
-    for div in root.findall(".//tei:text//tei:body//tei:div", ns):
-        head = div.findtext("./tei:head", default="", namespaces=ns)
-        paras = []
-        for p in div.findall(".//tei:p", ns):
-            para_text = norm_text(" ".join(list(p.itertext())))
-            if para_text:
-                paras.append(para_text)
-        joined = "\n".join(paras)
-        if head or joined:
-            sections.append({"heading": head or "Section", "text": joined[:4000]})
-            if joined:
-                text_pool.append(joined)
-
-    references = []
-    for bibl in root.findall(".//tei:listBibl//tei:biblStruct", ns)[:40]:
-        ref_title = bibl.findtext(".//tei:title", default="", namespaces=ns)
-        ref_doi = ""
-        for idno in bibl.findall(".//tei:idno", ns):
-            if (idno.attrib.get("type") or "").lower() == "doi":
-                ref_doi = norm_text(" ".join(idno.itertext()))
-                break
-        references.append({"title": norm_text(ref_title), "doi": normalize_doi(ref_doi)})
-
-    semantic_text = " ".join([abstract] + text_pool[:4])
-
-    return {
-        "parser": "grobid",
-        "title": norm_text(title),
-        "abstract": abstract,
-        "authors": authors[:12],
-        "sections": sections[:12],
-        "references": references[:40],
-        "claims": extract_claim_like_sentences(semantic_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(semantic_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": "",
-    }
-
-
-def parse_pdf_with_pymupdf(pdf_path):
-    if fitz is None:
-        raise RuntimeError("PyMuPDF not installed")
-
-    doc = fitz.open(pdf_path)
-    raw_text = "\n".join(page.get_text("text") for page in doc).strip()
-    first_page = raw_text[:4000]
-    lines = [x.strip() for x in first_page.splitlines() if x.strip()]
-    title = lines[0][:300] if lines else Path(pdf_path).name
-
-    abstract = ""
-    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n1[\.\s]|introduction)", raw_text, re.I | re.S)
-    if match:
-        abstract = norm_text(match.group(1))[:2500]
-
-    return {
-        "parser": "pymupdf",
-        "title": title,
-        "abstract": abstract,
-        "authors": [],
-        "sections": [{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(raw_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(raw_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": raw_text[:50000],
-    }
-
-
-def parse_pdf_with_docling(pdf_path):
-    try:
-        from docling.document_converter import DocumentConverter
-    except Exception as e:
-        raise RuntimeError(f"Docling import failed: {e}")
-
-    converter = DocumentConverter()
-    result = converter.convert(pdf_path)
-    doc = result.document
-    markdown = doc.export_to_markdown()
-
-    title = Path(pdf_path).name
-    first_nonempty = next((line.strip("# ").strip() for line in markdown.splitlines() if line.strip()), "")
-    if first_nonempty:
-        title = first_nonempty[:300]
-
-    return {
-        "parser": "docling",
-        "title": title,
-        "abstract": "",
-        "authors": [],
-        "sections": [{"heading": "Document", "text": markdown[:12000]}] if markdown else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(markdown, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(markdown, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": markdown[:50000],
-    }
-
-
-def parse_uploaded_pdf(file_obj, parser_order):
-    if not file_obj:
-        return "### PDF parse status\n\nNo PDF uploaded yet.", {}
-
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    parser_order = ensure_list(parser_order) or ["grobid", "docling", "pymupdf"]
-    errors = []
-
-    for parser_name in parser_order:
-        try:
-            if parser_name == "grobid":
-                result = parse_pdf_with_grobid(path)
-            elif parser_name == "docling":
-                result = parse_pdf_with_docling(path)
-            elif parser_name == "pymupdf":
-                result = parse_pdf_with_pymupdf(path)
-            else:
-                continue
-
-            summary = (
-                f"### PDF parse status\n\n"
-                f"- Parser used: {result['parser']}\n"
-                f"- Title: {result.get('title') or 'Unknown'}\n"
-                f"- Authors: {', '.join(result.get('authors')[:6]) if result.get('authors') else 'Unknown'}\n"
-                f"- Abstract found: {'Yes' if result.get('abstract') else 'No'}\n"
-                f"- Sections extracted: {len(result.get('sections') or [])}\n"
-                f"- References extracted: {len(result.get('references') or [])}\n"
-                f"- Concepts extracted: {len(result.get('concepts') or [])}\n"
-                f"- Claims extracted: {len(result.get('claims') or [])}\n"
-            )
-            return summary, result
-        except Exception as e:
-            errors.append(f"{parser_name}: {e}")
-
-    fail_summary = "### PDF parse status\n\n" + "\n".join([f"- {x}" for x in errors])
-    return fail_summary, {"parser": None, "errors": errors}
-
-
-def render_parse_result(parsed):
-    if not parsed or not isinstance(parsed, dict) or (not parsed.get("title") and not parsed.get("sections")):
-        return '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-
-    sections_html = []
-    for section in parsed.get("sections", [])[:6]:
-        sections_html.append(
-            f"""
-            <details class="agent-step">
-              <summary class="agent-summary">
-                <div class="agent-index">§</div>
-                <div class="agent-head">
-                  <h4>{safe_text(section.get('heading', 'Section'))}</h4>
-                  <span>section</span>
-                </div>
-              </summary>
-              <div class="agent-copy">
-                <p>{safe_text(section.get('text', '')[:1800])}</p>
-              </div>
-            </details>
-            """
-        )
-
-    refs = parsed.get("references", [])[:12]
-    refs_html = "".join(
-        f"<li>{safe_text(r.get('title') or 'Untitled')} {'· DOI ' + safe_text(r.get('doi')) if r.get('doi') else ''}</li>"
-        for r in refs
-    ) or "<li>No references extracted.</li>"
-
-    concepts = parsed.get("concepts", [])[:10]
-    claims = parsed.get("claims", [])[:6]
-    concepts_html = "".join(f"<li>{safe_text(x)}</li>" for x in concepts) or "<li>No concepts extracted.</li>"
-    claims_html = "".join(f"<li>{safe_text(x)}</li>" for x in claims) or "<li>No claims extracted.</li>"
-
-    title = safe_text(parsed.get("title") or "Parsed document")
-    abstract = safe_text((parsed.get("abstract") or "")[:2400]) or "No abstract extracted."
-    parser_name = safe_text(parsed.get("parser") or "unknown")
-
-    return f"""
-    <div class="panel" style="padding:18px">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">PDF Parse</p>
-          <h3>{title}</h3>
-        </div>
-        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name}</span></div>
-      </div>
-      <div class="parse-grid">
-        <div class="parse-card">
-          <h4>Abstract</h4>
-          <p>{abstract}</p>
-        </div>
-        <div class="parse-card">
-          <h4>References</h4>
-          <ul class="ref-list">{refs_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Concepts</h4>
-          <ul class="ref-list">{concepts_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Claims</h4>
-          <ul class="ref-list">{claims_html}</ul>
-        </div>
-      </div>
-      <div class="timeline" style="margin-top:14px;">
-        {''.join(sections_html) if sections_html else '<div class="panel" style="padding:16px;"><p>No sections extracted.</p></div>'}
-      </div>
-    </div>
-    """
-
-
-def add_node(nodes_by_id: Dict[str, Dict], node_id: str, node_type: str, label: str = "", **attrs):
-    if not node_id:
-        return
-    current = nodes_by_id.get(node_id, {})
-    merged = {"id": node_id, "type": node_type, "label": label or current.get("label", node_id)}
-    merged.update(current)
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            merged[key] = value
-    nodes_by_id[node_id] = merged
-
-
-def add_edge(edges: List[Dict], source: str, target: str, edge_type: str, **attrs):
-    if not source or not target or source == target:
-        return
-    edge = {"source": source, "target": target, "type": edge_type}
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            edge[key] = value
-    edges.append(edge)
-
-
-def build_ingest_payload(query, selected_papers, parsed_pdf=None):
-    nodes_by_id = {}
-    edges = []
-
-    topic_id = "topic:query"
-    add_node(nodes_by_id, topic_id, "Topic", label=query or "Research topic", query=query or "")
-
-    for i, paper in enumerate(selected_papers, start=1):
-        paper_id = normalize_doi(paper.get("doi")) or (paper.get("external_ids") or {}).get("arxiv") or f"paper:{i}:{slugify(paper.get('title', 'paper'))[:32]}"
-        add_node(
-            nodes_by_id,
-            paper_id,
-            "Paper",
-            label=paper.get("title") or f"Paper {i}",
-            title=paper.get("title"),
-            year=paper.get("year"),
-            venue=paper.get("venue"),
-            doi=normalize_doi(paper.get("doi")),
-            source=paper.get("source"),
-            url=paper.get("url"),
-            pdf=paper.get("pdf"),
-            score=paper.get("score"),
-            learned_score=paper.get("learned_score", paper.get("score")),
-            open_access=paper.get("open_access"),
-        )
-        add_edge(edges, topic_id, paper_id, "ABOUT", weight=paper.get("learned_score", paper.get("score", 0)))
-
-        for author in paper.get("authors", [])[:6]:
-            author_id = f"author:{slugify(author)[:64]}"
-            add_node(nodes_by_id, author_id, "Author", label=author, name=author)
-            add_edge(edges, paper_id, author_id, "WRITTEN_BY")
-
-        for concept in (paper.get("concepts") or [])[:6]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, paper_id, concept_id, "MENTIONS")
-
-        for claim in (paper.get("claims") or [])[:3]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, paper_id, claim_id, "ASSERTS")
-
-    if parsed_pdf and isinstance(parsed_pdf, dict) and parsed_pdf.get("title"):
-        doc_id = "upload:pdf"
-        add_node(nodes_by_id, doc_id, "UploadedPDF", label=parsed_pdf.get("title"), title=parsed_pdf.get("title"), parser=parsed_pdf.get("parser"))
-        add_edge(edges, topic_id, doc_id, "UPLOADED_SOURCE")
-
-        for concept in (parsed_pdf.get("concepts") or [])[:6]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, doc_id, concept_id, "MENTIONS")
-
-        for claim in (parsed_pdf.get("claims") or [])[:4]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, doc_id, claim_id, "ASSERTS")
-
-        for idx, ref in enumerate(parsed_pdf.get("references", [])[:12], start=1):
-            ref_title = ref.get("title") or f"Reference {idx}"
-            ref_doi = normalize_doi(ref.get("doi") or "")
-            ref_id = ref_doi or f"ref:{idx}:{slugify(ref_title)[:32]}"
-            add_node(nodes_by_id, ref_id, "Reference", label=ref_title, title=ref_title, doi=ref_doi)
-            add_edge(edges, doc_id, ref_id, "CITES")
-
-    return {"status": "ok", "nodes": list(nodes_by_id.values()), "edges": edges}
-
-
-def learn_from_payload(payload: Dict, query: str = "") -> Dict:
-    if not payload:
-        return GRAPH_MEMORY
-
-    GRAPH_MEMORY["queries"].append(query or "")
-    GRAPH_MEMORY["events"].append({
-        "ts": time.time(),
-        "query": query or "",
-        "nodes": len(payload.get("nodes", [])),
-        "edges": len(payload.get("edges", [])),
-    })
-
-    for node in payload.get("nodes", []):
-        node_id = node.get("id")
-        if not node_id:
-            continue
-        GRAPH_MEMORY["nodes"][node_id] = node
-
-        node_type = (node.get("type") or "").lower()
-        if node_type == "paper":
-            GRAPH_MEMORY["papers"][node_id] = node
-        if node_type == "concept" and node.get("label"):
-            GRAPH_MEMORY["concept_counts"][node["label"].lower()] += 1
-        if node_type == "claim" and node.get("label"):
-            GRAPH_MEMORY["claim_counts"][node["label"].lower()] += 1
-
-    GRAPH_MEMORY["edges"].extend(payload.get("edges", []))
-    return GRAPH_MEMORY
-
-
-def resolve_selected_papers(selected_indices, papers_state):
-    papers = ensure_list(papers_state)
-    selected_indices = ensure_list(selected_indices)
-    selected = []
-
-    if not selected_indices:
-        return selected
-
-    value_map = {paper_choice_value(i, paper): paper for i, paper in enumerate(papers)}
-    label_map = {paper_choice_label(i, paper): paper for i, paper in enumerate(papers)}
-
-    for idx in selected_indices:
-        try:
-            if isinstance(idx, int):
-                if 0 <= idx < len(papers):
-                    selected.append(papers[idx])
-                continue
-
-            idx_str = str(idx)
-            if idx_str in value_map:
-                selected.append(value_map[idx_str])
-                continue
-
-            if idx_str.isdigit():
-                num = int(idx_str)
-                if 0 <= num < len(papers):
-                    selected.append(papers[num])
-                    continue
-
-            if "|" in idx_str:
-                left = idx_str.split("|", 1)[0]
-                if left.isdigit():
-                    num = int(left)
-                    if 0 <= num < len(papers):
-                        selected.append(papers[num])
-                        continue
-
-            if idx_str in label_map:
-                selected.append(label_map[idx_str])
-                continue
-        except Exception:
-            continue
-
-    out = []
-    seen = set()
-    for paper in selected:
-        key = paper_identity_key(paper)
-        if key not in seen:
-            seen.add(key)
-            out.append(paper)
-    return out
-
-
-def summarize_learning_state(query_text, papers, selected_sources):
-    concept_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:3])
-
-    top_concepts = [c for c, _ in Counter([c.lower() for c in concept_pool]).most_common(6)]
-
-    return (
-        "### Discovery results\n\n"
-        f"- Query: {query_text}\n"
-        f"- Sources: {', '.join(selected_sources)}\n"
-        f"- Candidates found: {len(papers)}\n"
-        f"- Top learned concepts: {', '.join(top_concepts) if top_concepts else 'None'}\n"
-        "- Select papers below, then click **Ingest selected into graph**.\n"
-    )
-
-
-def run_paper_discovery(query, search_mode, sources, pdf_file):
-    query_text = norm_text(query or "")
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-
-    if not query_text and not pdf_file:
-        empty_graph = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            empty_graph,
-            '<div class="panel papers-panel" style="padding:18px"><p>Enter a topic, title, DOI, link, or upload a PDF to start learning.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            "No PDF uploaded yet.",
-            gr.update(choices=[], value=[]),
-            [],
-            "### No discovery results yet.",
-        )
-
-    if not query_text and pdf_file:
-        uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name
-        graph_nodes, graph_edges = build_learning_graph_state("", [], uploaded_name)
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges, "Uploaded PDF Waiting for Parse"),
-            '<div class="panel papers-panel" style="padding:18px"><p>No query yet. Parse the uploaded PDF or enter a research topic to begin discovery.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            "### Upload detected.\n\n- Parse the PDF to extract structure.\n- Or enter a topic to start discovery.",
-        )
-
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    try:
-        papers = discover_papers(query_text, search_mode, selected_sources, max_results=GRAPH_MAX_RESULTS)
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:6], uploaded_name)
-        graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Self-Learning Knowledge Graph")
-        papers_html = format_papers_html(papers)
-        journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
-        pdf_summary = uploaded_pdf_summary(pdf_file)
-        choices = format_selection_choices(papers)
-        status_md = summarize_learning_state(query_text, papers, selected_sources)
-
-        return (
-            graph_html,
-            papers_html,
-            journals_html,
-            pdf_summary,
-            gr.update(choices=choices, value=[]),
-            papers,
-            status_md,
-        )
-
-    except Exception as e:
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, [], uploaded_name)
-        error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges),
-            error_html,
-            build_journal_html(query_text or "biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
-        )
-
-
-def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
-    papers = ensure_list(papers_state)
-    selected = resolve_selected_papers(selected_indices, papers)
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    if not selected and parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title") and papers:
-        selected = papers[:3]
-
-    if not selected and not (parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title")):
-        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            graph_html,
-            "### Graph ingest status\n\nSelect papers or parse an uploaded PDF first.",
-            {"status": "empty", "nodes": [], "edges": []},
-        )
-
-    query_text = norm_text(query or "")
-    if not query_text and isinstance(parsed_state, dict):
-        query_text = parsed_state.get("title") or "Research topic"
-    if not query_text:
-        query_text = "Research topic"
-
-    selected = [enrich_paper_semantics(query_text, paper) for paper in selected]
-    graph_nodes, graph_edges = graph_from_selected(query_text, selected, uploaded_name, parsed_state if isinstance(parsed_state, dict) else None)
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
-
-    payload = build_ingest_payload(query_text, selected, parsed_state if isinstance(parsed_state, dict) else None)
-    learn_from_payload(payload, query=query_text)
-
-    top_concepts = []
-    for paper in selected:
-        top_concepts.extend((paper.get("concepts") or [])[:3])
-    if isinstance(parsed_state, dict):
-        top_concepts.extend((parsed_state.get("concepts") or [])[:3])
-
-    summary_lines = [
-        "### Graph ingest status",
-        "",
-        f"- Topic: {query_text}",
-        f"- Selected papers: {len(selected)}",
-        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
-        f"- Nodes created: {len(payload['nodes'])}",
-        f"- Edges created: {len(payload['edges'])}",
-        f"- Learned concepts: {', '.join(unique_keep_order(top_concepts)[:8]) if top_concepts else 'None'}",
-        f"- Memory papers stored: {len(GRAPH_MEMORY['papers'])}",
-        f"- Memory concepts stored: {len(GRAPH_MEMORY['concept_counts'])}",
-    ]
-
-    return graph_html, "\n".join(summary_lines), payload

dvnc_ai_v2_hf/deprecated/self_learning_graph_old5.py

@@ -1,1490 +0,0 @@
-import html
-import json
-import os
-import re
-import time
-import urllib.parse
-import xml.etree.ElementTree as ET
-from collections import Counter
-from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
-
-import gradio as gr
-import requests
-
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-
-JOURNALS = [
-    {"name": "Nature", "url": "https://www.nature.com/search", "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science", "url": "https://www.science.org/search", "desc": "High-impact science journal and family."},
-    {"name": "Cell", "url": "https://www.cell.com/search", "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search", "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore", "url": "https://ieeexplore.ieee.org/search/searchresult.jsp", "desc": "Engineering, AI, signal processing, and systems."},
-]
-
-SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
-SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-DEFAULT_SOURCES = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "").strip()
-GROBID_URL = os.getenv("GROBID_URL", "").strip()
-OPENALEX_EMAIL = os.getenv("OPENALEX_EMAIL", "").strip()
-CROSSREF_MAILTO = os.getenv("CROSSREF_MAILTO", "").strip()
-REQUEST_TIMEOUT = int(os.getenv("REQUEST_TIMEOUT", "25"))
-GRAPH_MAX_CONCEPTS = int(os.getenv("GRAPH_MAX_CONCEPTS", "12"))
-GRAPH_MAX_CLAIMS = int(os.getenv("GRAPH_MAX_CLAIMS", "8"))
-GRAPH_MAX_RESULTS = int(os.getenv("GRAPH_MAX_RESULTS", "12"))
-GRAPH_MAX_EXPANSIONS = int(os.getenv("GRAPH_MAX_EXPANSIONS", "6"))
-GRAPH_MAX_NODES = int(os.getenv("GRAPH_MAX_NODES", "400"))
-GRAPH_MAX_EDGES = int(os.getenv("GRAPH_MAX_EDGES", "1200"))
-MAX_ABSTRACT_CHARS = int(os.getenv("MAX_ABSTRACT_CHARS", "4000"))
-MAX_RAW_TEXT_CHARS = int(os.getenv("MAX_RAW_TEXT_CHARS", "70000"))
-
-STOPWORDS = {
-    "a", "an", "and", "are", "as", "at", "be", "been", "being", "by", "can", "could", "did", "do", "does",
-    "for", "from", "had", "has", "have", "if", "in", "into", "is", "it", "its", "may", "might", "of", "on",
-    "or", "our", "such", "that", "the", "their", "there", "these", "this", "those", "to", "using", "use",
-    "used", "via", "was", "were", "will", "with", "within", "without", "we", "they", "you", "your", "study",
-    "paper", "research", "results", "method", "methods", "analysis", "approach", "toward", "towards",
-    "based", "new", "novel", "effect", "effects", "model", "models", "system", "systems", "show", "shows",
-    "introduction", "conclusion", "discussion", "figure", "table", "supplementary", "material", "materials",
-}
-
-GRAPH_MEMORY = {
-    "papers": {},
-    "nodes": {},
-    "edges": [],
-    "concept_counts": Counter(),
-    "claim_counts": Counter(),
-    "queries": [],
-    "events": [],
-    "frontier": [],
-    "payloads": [],
-}
-
-
-class ScholarlyClient:
-    def __init__(self):
-        self.session = requests.Session()
-        ua = "dvnc-ai-self-learning-graph/1.0"
-        if CROSSREF_MAILTO:
-            ua += f" (mailto:{CROSSREF_MAILTO})"
-        self.session.headers.update({"User-Agent": ua, "Accept": "application/json, text/xml, */*"})
-        self.session_timeout = REQUEST_TIMEOUT
-
-    def get(self, url: str, **kwargs):
-        timeout = kwargs.pop("timeout", self.session_timeout)
-        return self.session.get(url, timeout=timeout, **kwargs)
-
-    def post(self, url: str, **kwargs):
-        timeout = kwargs.pop("timeout", max(self.session_timeout, 120))
-        return self.session.post(url, timeout=timeout, **kwargs)
-
-
-HTTP = ScholarlyClient()
-
-
-def safe_text(x, default=""):
-    return html.escape(str(x if x is not None else default))
-
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def slugify(text: str) -> str:
-    return re.sub(r"[^a-z0-9]+", "-", (text or "").lower()).strip("-")
-
-
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-def truncate_text(text: str, limit: int) -> str:
-    text = norm_text(text)
-    return text if len(text) <= limit else text[: limit - 1].rstrip() + "…"
-
-
-def normalize_doi(text: str) -> str:
-    text = (text or "").strip()
-    text = re.sub(r"^https?://(dx\.)?doi\.org/", "", text, flags=re.I)
-    return text.strip().rstrip("/")
-
-
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    doi_pattern = r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$"
-    if re.match(doi_pattern, q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-
-
-def tokenize(text: str) -> List[str]:
-    return [t for t in re.findall(r"[a-zA-Z][a-zA-Z0-9\-]{2,}", (text or "").lower()) if t not in STOPWORDS]
-
-
-def unique_keep_order(items: List[str]) -> List[str]:
-    seen = set()
-    out = []
-    for item in items:
-        key = norm_text(item).lower()
-        if key and key not in seen:
-            seen.add(key)
-            out.append(norm_text(item))
-    return out
-
-
-def text_overlap_score(a: str, b: str) -> float:
-    sa = set(tokenize(a))
-    sb = set(tokenize(b))
-    if not sa or not sb:
-        return 0.0
-    return len(sa & sb) / max(1, len(sa | sb))
-
-
-def compute_recency_bonus(year: str) -> float:
-    try:
-        y = int(str(year)[:4])
-    except Exception:
-        return 0.0
-    current = time.gmtime().tm_year
-    age = max(current - y, 0)
-    return max(0.0, 0.14 - age * 0.015)
-
-
-def extract_candidate_phrases(text: str, max_terms: int = 20) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-    tokens = re.findall(r"[A-Za-z][A-Za-z0-9\-]{2,}", text)
-    phrases = []
-    for n in (3, 2, 1):
-        for i in range(len(tokens) - n + 1):
-            phrase = " ".join(tokens[i:i + n]).strip().lower()
-            if len(phrase) < 4:
-                continue
-            parts = phrase.split()
-            if any(p in STOPWORDS for p in parts):
-                continue
-            if all(len(p) <= 2 for p in parts):
-                continue
-            phrases.append(phrase)
-    counts = Counter(phrases)
-    ranked = [p for p, _ in counts.most_common(max_terms * 4)]
-    filtered = []
-    for phrase in ranked:
-        if phrase in filtered:
-            continue
-        if any(phrase != other and phrase in other for other in filtered):
-            continue
-        filtered.append(phrase)
-        if len(filtered) >= max_terms:
-            break
-    return filtered
-
-
-def extract_concepts_from_text(text: str, max_terms: int = GRAPH_MAX_CONCEPTS) -> List[str]:
-    return extract_candidate_phrases(text, max_terms=max_terms)
-
-
-def extract_claim_like_sentences(text: str, max_items: int = GRAPH_MAX_CLAIMS) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-    parts = re.split(r"(?<=[\.\!\?])\s+", text)
-    scored = []
-    for sentence in parts:
-        s = norm_text(sentence)
-        if len(s) < 40 or len(s) > 320:
-            continue
-        lower = s.lower()
-        score = 0.0
-        if any(k in lower for k in ["improves", "reduces", "increases", "suggests", "demonstrates", "shows", "reveals", "predicts", "achieves", "outperforms", "enables", "supports"]):
-            score += 2.0
-        if any(k in lower for k in ["significant", "associated", "correlated", "effective", "robust", "accurate", "validated", "statistically"]):
-            score += 1.0
-        if any(k in lower for k in ["compared", "versus", "baseline", "state-of-the-art", "sota"]):
-            score += 1.0
-        score += min(len(tokenize(s)) / 15.0, 2.0)
-        scored.append((score, s))
-    return [s for _, s in sorted(scored, key=lambda x: x[0], reverse=True)[:max_items]]
-
-
-def parse_openalex_abstract(inverted_index) -> str:
-    if not inverted_index or not isinstance(inverted_index, dict):
-        return ""
-    pos_to_word = {}
-    for word, positions in inverted_index.items():
-        for pos in positions:
-            pos_to_word[pos] = word
-    if not pos_to_word:
-        return ""
-    return " ".join(pos_to_word[i] for i in sorted(pos_to_word))
-
-
-def score_frontier_candidate(query: str, seed_concepts: List[str], paper: Dict) -> Dict:
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue])
-    rel = text_overlap_score(query, base_text)
-    concept_overlap = 0.0
-    if seed_concepts:
-        concept_overlap = text_overlap_score(" ".join(seed_concepts), " ".join(paper.get("concepts") or []))
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    score = float(paper.get("score", 0)) + rel * 0.45 + concept_overlap * 0.2 + recency + doi_bonus + oa_bonus
-    paper["frontier_score"] = round(score, 4)
-    paper["frontier_relevance"] = round(rel, 4)
-    paper["frontier_concept_overlap"] = round(concept_overlap, 4)
-    return paper
-
-
-def enrich_paper_semantics(query: str, paper: Dict) -> Dict:
-    paper = dict(paper)
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue]).strip()
-    concepts = extract_concepts_from_text(base_text, max_terms=GRAPH_MAX_CONCEPTS)
-    claims = extract_claim_like_sentences(abstract, max_items=GRAPH_MAX_CLAIMS)
-    rel = text_overlap_score(query, f"{title} {abstract}")
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    concept_bonus = min(len(concepts), 8) * 0.01
-    learned_score = float(paper.get("score", 0)) + rel * 0.5 + recency + doi_bonus + oa_bonus + concept_bonus
-    paper["concepts"] = concepts[:GRAPH_MAX_CONCEPTS]
-    paper["claims"] = claims[:GRAPH_MAX_CLAIMS]
-    paper["relevance"] = round(rel, 4)
-    paper["learned_score"] = round(learned_score, 4)
-    return paper
-
-
-def paper_identity_key(paper: Dict) -> str:
-    return (
-        normalize_doi(paper.get("doi") or "")
-        or (paper.get("external_ids") or {}).get("arxiv")
-        or (paper.get("external_ids") or {}).get("pmcid")
-        or norm_text(paper.get("title", "")).lower()
-        or str(paper.get("id"))
-    )
-
-
-def journal_query_links(query: str):
-    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
-    rows = []
-    for journal in JOURNALS:
-        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
-        if "ieeexplore" in journal["url"]:
-            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
-        rows.append({"name": journal["name"], "desc": journal["desc"], "url": url})
-    return rows
-
-
-def build_journal_html(query):
-    rows = []
-    for journal in journal_query_links(query):
-        rows.append(
-            f"""
-            <a class="journal-card" href="{safe_text(journal['url'])}" target="_blank" rel="noopener noreferrer">
-              <div>
-                <h4>{safe_text(journal['name'])}</h4>
-                <p>{safe_text(journal['desc'])}</p>
-              </div>
-              <span>Open</span>
-            </a>
-            """
-        )
-    return '<div class="journal-grid">' + ''.join(rows) + '</div>'
-
-
-def search_arxiv(query, max_results=8):
-    encoded = urllib.parse.quote(query)
-    url = (
-        "http://export.arxiv.org/api/query?search_query=all:"
-        f"{encoded}&start=0&max_results={max_results}&sortBy=relevance&sortOrder=descending"
-    )
-    response = HTTP.get(url)
-    response.raise_for_status()
-    root = ET.fromstring(response.text)
-    ns = {"atom": "http://www.w3.org/2005/Atom"}
-    papers = []
-    for entry in root.findall("atom:entry", ns):
-        title = " ".join((entry.findtext("atom:title", default="", namespaces=ns) or "").split())
-        summary = truncate_text(" ".join((entry.findtext("atom:summary", default="", namespaces=ns) or "").split()), MAX_ABSTRACT_CHARS)
-        published = entry.findtext("atom:published", default="", namespaces=ns)
-        paper_id = entry.findtext("atom:id", default="", namespaces=ns)
-        authors = [a.findtext("atom:name", default="", namespaces=ns) for a in entry.findall("atom:author", ns)]
-        pdf_url = ""
-        for link in entry.findall("atom:link", ns):
-            if link.attrib.get("title") == "pdf":
-                pdf_url = link.attrib.get("href", "")
-                break
-        papers.append({
-            "id": paper_id or title,
-            "title": title,
-            "summary": summary,
-            "abstract": summary,
-            "published": published[:10],
-            "authors": [a for a in authors[:8] if a],
-            "authors_text": ", ".join([a for a in authors[:4] if a]) or "Unknown authors",
-            "url": paper_id,
-            "pdf": pdf_url,
-            "doi": "",
-            "venue": "arXiv",
-            "year": published[:4] if published else "",
-            "source": "arxiv",
-            "score": 0.76,
-            "open_access": True,
-            "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
-        })
-    return papers
-
-
-def search_crossref(query, mode="topic", max_results=8):
-    params = {}
-    if CROSSREF_MAILTO:
-        params["mailto"] = CROSSREF_MAILTO
-    if mode == "doi":
-        url = f"https://api.crossref.org/works/{urllib.parse.quote(query)}"
-        response = HTTP.get(url, params=params)
-        if response.status_code != 200:
-            return []
-        items = [response.json().get("message", {})]
-    else:
-        params["rows"] = max_results
-        if mode in ("title", "paper_name"):
-            params["query.title"] = query
-        else:
-            params["query.bibliographic"] = query
-        response = HTTP.get("https://api.crossref.org/works", params=params)
-        response.raise_for_status()
-        items = response.json().get("message", {}).get("items", [])
-    out = []
-    for item in items:
-        authors = []
-        for a in item.get("author", []) or []:
-            name = " ".join(filter(None, [a.get("given"), a.get("family")])).strip()
-            if name:
-                authors.append(name)
-        title = (item.get("title") or ["Untitled"])[0]
-        year = ""
-        for key in ["published-print", "published-online", "created"]:
-            if item.get(key, {}).get("date-parts"):
-                year = str(item[key]["date-parts"][0][0])
-                break
-        abstract = truncate_text(re.sub("<.*?>", "", item.get("abstract") or ""), MAX_ABSTRACT_CHARS)
-        doi = normalize_doi(item.get("DOI", ""))
-        out.append({
-            "id": doi or title,
-            "title": norm_text(title),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": year,
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("URL", ""),
-            "pdf": "",
-            "doi": doi,
-            "venue": (item.get("container-title") or [""])[0],
-            "year": year,
-            "source": "crossref",
-            "score": 0.72,
-            "open_access": None,
-            "external_ids": {"crossref": doi} if doi else {},
-        })
-    return out
-
-
-def search_openalex(query, mode="topic", max_results=8):
-    params = {"per-page": max_results}
-    if OPENALEX_EMAIL:
-        params["mailto"] = OPENALEX_EMAIL
-    if mode == "doi":
-        doi = normalize_doi(query)
-        params["filter"] = f"doi:https://doi.org/{doi}"
-    else:
-        params["search"] = query
-    response = HTTP.get("https://api.openalex.org/works", params=params)
-    response.raise_for_status()
-    items = response.json().get("results", [])
-    out = []
-    for item in items:
-        authors = []
-        for auth in item.get("authorships", [])[:8]:
-            author = auth.get("author") or {}
-            if author.get("display_name"):
-                authors.append(author["display_name"])
-        oa = item.get("open_access") or {}
-        doi = normalize_doi(item.get("doi") or "")
-        abstract = truncate_text(parse_openalex_abstract(item.get("abstract_inverted_index")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": item.get("id") or doi or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("publication_year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": (item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or "",
-            "pdf": oa.get("oa_url") or "",
-            "doi": doi,
-            "venue": ((item.get("primary_location") or {}).get("source") or {}).get("display_name") or "",
-            "year": str(item.get("publication_year") or ""),
-            "source": "openalex",
-            "score": 0.80,
-            "open_access": oa.get("is_oa"),
-            "external_ids": item.get("ids") or {},
-        })
-    return out
-
-
-def search_semantic_scholar(query, mode="topic", max_results=8):
-    headers = {}
-    if SEMANTIC_SCHOLAR_API_KEY:
-        headers["x-api-key"] = SEMANTIC_SCHOLAR_API_KEY
-    fields = "title,authors,year,abstract,venue,externalIds,url,openAccessPdf"
-    if mode == "doi":
-        doi = normalize_doi(query)
-        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{urllib.parse.quote(doi)}"
-        response = HTTP.get(url, params={"fields": fields}, headers=headers)
-        if response.status_code != 200:
-            return []
-        items = [response.json()]
-    else:
-        response = HTTP.get(
-            "https://api.semanticscholar.org/graph/v1/paper/search",
-            params={"query": query, "limit": max_results, "fields": fields},
-            headers=headers,
-        )
-        if response.status_code != 200:
-            return []
-        items = response.json().get("data", [])
-    out = []
-    for item in items:
-        external = item.get("externalIds") or {}
-        authors = [a.get("name") for a in item.get("authors", []) if a.get("name")]
-        abstract = truncate_text(norm_text(item.get("abstract", "")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": external.get("CorpusId") or external.get("DOI") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("url") or "",
-            "pdf": (item.get("openAccessPdf") or {}).get("url") or "",
-            "doi": normalize_doi(external.get("DOI", "")),
-            "venue": item.get("venue") or "",
-            "year": str(item.get("year") or ""),
-            "source": "semantic_scholar",
-            "score": 0.84,
-            "open_access": bool((item.get("openAccessPdf") or {}).get("url")),
-            "external_ids": external,
-        })
-    return out
-
-
-def search_europe_pmc(query, mode="topic", max_results=8):
-    epmc_query = f'DOI:"{query}"' if mode == "doi" else query
-    params = {"query": epmc_query, "format": "json", "pageSize": max_results, "resultType": "core"}
-    response = HTTP.get("https://www.ebi.ac.uk/europepmc/webservices/rest/search", params=params)
-    if response.status_code != 200:
-        return []
-    items = response.json().get("resultList", {}).get("result", [])
-    out = []
-    for item in items:
-        author_string = item.get("authorString", "")
-        authors = [x.strip() for x in author_string.split(",")[:8] if x.strip()]
-        pmcid = item.get("pmcid", "")
-        pdf_url = f"https://europepmc.org/articles/{pmcid}?pdf=render" if pmcid else ""
-        landing_url = f"https://europepmc.org/article/PMC/{pmcid}" if pmcid else ""
-        abstract = truncate_text(norm_text(item.get("abstractText", "")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": item.get("id") or item.get("doi") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("pubYear") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": landing_url,
-            "pdf": pdf_url,
-            "doi": normalize_doi(item.get("doi", "")),
-            "venue": item.get("journalTitle", ""),
-            "year": str(item.get("pubYear") or ""),
-            "source": "europe_pmc",
-            "score": 0.78,
-            "open_access": bool(pmcid),
-            "external_ids": {"pmid": item.get("pmid"), "pmcid": pmcid},
-        })
-    return out
-
-
-def resolve_link(query):
-    url = (query or "").strip()
-    if not url:
-        return []
-    try:
-        response = HTTP.get(url, allow_redirects=True, headers={"User-Agent": "dvnc-ai-space/1.0"})
-        content_type = response.headers.get("content-type", "")
-        if "pdf" in content_type or url.lower().endswith(".pdf"):
-            name = Path(url.split("?")[0]).name or "linked-paper.pdf"
-            return [{
-                "id": url,
-                "title": name,
-                "summary": "Direct PDF link detected.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": url,
-                "doi": "",
-                "venue": "Direct PDF",
-                "year": "",
-                "source": "link",
-                "score": 0.66,
-                "open_access": True,
-                "external_ids": {},
-            }]
-        doi = ""
-        title = url
-        pdf_link = ""
-        if BeautifulSoup is not None:
-            soup = BeautifulSoup(response.text, "html.parser")
-            title = soup.title.text.strip() if soup.title else url
-            for meta_name in ["citation_doi", "dc.identifier", "dc.Identifier"]:
-                tag = soup.find("meta", attrs={"name": meta_name})
-                if tag and tag.get("content"):
-                    doi = normalize_doi(tag["content"].strip())
-                    break
-            for a in soup.find_all("a", href=True):
-                href = a["href"]
-                if ".pdf" in href.lower():
-                    pdf_link = href if href.startswith("http") else urllib.parse.urljoin(url, href)
-                    break
-        if doi:
-            results = search_crossref(doi, mode="doi", max_results=1)
-            if results:
-                if pdf_link and not results[0].get("pdf"):
-                    results[0]["pdf"] = pdf_link
-                if url and not results[0].get("url"):
-                    results[0]["url"] = url
-                return results
-        return [{
-            "id": url,
-            "title": title,
-            "summary": "Landing page resolved from direct link.",
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": pdf_link,
-            "doi": doi,
-            "venue": "Web Link",
-            "year": "",
-            "source": "link",
-            "score": 0.54,
-            "open_access": bool(pdf_link),
-            "external_ids": {},
-        }]
-    except Exception as e:
-        return [{
-            "id": url,
-            "title": "Link resolution error",
-            "summary": str(e),
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": "",
-            "doi": "",
-            "venue": "Link",
-            "year": "",
-            "source": "link",
-            "score": 0.20,
-            "open_access": None,
-            "external_ids": {},
-        }]
-
-
-def dedupe_papers(items: List[Dict]) -> List[Dict]:
-    seen = {}
-    for item in items:
-        key = paper_identity_key(item) or f"{item.get('source', 'src')}::{item.get('title', 'paper')}"
-        current = seen.get(key)
-        candidate_score = float(item.get("learned_score", item.get("score", 0)))
-        current_score = float(current.get("learned_score", current.get("score", 0))) if current else -1
-        if current is None or candidate_score > current_score:
-            seen[key] = item
-    return sorted(seen.values(), key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-
-
-def discover_papers(query, mode, sources, max_results=10):
-    query = (query or "").strip()
-    if not query:
-        return []
-    mode = detect_query_type(query) if mode == "autonomous_web" else mode
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    results = []
-    if mode == "link":
-        return dedupe_papers(resolve_link(query))
-    if "arxiv" in selected_sources and mode != "doi":
-        try:
-            results.extend(search_arxiv(query, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "crossref" in selected_sources:
-        try:
-            results.extend(search_crossref(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "openalex" in selected_sources:
-        try:
-            results.extend(search_openalex(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "semantic_scholar" in selected_sources:
-        try:
-            results.extend(search_semantic_scholar(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "europe_pmc" in selected_sources:
-        try:
-            results.extend(search_europe_pmc(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    papers = [enrich_paper_semantics(query, p) for p in dedupe_papers(results)]
-    papers = sorted(papers, key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-    return papers[:max_results]
-
-
-def propose_expansion_queries(query: str, papers: List[Dict], parsed_state: Optional[Dict] = None, limit: int = GRAPH_MAX_EXPANSIONS) -> List[str]:
-    concept_pool = []
-    venue_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:4])
-        if paper.get("venue"):
-            venue_pool.append(paper["venue"])
-    if parsed_state and isinstance(parsed_state, dict):
-        concept_pool.extend((parsed_state.get("concepts") or [])[:6])
-    ranked_concepts = [c for c, _ in Counter([norm_text(c).lower() for c in concept_pool if c]).most_common(limit * 2)]
-    expansions = [norm_text(query)] if query else []
-    for concept in ranked_concepts:
-        if not concept:
-            continue
-        if query:
-            expansions.append(f"{query} {concept}")
-        else:
-            expansions.append(concept)
-    for venue in unique_keep_order(venue_pool)[:2]:
-        if query and venue:
-            expansions.append(f"{query} {venue}")
-    return unique_keep_order(expansions)[:limit]
-
-
-def frontier_expand(query: str, sources: List[str], selected_papers: List[Dict], parsed_state: Optional[Dict] = None, per_query: int = 4) -> List[Dict]:
-    seed_concepts = []
-    for p in selected_papers[:6]:
-        seed_concepts.extend((p.get("concepts") or [])[:4])
-    if parsed_state and isinstance(parsed_state, dict):
-        seed_concepts.extend((parsed_state.get("concepts") or [])[:6])
-    expansion_queries = propose_expansion_queries(query, selected_papers, parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    frontier = []
-    for eq in expansion_queries:
-        try:
-            items = discover_papers(eq, "topic", sources, max_results=per_query)
-            for item in items:
-                frontier.append(score_frontier_candidate(query or eq, seed_concepts, item))
-        except Exception:
-            continue
-    frontier = dedupe_papers(frontier)
-    frontier.sort(key=lambda x: float(x.get("frontier_score", x.get("learned_score", x.get("score", 0))),), reverse=True)
-    GRAPH_MEMORY["frontier"] = frontier[: GRAPH_MAX_EXPANSIONS * per_query]
-    return GRAPH_MEMORY["frontier"]
-
-
-def paper_choice_value(index: int, paper: Dict) -> str:
-    doi = normalize_doi(paper.get("doi") or "")
-    title_slug = slugify(paper.get("title", ""))[:40]
-    return f"{index}|{doi}|{title_slug}"
-
-
-def paper_choice_label(index: int, paper: Dict) -> str:
-    score = round(float(paper.get("learned_score", paper.get("score", 0))), 3)
-    title = paper.get("title", "Untitled")
-    authors_text = paper.get("authors_text", "Unknown authors")[:80]
-    source = paper.get("source", "src")
-    return f"[{source}] {title} — {authors_text} — score {score}"
-
-
-def format_selection_choices(papers):
-    return [(paper_choice_label(i, paper), paper_choice_value(i, paper)) for i, paper in enumerate(papers)]
-
-
-def format_papers_html(papers):
-    if not papers:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No papers found yet.</p></div>'
-    items = []
-    for i, paper in enumerate(papers, start=1):
-        summary = safe_text((paper.get("summary") or paper.get("abstract") or "")[:280])
-        doi_line = f'<span class="paper-badge doi-badge">{safe_text(paper.get("doi"))}</span>' if paper.get("doi") else ""
-        pdf_link = paper.get("pdf") or "#"
-        abs_link = paper.get("url") or "#"
-        concepts_text = ", ".join((paper.get("concepts") or [])[:4])
-        items.append(
-            f"""
-            <article class="paper-card">
-              <div class="paper-topline">
-                <span class="paper-badge">{safe_text(paper.get('source', 'paper'))}</span>
-                <span class="paper-badge alt">{safe_text(paper.get('published', '') or 'Paper')}</span>
-                {doi_line}
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{summary or 'No abstract snippet available.'}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-                <div><strong>Venue:</strong> {safe_text(paper.get('venue', 'Unknown venue'))}</div>
-                <div><strong>Learned score:</strong> {safe_text(round(float(paper.get('learned_score', paper.get('score', 0))), 3))}</div>
-                <div><strong>Concepts:</strong> {safe_text(concepts_text or 'None extracted')}</div>
-              </div>
-              <div class="paper-links">
-                <a href="{safe_text(abs_link)}" target="_blank" rel="noopener noreferrer">Abstract</a>
-                <a href="{safe_text(pdf_link)}" target="_blank" rel="noopener noreferrer">PDF</a>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(items) + '</div>'
-
-
-def format_frontier_html(frontier):
-    if not frontier:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No autonomous expansion candidates yet.</p></div>'
-    cards = []
-    for i, paper in enumerate(frontier[:12], start=1):
-        cards.append(
-            f"""
-            <article class="paper-card frontier-card">
-              <div class="paper-topline">
-                <span class="paper-badge">frontier</span>
-                <span class="paper-badge alt">{safe_text(paper.get('source', 'paper'))}</span>
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{safe_text((paper.get('summary') or paper.get('abstract') or '')[:260])}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Frontier score:</strong> {safe_text(paper.get('frontier_score', paper.get('learned_score', paper.get('score', 0))))}</div>
-                <div><strong>Concept overlap:</strong> {safe_text(paper.get('frontier_concept_overlap', 0))}</div>
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(cards) + '</div>'
-
-
-def uploaded_pdf_summary(file_obj):
-    if not file_obj:
-        return "No PDF uploaded yet."
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    p = Path(path)
-    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, references, concepts, and claims."
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    if not nodes:
-        return """
-        <div class="panel brain-shell">
-          <div class="brain-header">
-            <div>
-              <p class="eyebrow">Learning Graph</p>
-              <h3>Self-Learning Knowledge Graph</h3>
-            </div>
-          </div>
-          <div class="brain-stage learning-empty">
-            <div class="empty-graph-copy">
-              <h4>No papers mapped yet</h4>
-              <p>Search papers, pick a topic, select candidates, or upload a PDF to grow the graph in real time.</p>
-            </div>
-          </div>
-        </div>
-        """
-    coords = [
-        (100, 90), (250, 60), (420, 75), (590, 115), (690, 250), (620, 395),
-        (455, 455), (280, 455), (110, 395), (60, 250), (215, 250), (365, 245),
-        (525, 250), (300, 145), (480, 340), (180, 340), (545, 175), (130, 170)
-    ]
-    graph_nodes = [dict(n) for n in nodes[:18]]
-    for i, node in enumerate(graph_nodes):
-        x, y = coords[i % len(coords)]
-        node["sx"] = x
-        node["sy"] = y
-    node_map = {n["id"]: n for n in graph_nodes}
-    edge_items, node_items, label_items = [], [], []
-    for edge in edges[:80]:
-        source = edge.get("source")
-        target = edge.get("target")
-        edge_type = edge.get("type", "")
-        if source in node_map and target in node_map:
-            a = node_map[source]
-            b = node_map[target]
-            edge_items.append(
-                f'<line class="learn-edge edge-{safe_text(edge_type.lower())}" x1="{a["sx"]}" y1="{a["sy"]}" x2="{b["sx"]}" y2="{b["sy"]}" />'
-            )
-    for node in graph_nodes:
-        kind = (node.get("kind") or node.get("type") or "paper").lower()
-        if kind == "topic":
-            kind = "query"
-        if kind == "uploadedpdf":
-            kind = "upload"
-        radius = 25 if kind == "query" else 18 if kind in {"concept", "author", "claim", "reference"} else 20
-        css_class = f"learn-node {kind}"
-        node_items.append(f'<circle class="{css_class}" cx="{node["sx"]}" cy="{node["sy"]}" r="{radius}" />')
-        label = node.get("label") or node.get("title") or node.get("id")
-        label_items.append(f'<text class="learn-label" x="{node["sx"] + 26}" y="{node["sy"] - 8}">{safe_text(str(label)[:46])}</text>')
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Learning Graph</p>
-          <h3>{safe_text(title)}</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-query"></i> topic</span>
-          <span><i class="dot dot-paper"></i> paper</span>
-          <span><i class="dot dot-upload"></i> uploaded PDF</span>
-          <span><i class="dot dot-concept"></i> concept</span>
-          <span><i class="dot dot-author"></i> author</span>
-          <span><i class="dot dot-ref"></i> reference</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 760 520" class="brain-svg" role="img" aria-label="Self-learning knowledge graph">
-          {''.join(edge_items)}
-          {''.join(node_items)}
-          {''.join(label_items)}
-        </svg>
-      </div>
-    </div>
-    """
-
-
-def build_learning_graph_state(query, papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-    for i, paper in enumerate(papers[:5], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper.get("title", f"Paper {i}"), "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{i}_{slugify(concept)[:20]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-    return nodes, edges
-
-
-def graph_from_selected(query, selected_papers, uploaded_name=None, parsed_state=None, frontier=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-    for i, paper in enumerate(selected_papers[:6], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper.get("title", f"Paper {i}"), "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-        for author in paper.get("authors", [])[:2]:
-            aid = f"author_{i}_{slugify(author)[:24]}"
-            nodes.append({"id": aid, "label": author, "kind": "author"})
-            edges.append({"source": pid, "target": aid, "type": "WRITTEN_BY"})
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{i}_{slugify(concept)[:24]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-        for claim in (paper.get("claims") or [])[:1]:
-            cid = f"claim_{i}_{slugify(claim)[:24]}"
-            nodes.append({"id": cid, "label": claim[:42], "kind": "claim"})
-            edges.append({"source": pid, "target": cid, "type": "ASSERTS"})
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-        if parsed_state and isinstance(parsed_state, dict):
-            for concept in (parsed_state.get("concepts") or [])[:3]:
-                cid = f"upload_concept_{slugify(concept)[:24]}"
-                nodes.append({"id": cid, "label": concept, "kind": "concept"})
-                edges.append({"source": "upload", "target": cid, "type": "MENTIONS"})
-    for j, fp in enumerate(ensure_list(frontier)[:3], start=1):
-        fid = f"frontier_{j}"
-        nodes.append({"id": fid, "label": fp.get("title", f"Frontier {j}"), "kind": "reference"})
-        edges.append({"source": "query", "target": fid, "type": "FRONTIER"})
-    return nodes, edges
-
-
-def parse_pdf_with_grobid(pdf_path):
-    if not GROBID_URL:
-        raise RuntimeError("GROBID_URL is not set")
-    with open(pdf_path, "rb") as f:
-        files = {"input": (Path(pdf_path).name, f, "application/pdf")}
-        response = HTTP.post(
-            f"{GROBID_URL.rstrip('/')}/api/processFulltextDocument",
-            files=files,
-            data={"includeRawAffiliations": "1", "segmentSentences": "1"},
-            timeout=180,
-        )
-    response.raise_for_status()
-    tei_xml = response.text
-    root = ET.fromstring(tei_xml)
-    ns = {"tei": "http://www.tei-c.org/ns/1.0"}
-    title = root.findtext(".//tei:titleStmt/tei:title", default="", namespaces=ns) or Path(pdf_path).name
-    abstract_parts = []
-    for p in root.findall(".//tei:profileDesc/tei:abstract//tei:p", ns):
-        abstract_parts.append(" ".join(list(p.itertext())))
-    abstract = truncate_text(" ".join(abstract_parts), MAX_ABSTRACT_CHARS)
-    authors = []
-    for author in root.findall(".//tei:sourceDesc//tei:author", ns):
-        parts = []
-        forename = author.findall(".//tei:forename", ns)
-        surname = author.findall(".//tei:surname", ns)
-        parts.extend([norm_text(" ".join(x.itertext())) for x in forename])
-        parts.extend([norm_text(" ".join(x.itertext())) for x in surname])
-        name = norm_text(" ".join(parts))
-        if name:
-            authors.append(name)
-    sections = []
-    text_pool = []
-    for div in root.findall(".//tei:text//tei:body//tei:div", ns):
-        head = div.findtext("./tei:head", default="", namespaces=ns)
-        paras = []
-        for p in div.findall(".//tei:p", ns):
-            para_text = norm_text(" ".join(list(p.itertext())))
-            if para_text:
-                paras.append(para_text)
-        joined = "\n".join(paras)
-        if head or joined:
-            sections.append({"heading": head or "Section", "text": truncate_text(joined, 4000)})
-            if joined:
-                text_pool.append(joined)
-    references = []
-    for bibl in root.findall(".//tei:listBibl//tei:biblStruct", ns)[:60]:
-        ref_title = bibl.findtext(".//tei:title", default="", namespaces=ns)
-        ref_doi = ""
-        for idno in bibl.findall(".//tei:idno", ns):
-            if (idno.attrib.get("type") or "").lower() == "doi":
-                ref_doi = norm_text(" ".join(idno.itertext()))
-                break
-        references.append({"title": norm_text(ref_title), "doi": normalize_doi(ref_doi)})
-    semantic_text = truncate_text(" ".join([abstract] + text_pool[:5]), MAX_RAW_TEXT_CHARS)
-    return {
-        "parser": "grobid",
-        "title": norm_text(title),
-        "abstract": abstract,
-        "authors": authors[:12],
-        "sections": sections[:14],
-        "references": references[:60],
-        "claims": extract_claim_like_sentences(semantic_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(semantic_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": "",
-        "parser_quality": "scholarly-structured",
-    }
-
-
-def parse_pdf_with_pymupdf(pdf_path):
-    if fitz is None:
-        raise RuntimeError("PyMuPDF not installed")
-    doc = fitz.open(pdf_path)
-    raw_text = truncate_text("\n".join(page.get_text("text") for page in doc).strip(), MAX_RAW_TEXT_CHARS)
-    first_page = raw_text[:4000]
-    lines = [x.strip() for x in first_page.splitlines() if x.strip()]
-    title = lines[0][:300] if lines else Path(pdf_path).name
-    abstract = ""
-    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n1[\.\s]|introduction)", raw_text, re.I | re.S)
-    if match:
-        abstract = truncate_text(match.group(1), 2500)
-    return {
-        "parser": "pymupdf",
-        "title": title,
-        "abstract": abstract,
-        "authors": [],
-        "sections": [{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(raw_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(raw_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": raw_text,
-        "parser_quality": "text-fallback",
-    }
-
-
-def parse_pdf_with_docling(pdf_path):
-    try:
-        from docling.document_converter import DocumentConverter
-    except Exception as e:
-        raise RuntimeError(f"Docling import failed: {e}")
-    converter = DocumentConverter()
-    result = converter.convert(pdf_path)
-    doc = result.document
-    markdown = truncate_text(doc.export_to_markdown(), MAX_RAW_TEXT_CHARS)
-    title = Path(pdf_path).name
-    first_nonempty = next((line.strip("# ").strip() for line in markdown.splitlines() if line.strip()), "")
-    if first_nonempty:
-        title = first_nonempty[:300]
-    return {
-        "parser": "docling",
-        "title": title,
-        "abstract": "",
-        "authors": [],
-        "sections": [{"heading": "Document", "text": markdown[:12000]}] if markdown else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(markdown, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(markdown, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": markdown,
-        "parser_quality": "layout-aware",
-    }
-
-
-def parse_uploaded_pdf(file_obj, parser_order):
-    if not file_obj:
-        return "### PDF parse status\n\nNo PDF uploaded yet.", {}
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    parser_order = ensure_list(parser_order) or ["grobid", "docling", "pymupdf"]
-    errors = []
-    for parser_name in parser_order:
-        try:
-            if parser_name == "grobid":
-                result = parse_pdf_with_grobid(path)
-            elif parser_name == "docling":
-                result = parse_pdf_with_docling(path)
-            elif parser_name == "pymupdf":
-                result = parse_pdf_with_pymupdf(path)
-            else:
-                continue
-            summary = (
-                f"### PDF parse status\n\n"
-                f"- Parser used: {result['parser']}\n"
-                f"- Parser quality: {result.get('parser_quality', 'unknown')}\n"
-                f"- Title: {result.get('title') or 'Unknown'}\n"
-                f"- Authors: {', '.join(result.get('authors')[:6]) if result.get('authors') else 'Unknown'}\n"
-                f"- Abstract found: {'Yes' if result.get('abstract') else 'No'}\n"
-                f"- Sections extracted: {len(result.get('sections') or [])}\n"
-                f"- References extracted: {len(result.get('references') or [])}\n"
-                f"- Concepts extracted: {len(result.get('concepts') or [])}\n"
-                f"- Claims extracted: {len(result.get('claims') or [])}\n"
-            )
-            return summary, result
-        except Exception as e:
-            errors.append(f"{parser_name}: {e}")
-    fail_summary = "### PDF parse status\n\n" + "\n".join([f"- {x}" for x in errors])
-    return fail_summary, {"parser": None, "errors": errors}
-
-
-def render_parse_result(parsed):
-    if not parsed or not isinstance(parsed, dict) or (not parsed.get("title") and not parsed.get("sections")):
-        return '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-    sections_html = []
-    for section in parsed.get("sections", [])[:6]:
-        sections_html.append(
-            f"""
-            <details class="agent-step">
-              <summary class="agent-summary">
-                <div class="agent-index">§</div>
-                <div class="agent-head">
-                  <h4>{safe_text(section.get('heading', 'Section'))}</h4>
-                  <span>section</span>
-                </div>
-              </summary>
-              <div class="agent-copy">
-                <p>{safe_text(section.get('text', '')[:1800])}</p>
-              </div>
-            </details>
-            """
-        )
-    refs = parsed.get("references", [])[:12]
-    refs_html = "".join(
-        f"<li>{safe_text(r.get('title') or 'Untitled')} {'· DOI ' + safe_text(r.get('doi')) if r.get('doi') else ''}</li>"
-        for r in refs
-    ) or "<li>No references extracted.</li>"
-    concepts = parsed.get("concepts", [])[:10]
-    claims = parsed.get("claims", [])[:6]
-    concepts_html = "".join(f"<li>{safe_text(x)}</li>" for x in concepts) or "<li>No concepts extracted.</li>"
-    claims_html = "".join(f"<li>{safe_text(x)}</li>" for x in claims) or "<li>No claims extracted.</li>"
-    title = safe_text(parsed.get("title") or "Parsed document")
-    abstract = safe_text((parsed.get("abstract") or "")[:2400]) or "No abstract extracted."
-    parser_name = safe_text(parsed.get("parser") or "unknown")
-    parser_quality = safe_text(parsed.get("parser_quality") or "unknown")
-    return f"""
-    <div class="panel" style="padding:18px">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">PDF Parse</p>
-          <h3>{title}</h3>
-        </div>
-        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name} · {parser_quality}</span></div>
-      </div>
-      <div class="parse-grid">
-        <div class="parse-card">
-          <h4>Abstract</h4>
-          <p>{abstract}</p>
-        </div>
-        <div class="parse-card">
-          <h4>References</h4>
-          <ul class="ref-list">{refs_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Concepts</h4>
-          <ul class="ref-list">{concepts_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Claims</h4>
-          <ul class="ref-list">{claims_html}</ul>
-        </div>
-      </div>
-      <div class="timeline" style="margin-top:14px;">
-        {''.join(sections_html) if sections_html else '<div class="panel" style="padding:16px;"><p>No sections extracted.</p></div>'}
-      </div>
-    </div>
-    """
-
-
-def add_node(nodes_by_id: Dict[str, Dict], node_id: str, node_type: str, label: str = "", **attrs):
-    if not node_id:
-        return
-    current = nodes_by_id.get(node_id, {})
-    merged = {"id": node_id, "type": node_type, "label": label or current.get("label", node_id)}
-    merged.update(current)
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            merged[key] = value
-    nodes_by_id[node_id] = merged
-
-
-def add_edge(edges: List[Dict], source: str, target: str, edge_type: str, **attrs):
-    if not source or not target or source == target:
-        return
-    edge = {"source": source, "target": target, "type": edge_type}
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            edge[key] = value
-    edges.append(edge)
-
-
-def build_ingest_payload(query, selected_papers, parsed_pdf=None, frontier=None):
-    nodes_by_id = {}
-    edges = []
-    topic_id = "topic:query"
-    add_node(nodes_by_id, topic_id, "Topic", label=query or "Research topic", query=query or "")
-    for i, paper in enumerate(selected_papers, start=1):
-        paper_id = normalize_doi(paper.get("doi")) or (paper.get("external_ids") or {}).get("arxiv") or f"paper:{i}:{slugify(paper.get('title', 'paper'))[:32]}"
-        add_node(
-            nodes_by_id,
-            paper_id,
-            "Paper",
-            label=paper.get("title") or f"Paper {i}",
-            title=paper.get("title"),
-            year=paper.get("year"),
-            venue=paper.get("venue"),
-            doi=normalize_doi(paper.get("doi")),
-            source=paper.get("source"),
-            url=paper.get("url"),
-            pdf=paper.get("pdf"),
-            score=paper.get("score"),
-            learned_score=paper.get("learned_score", paper.get("score")),
-            open_access=paper.get("open_access"),
-            authors_text=paper.get("authors_text"),
-        )
-        add_edge(edges, topic_id, paper_id, "ABOUT", weight=paper.get("learned_score", paper.get("score", 0)))
-        for author in paper.get("authors", [])[:6]:
-            author_id = f"author:{slugify(author)[:64]}"
-            add_node(nodes_by_id, author_id, "Author", label=author, name=author)
-            add_edge(edges, paper_id, author_id, "WRITTEN_BY")
-        for concept in (paper.get("concepts") or [])[:6]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, paper_id, concept_id, "MENTIONS")
-        for claim in (paper.get("claims") or [])[:3]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, paper_id, claim_id, "ASSERTS")
-    if parsed_pdf and isinstance(parsed_pdf, dict) and parsed_pdf.get("title"):
-        doc_id = "upload:pdf"
-        add_node(nodes_by_id, doc_id, "UploadedPDF", label=parsed_pdf.get("title"), title=parsed_pdf.get("title"), parser=parsed_pdf.get("parser"))
-        add_edge(edges, topic_id, doc_id, "UPLOADED_SOURCE")
-        for concept in (parsed_pdf.get("concepts") or [])[:6]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, doc_id, concept_id, "MENTIONS")
-        for claim in (parsed_pdf.get("claims") or [])[:4]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, doc_id, claim_id, "ASSERTS")
-        for idx, ref in enumerate(parsed_pdf.get("references", [])[:12], start=1):
-            ref_title = ref.get("title") or f"Reference {idx}"
-            ref_doi = normalize_doi(ref.get("doi") or "")
-            ref_id = ref_doi or f"ref:{idx}:{slugify(ref_title)[:32]}"
-            add_node(nodes_by_id, ref_id, "Reference", label=ref_title, title=ref_title, doi=ref_doi)
-            add_edge(edges, doc_id, ref_id, "CITES")
-    for idx, item in enumerate(ensure_list(frontier)[:12], start=1):
-        fid = normalize_doi(item.get("doi")) or f"frontier:{idx}:{slugify(item.get('title', 'paper'))[:32]}"
-        add_node(nodes_by_id, fid, "FrontierPaper", label=item.get("title") or f"Frontier {idx}", title=item.get("title"), frontier_score=item.get("frontier_score"), url=item.get("url"))
-        add_edge(edges, topic_id, fid, "FRONTIER_CANDIDATE", weight=item.get("frontier_score", item.get("learned_score", item.get("score", 0))))
-    return {"status": "ok", "nodes": list(nodes_by_id.values())[:GRAPH_MAX_NODES], "edges": edges[:GRAPH_MAX_EDGES]}
-
-
-def learn_from_payload(payload: Dict, query: str = "") -> Dict:
-    if not payload:
-        return GRAPH_MEMORY
-    GRAPH_MEMORY["queries"].append(query or "")
-    GRAPH_MEMORY["events"].append({
-        "ts": time.time(),
-        "query": query or "",
-        "nodes": len(payload.get("nodes", [])),
-        "edges": len(payload.get("edges", [])),
-    })
-    GRAPH_MEMORY["payloads"].append(payload)
-    for node in payload.get("nodes", []):
-        node_id = node.get("id")
-        if not node_id:
-            continue
-        GRAPH_MEMORY["nodes"][node_id] = node
-        node_type = (node.get("type") or "").lower()
-        if node_type in {"paper", "frontierpaper"}:
-            GRAPH_MEMORY["papers"][node_id] = node
-        if node_type == "concept" and node.get("label"):
-            GRAPH_MEMORY["concept_counts"][node["label"].lower()] += 1
-        if node_type == "claim" and node.get("label"):
-            GRAPH_MEMORY["claim_counts"][node["label"].lower()] += 1
-    GRAPH_MEMORY["edges"].extend(payload.get("edges", []))
-    GRAPH_MEMORY["edges"] = GRAPH_MEMORY["edges"][:GRAPH_MAX_EDGES]
-    return GRAPH_MEMORY
-
-
-def export_learning_state() -> str:
-    snapshot = {
-        "papers": list(GRAPH_MEMORY["papers"].values())[:50],
-        "nodes": list(GRAPH_MEMORY["nodes"].values())[:200],
-        "edges": GRAPH_MEMORY["edges"][:400],
-        "top_concepts": GRAPH_MEMORY["concept_counts"].most_common(20),
-        "top_claims": GRAPH_MEMORY["claim_counts"].most_common(20),
-        "queries": GRAPH_MEMORY["queries"][-20:],
-        "events": GRAPH_MEMORY["events"][-20:],
-        "frontier": GRAPH_MEMORY["frontier"][:20],
-    }
-    return json.dumps(snapshot, indent=2, ensure_ascii=False)
-
-
-def resolve_selected_papers(selected_indices, papers_state):
-    papers = ensure_list(papers_state)
-    selected_indices = ensure_list(selected_indices)
-    selected = []
-    if not selected_indices:
-        return selected
-    value_map = {paper_choice_value(i, paper): paper for i, paper in enumerate(papers)}
-    label_map = {paper_choice_label(i, paper): paper for i, paper in enumerate(papers)}
-    for idx in selected_indices:
-        try:
-            if isinstance(idx, int):
-                if 0 <= idx < len(papers):
-                    selected.append(papers[idx])
-                continue
-            idx_str = str(idx)
-            if idx_str in value_map:
-                selected.append(value_map[idx_str])
-                continue
-            if idx_str.isdigit():
-                num = int(idx_str)
-                if 0 <= num < len(papers):
-                    selected.append(papers[num])
-                    continue
-            if "|" in idx_str:
-                left = idx_str.split("|", 1)[0]
-                if left.isdigit():
-                    num = int(left)
-                    if 0 <= num < len(papers):
-                        selected.append(papers[num])
-                        continue
-            if idx_str in label_map:
-                selected.append(label_map[idx_str])
-                continue
-        except Exception:
-            continue
-    out = []
-    seen = set()
-    for paper in selected:
-        key = paper_identity_key(paper)
-        if key not in seen:
-            seen.add(key)
-            out.append(paper)
-    return out
-
-
-def summarize_learning_state(query_text, papers, selected_sources):
-    concept_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:3])
-    top_concepts = [c for c, _ in Counter([c.lower() for c in concept_pool]).most_common(6)]
-    return (
-        "### Discovery results\n\n"
-        f"- Query: {query_text}\n"
-        f"- Sources: {', '.join(selected_sources)}\n"
-        f"- Candidates found: {len(papers)}\n"
-        f"- Top learned concepts: {', '.join(top_concepts) if top_concepts else 'None'}\n"
-        "- Select papers below, then click **Ingest selected into graph**.\n"
-    )
-
-
-def run_paper_discovery(query, search_mode, sources, pdf_file):
-    query_text = norm_text(query or "")
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    if not query_text and not pdf_file:
-        empty_graph = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            empty_graph,
-            '<div class="panel papers-panel" style="padding:18px"><p>Enter a topic, title, DOI, link, or upload a PDF to start learning.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            "No PDF uploaded yet.",
-            gr.update(choices=[], value=[]),
-            [],
-            "### No discovery results yet.",
-        )
-    if not query_text and pdf_file:
-        uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name
-        graph_nodes, graph_edges = build_learning_graph_state("", [], uploaded_name)
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges, "Uploaded PDF Waiting for Parse"),
-            '<div class="panel papers-panel" style="padding:18px"><p>No query yet. Parse the uploaded PDF or enter a research topic to begin discovery.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            "### Upload detected.\n\n- Parse the PDF to extract structure.\n- Or enter a topic to start discovery.",
-        )
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-    try:
-        papers = discover_papers(query_text, search_mode, selected_sources, max_results=GRAPH_MAX_RESULTS)
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:6], uploaded_name)
-        graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Self-Learning Knowledge Graph")
-        papers_html = format_papers_html(papers)
-        journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
-        pdf_summary = uploaded_pdf_summary(pdf_file)
-        choices = format_selection_choices(papers)
-        status_md = summarize_learning_state(query_text, papers, selected_sources)
-        return (
-            graph_html,
-            papers_html,
-            journals_html,
-            pdf_summary,
-            gr.update(choices=choices, value=[]),
-            papers,
-            status_md,
-        )
-    except Exception as e:
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, [], uploaded_name)
-        error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges),
-            error_html,
-            build_journal_html(query_text or "biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
-        )
-
-
-def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
-    papers = ensure_list(papers_state)
-    selected = resolve_selected_papers(selected_indices, papers)
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-    if not selected and parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title") and papers:
-        selected = papers[:3]
-    if not selected and not (parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title")):
-        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            graph_html,
-            "### Graph ingest status\n\nSelect papers or parse an uploaded PDF first.",
-            {"status": "empty", "nodes": [], "edges": []},
-        )
-    query_text = norm_text(query or "")
-    if not query_text and isinstance(parsed_state, dict):
-        query_text = parsed_state.get("title") or "Research topic"
-    if not query_text:
-        query_text = "Research topic"
-    selected = [enrich_paper_semantics(query_text, paper) for paper in selected]
-    frontier = frontier_expand(query_text, DEFAULT_SOURCES, selected, parsed_state=parsed_state if isinstance(parsed_state, dict) else None, per_query=3)
-    graph_nodes, graph_edges = graph_from_selected(query_text, selected, uploaded_name, parsed_state if isinstance(parsed_state, dict) else None, frontier=frontier)
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
-    payload = build_ingest_payload(query_text, selected, parsed_state if isinstance(parsed_state, dict) else None, frontier=frontier)
-    learn_from_payload(payload, query=query_text)
-    top_concepts = []
-    for paper in selected:
-        top_concepts.extend((paper.get("concepts") or [])[:3])
-    if isinstance(parsed_state, dict):
-        top_concepts.extend((parsed_state.get("concepts") or [])[:3])
-    summary_lines = [
-        "### Graph ingest status",
-        "",
-        f"- Topic: {query_text}",
-        f"- Selected papers: {len(selected)}",
-        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
-        f"- Frontier candidates proposed: {len(frontier)}",
-        f"- Nodes created: {len(payload['nodes'])}",
-        f"- Edges created: {len(payload['edges'])}",
-        f"- Learned concepts: {', '.join(unique_keep_order(top_concepts)[:8]) if top_concepts else 'None'}",
-        f"- Memory papers stored: {len(GRAPH_MEMORY['papers'])}",
-        f"- Memory concepts stored: {len(GRAPH_MEMORY['concept_counts'])}",
-    ]
-    return graph_html, "\n".join(summary_lines), payload
-
-
-def autonomous_expand_into_markdown(query, payload, parsed_state=None):
-    frontier = GRAPH_MEMORY.get("frontier") or []
-    lines = [
-        "### Autonomous expansion plan",
-        "",
-        f"- Seed query: {query or 'Research topic'}",
-        f"- Current nodes: {len(payload.get('nodes', [])) if isinstance(payload, dict) else 0}",
-        f"- Current edges: {len(payload.get('edges', [])) if isinstance(payload, dict) else 0}",
-        f"- Frontier candidates: {len(frontier)}",
-    ]
-    proposed = propose_expansion_queries(query or "", list(GRAPH_MEMORY.get("papers", {}).values())[:8], parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    if proposed:
-        lines.extend(["", "#### Proposed next queries", ""])
-        lines.extend([f"- {q}" for q in proposed])
-    if frontier:
-        lines.extend(["", "#### Top frontier papers", ""])
-        for item in frontier[:8]:
-            lines.append(f"- {item.get('title', 'Untitled')} ({item.get('source', 'unknown')}) — frontier score {item.get('frontier_score', item.get('learned_score', item.get('score', 0)))}")
-    return "\n".join(lines)
-
-
-__all__ = [
-    "SEARCH_MODES",
-    "SOURCE_OPTIONS",
-    "DEFAULT_SOURCES",
-    "GRAPH_MEMORY",
-    "discover_papers",
-    "run_paper_discovery",
-    "parse_uploaded_pdf",
-    "render_parse_result",
-    "ingest_selected_papers",
-    "build_ingest_payload",
-    "learn_from_payload",
-    "frontier_expand",
-    "autonomous_expand_into_markdown",
-    "export_learning_state",
-    "format_frontier_html",
-]

dvnc_ai_v2_hf/deprecated/self_learning_graph_old6.py

@@ -1,1490 +0,0 @@
-import html
-import json
-import os
-import re
-import time
-import urllib.parse
-import xml.etree.ElementTree as ET
-from collections import Counter
-from pathlib import Path
-from typing import Any, Dict, List, Optional, Tuple
-
-import gradio as gr
-import requests
-
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-
-JOURNALS = [
-    {"name": "Nature", "url": "https://www.nature.com/search", "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science", "url": "https://www.science.org/search", "desc": "High-impact science journal and family."},
-    {"name": "Cell", "url": "https://www.cell.com/search", "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search", "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore", "url": "https://ieeexplore.ieee.org/search/searchresult.jsp", "desc": "Engineering, AI, signal processing, and systems."},
-]
-
-SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
-SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-DEFAULT_SOURCES = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "").strip()
-GROBID_URL = os.getenv("GROBID_URL", "").strip()
-OPENALEX_EMAIL = os.getenv("OPENALEX_EMAIL", "").strip()
-CROSSREF_MAILTO = os.getenv("CROSSREF_MAILTO", "").strip()
-REQUEST_TIMEOUT = int(os.getenv("REQUEST_TIMEOUT", "25"))
-GRAPH_MAX_CONCEPTS = int(os.getenv("GRAPH_MAX_CONCEPTS", "12"))
-GRAPH_MAX_CLAIMS = int(os.getenv("GRAPH_MAX_CLAIMS", "8"))
-GRAPH_MAX_RESULTS = int(os.getenv("GRAPH_MAX_RESULTS", "12"))
-GRAPH_MAX_EXPANSIONS = int(os.getenv("GRAPH_MAX_EXPANSIONS", "6"))
-GRAPH_MAX_NODES = int(os.getenv("GRAPH_MAX_NODES", "400"))
-GRAPH_MAX_EDGES = int(os.getenv("GRAPH_MAX_EDGES", "1200"))
-MAX_ABSTRACT_CHARS = int(os.getenv("MAX_ABSTRACT_CHARS", "4000"))
-MAX_RAW_TEXT_CHARS = int(os.getenv("MAX_RAW_TEXT_CHARS", "70000"))
-
-STOPWORDS = {
-    "a", "an", "and", "are", "as", "at", "be", "been", "being", "by", "can", "could", "did", "do", "does",
-    "for", "from", "had", "has", "have", "if", "in", "into", "is", "it", "its", "may", "might", "of", "on",
-    "or", "our", "such", "that", "the", "their", "there", "these", "this", "those", "to", "using", "use",
-    "used", "via", "was", "were", "will", "with", "within", "without", "we", "they", "you", "your", "study",
-    "paper", "research", "results", "method", "methods", "analysis", "approach", "toward", "towards",
-    "based", "new", "novel", "effect", "effects", "model", "models", "system", "systems", "show", "shows",
-    "introduction", "conclusion", "discussion", "figure", "table", "supplementary", "material", "materials",
-}
-
-GRAPH_MEMORY = {
-    "papers": {},
-    "nodes": {},
-    "edges": [],
-    "concept_counts": Counter(),
-    "claim_counts": Counter(),
-    "queries": [],
-    "events": [],
-    "frontier": [],
-    "payloads": [],
-}
-
-
-class ScholarlyClient:
-    def __init__(self):
-        self.session = requests.Session()
-        ua = "dvnc-ai-self-learning-graph/1.0"
-        if CROSSREF_MAILTO:
-            ua += f" (mailto:{CROSSREF_MAILTO})"
-        self.session.headers.update({"User-Agent": ua, "Accept": "application/json, text/xml, */*"})
-        self.session_timeout = REQUEST_TIMEOUT
-
-    def get(self, url: str, **kwargs):
-        timeout = kwargs.pop("timeout", self.session_timeout)
-        return self.session.get(url, timeout=timeout, **kwargs)
-
-    def post(self, url: str, **kwargs):
-        timeout = kwargs.pop("timeout", max(self.session_timeout, 120))
-        return self.session.post(url, timeout=timeout, **kwargs)
-
-
-HTTP = ScholarlyClient()
-
-
-def safe_text(x, default=""):
-    return html.escape(str(x if x is not None else default))
-
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def slugify(text: str) -> str:
-    return re.sub(r"[^a-z0-9]+", "-", (text or "").lower()).strip("-")
-
-
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-def truncate_text(text: str, limit: int) -> str:
-    text = norm_text(text)
-    return text if len(text) <= limit else text[: limit - 1].rstrip() + "…"
-
-
-def normalize_doi(text: str) -> str:
-    text = (text or "").strip()
-    text = re.sub(r"^https?://(dx\.)?doi\.org/", "", text, flags=re.I)
-    return text.strip().rstrip("/")
-
-
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    doi_pattern = r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$"
-    if re.match(doi_pattern, q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-
-
-def tokenize(text: str) -> List[str]:
-    return [t for t in re.findall(r"[a-zA-Z][a-zA-Z0-9\-]{2,}", (text or "").lower()) if t not in STOPWORDS]
-
-
-def unique_keep_order(items: List[str]) -> List[str]:
-    seen = set()
-    out = []
-    for item in items:
-        key = norm_text(item).lower()
-        if key and key not in seen:
-            seen.add(key)
-            out.append(norm_text(item))
-    return out
-
-
-def text_overlap_score(a: str, b: str) -> float:
-    sa = set(tokenize(a))
-    sb = set(tokenize(b))
-    if not sa or not sb:
-        return 0.0
-    return len(sa & sb) / max(1, len(sa | sb))
-
-
-def compute_recency_bonus(year: str) -> float:
-    try:
-        y = int(str(year)[:4])
-    except Exception:
-        return 0.0
-    current = time.gmtime().tm_year
-    age = max(current - y, 0)
-    return max(0.0, 0.14 - age * 0.015)
-
-
-def extract_candidate_phrases(text: str, max_terms: int = 20) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-    tokens = re.findall(r"[A-Za-z][A-Za-z0-9\-]{2,}", text)
-    phrases = []
-    for n in (3, 2, 1):
-        for i in range(len(tokens) - n + 1):
-            phrase = " ".join(tokens[i:i + n]).strip().lower()
-            if len(phrase) < 4:
-                continue
-            parts = phrase.split()
-            if any(p in STOPWORDS for p in parts):
-                continue
-            if all(len(p) <= 2 for p in parts):
-                continue
-            phrases.append(phrase)
-    counts = Counter(phrases)
-    ranked = [p for p, _ in counts.most_common(max_terms * 4)]
-    filtered = []
-    for phrase in ranked:
-        if phrase in filtered:
-            continue
-        if any(phrase != other and phrase in other for other in filtered):
-            continue
-        filtered.append(phrase)
-        if len(filtered) >= max_terms:
-            break
-    return filtered
-
-
-def extract_concepts_from_text(text: str, max_terms: int = GRAPH_MAX_CONCEPTS) -> List[str]:
-    return extract_candidate_phrases(text, max_terms=max_terms)
-
-
-def extract_claim_like_sentences(text: str, max_items: int = GRAPH_MAX_CLAIMS) -> List[str]:
-    text = norm_text(text)
-    if not text:
-        return []
-    parts = re.split(r"(?<=[\.\!\?])\s+", text)
-    scored = []
-    for sentence in parts:
-        s = norm_text(sentence)
-        if len(s) < 40 or len(s) > 320:
-            continue
-        lower = s.lower()
-        score = 0.0
-        if any(k in lower for k in ["improves", "reduces", "increases", "suggests", "demonstrates", "shows", "reveals", "predicts", "achieves", "outperforms", "enables", "supports"]):
-            score += 2.0
-        if any(k in lower for k in ["significant", "associated", "correlated", "effective", "robust", "accurate", "validated", "statistically"]):
-            score += 1.0
-        if any(k in lower for k in ["compared", "versus", "baseline", "state-of-the-art", "sota"]):
-            score += 1.0
-        score += min(len(tokenize(s)) / 15.0, 2.0)
-        scored.append((score, s))
-    return [s for _, s in sorted(scored, key=lambda x: x[0], reverse=True)[:max_items]]
-
-
-def parse_openalex_abstract(inverted_index) -> str:
-    if not inverted_index or not isinstance(inverted_index, dict):
-        return ""
-    pos_to_word = {}
-    for word, positions in inverted_index.items():
-        for pos in positions:
-            pos_to_word[pos] = word
-    if not pos_to_word:
-        return ""
-    return " ".join(pos_to_word[i] for i in sorted(pos_to_word))
-
-
-def score_frontier_candidate(query: str, seed_concepts: List[str], paper: Dict) -> Dict:
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue])
-    rel = text_overlap_score(query, base_text)
-    concept_overlap = 0.0
-    if seed_concepts:
-        concept_overlap = text_overlap_score(" ".join(seed_concepts), " ".join(paper.get("concepts") or []))
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    score = float(paper.get("score", 0)) + rel * 0.45 + concept_overlap * 0.2 + recency + doi_bonus + oa_bonus
-    paper["frontier_score"] = round(score, 4)
-    paper["frontier_relevance"] = round(rel, 4)
-    paper["frontier_concept_overlap"] = round(concept_overlap, 4)
-    return paper
-
-
-def enrich_paper_semantics(query: str, paper: Dict) -> Dict:
-    paper = dict(paper)
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue]).strip()
-    concepts = extract_concepts_from_text(base_text, max_terms=GRAPH_MAX_CONCEPTS)
-    claims = extract_claim_like_sentences(abstract, max_items=GRAPH_MAX_CLAIMS)
-    rel = text_overlap_score(query, f"{title} {abstract}")
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    concept_bonus = min(len(concepts), 8) * 0.01
-    learned_score = float(paper.get("score", 0)) + rel * 0.5 + recency + doi_bonus + oa_bonus + concept_bonus
-    paper["concepts"] = concepts[:GRAPH_MAX_CONCEPTS]
-    paper["claims"] = claims[:GRAPH_MAX_CLAIMS]
-    paper["relevance"] = round(rel, 4)
-    paper["learned_score"] = round(learned_score, 4)
-    return paper
-
-
-def paper_identity_key(paper: Dict) -> str:
-    return (
-        normalize_doi(paper.get("doi") or "")
-        or (paper.get("external_ids") or {}).get("arxiv")
-        or (paper.get("external_ids") or {}).get("pmcid")
-        or norm_text(paper.get("title", "")).lower()
-        or str(paper.get("id"))
-    )
-
-
-def journal_query_links(query: str):
-    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
-    rows = []
-    for journal in JOURNALS:
-        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
-        if "ieeexplore" in journal["url"]:
-            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
-        rows.append({"name": journal["name"], "desc": journal["desc"], "url": url})
-    return rows
-
-
-def build_journal_html(query):
-    rows = []
-    for journal in journal_query_links(query):
-        rows.append(
-            f"""
-            <a class="journal-card" href="{safe_text(journal['url'])}" target="_blank" rel="noopener noreferrer">
-              <div>
-                <h4>{safe_text(journal['name'])}</h4>
-                <p>{safe_text(journal['desc'])}</p>
-              </div>
-              <span>Open</span>
-            </a>
-            """
-        )
-    return '<div class="journal-grid">' + ''.join(rows) + '</div>'
-
-
-def search_arxiv(query, max_results=8):
-    encoded = urllib.parse.quote(query)
-    url = (
-        "http://export.arxiv.org/api/query?search_query=all:"
-        f"{encoded}&start=0&max_results={max_results}&sortBy=relevance&sortOrder=descending"
-    )
-    response = HTTP.get(url)
-    response.raise_for_status()
-    root = ET.fromstring(response.text)
-    ns = {"atom": "http://www.w3.org/2005/Atom"}
-    papers = []
-    for entry in root.findall("atom:entry", ns):
-        title = " ".join((entry.findtext("atom:title", default="", namespaces=ns) or "").split())
-        summary = truncate_text(" ".join((entry.findtext("atom:summary", default="", namespaces=ns) or "").split()), MAX_ABSTRACT_CHARS)
-        published = entry.findtext("atom:published", default="", namespaces=ns)
-        paper_id = entry.findtext("atom:id", default="", namespaces=ns)
-        authors = [a.findtext("atom:name", default="", namespaces=ns) for a in entry.findall("atom:author", ns)]
-        pdf_url = ""
-        for link in entry.findall("atom:link", ns):
-            if link.attrib.get("title") == "pdf":
-                pdf_url = link.attrib.get("href", "")
-                break
-        papers.append({
-            "id": paper_id or title,
-            "title": title,
-            "summary": summary,
-            "abstract": summary,
-            "published": published[:10],
-            "authors": [a for a in authors[:8] if a],
-            "authors_text": ", ".join([a for a in authors[:4] if a]) or "Unknown authors",
-            "url": paper_id,
-            "pdf": pdf_url,
-            "doi": "",
-            "venue": "arXiv",
-            "year": published[:4] if published else "",
-            "source": "arxiv",
-            "score": 0.76,
-            "open_access": True,
-            "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
-        })
-    return papers
-
-
-def search_crossref(query, mode="topic", max_results=8):
-    params = {}
-    if CROSSREF_MAILTO:
-        params["mailto"] = CROSSREF_MAILTO
-    if mode == "doi":
-        url = f"https://api.crossref.org/works/{urllib.parse.quote(query)}"
-        response = HTTP.get(url, params=params)
-        if response.status_code != 200:
-            return []
-        items = [response.json().get("message", {})]
-    else:
-        params["rows"] = max_results
-        if mode in ("title", "paper_name"):
-            params["query.title"] = query
-        else:
-            params["query.bibliographic"] = query
-        response = HTTP.get("https://api.crossref.org/works", params=params)
-        response.raise_for_status()
-        items = response.json().get("message", {}).get("items", [])
-    out = []
-    for item in items:
-        authors = []
-        for a in item.get("author", []) or []:
-            name = " ".join(filter(None, [a.get("given"), a.get("family")])).strip()
-            if name:
-                authors.append(name)
-        title = (item.get("title") or ["Untitled"])[0]
-        year = ""
-        for key in ["published-print", "published-online", "created"]:
-            if item.get(key, {}).get("date-parts"):
-                year = str(item[key]["date-parts"][0][0])
-                break
-        abstract = truncate_text(re.sub("<.*?>", "", item.get("abstract") or ""), MAX_ABSTRACT_CHARS)
-        doi = normalize_doi(item.get("DOI", ""))
-        out.append({
-            "id": doi or title,
-            "title": norm_text(title),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": year,
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("URL", ""),
-            "pdf": "",
-            "doi": doi,
-            "venue": (item.get("container-title") or [""])[0],
-            "year": year,
-            "source": "crossref",
-            "score": 0.72,
-            "open_access": None,
-            "external_ids": {"crossref": doi} if doi else {},
-        })
-    return out
-
-
-def search_openalex(query, mode="topic", max_results=8):
-    params = {"per-page": max_results}
-    if OPENALEX_EMAIL:
-        params["mailto"] = OPENALEX_EMAIL
-    if mode == "doi":
-        doi = normalize_doi(query)
-        params["filter"] = f"doi:https://doi.org/{doi}"
-    else:
-        params["search"] = query
-    response = HTTP.get("https://api.openalex.org/works", params=params)
-    response.raise_for_status()
-    items = response.json().get("results", [])
-    out = []
-    for item in items:
-        authors = []
-        for auth in item.get("authorships", [])[:8]:
-            author = auth.get("author") or {}
-            if author.get("display_name"):
-                authors.append(author["display_name"])
-        oa = item.get("open_access") or {}
-        doi = normalize_doi(item.get("doi") or "")
-        abstract = truncate_text(parse_openalex_abstract(item.get("abstract_inverted_index")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": item.get("id") or doi or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("publication_year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": (item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or "",
-            "pdf": oa.get("oa_url") or "",
-            "doi": doi,
-            "venue": ((item.get("primary_location") or {}).get("source") or {}).get("display_name") or "",
-            "year": str(item.get("publication_year") or ""),
-            "source": "openalex",
-            "score": 0.80,
-            "open_access": oa.get("is_oa"),
-            "external_ids": item.get("ids") or {},
-        })
-    return out
-
-
-def search_semantic_scholar(query, mode="topic", max_results=8):
-    headers = {}
-    if SEMANTIC_SCHOLAR_API_KEY:
-        headers["x-api-key"] = SEMANTIC_SCHOLAR_API_KEY
-    fields = "title,authors,year,abstract,venue,externalIds,url,openAccessPdf"
-    if mode == "doi":
-        doi = normalize_doi(query)
-        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{urllib.parse.quote(doi)}"
-        response = HTTP.get(url, params={"fields": fields}, headers=headers)
-        if response.status_code != 200:
-            return []
-        items = [response.json()]
-    else:
-        response = HTTP.get(
-            "https://api.semanticscholar.org/graph/v1/paper/search",
-            params={"query": query, "limit": max_results, "fields": fields},
-            headers=headers,
-        )
-        if response.status_code != 200:
-            return []
-        items = response.json().get("data", [])
-    out = []
-    for item in items:
-        external = item.get("externalIds") or {}
-        authors = [a.get("name") for a in item.get("authors", []) if a.get("name")]
-        abstract = truncate_text(norm_text(item.get("abstract", "")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": external.get("CorpusId") or external.get("DOI") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("url") or "",
-            "pdf": (item.get("openAccessPdf") or {}).get("url") or "",
-            "doi": normalize_doi(external.get("DOI", "")),
-            "venue": item.get("venue") or "",
-            "year": str(item.get("year") or ""),
-            "source": "semantic_scholar",
-            "score": 0.84,
-            "open_access": bool((item.get("openAccessPdf") or {}).get("url")),
-            "external_ids": external,
-        })
-    return out
-
-
-def search_europe_pmc(query, mode="topic", max_results=8):
-    epmc_query = f'DOI:"{query}"' if mode == "doi" else query
-    params = {"query": epmc_query, "format": "json", "pageSize": max_results, "resultType": "core"}
-    response = HTTP.get("https://www.ebi.ac.uk/europepmc/webservices/rest/search", params=params)
-    if response.status_code != 200:
-        return []
-    items = response.json().get("resultList", {}).get("result", [])
-    out = []
-    for item in items:
-        author_string = item.get("authorString", "")
-        authors = [x.strip() for x in author_string.split(",")[:8] if x.strip()]
-        pmcid = item.get("pmcid", "")
-        pdf_url = f"https://europepmc.org/articles/{pmcid}?pdf=render" if pmcid else ""
-        landing_url = f"https://europepmc.org/article/PMC/{pmcid}" if pmcid else ""
-        abstract = truncate_text(norm_text(item.get("abstractText", "")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": item.get("id") or item.get("doi") or item.get("title"),
-            "title": norm_text(item.get("title")),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("pubYear") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": landing_url,
-            "pdf": pdf_url,
-            "doi": normalize_doi(item.get("doi", "")),
-            "venue": item.get("journalTitle", ""),
-            "year": str(item.get("pubYear") or ""),
-            "source": "europe_pmc",
-            "score": 0.78,
-            "open_access": bool(pmcid),
-            "external_ids": {"pmid": item.get("pmid"), "pmcid": pmcid},
-        })
-    return out
-
-
-def resolve_link(query):
-    url = (query or "").strip()
-    if not url:
-        return []
-    try:
-        response = HTTP.get(url, allow_redirects=True, headers={"User-Agent": "dvnc-ai-space/1.0"})
-        content_type = response.headers.get("content-type", "")
-        if "pdf" in content_type or url.lower().endswith(".pdf"):
-            name = Path(url.split("?")[0]).name or "linked-paper.pdf"
-            return [{
-                "id": url,
-                "title": name,
-                "summary": "Direct PDF link detected.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": url,
-                "doi": "",
-                "venue": "Direct PDF",
-                "year": "",
-                "source": "link",
-                "score": 0.66,
-                "open_access": True,
-                "external_ids": {},
-            }]
-        doi = ""
-        title = url
-        pdf_link = ""
-        if BeautifulSoup is not None:
-            soup = BeautifulSoup(response.text, "html.parser")
-            title = soup.title.text.strip() if soup.title else url
-            for meta_name in ["citation_doi", "dc.identifier", "dc.Identifier"]:
-                tag = soup.find("meta", attrs={"name": meta_name})
-                if tag and tag.get("content"):
-                    doi = normalize_doi(tag["content"].strip())
-                    break
-            for a in soup.find_all("a", href=True):
-                href = a["href"]
-                if ".pdf" in href.lower():
-                    pdf_link = href if href.startswith("http") else urllib.parse.urljoin(url, href)
-                    break
-        if doi:
-            results = search_crossref(doi, mode="doi", max_results=1)
-            if results:
-                if pdf_link and not results[0].get("pdf"):
-                    results[0]["pdf"] = pdf_link
-                if url and not results[0].get("url"):
-                    results[0]["url"] = url
-                return results
-        return [{
-            "id": url,
-            "title": title,
-            "summary": "Landing page resolved from direct link.",
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": pdf_link,
-            "doi": doi,
-            "venue": "Web Link",
-            "year": "",
-            "source": "link",
-            "score": 0.54,
-            "open_access": bool(pdf_link),
-            "external_ids": {},
-        }]
-    except Exception as e:
-        return [{
-            "id": url,
-            "title": "Link resolution error",
-            "summary": str(e),
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": "",
-            "doi": "",
-            "venue": "Link",
-            "year": "",
-            "source": "link",
-            "score": 0.20,
-            "open_access": None,
-            "external_ids": {},
-        }]
-
-
-def dedupe_papers(items: List[Dict]) -> List[Dict]:
-    seen = {}
-    for item in items:
-        key = paper_identity_key(item) or f"{item.get('source', 'src')}::{item.get('title', 'paper')}"
-        current = seen.get(key)
-        candidate_score = float(item.get("learned_score", item.get("score", 0)))
-        current_score = float(current.get("learned_score", current.get("score", 0))) if current else -1
-        if current is None or candidate_score > current_score:
-            seen[key] = item
-    return sorted(seen.values(), key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-
-
-def discover_papers(query, mode, sources, max_results=10):
-    query = (query or "").strip()
-    if not query:
-        return []
-    mode = detect_query_type(query) if mode == "autonomous_web" else mode
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    results = []
-    if mode == "link":
-        return dedupe_papers(resolve_link(query))
-    if "arxiv" in selected_sources and mode != "doi":
-        try:
-            results.extend(search_arxiv(query, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "crossref" in selected_sources:
-        try:
-            results.extend(search_crossref(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "openalex" in selected_sources:
-        try:
-            results.extend(search_openalex(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "semantic_scholar" in selected_sources:
-        try:
-            results.extend(search_semantic_scholar(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "europe_pmc" in selected_sources:
-        try:
-            results.extend(search_europe_pmc(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    papers = [enrich_paper_semantics(query, p) for p in dedupe_papers(results)]
-    papers = sorted(papers, key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-    return papers[:max_results]
-
-
-def propose_expansion_queries(query: str, papers: List[Dict], parsed_state: Optional[Dict] = None, limit: int = GRAPH_MAX_EXPANSIONS) -> List[str]:
-    concept_pool = []
-    venue_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:4])
-        if paper.get("venue"):
-            venue_pool.append(paper["venue"])
-    if parsed_state and isinstance(parsed_state, dict):
-        concept_pool.extend((parsed_state.get("concepts") or [])[:6])
-    ranked_concepts = [c for c, _ in Counter([norm_text(c).lower() for c in concept_pool if c]).most_common(limit * 2)]
-    expansions = [norm_text(query)] if query else []
-    for concept in ranked_concepts:
-        if not concept:
-            continue
-        if query:
-            expansions.append(f"{query} {concept}")
-        else:
-            expansions.append(concept)
-    for venue in unique_keep_order(venue_pool)[:2]:
-        if query and venue:
-            expansions.append(f"{query} {venue}")
-    return unique_keep_order(expansions)[:limit]
-
-
-def frontier_expand(query: str, sources: List[str], selected_papers: List[Dict], parsed_state: Optional[Dict] = None, per_query: int = 4) -> List[Dict]:
-    seed_concepts = []
-    for p in selected_papers[:6]:
-        seed_concepts.extend((p.get("concepts") or [])[:4])
-    if parsed_state and isinstance(parsed_state, dict):
-        seed_concepts.extend((parsed_state.get("concepts") or [])[:6])
-    expansion_queries = propose_expansion_queries(query, selected_papers, parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    frontier = []
-    for eq in expansion_queries:
-        try:
-            items = discover_papers(eq, "topic", sources, max_results=per_query)
-            for item in items:
-                frontier.append(score_frontier_candidate(query or eq, seed_concepts, item))
-        except Exception:
-            continue
-    frontier = dedupe_papers(frontier)
-    frontier.sort(key=lambda x: float(x.get("frontier_score", x.get("learned_score", x.get("score", 0))),), reverse=True)
-    GRAPH_MEMORY["frontier"] = frontier[: GRAPH_MAX_EXPANSIONS * per_query]
-    return GRAPH_MEMORY["frontier"]
-
-
-def paper_choice_value(index: int, paper: Dict) -> str:
-    doi = normalize_doi(paper.get("doi") or "")
-    title_slug = slugify(paper.get("title", ""))[:40]
-    return f"{index}|{doi}|{title_slug}"
-
-
-def paper_choice_label(index: int, paper: Dict) -> str:
-    score = round(float(paper.get("learned_score", paper.get("score", 0))), 3)
-    title = paper.get("title", "Untitled")
-    authors_text = paper.get("authors_text", "Unknown authors")[:80]
-    source = paper.get("source", "src")
-    return f"[{source}] {title} — {authors_text} — score {score}"
-
-
-def format_selection_choices(papers):
-    return [(paper_choice_label(i, paper), paper_choice_value(i, paper)) for i, paper in enumerate(papers)]
-
-
-def format_papers_html(papers):
-    if not papers:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No papers found yet.</p></div>'
-    items = []
-    for i, paper in enumerate(papers, start=1):
-        summary = safe_text((paper.get("summary") or paper.get("abstract") or "")[:280])
-        doi_line = f'<span class="paper-badge doi-badge">{safe_text(paper.get("doi"))}</span>' if paper.get("doi") else ""
-        pdf_link = paper.get("pdf") or "#"
-        abs_link = paper.get("url") or "#"
-        concepts_text = ", ".join((paper.get("concepts") or [])[:4])
-        items.append(
-            f"""
-            <article class="paper-card">
-              <div class="paper-topline">
-                <span class="paper-badge">{safe_text(paper.get('source', 'paper'))}</span>
-                <span class="paper-badge alt">{safe_text(paper.get('published', '') or 'Paper')}</span>
-                {doi_line}
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{summary or 'No abstract snippet available.'}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-                <div><strong>Venue:</strong> {safe_text(paper.get('venue', 'Unknown venue'))}</div>
-                <div><strong>Learned score:</strong> {safe_text(round(float(paper.get('learned_score', paper.get('score', 0))), 3))}</div>
-                <div><strong>Concepts:</strong> {safe_text(concepts_text or 'None extracted')}</div>
-              </div>
-              <div class="paper-links">
-                <a href="{safe_text(abs_link)}" target="_blank" rel="noopener noreferrer">Abstract</a>
-                <a href="{safe_text(pdf_link)}" target="_blank" rel="noopener noreferrer">PDF</a>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(items) + '</div>'
-
-
-def format_frontier_html(frontier):
-    if not frontier:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No autonomous expansion candidates yet.</p></div>'
-    cards = []
-    for i, paper in enumerate(frontier[:12], start=1):
-        cards.append(
-            f"""
-            <article class="paper-card frontier-card">
-              <div class="paper-topline">
-                <span class="paper-badge">frontier</span>
-                <span class="paper-badge alt">{safe_text(paper.get('source', 'paper'))}</span>
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{safe_text((paper.get('summary') or paper.get('abstract') or '')[:260])}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Frontier score:</strong> {safe_text(paper.get('frontier_score', paper.get('learned_score', paper.get('score', 0))))}</div>
-                <div><strong>Concept overlap:</strong> {safe_text(paper.get('frontier_concept_overlap', 0))}</div>
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(cards) + '</div>'
-
-
-def uploaded_pdf_summary(file_obj):
-    if not file_obj:
-        return "No PDF uploaded yet."
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    p = Path(path)
-    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, references, concepts, and claims."
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    if not nodes:
-        return """
-        <div class="panel brain-shell">
-          <div class="brain-header">
-            <div>
-              <p class="eyebrow">Learning Graph</p>
-              <h3>Self-Learning Knowledge Graph</h3>
-            </div>
-          </div>
-          <div class="brain-stage learning-empty">
-            <div class="empty-graph-copy">
-              <h4>No papers mapped yet</h4>
-              <p>Search papers, pick a topic, select candidates, or upload a PDF to grow the graph in real time.</p>
-            </div>
-          </div>
-        </div>
-        """
-    coords = [
-        (100, 90), (250, 60), (420, 75), (590, 115), (690, 250), (620, 395),
-        (455, 455), (280, 455), (110, 395), (60, 250), (215, 250), (365, 245),
-        (525, 250), (300, 145), (480, 340), (180, 340), (545, 175), (130, 170)
-    ]
-    graph_nodes = [dict(n) for n in nodes[:18]]
-    for i, node in enumerate(graph_nodes):
-        x, y = coords[i % len(coords)]
-        node["sx"] = x
-        node["sy"] = y
-    node_map = {n["id"]: n for n in graph_nodes}
-    edge_items, node_items, label_items = [], [], []
-    for edge in edges[:80]:
-        source = edge.get("source")
-        target = edge.get("target")
-        edge_type = edge.get("type", "")
-        if source in node_map and target in node_map:
-            a = node_map[source]
-            b = node_map[target]
-            edge_items.append(
-                f'<line class="learn-edge edge-{safe_text(edge_type.lower())}" x1="{a["sx"]}" y1="{a["sy"]}" x2="{b["sx"]}" y2="{b["sy"]}" />'
-            )
-    for node in graph_nodes:
-        kind = (node.get("kind") or node.get("type") or "paper").lower()
-        if kind == "topic":
-            kind = "query"
-        if kind == "uploadedpdf":
-            kind = "upload"
-        radius = 25 if kind == "query" else 18 if kind in {"concept", "author", "claim", "reference"} else 20
-        css_class = f"learn-node {kind}"
-        node_items.append(f'<circle class="{css_class}" cx="{node["sx"]}" cy="{node["sy"]}" r="{radius}" />')
-        label = node.get("label") or node.get("title") or node.get("id")
-        label_items.append(f'<text class="learn-label" x="{node["sx"] + 26}" y="{node["sy"] - 8}">{safe_text(str(label)[:46])}</text>')
-    return f"""
-    <div class="panel brain-shell">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">Learning Graph</p>
-          <h3>{safe_text(title)}</h3>
-        </div>
-        <div class="brain-legend">
-          <span><i class="dot dot-query"></i> topic</span>
-          <span><i class="dot dot-paper"></i> paper</span>
-          <span><i class="dot dot-upload"></i> uploaded PDF</span>
-          <span><i class="dot dot-concept"></i> concept</span>
-          <span><i class="dot dot-author"></i> author</span>
-          <span><i class="dot dot-ref"></i> reference</span>
-        </div>
-      </div>
-      <div class="brain-stage">
-        <svg viewBox="0 0 760 520" class="brain-svg" role="img" aria-label="Self-learning knowledge graph">
-          {''.join(edge_items)}
-          {''.join(node_items)}
-          {''.join(label_items)}
-        </svg>
-      </div>
-    </div>
-    """
-
-
-def build_learning_graph_state(query, papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-    for i, paper in enumerate(papers[:5], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper.get("title", f"Paper {i}"), "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{i}_{slugify(concept)[:20]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-    return nodes, edges
-
-
-def graph_from_selected(query, selected_papers, uploaded_name=None, parsed_state=None, frontier=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-    for i, paper in enumerate(selected_papers[:6], start=1):
-        pid = f"paper_{i}"
-        nodes.append({"id": pid, "label": paper.get("title", f"Paper {i}"), "kind": "paper"})
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-        for author in paper.get("authors", [])[:2]:
-            aid = f"author_{i}_{slugify(author)[:24]}"
-            nodes.append({"id": aid, "label": author, "kind": "author"})
-            edges.append({"source": pid, "target": aid, "type": "WRITTEN_BY"})
-        for concept in (paper.get("concepts") or [])[:2]:
-            cid = f"concept_{i}_{slugify(concept)[:24]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-        for claim in (paper.get("claims") or [])[:1]:
-            cid = f"claim_{i}_{slugify(claim)[:24]}"
-            nodes.append({"id": cid, "label": claim[:42], "kind": "claim"})
-            edges.append({"source": pid, "target": cid, "type": "ASSERTS"})
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-        if parsed_state and isinstance(parsed_state, dict):
-            for concept in (parsed_state.get("concepts") or [])[:3]:
-                cid = f"upload_concept_{slugify(concept)[:24]}"
-                nodes.append({"id": cid, "label": concept, "kind": "concept"})
-                edges.append({"source": "upload", "target": cid, "type": "MENTIONS"})
-    for j, fp in enumerate(ensure_list(frontier)[:3], start=1):
-        fid = f"frontier_{j}"
-        nodes.append({"id": fid, "label": fp.get("title", f"Frontier {j}"), "kind": "reference"})
-        edges.append({"source": "query", "target": fid, "type": "FRONTIER"})
-    return nodes, edges
-
-
-def parse_pdf_with_grobid(pdf_path):
-    if not GROBID_URL:
-        raise RuntimeError("GROBID_URL is not set")
-    with open(pdf_path, "rb") as f:
-        files = {"input": (Path(pdf_path).name, f, "application/pdf")}
-        response = HTTP.post(
-            f"{GROBID_URL.rstrip('/')}/api/processFulltextDocument",
-            files=files,
-            data={"includeRawAffiliations": "1", "segmentSentences": "1"},
-            timeout=180,
-        )
-    response.raise_for_status()
-    tei_xml = response.text
-    root = ET.fromstring(tei_xml)
-    ns = {"tei": "http://www.tei-c.org/ns/1.0"}
-    title = root.findtext(".//tei:titleStmt/tei:title", default="", namespaces=ns) or Path(pdf_path).name
-    abstract_parts = []
-    for p in root.findall(".//tei:profileDesc/tei:abstract//tei:p", ns):
-        abstract_parts.append(" ".join(list(p.itertext())))
-    abstract = truncate_text(" ".join(abstract_parts), MAX_ABSTRACT_CHARS)
-    authors = []
-    for author in root.findall(".//tei:sourceDesc//tei:author", ns):
-        parts = []
-        forename = author.findall(".//tei:forename", ns)
-        surname = author.findall(".//tei:surname", ns)
-        parts.extend([norm_text(" ".join(x.itertext())) for x in forename])
-        parts.extend([norm_text(" ".join(x.itertext())) for x in surname])
-        name = norm_text(" ".join(parts))
-        if name:
-            authors.append(name)
-    sections = []
-    text_pool = []
-    for div in root.findall(".//tei:text//tei:body//tei:div", ns):
-        head = div.findtext("./tei:head", default="", namespaces=ns)
-        paras = []
-        for p in div.findall(".//tei:p", ns):
-            para_text = norm_text(" ".join(list(p.itertext())))
-            if para_text:
-                paras.append(para_text)
-        joined = "\n".join(paras)
-        if head or joined:
-            sections.append({"heading": head or "Section", "text": truncate_text(joined, 4000)})
-            if joined:
-                text_pool.append(joined)
-    references = []
-    for bibl in root.findall(".//tei:listBibl//tei:biblStruct", ns)[:60]:
-        ref_title = bibl.findtext(".//tei:title", default="", namespaces=ns)
-        ref_doi = ""
-        for idno in bibl.findall(".//tei:idno", ns):
-            if (idno.attrib.get("type") or "").lower() == "doi":
-                ref_doi = norm_text(" ".join(idno.itertext()))
-                break
-        references.append({"title": norm_text(ref_title), "doi": normalize_doi(ref_doi)})
-    semantic_text = truncate_text(" ".join([abstract] + text_pool[:5]), MAX_RAW_TEXT_CHARS)
-    return {
-        "parser": "grobid",
-        "title": norm_text(title),
-        "abstract": abstract,
-        "authors": authors[:12],
-        "sections": sections[:14],
-        "references": references[:60],
-        "claims": extract_claim_like_sentences(semantic_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(semantic_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": "",
-        "parser_quality": "scholarly-structured",
-    }
-
-
-def parse_pdf_with_pymupdf(pdf_path):
-    if fitz is None:
-        raise RuntimeError("PyMuPDF not installed")
-    doc = fitz.open(pdf_path)
-    raw_text = truncate_text("\n".join(page.get_text("text") for page in doc).strip(), MAX_RAW_TEXT_CHARS)
-    first_page = raw_text[:4000]
-    lines = [x.strip() for x in first_page.splitlines() if x.strip()]
-    title = lines[0][:300] if lines else Path(pdf_path).name
-    abstract = ""
-    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n1[\.\s]|introduction)", raw_text, re.I | re.S)
-    if match:
-        abstract = truncate_text(match.group(1), 2500)
-    return {
-        "parser": "pymupdf",
-        "title": title,
-        "abstract": abstract,
-        "authors": [],
-        "sections": [{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(raw_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(raw_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": raw_text,
-        "parser_quality": "text-fallback",
-    }
-
-
-def parse_pdf_with_docling(pdf_path):
-    try:
-        from docling.document_converter import DocumentConverter
-    except Exception as e:
-        raise RuntimeError(f"Docling import failed: {e}")
-    converter = DocumentConverter()
-    result = converter.convert(pdf_path)
-    doc = result.document
-    markdown = truncate_text(doc.export_to_markdown(), MAX_RAW_TEXT_CHARS)
-    title = Path(pdf_path).name
-    first_nonempty = next((line.strip("# ").strip() for line in markdown.splitlines() if line.strip()), "")
-    if first_nonempty:
-        title = first_nonempty[:300]
-    return {
-        "parser": "docling",
-        "title": title,
-        "abstract": "",
-        "authors": [],
-        "sections": [{"heading": "Document", "text": markdown[:12000]}] if markdown else [],
-        "references": [],
-        "claims": extract_claim_like_sentences(markdown, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(markdown, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": markdown,
-        "parser_quality": "layout-aware",
-    }
-
-
-def parse_uploaded_pdf(file_obj, parser_order):
-    if not file_obj:
-        return "### PDF parse status\n\nNo PDF uploaded yet.", {}
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    parser_order = ensure_list(parser_order) or ["grobid", "docling", "pymupdf"]
-    errors = []
-    for parser_name in parser_order:
-        try:
-            if parser_name == "grobid":
-                result = parse_pdf_with_grobid(path)
-            elif parser_name == "docling":
-                result = parse_pdf_with_docling(path)
-            elif parser_name == "pymupdf":
-                result = parse_pdf_with_pymupdf(path)
-            else:
-                continue
-            summary = (
-                f"### PDF parse status\n\n"
-                f"- Parser used: {result['parser']}\n"
-                f"- Parser quality: {result.get('parser_quality', 'unknown')}\n"
-                f"- Title: {result.get('title') or 'Unknown'}\n"
-                f"- Authors: {', '.join(result.get('authors')[:6]) if result.get('authors') else 'Unknown'}\n"
-                f"- Abstract found: {'Yes' if result.get('abstract') else 'No'}\n"
-                f"- Sections extracted: {len(result.get('sections') or [])}\n"
-                f"- References extracted: {len(result.get('references') or [])}\n"
-                f"- Concepts extracted: {len(result.get('concepts') or [])}\n"
-                f"- Claims extracted: {len(result.get('claims') or [])}\n"
-            )
-            return summary, result
-        except Exception as e:
-            errors.append(f"{parser_name}: {e}")
-    fail_summary = "### PDF parse status\n\n" + "\n".join([f"- {x}" for x in errors])
-    return fail_summary, {"parser": None, "errors": errors}
-
-
-def render_parse_result(parsed):
-    if not parsed or not isinstance(parsed, dict) or (not parsed.get("title") and not parsed.get("sections")):
-        return '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-    sections_html = []
-    for section in parsed.get("sections", [])[:6]:
-        sections_html.append(
-            f"""
-            <details class="agent-step">
-              <summary class="agent-summary">
-                <div class="agent-index">§</div>
-                <div class="agent-head">
-                  <h4>{safe_text(section.get('heading', 'Section'))}</h4>
-                  <span>section</span>
-                </div>
-              </summary>
-              <div class="agent-copy">
-                <p>{safe_text(section.get('text', '')[:1800])}</p>
-              </div>
-            </details>
-            """
-        )
-    refs = parsed.get("references", [])[:12]
-    refs_html = "".join(
-        f"<li>{safe_text(r.get('title') or 'Untitled')} {'· DOI ' + safe_text(r.get('doi')) if r.get('doi') else ''}</li>"
-        for r in refs
-    ) or "<li>No references extracted.</li>"
-    concepts = parsed.get("concepts", [])[:10]
-    claims = parsed.get("claims", [])[:6]
-    concepts_html = "".join(f"<li>{safe_text(x)}</li>" for x in concepts) or "<li>No concepts extracted.</li>"
-    claims_html = "".join(f"<li>{safe_text(x)}</li>" for x in claims) or "<li>No claims extracted.</li>"
-    title = safe_text(parsed.get("title") or "Parsed document")
-    abstract = safe_text((parsed.get("abstract") or "")[:2400]) or "No abstract extracted."
-    parser_name = safe_text(parsed.get("parser") or "unknown")
-    parser_quality = safe_text(parsed.get("parser_quality") or "unknown")
-    return f"""
-    <div class="panel" style="padding:18px">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">PDF Parse</p>
-          <h3>{title}</h3>
-        </div>
-        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name} · {parser_quality}</span></div>
-      </div>
-      <div class="parse-grid">
-        <div class="parse-card">
-          <h4>Abstract</h4>
-          <p>{abstract}</p>
-        </div>
-        <div class="parse-card">
-          <h4>References</h4>
-          <ul class="ref-list">{refs_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Concepts</h4>
-          <ul class="ref-list">{concepts_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Claims</h4>
-          <ul class="ref-list">{claims_html}</ul>
-        </div>
-      </div>
-      <div class="timeline" style="margin-top:14px;">
-        {''.join(sections_html) if sections_html else '<div class="panel" style="padding:16px;"><p>No sections extracted.</p></div>'}
-      </div>
-    </div>
-    """
-
-
-def add_node(nodes_by_id: Dict[str, Dict], node_id: str, node_type: str, label: str = "", **attrs):
-    if not node_id:
-        return
-    current = nodes_by_id.get(node_id, {})
-    merged = {"id": node_id, "type": node_type, "label": label or current.get("label", node_id)}
-    merged.update(current)
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            merged[key] = value
-    nodes_by_id[node_id] = merged
-
-
-def add_edge(edges: List[Dict], source: str, target: str, edge_type: str, **attrs):
-    if not source or not target or source == target:
-        return
-    edge = {"source": source, "target": target, "type": edge_type}
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            edge[key] = value
-    edges.append(edge)
-
-
-def build_ingest_payload(query, selected_papers, parsed_pdf=None, frontier=None):
-    nodes_by_id = {}
-    edges = []
-    topic_id = "topic:query"
-    add_node(nodes_by_id, topic_id, "Topic", label=query or "Research topic", query=query or "")
-    for i, paper in enumerate(selected_papers, start=1):
-        paper_id = normalize_doi(paper.get("doi")) or (paper.get("external_ids") or {}).get("arxiv") or f"paper:{i}:{slugify(paper.get('title', 'paper'))[:32]}"
-        add_node(
-            nodes_by_id,
-            paper_id,
-            "Paper",
-            label=paper.get("title") or f"Paper {i}",
-            title=paper.get("title"),
-            year=paper.get("year"),
-            venue=paper.get("venue"),
-            doi=normalize_doi(paper.get("doi")),
-            source=paper.get("source"),
-            url=paper.get("url"),
-            pdf=paper.get("pdf"),
-            score=paper.get("score"),
-            learned_score=paper.get("learned_score", paper.get("score")),
-            open_access=paper.get("open_access"),
-            authors_text=paper.get("authors_text"),
-        )
-        add_edge(edges, topic_id, paper_id, "ABOUT", weight=paper.get("learned_score", paper.get("score", 0)))
-        for author in paper.get("authors", [])[:6]:
-            author_id = f"author:{slugify(author)[:64]}"
-            add_node(nodes_by_id, author_id, "Author", label=author, name=author)
-            add_edge(edges, paper_id, author_id, "WRITTEN_BY")
-        for concept in (paper.get("concepts") or [])[:6]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, paper_id, concept_id, "MENTIONS")
-        for claim in (paper.get("claims") or [])[:3]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, paper_id, claim_id, "ASSERTS")
-    if parsed_pdf and isinstance(parsed_pdf, dict) and parsed_pdf.get("title"):
-        doc_id = "upload:pdf"
-        add_node(nodes_by_id, doc_id, "UploadedPDF", label=parsed_pdf.get("title"), title=parsed_pdf.get("title"), parser=parsed_pdf.get("parser"))
-        add_edge(edges, topic_id, doc_id, "UPLOADED_SOURCE")
-        for concept in (parsed_pdf.get("concepts") or [])[:6]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, doc_id, concept_id, "MENTIONS")
-        for claim in (parsed_pdf.get("claims") or [])[:4]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:120], text=claim)
-            add_edge(edges, doc_id, claim_id, "ASSERTS")
-        for idx, ref in enumerate(parsed_pdf.get("references", [])[:12], start=1):
-            ref_title = ref.get("title") or f"Reference {idx}"
-            ref_doi = normalize_doi(ref.get("doi") or "")
-            ref_id = ref_doi or f"ref:{idx}:{slugify(ref_title)[:32]}"
-            add_node(nodes_by_id, ref_id, "Reference", label=ref_title, title=ref_title, doi=ref_doi)
-            add_edge(edges, doc_id, ref_id, "CITES")
-    for idx, item in enumerate(ensure_list(frontier)[:12], start=1):
-        fid = normalize_doi(item.get("doi")) or f"frontier:{idx}:{slugify(item.get('title', 'paper'))[:32]}"
-        add_node(nodes_by_id, fid, "FrontierPaper", label=item.get("title") or f"Frontier {idx}", title=item.get("title"), frontier_score=item.get("frontier_score"), url=item.get("url"))
-        add_edge(edges, topic_id, fid, "FRONTIER_CANDIDATE", weight=item.get("frontier_score", item.get("learned_score", item.get("score", 0))))
-    return {"status": "ok", "nodes": list(nodes_by_id.values())[:GRAPH_MAX_NODES], "edges": edges[:GRAPH_MAX_EDGES]}
-
-
-def learn_from_payload(payload: Dict, query: str = "") -> Dict:
-    if not payload:
-        return GRAPH_MEMORY
-    GRAPH_MEMORY["queries"].append(query or "")
-    GRAPH_MEMORY["events"].append({
-        "ts": time.time(),
-        "query": query or "",
-        "nodes": len(payload.get("nodes", [])),
-        "edges": len(payload.get("edges", [])),
-    })
-    GRAPH_MEMORY["payloads"].append(payload)
-    for node in payload.get("nodes", []):
-        node_id = node.get("id")
-        if not node_id:
-            continue
-        GRAPH_MEMORY["nodes"][node_id] = node
-        node_type = (node.get("type") or "").lower()
-        if node_type in {"paper", "frontierpaper"}:
-            GRAPH_MEMORY["papers"][node_id] = node
-        if node_type == "concept" and node.get("label"):
-            GRAPH_MEMORY["concept_counts"][node["label"].lower()] += 1
-        if node_type == "claim" and node.get("label"):
-            GRAPH_MEMORY["claim_counts"][node["label"].lower()] += 1
-    GRAPH_MEMORY["edges"].extend(payload.get("edges", []))
-    GRAPH_MEMORY["edges"] = GRAPH_MEMORY["edges"][:GRAPH_MAX_EDGES]
-    return GRAPH_MEMORY
-
-
-def export_learning_state() -> str:
-    snapshot = {
-        "papers": list(GRAPH_MEMORY["papers"].values())[:50],
-        "nodes": list(GRAPH_MEMORY["nodes"].values())[:200],
-        "edges": GRAPH_MEMORY["edges"][:400],
-        "top_concepts": GRAPH_MEMORY["concept_counts"].most_common(20),
-        "top_claims": GRAPH_MEMORY["claim_counts"].most_common(20),
-        "queries": GRAPH_MEMORY["queries"][-20:],
-        "events": GRAPH_MEMORY["events"][-20:],
-        "frontier": GRAPH_MEMORY["frontier"][:20],
-    }
-    return json.dumps(snapshot, indent=2, ensure_ascii=False)
-
-
-def resolve_selected_papers(selected_indices, papers_state):
-    papers = ensure_list(papers_state)
-    selected_indices = ensure_list(selected_indices)
-    selected = []
-    if not selected_indices:
-        return selected
-    value_map = {paper_choice_value(i, paper): paper for i, paper in enumerate(papers)}
-    label_map = {paper_choice_label(i, paper): paper for i, paper in enumerate(papers)}
-    for idx in selected_indices:
-        try:
-            if isinstance(idx, int):
-                if 0 <= idx < len(papers):
-                    selected.append(papers[idx])
-                continue
-            idx_str = str(idx)
-            if idx_str in value_map:
-                selected.append(value_map[idx_str])
-                continue
-            if idx_str.isdigit():
-                num = int(idx_str)
-                if 0 <= num < len(papers):
-                    selected.append(papers[num])
-                    continue
-            if "|" in idx_str:
-                left = idx_str.split("|", 1)[0]
-                if left.isdigit():
-                    num = int(left)
-                    if 0 <= num < len(papers):
-                        selected.append(papers[num])
-                        continue
-            if idx_str in label_map:
-                selected.append(label_map[idx_str])
-                continue
-        except Exception:
-            continue
-    out = []
-    seen = set()
-    for paper in selected:
-        key = paper_identity_key(paper)
-        if key not in seen:
-            seen.add(key)
-            out.append(paper)
-    return out
-
-
-def summarize_learning_state(query_text, papers, selected_sources):
-    concept_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:3])
-    top_concepts = [c for c, _ in Counter([c.lower() for c in concept_pool]).most_common(6)]
-    return (
-        "### Discovery results\n\n"
-        f"- Query: {query_text}\n"
-        f"- Sources: {', '.join(selected_sources)}\n"
-        f"- Candidates found: {len(papers)}\n"
-        f"- Top learned concepts: {', '.join(top_concepts) if top_concepts else 'None'}\n"
-        "- Select papers below, then click **Ingest selected into graph**.\n"
-    )
-
-
-def run_paper_discovery(query, search_mode, sources, pdf_file):
-    query_text = norm_text(query or "")
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    if not query_text and not pdf_file:
-        empty_graph = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            empty_graph,
-            '<div class="panel papers-panel" style="padding:18px"><p>Enter a topic, title, DOI, link, or upload a PDF to start learning.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            "No PDF uploaded yet.",
-            gr.update(choices=[], value=[]),
-            [],
-            "### No discovery results yet.",
-        )
-    if not query_text and pdf_file:
-        uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name
-        graph_nodes, graph_edges = build_learning_graph_state("", [], uploaded_name)
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges, "Uploaded PDF Waiting for Parse"),
-            '<div class="panel papers-panel" style="padding:18px"><p>No query yet. Parse the uploaded PDF or enter a research topic to begin discovery.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            "### Upload detected.\n\n- Parse the PDF to extract structure.\n- Or enter a topic to start discovery.",
-        )
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-    try:
-        papers = discover_papers(query_text, search_mode, selected_sources, max_results=GRAPH_MAX_RESULTS)
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:6], uploaded_name)
-        graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Self-Learning Knowledge Graph")
-        papers_html = format_papers_html(papers)
-        journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
-        pdf_summary = uploaded_pdf_summary(pdf_file)
-        choices = format_selection_choices(papers)
-        status_md = summarize_learning_state(query_text, papers, selected_sources)
-        return (
-            graph_html,
-            papers_html,
-            journals_html,
-            pdf_summary,
-            gr.update(choices=choices, value=[]),
-            papers,
-            status_md,
-        )
-    except Exception as e:
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, [], uploaded_name)
-        error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges),
-            error_html,
-            build_journal_html(query_text or "biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
-        )
-
-
-def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
-    papers = ensure_list(papers_state)
-    selected = resolve_selected_papers(selected_indices, papers)
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-    if not selected and parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title") and papers:
-        selected = papers[:3]
-    if not selected and not (parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title")):
-        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            graph_html,
-            "### Graph ingest status\n\nSelect papers or parse an uploaded PDF first.",
-            {"status": "empty", "nodes": [], "edges": []},
-        )
-    query_text = norm_text(query or "")
-    if not query_text and isinstance(parsed_state, dict):
-        query_text = parsed_state.get("title") or "Research topic"
-    if not query_text:
-        query_text = "Research topic"
-    selected = [enrich_paper_semantics(query_text, paper) for paper in selected]
-    frontier = frontier_expand(query_text, DEFAULT_SOURCES, selected, parsed_state=parsed_state if isinstance(parsed_state, dict) else None, per_query=3)
-    graph_nodes, graph_edges = graph_from_selected(query_text, selected, uploaded_name, parsed_state if isinstance(parsed_state, dict) else None, frontier=frontier)
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
-    payload = build_ingest_payload(query_text, selected, parsed_state if isinstance(parsed_state, dict) else None, frontier=frontier)
-    learn_from_payload(payload, query=query_text)
-    top_concepts = []
-    for paper in selected:
-        top_concepts.extend((paper.get("concepts") or [])[:3])
-    if isinstance(parsed_state, dict):
-        top_concepts.extend((parsed_state.get("concepts") or [])[:3])
-    summary_lines = [
-        "### Graph ingest status",
-        "",
-        f"- Topic: {query_text}",
-        f"- Selected papers: {len(selected)}",
-        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
-        f"- Frontier candidates proposed: {len(frontier)}",
-        f"- Nodes created: {len(payload['nodes'])}",
-        f"- Edges created: {len(payload['edges'])}",
-        f"- Learned concepts: {', '.join(unique_keep_order(top_concepts)[:8]) if top_concepts else 'None'}",
-        f"- Memory papers stored: {len(GRAPH_MEMORY['papers'])}",
-        f"- Memory concepts stored: {len(GRAPH_MEMORY['concept_counts'])}",
-    ]
-    return graph_html, "\n".join(summary_lines), payload
-
-
-def autonomous_expand_into_markdown(query, payload, parsed_state=None):
-    frontier = GRAPH_MEMORY.get("frontier") or []
-    lines = [
-        "### Autonomous expansion plan",
-        "",
-        f"- Seed query: {query or 'Research topic'}",
-        f"- Current nodes: {len(payload.get('nodes', [])) if isinstance(payload, dict) else 0}",
-        f"- Current edges: {len(payload.get('edges', [])) if isinstance(payload, dict) else 0}",
-        f"- Frontier candidates: {len(frontier)}",
-    ]
-    proposed = propose_expansion_queries(query or "", list(GRAPH_MEMORY.get("papers", {}).values())[:8], parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    if proposed:
-        lines.extend(["", "#### Proposed next queries", ""])
-        lines.extend([f"- {q}" for q in proposed])
-    if frontier:
-        lines.extend(["", "#### Top frontier papers", ""])
-        for item in frontier[:8]:
-            lines.append(f"- {item.get('title', 'Untitled')} ({item.get('source', 'unknown')}) — frontier score {item.get('frontier_score', item.get('learned_score', item.get('score', 0)))}")
-    return "\n".join(lines)
-
-
-__all__ = [
-    "SEARCH_MODES",
-    "SOURCE_OPTIONS",
-    "DEFAULT_SOURCES",
-    "GRAPH_MEMORY",
-    "discover_papers",
-    "run_paper_discovery",
-    "parse_uploaded_pdf",
-    "render_parse_result",
-    "ingest_selected_papers",
-    "build_ingest_payload",
-    "learn_from_payload",
-    "frontier_expand",
-    "autonomous_expand_into_markdown",
-    "export_learning_state",
-    "format_frontier_html",
-]

dvnc_ai_v2_hf/deprecated/self_learning_graph_old_2.py

@@ -1,73 +0,0 @@
-Core constants and state
-- JOURNALS
-- SEARCH_MODES
-- SOURCE_OPTIONS
-- DEFAULT_SOURCES
-- GRAPH_MEMORY
-
-Text and identity helpers
-- safe_text(x, default="")
-- norm_text(x)
-- slugify(text)
-- ensure_list(x)
-- normalize_doi(text)
-- detect_query_type(query)
-- tokenize(text)
-- paper_identity_key(paper)
-
-Semantic helpers
-- extract_candidate_phrases(text, max_terms=...)
-- extract_concepts_from_text(text, max_terms=...)
-- extract_claim_like_sentences(text, max_items=...)
-- text_overlap_score(a, b)
-- compute_recency_bonus(year)
-- parse_openalex_abstract(inverted_index)
-- enrich_paper_semantics(query, paper)
-
-Retrieval layer
-- search_arxiv(query, max_results=8)
-- search_crossref(query, mode="topic", max_results=8)
-- search_openalex(query, mode="topic", max_results=8)
-- search_semantic_scholar(query, mode="topic", max_results=8)
-- search_europe_pmc(query, mode="topic", max_results=8)
-- resolve_link(query)
-- dedupe_papers(items)
-- discover_papers(query, mode, sources, max_results=10)
-
-Self-learning layer
-- extract_reference_queries_from_parsed(parsed_pdf, max_items=8)
-- propose_expansion_queries(seed_query, papers, parsed_pdf=None, max_items=8)
-- autonomous_expand_graph(query, sources, parsed_pdf=None, max_rounds=2, max_results=8)
-
-UI formatting
-- journal_query_links(query)
-- build_journal_html(query)
-- paper_choice_value(index, paper)
-- paper_choice_label(index, paper)
-- format_papers_html(papers)
-- format_selection_choices(papers)
-- uploaded_pdf_summary(file_obj)
-- build_learning_graph_html(nodes, edges, title=...)
-- build_learning_graph_state(query, papers, uploaded_name=None)
-- graph_from_selected(query, selected_papers, uploaded_name=None)
-
-PDF parsing
-- parse_pdf_with_grobid(pdf_path)
-- parse_pdf_with_pymupdf(pdf_path)
-- parse_pdf_with_docling(pdf_path)
-- parse_uploaded_pdf(file_obj, parser_order)
-- render_parse_result(parsed)
-
-Graph payload and memory
-- add_node(nodes_by_id, node_id, node_type, label="", **attrs)
-- add_edge(edges, source, target, edge_type, **attrs)
-- build_ingest_payload(query, selected_papers, parsed_pdf=None)
-- build_autonomous_payload(query, papers, parsed_pdf, visited_queries, rounds)
-- learn_from_payload(payload, query="")
-- get_graph_memory_snapshot()
-- reset_graph_memory()
-
-Main Gradio callbacks
-- run_paper_discovery(query, search_mode, sources, pdf_file)
-- ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state)
-- run_self_learning_cycle(query, search_mode, sources, pdf_file, parser_order, selected_indices, papers_state, parsed_state)

dvnc_ai_v2_hf/deprecated/self_learning_graph_old_8.py

@@ -1,2070 +0,0 @@
-import html
-import json
-import os
-import re
-import time
-import urllib.parse
-import xml.etree.ElementTree as ET
-from collections import Counter
-from pathlib import Path
-from typing import Any, Dict, List, Optional
-
-import gradio as gr
-import requests
-
-try:
-    import fitz  # PyMuPDF
-except Exception:
-    fitz = None
-
-try:
-    from bs4 import BeautifulSoup
-except Exception:
-    BeautifulSoup = None
-
-
-JOURNALS = [
-    {"name": "Nature", "url": "https://www.nature.com/search", "desc": "Flagship multidisciplinary research journal."},
-    {"name": "Science", "url": "https://www.science.org/search", "desc": "High-impact science journal and family."},
-    {"name": "Cell", "url": "https://www.cell.com/search", "desc": "Life sciences and translational biology."},
-    {"name": "The Lancet", "url": "https://www.thelancet.com/search", "desc": "Clinical and medical research."},
-    {"name": "IEEE Xplore", "url": "https://ieeexplore.ieee.org/search/searchresult.jsp", "desc": "Engineering, AI, signal processing, and systems."},
-]
-
-SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
-SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-DEFAULT_SOURCES = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-PDF_PARSERS = ["grobid", "docling", "pymupdf"]
-
-SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "").strip()
-GROBID_URL = os.getenv("GROBID_URL", "").strip()
-OPENALEX_EMAIL = os.getenv("OPENALEX_EMAIL", "").strip()
-CROSSREF_MAILTO = os.getenv("CROSSREF_MAILTO", "").strip()
-REQUEST_TIMEOUT = int(os.getenv("REQUEST_TIMEOUT", "25"))
-GRAPH_MAX_CONCEPTS = int(os.getenv("GRAPH_MAX_CONCEPTS", "14"))
-GRAPH_MAX_CLAIMS = int(os.getenv("GRAPH_MAX_CLAIMS", "8"))
-GRAPH_MAX_RESULTS = int(os.getenv("GRAPH_MAX_RESULTS", "12"))
-GRAPH_MAX_EXPANSIONS = int(os.getenv("GRAPH_MAX_EXPANSIONS", "6"))
-GRAPH_MAX_NODES = int(os.getenv("GRAPH_MAX_NODES", "500"))
-GRAPH_MAX_EDGES = int(os.getenv("GRAPH_MAX_EDGES", "1600"))
-MAX_ABSTRACT_CHARS = int(os.getenv("MAX_ABSTRACT_CHARS", "4500"))
-MAX_RAW_TEXT_CHARS = int(os.getenv("MAX_RAW_TEXT_CHARS", "90000"))
-GRAPH_IFRAME_HEIGHT = int(os.getenv("GRAPH_IFRAME_HEIGHT", "760"))
-
-STOPWORDS = {
-    "a", "an", "and", "are", "as", "at", "be", "been", "being", "by", "can", "could", "did", "do", "does",
-    "for", "from", "had", "has", "have", "if", "in", "into", "is", "it", "its", "may", "might", "of", "on",
-    "or", "our", "such", "that", "the", "their", "there", "these", "this", "those", "to", "using", "use",
-    "used", "via", "was", "were", "will", "with", "within", "without", "we", "they", "you", "your", "study",
-    "paper", "research", "results", "result", "method", "methods", "analysis", "approach", "toward", "towards",
-    "based", "new", "novel", "effect", "effects", "model", "models", "system", "systems", "show", "shows",
-    "shown", "introduction", "conclusion", "discussion", "figure", "table", "supplementary", "material",
-    "materials", "however", "therefore", "furthermore", "among", "across", "between", "also", "than",
-    "both", "et", "al",
-}
-
-COMMON_SECTION_HEADINGS = {
-    "abstract", "introduction", "background", "methods", "materials", "results", "discussion", "conclusion",
-    "references", "acknowledgements", "acknowledgments", "keywords", "supplementary", "appendix"
-}
-
-GRAPH_MEMORY = {
-    "papers": {},
-    "nodes": {},
-    "edges": [],
-    "concept_counts": Counter(),
-    "claim_counts": Counter(),
-    "queries": [],
-    "events": [],
-    "frontier": [],
-    "payloads": [],
-}
-
-
-class ScholarlyClient:
-    def __init__(self):
-        self.session = requests.Session()
-        ua = "dvnc-ai-self-learning-graph/2.0"
-        if CROSSREF_MAILTO:
-            ua += f" (mailto:{CROSSREF_MAILTO})"
-        self.session.headers.update({"User-Agent": ua, "Accept": "application/json, text/xml, */*"})
-        self.session_timeout = REQUEST_TIMEOUT
-
-    def get(self, url: str, **kwargs):
-        timeout = kwargs.pop("timeout", self.session_timeout)
-        return self.session.get(url, timeout=timeout, **kwargs)
-
-    def post(self, url: str, **kwargs):
-        timeout = kwargs.pop("timeout", max(self.session_timeout, 120))
-        return self.session.post(url, timeout=timeout, **kwargs)
-
-
-HTTP = ScholarlyClient()
-
-
-def safe_text(x, default=""):
-    return html.escape(str(x if x is not None else default))
-
-
-def norm_text(x: Optional[str]) -> str:
-    return re.sub(r"\s+", " ", (x or "")).strip()
-
-
-def slugify(text: str) -> str:
-    return re.sub(r"[^a-z0-9]+", "-", (text or "").lower()).strip("-")
-
-
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-def truncate_text(text: str, limit: int) -> str:
-    text = norm_text(text)
-    return text if len(text) <= limit else text[: limit - 1].rstrip() + "…"
-
-
-def normalize_doi(text: str) -> str:
-    text = (text or "").strip()
-    text = re.sub(r"^https?://(dx\.)?doi\.org/", "", text, flags=re.I)
-    return text.strip().rstrip("/")
-
-
-def detect_query_type(query: str) -> str:
-    q = (query or "").strip()
-    doi_pattern = r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$"
-    if re.match(doi_pattern, q, flags=re.I):
-        return "doi"
-    if q.startswith("http://") or q.startswith("https://"):
-        return "link"
-    return "topic"
-
-
-def unique_keep_order(items: List[str]) -> List[str]:
-    seen = set()
-    out = []
-    for item in items:
-        key = norm_text(item).lower()
-        if key and key not in seen:
-            seen.add(key)
-            out.append(norm_text(item))
-    return out
-
-
-def tokenize(text: str) -> List[str]:
-    return [t for t in re.findall(r"[A-Za-z][A-Za-z0-9\-/+]{2,}", (text or "")) if t.lower() not in STOPWORDS]
-
-
-def text_overlap_score(a: str, b: str) -> float:
-    sa = {x.lower() for x in tokenize(a)}
-    sb = {x.lower() for x in tokenize(b)}
-    if not sa or not sb:
-        return 0.0
-    return len(sa & sb) / max(1, len(sa | sb))
-
-
-def compute_recency_bonus(year: str) -> float:
-    try:
-        y = int(str(year)[:4])
-    except Exception:
-        return 0.0
-    current = time.gmtime().tm_year
-    age = max(current - y, 0)
-    return max(0.0, 0.16 - age * 0.018)
-
-
-def dehyphenate_text(text: str) -> str:
-    text = re.sub(r"(\w)-\s*\n\s*(\w)", r"\1\2", text)
-    text = re.sub(r"([a-z])\n([a-z])", r"\1 \2", text)
-    return text
-
-
-def clean_extracted_text(text: str) -> str:
-    text = text or ""
-    replacements = {
-        "\u00ad": "",
-        "\ufb01": "fi",
-        "\ufb02": "fl",
-        "\u2010": "-",
-        "\u2011": "-",
-        "\u2012": "-",
-        "\u2013": "-",
-        "\u2014": "-",
-        "\u2212": "-",
-        "\u00a0": " ",
-    }
-    for old, new in replacements.items():
-        text = text.replace(old, new)
-
-    text = dehyphenate_text(text)
-    text = re.sub(r"[ \t]+", " ", text)
-    text = re.sub(r"\n{3,}", "\n\n", text)
-
-    text = re.sub(r"(?i)\b([a-z]{4,})\s+([a-z]?\1)\b", r"\1", text)
-    text = re.sub(r"(?i)\b([a-z])([A-Z][a-z]+)\b", r"\1 \2", text)
-    text = re.sub(
-        r"(?i)(\b[a-z]{1})(hydrogel|conductive|responsive|injectable|biomaterial|scaffold|tissue|cardiac|patch|repair)\b",
-        r"\2",
-        text,
-    )
-    text = re.sub(r"(?i)\b([a-z]{3,})\s+([a-z]\1)\b", r"\1", text)
-
-    return text.strip()
-
-
-def line_quality_score(line: str) -> float:
-    line = norm_text(line)
-    if not line:
-        return -10.0
-    lower = line.lower().strip(":")
-    if lower in COMMON_SECTION_HEADINGS:
-        return -5.0
-    words = line.split()
-    if len(words) < 3:
-        return -2.0
-    score = 0.0
-    score += min(len(words), 18) * 0.2
-    score += sum(1 for w in words if w[:1].isupper()) * 0.15
-    if len(line) > 180:
-        score -= 2.0
-    if re.search(r"doi|http|www\.|@", lower):
-        score -= 1.5
-    if re.search(r"^[0-9\.\-\s]+$", line):
-        score -= 3.0
-    return score
-
-
-def extract_title_from_text(raw_text: str, fallback: str = "Uploaded PDF") -> str:
-    raw_text = clean_extracted_text(raw_text or "")
-    lines = [norm_text(x) for x in raw_text.splitlines() if norm_text(x)]
-    head = lines[:18]
-    if not head:
-        return fallback
-    best = sorted(head, key=line_quality_score, reverse=True)[0]
-    best = re.sub(r"\s{2,}", " ", best).strip(" -:;")
-    return truncate_text(best or fallback, 260)
-
-
-def sentence_split(text: str) -> List[str]:
-    text = clean_extracted_text(text)
-    return [norm_text(s) for s in re.split(r"(?<=[\.\!\?])\s+", text) if norm_text(s)]
-
-
-def looks_like_bad_phrase(phrase: str) -> bool:
-    phrase = norm_text(phrase)
-    if not phrase:
-        return True
-    if len(phrase) < 4 or len(phrase) > 90:
-        return True
-    tokens_ = phrase.split()
-    if len(tokens_) > 6:
-        return True
-    for tok in tokens_:
-        t = tok.strip("-.,;:()[]{}")
-        if not t:
-            return True
-        if len(t) == 1 and t.lower() not in {"p", "h"}:
-            return True
-        if re.search(r"[^A-Za-z0-9\-/+]", t):
-            return True
-        if re.search(r"(.)\1\1\1", t.lower()):
-            return True
-        if re.match(r"^[bcdfghjklmnpqrstvwxyz]{5,}$", t.lower()):
-            return True
-    return False
-
-
-def normalize_concept_label(phrase: str) -> str:
-    phrase = norm_text(phrase)
-    mapping = {"ph": "pH", "ecg": "ECG", "mri": "MRI", "ai": "AI", "ml": "ML", "3d": "3D"}
-    parts = []
-    for part in phrase.split():
-        low = part.lower()
-        parts.append(mapping.get(low, part))
-    return " ".join(parts)
-
-
-def dedupe_similar_strings(items: List[str], max_items: int) -> List[str]:
-    result = []
-    seen = set()
-    for item in items:
-        cleaned = normalize_concept_label(item)
-        low = cleaned.lower()
-        if low in seen:
-            continue
-        if any(low != s and (low in s or s in low) and abs(len(low) - len(s)) <= 2 for s in seen):
-            continue
-        seen.add(low)
-        result.append(cleaned)
-        if len(result) >= max_items:
-            break
-    return result
-
-
-def extract_candidate_phrases(text: str, max_terms: int = 20) -> List[str]:
-    text = clean_extracted_text(text)
-    words = re.findall(r"[A-Za-z][A-Za-z0-9\-/+]{2,}", text)
-    phrases = []
-
-    for n in (3, 2, 1):
-        for i in range(len(words) - n + 1):
-            phrase = " ".join(words[i:i + n])
-            low = phrase.lower()
-            parts = low.split()
-            if any(p in STOPWORDS for p in parts):
-                continue
-            if looks_like_bad_phrase(phrase):
-                continue
-            if len(parts) == 1 and len(parts[0]) < 5:
-                continue
-            phrases.append(low)
-
-    counts = Counter(phrases)
-    ranked = []
-    for phrase, count in counts.most_common(max_terms * 8):
-        score = float(count)
-        token_count = len(phrase.split())
-        score += 0.25 * token_count
-        if any(x in phrase for x in ["hydrogel", "scaffold", "conductive", "regeneration", "learning", "graph", "neural", "cardiac", "biomaterial"]):
-            score += 0.5
-        ranked.append((score, phrase))
-
-    ranked_phrases = [p for _, p in sorted(ranked, key=lambda x: x[0], reverse=True)]
-    return dedupe_similar_strings(ranked_phrases, max_terms)
-
-
-def extract_concepts_from_text(text: str, max_terms: int = GRAPH_MAX_CONCEPTS) -> List[str]:
-    return extract_candidate_phrases(text, max_terms=max_terms)
-
-
-def extract_claim_like_sentences(text: str, max_items: int = GRAPH_MAX_CLAIMS) -> List[str]:
-    scored = []
-    for sentence in sentence_split(text):
-        if len(sentence) < 45 or len(sentence) > 320:
-            continue
-        lower = sentence.lower()
-        score = 0.0
-        if any(k in lower for k in ["improves", "reduces", "increases", "demonstrates", "shows", "reveals", "predicts", "achieves", "outperforms", "enables", "supports"]):
-            score += 2.0
-        if any(k in lower for k in ["significant", "associated", "correlated", "effective", "robust", "accurate", "validated", "statistically"]):
-            score += 1.0
-        if any(k in lower for k in ["compared", "versus", "baseline", "state-of-the-art", "sota"]):
-            score += 1.0
-        score += min(len(tokenize(sentence)) / 18.0, 2.0)
-        scored.append((score, sentence))
-    best = [s for _, s in sorted(scored, key=lambda x: x[0], reverse=True)]
-    return dedupe_similar_strings(best, max_items)
-
-
-def parse_openalex_abstract(inverted_index) -> str:
-    if not inverted_index or not isinstance(inverted_index, dict):
-        return ""
-    pos_to_word = {}
-    for word, positions in inverted_index.items():
-        for pos in positions:
-            pos_to_word[pos] = word
-    if not pos_to_word:
-        return ""
-    return " ".join(pos_to_word[i] for i in sorted(pos_to_word))
-
-
-def paper_identity_key(paper: Dict) -> str:
-    return (
-        normalize_doi(paper.get("doi") or "")
-        or (paper.get("external_ids") or {}).get("arxiv")
-        or (paper.get("external_ids") or {}).get("pmcid")
-        or norm_text(paper.get("title", "")).lower()
-        or str(paper.get("id"))
-    )
-
-
-def score_frontier_candidate(query: str, seed_concepts: List[str], paper: Dict) -> Dict:
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue])
-    rel = text_overlap_score(query, base_text)
-    concept_overlap = 0.0
-    if seed_concepts:
-        concept_overlap = text_overlap_score(" ".join(seed_concepts), " ".join(paper.get("concepts") or []))
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    score = float(paper.get("learned_score", paper.get("score", 0))) + rel * 0.45 + concept_overlap * 0.22 + recency + doi_bonus + oa_bonus
-    paper["frontier_score"] = round(score, 4)
-    paper["frontier_relevance"] = round(rel, 4)
-    paper["frontier_concept_overlap"] = round(concept_overlap, 4)
-    return paper
-
-
-def enrich_paper_semantics(query: str, paper: Dict) -> Dict:
-    paper = dict(paper)
-    title = clean_extracted_text(paper.get("title", ""))
-    abstract = clean_extracted_text(paper.get("abstract", "") or paper.get("summary", ""))
-    venue = clean_extracted_text(paper.get("venue", ""))
-    base_text = " ".join([title, abstract, venue]).strip()
-
-    concepts = extract_concepts_from_text(base_text, max_terms=GRAPH_MAX_CONCEPTS)
-    claims = extract_claim_like_sentences(abstract, max_items=GRAPH_MAX_CLAIMS)
-
-    rel = text_overlap_score(query, f"{title} {abstract}")
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    concept_bonus = min(len(concepts), 8) * 0.012
-
-    learned_score = float(paper.get("score", 0)) + rel * 0.52 + recency + doi_bonus + oa_bonus + concept_bonus
-
-    paper["title"] = title or paper.get("title", "Untitled")
-    paper["abstract"] = abstract
-    paper["summary"] = truncate_text(abstract or paper.get("summary", ""), 520)
-    paper["venue"] = venue
-    paper["concepts"] = concepts[:GRAPH_MAX_CONCEPTS]
-    paper["claims"] = claims[:GRAPH_MAX_CLAIMS]
-    paper["relevance"] = round(rel, 4)
-    paper["learned_score"] = round(learned_score, 4)
-    return paper
-
-
-def journal_query_links(query: str):
-    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
-    rows = []
-    for journal in JOURNALS:
-        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
-        if "ieeexplore" in journal["url"]:
-            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
-        rows.append({"name": journal["name"], "desc": journal["desc"], "url": url})
-    return rows
-
-
-def build_journal_html(query):
-    rows = []
-    for journal in journal_query_links(query):
-        rows.append(
-            f"""
-            <a class="journal-card" href="{safe_text(journal['url'])}" target="_blank" rel="noopener noreferrer">
-              <div>
-                <h4>{safe_text(journal['name'])}</h4>
-                <p>{safe_text(journal['desc'])}</p>
-              </div>
-              <span>Open</span>
-            </a>
-            """
-        )
-    return '<div class="journal-grid">' + ''.join(rows) + '</div>'
-
-
-def search_arxiv(query, max_results=8):
-    encoded = urllib.parse.quote(query)
-    url = (
-        "http://export.arxiv.org/api/query?search_query=all:"
-        f"{encoded}&start=0&max_results={max_results}&sortBy=relevance&sortOrder=descending"
-    )
-    response = HTTP.get(url)
-    response.raise_for_status()
-    root = ET.fromstring(response.text)
-    ns = {"atom": "http://www.w3.org/2005/Atom"}
-    papers = []
-    for entry in root.findall("atom:entry", ns):
-        title = clean_extracted_text(entry.findtext("atom:title", default="", namespaces=ns) or "")
-        summary = truncate_text(clean_extracted_text(entry.findtext("atom:summary", default="", namespaces=ns) or ""), MAX_ABSTRACT_CHARS)
-        published = entry.findtext("atom:published", default="", namespaces=ns)
-        paper_id = entry.findtext("atom:id", default="", namespaces=ns)
-        authors = [clean_extracted_text(a.findtext("atom:name", default="", namespaces=ns)) for a in entry.findall("atom:author", ns)]
-        pdf_url = ""
-        for link in entry.findall("atom:link", ns):
-            if link.attrib.get("title") == "pdf":
-                pdf_url = link.attrib.get("href", "")
-                break
-        papers.append({
-            "id": paper_id or title,
-            "title": title,
-            "summary": summary,
-            "abstract": summary,
-            "published": published[:10],
-            "authors": [a for a in authors[:8] if a],
-            "authors_text": ", ".join([a for a in authors[:4] if a]) or "Unknown authors",
-            "url": paper_id,
-            "pdf": pdf_url,
-            "doi": "",
-            "venue": "arXiv",
-            "year": published[:4] if published else "",
-            "source": "arxiv",
-            "score": 0.76,
-            "open_access": True,
-            "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
-        })
-    return papers
-
-
-def search_crossref(query, mode="topic", max_results=8):
-    params = {}
-    if CROSSREF_MAILTO:
-        params["mailto"] = CROSSREF_MAILTO
-    if mode == "doi":
-        url = f"https://api.crossref.org/works/{urllib.parse.quote(query)}"
-        response = HTTP.get(url, params=params)
-        if response.status_code != 200:
-            return []
-        items = [response.json().get("message", {})]
-    else:
-        params["rows"] = max_results
-        if mode in ("title", "paper_name"):
-            params["query.title"] = query
-        else:
-            params["query.bibliographic"] = query
-        response = HTTP.get("https://api.crossref.org/works", params=params)
-        response.raise_for_status()
-        items = response.json().get("message", {}).get("items", [])
-    out = []
-    for item in items:
-        authors = []
-        for a in item.get("author", []) or []:
-            name = " ".join(filter(None, [a.get("given"), a.get("family")])).strip()
-            if name:
-                authors.append(clean_extracted_text(name))
-        title = clean_extracted_text((item.get("title") or ["Untitled"])[0])
-        year = ""
-        for key in ["published-print", "published-online", "created"]:
-            if item.get(key, {}).get("date-parts"):
-                year = str(item[key]["date-parts"][0][0])
-                break
-        abstract = truncate_text(clean_extracted_text(re.sub("<.*?>", " ", item.get("abstract") or "")), MAX_ABSTRACT_CHARS)
-        doi = normalize_doi(item.get("DOI", ""))
-        out.append({
-            "id": doi or title,
-            "title": title,
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": year,
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("URL", ""),
-            "pdf": "",
-            "doi": doi,
-            "venue": clean_extracted_text((item.get("container-title") or [""])[0]),
-            "year": year,
-            "source": "crossref",
-            "score": 0.72,
-            "open_access": None,
-            "external_ids": {"crossref": doi} if doi else {},
-        })
-    return out
-
-
-def search_openalex(query, mode="topic", max_results=8):
-    params = {"per-page": max_results}
-    if OPENALEX_EMAIL:
-        params["mailto"] = OPENALEX_EMAIL
-    if mode == "doi":
-        doi = normalize_doi(query)
-        params["filter"] = f"doi:https://doi.org/{doi}"
-    else:
-        params["search"] = query
-    response = HTTP.get("https://api.openalex.org/works", params=params)
-    response.raise_for_status()
-    items = response.json().get("results", [])
-    out = []
-    for item in items:
-        authors = []
-        for auth in item.get("authorships", [])[:8]:
-            author = auth.get("author") or {}
-            if author.get("display_name"):
-                authors.append(clean_extracted_text(author["display_name"]))
-        oa = item.get("open_access") or {}
-        doi = normalize_doi(item.get("doi") or "")
-        abstract = truncate_text(clean_extracted_text(parse_openalex_abstract(item.get("abstract_inverted_index"))), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": item.get("id") or doi or item.get("title"),
-            "title": clean_extracted_text(item.get("title") or ""),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("publication_year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": (item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or "",
-            "pdf": oa.get("oa_url") or "",
-            "doi": doi,
-            "venue": clean_extracted_text(((item.get("primary_location") or {}).get("source") or {}).get("display_name") or ""),
-            "year": str(item.get("publication_year") or ""),
-            "source": "openalex",
-            "score": 0.80,
-            "open_access": oa.get("is_oa"),
-            "external_ids": item.get("ids") or {},
-        })
-    return out
-
-
-def search_semantic_scholar(query, mode="topic", max_results=8):
-    headers = {}
-    if SEMANTIC_SCHOLAR_API_KEY:
-        headers["x-api-key"] = SEMANTIC_SCHOLAR_API_KEY
-    fields = "title,authors,year,abstract,venue,externalIds,url,openAccessPdf"
-    if mode == "doi":
-        doi = normalize_doi(query)
-        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{urllib.parse.quote(doi)}"
-        response = HTTP.get(url, params={"fields": fields}, headers=headers)
-        if response.status_code != 200:
-            return []
-        items = [response.json()]
-    else:
-        response = HTTP.get(
-            "https://api.semanticscholar.org/graph/v1/paper/search",
-            params={"query": query, "limit": max_results, "fields": fields},
-            headers=headers,
-        )
-        if response.status_code != 200:
-            return []
-        items = response.json().get("data", [])
-    out = []
-    for item in items:
-        external = item.get("externalIds") or {}
-        authors = [clean_extracted_text(a.get("name")) for a in item.get("authors", []) if a.get("name")]
-        abstract = truncate_text(clean_extracted_text(item.get("abstract", "")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": external.get("CorpusId") or external.get("DOI") or item.get("title"),
-            "title": clean_extracted_text(item.get("title") or ""),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("year") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": item.get("url") or "",
-            "pdf": (item.get("openAccessPdf") or {}).get("url") or "",
-            "doi": normalize_doi(external.get("DOI", "")),
-            "venue": clean_extracted_text(item.get("venue") or ""),
-            "year": str(item.get("year") or ""),
-            "source": "semantic_scholar",
-            "score": 0.84,
-            "open_access": bool((item.get("openAccessPdf") or {}).get("url")),
-            "external_ids": external,
-        })
-    return out
-
-
-def search_europe_pmc(query, mode="topic", max_results=8):
-    epmc_query = f'DOI:"{query}"' if mode == "doi" else query
-    params = {"query": epmc_query, "format": "json", "pageSize": max_results, "resultType": "core"}
-    response = HTTP.get("https://www.ebi.ac.uk/europepmc/webservices/rest/search", params=params)
-    if response.status_code != 200:
-        return []
-    items = response.json().get("resultList", {}).get("result", [])
-    out = []
-    for item in items:
-        author_string = item.get("authorString", "")
-        authors = [clean_extracted_text(x) for x in author_string.split(",")[:8] if norm_text(x)]
-        pmcid = item.get("pmcid", "")
-        pdf_url = f"https://europepmc.org/articles/{pmcid}?pdf=render" if pmcid else ""
-        landing_url = f"https://europepmc.org/article/PMC/{pmcid}" if pmcid else ""
-        abstract = truncate_text(clean_extracted_text(item.get("abstractText", "")), MAX_ABSTRACT_CHARS)
-        out.append({
-            "id": item.get("id") or item.get("doi") or item.get("title"),
-            "title": clean_extracted_text(item.get("title") or ""),
-            "summary": abstract[:500],
-            "abstract": abstract,
-            "published": str(item.get("pubYear") or ""),
-            "authors": authors,
-            "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": landing_url,
-            "pdf": pdf_url,
-            "doi": normalize_doi(item.get("doi", "")),
-            "venue": clean_extracted_text(item.get("journalTitle", "")),
-            "year": str(item.get("pubYear") or ""),
-            "source": "europe_pmc",
-            "score": 0.78,
-            "open_access": bool(pmcid),
-            "external_ids": {"pmid": item.get("pmid"), "pmcid": pmcid},
-        })
-    return out
-
-
-def resolve_link(query):
-    url = (query or "").strip()
-    if not url:
-        return []
-    try:
-        response = HTTP.get(url, allow_redirects=True, headers={"User-Agent": "dvnc-ai-space/2.0"})
-        content_type = response.headers.get("content-type", "")
-        if "pdf" in content_type or url.lower().endswith(".pdf"):
-            name = Path(url.split("?")[0]).name or "linked-paper.pdf"
-            return [{
-                "id": url,
-                "title": name,
-                "summary": "Direct PDF link detected.",
-                "abstract": "",
-                "published": "",
-                "authors": [],
-                "authors_text": "Unknown authors",
-                "url": url,
-                "pdf": url,
-                "doi": "",
-                "venue": "Direct PDF",
-                "year": "",
-                "source": "link",
-                "score": 0.66,
-                "open_access": True,
-                "external_ids": {},
-            }]
-        doi = ""
-        title = url
-        pdf_link = ""
-        if BeautifulSoup is not None:
-            soup = BeautifulSoup(response.text, "html.parser")
-            title = clean_extracted_text(soup.title.text.strip()) if soup.title else url
-            for meta_name in ["citation_doi", "dc.identifier", "dc.Identifier"]:
-                tag = soup.find("meta", attrs={"name": meta_name})
-                if tag and tag.get("content"):
-                    doi = normalize_doi(tag["content"].strip())
-                    break
-            for a in soup.find_all("a", href=True):
-                href = a["href"]
-                if ".pdf" in href.lower():
-                    pdf_link = href if href.startswith("http") else urllib.parse.urljoin(url, href)
-                    break
-        if doi:
-            results = search_crossref(doi, mode="doi", max_results=1)
-            if results:
-                if pdf_link and not results[0].get("pdf"):
-                    results[0]["pdf"] = pdf_link
-                if url and not results[0].get("url"):
-                    results[0]["url"] = url
-                return results
-        return [{
-            "id": url,
-            "title": title,
-            "summary": "Landing page resolved from direct link.",
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": pdf_link,
-            "doi": doi,
-            "venue": "Web Link",
-            "year": "",
-            "source": "link",
-            "score": 0.54,
-            "open_access": bool(pdf_link),
-            "external_ids": {},
-        }]
-    except Exception as e:
-        return [{
-            "id": url,
-            "title": "Link resolution error",
-            "summary": str(e),
-            "abstract": "",
-            "published": "",
-            "authors": [],
-            "authors_text": "Unknown authors",
-            "url": url,
-            "pdf": "",
-            "doi": "",
-            "venue": "Link",
-            "year": "",
-            "source": "link",
-            "score": 0.20,
-            "open_access": None,
-            "external_ids": {},
-        }]
-
-
-def dedupe_papers(items: List[Dict]) -> List[Dict]:
-    seen = {}
-    for item in items:
-        key = paper_identity_key(item) or f"{item.get('source', 'src')}::{item.get('title', 'paper')}"
-        current = seen.get(key)
-        candidate_score = float(item.get("learned_score", item.get("score", 0)))
-        current_score = float(current.get("learned_score", current.get("score", 0))) if current else -1
-        if current is None or candidate_score > current_score:
-            seen[key] = item
-    return sorted(seen.values(), key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-
-
-def discover_papers(query, mode, sources, max_results=10):
-    query = (query or "").strip()
-    if not query:
-        return []
-    mode = detect_query_type(query) if mode == "autonomous_web" else mode
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-    results = []
-
-    if mode == "link":
-        return dedupe_papers(resolve_link(query))
-
-    if "arxiv" in selected_sources and mode != "doi":
-        try:
-            results.extend(search_arxiv(query, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "crossref" in selected_sources:
-        try:
-            results.extend(search_crossref(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "openalex" in selected_sources:
-        try:
-            results.extend(search_openalex(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "semantic_scholar" in selected_sources:
-        try:
-            results.extend(search_semantic_scholar(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-    if "europe_pmc" in selected_sources:
-        try:
-            results.extend(search_europe_pmc(query, mode=mode, max_results=min(max_results, GRAPH_MAX_RESULTS)))
-        except Exception:
-            pass
-
-    papers = [enrich_paper_semantics(query, p) for p in dedupe_papers(results)]
-    papers = sorted(papers, key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-    return papers[:max_results]
-
-
-def propose_expansion_queries(query: str, papers: List[Dict], parsed_state: Optional[Dict] = None, limit: int = GRAPH_MAX_EXPANSIONS) -> List[str]:
-    concept_pool = []
-    venue_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:4])
-        if paper.get("venue"):
-            venue_pool.append(paper["venue"])
-    if parsed_state and isinstance(parsed_state, dict):
-        concept_pool.extend((parsed_state.get("concepts") or [])[:6])
-
-    ranked_concepts = [c for c, _ in Counter([norm_text(c).lower() for c in concept_pool if c]).most_common(limit * 2)]
-    expansions = [norm_text(query)] if query else []
-    for concept in ranked_concepts:
-        if concept:
-            expansions.append(f"{query} {concept}".strip())
-    for venue in unique_keep_order(venue_pool)[:2]:
-        if venue:
-            expansions.append(f"{query} {venue}".strip())
-    return unique_keep_order(expansions)[:limit]
-
-
-def frontier_expand(query: str, sources: List[str], selected_papers: List[Dict], parsed_state: Optional[Dict] = None, per_query: int = 4) -> List[Dict]:
-    seed_concepts = []
-    for p in selected_papers[:6]:
-        seed_concepts.extend((p.get("concepts") or [])[:4])
-    if parsed_state and isinstance(parsed_state, dict):
-        seed_concepts.extend((parsed_state.get("concepts") or [])[:6])
-
-    expansion_queries = propose_expansion_queries(query, selected_papers, parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    frontier = []
-    for eq in expansion_queries:
-        try:
-            items = discover_papers(eq, "topic", sources, max_results=per_query)
-            for item in items:
-                frontier.append(score_frontier_candidate(query or eq, seed_concepts, item))
-        except Exception:
-            continue
-
-    frontier = dedupe_papers(frontier)
-    frontier.sort(key=lambda x: float(x.get("frontier_score", x.get("learned_score", x.get("score", 0)))), reverse=True)
-    GRAPH_MEMORY["frontier"] = frontier[: GRAPH_MAX_EXPANSIONS * per_query]
-    return GRAPH_MEMORY["frontier"]
-
-
-def paper_choice_value(index: int, paper: Dict) -> str:
-    doi = normalize_doi(paper.get("doi") or "")
-    title_slug = slugify(paper.get("title", ""))[:40]
-    return f"{index}|{doi}|{title_slug}"
-
-
-def paper_choice_label(index: int, paper: Dict) -> str:
-    score = round(float(paper.get("learned_score", paper.get("score", 0))), 3)
-    title = paper.get("title", "Untitled")
-    authors_text = paper.get("authors_text", "Unknown authors")[:90]
-    source = paper.get("source", "src")
-    return f"[{source}] {title} — {authors_text} — score {score}"
-
-
-def format_selection_choices(papers):
-    return [(paper_choice_label(i, paper), paper_choice_value(i, paper)) for i, paper in enumerate(papers)]
-
-
-def format_papers_html(papers):
-    if not papers:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No papers found yet.</p></div>'
-
-    items = []
-    for i, paper in enumerate(papers, start=1):
-        summary = safe_text((paper.get("summary") or paper.get("abstract") or "")[:320])
-        doi_line = f'<span class="paper-badge doi-badge">{safe_text(paper.get("doi"))}</span>' if paper.get("doi") else ""
-        pdf_link = paper.get("pdf") or "#"
-        abs_link = paper.get("url") or "#"
-        concepts_text = ", ".join((paper.get("concepts") or [])[:5])
-
-        items.append(
-            f"""
-            <article class="paper-card">
-              <div class="paper-topline">
-                <span class="paper-badge">{safe_text(paper.get('source', 'paper'))}</span>
-                <span class="paper-badge alt">{safe_text(paper.get('published', '') or 'Paper')}</span>
-                {doi_line}
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{summary or 'No abstract snippet available.'}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-                <div><strong>Venue:</strong> {safe_text(paper.get('venue', 'Unknown venue'))}</div>
-                <div><strong>Learned score:</strong> {safe_text(round(float(paper.get('learned_score', paper.get('score', 0))), 3))}</div>
-                <div><strong>Concepts:</strong> {safe_text(concepts_text or 'None extracted')}</div>
-              </div>
-              <div class="paper-links">
-                <a href="{safe_text(abs_link)}" target="_blank" rel="noopener noreferrer">Abstract</a>
-                <a href="{safe_text(pdf_link)}" target="_blank" rel="noopener noreferrer">PDF</a>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(items) + '</div>'
-
-
-def format_frontier_html(frontier):
-    if not frontier:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No autonomous expansion candidates yet.</p></div>'
-    cards = []
-    for i, paper in enumerate(frontier[:12], start=1):
-        cards.append(
-            f"""
-            <article class="paper-card frontier-card">
-              <div class="paper-topline">
-                <span class="paper-badge">frontier</span>
-                <span class="paper-badge alt">{safe_text(paper.get('source', 'paper'))}</span>
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{safe_text((paper.get('summary') or paper.get('abstract') or '')[:280])}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Frontier score:</strong> {safe_text(paper.get('frontier_score', paper.get('learned_score', paper.get('score', 0))))}</div>
-                <div><strong>Concept overlap:</strong> {safe_text(paper.get('frontier_concept_overlap', 0))}</div>
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(cards) + '</div>'
-
-
-def uploaded_pdf_summary(file_obj):
-    if not file_obj:
-        return "No PDF uploaded yet."
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    p = Path(path)
-    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, references, concepts, and claims."
-
-
-def graph_kind_style(kind: str) -> Dict[str, Any]:
-    palette = {
-        "query": {"color": "#1f8ef1", "size": 14, "label": "Research topic"},
-        "paper": {"color": "#00c49a", "size": 10, "label": "Paper"},
-        "upload": {"color": "#ff9f43", "size": 11, "label": "Uploaded PDF"},
-        "concept": {"color": "#a66cff", "size": 8, "label": "Concept"},
-        "author": {"color": "#f368e0", "size": 7, "label": "Author"},
-        "claim": {"color": "#ff6b6b", "size": 8, "label": "Claim"},
-        "reference": {"color": "#6c757d", "size": 7, "label": "Reference"},
-        "frontier": {"color": "#ffd166", "size": 8, "label": "Frontier candidate"},
-    }
-    return palette.get(kind, {"color": "#9aa0a6", "size": 7, "label": kind.title()})
-
-
-def summarize_graph(nodes: List[Dict], edges: List[Dict]) -> Dict[str, Any]:
-    counts = Counter((n.get("kind") or n.get("type") or "unknown").lower() for n in nodes)
-    return {"nodes": len(nodes), "edges": len(edges), "counts": dict(counts)}
-
-
-def _prepare_3d_graph_data(nodes: List[Dict], edges: List[Dict], title: str) -> Dict[str, Any]:
-    node_out = []
-    for node in nodes:
-        kind = (node.get("kind") or node.get("type") or "paper").lower()
-        if kind == "topic":
-            kind = "query"
-        if kind == "uploadedpdf":
-            kind = "upload"
-        if kind == "frontierpaper":
-            kind = "frontier"
-        style = graph_kind_style(kind)
-        node_out.append({
-            "id": node.get("id"),
-            "label": truncate_text(node.get("label") or node.get("title") or node.get("id") or "node", 120),
-            "kind": kind,
-            "color": style["color"],
-            "val": style["size"],
-            "detail": {
-                "kind": style["label"],
-                "title": node.get("title") or node.get("label") or node.get("id"),
-                "venue": node.get("venue") or "",
-                "year": node.get("year") or "",
-                "doi": node.get("doi") or "",
-                "source": node.get("source") or "",
-                "authors_text": node.get("authors_text") or "",
-                "text": node.get("text") or "",
-            },
-        })
-    edge_out = []
-    for edge in edges:
-        edge_out.append({
-            "source": edge.get("source"),
-            "target": edge.get("target"),
-            "type": edge.get("type") or "RELATES_TO",
-            "label": edge.get("type") or "RELATES_TO",
-        })
-    return {"title": title, "nodes": node_out, "links": edge_out, "summary": summarize_graph(nodes, edges)}
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    if not nodes:
-        return """
-        <div class="panel brain-shell" style="overflow:auto; max-width:100%;">
-          <div class="brain-header">
-            <div>
-              <p class="eyebrow">Learning Graph</p>
-              <h3>Self-Learning Knowledge Graph</h3>
-            </div>
-          </div>
-          <div class="brain-stage learning-empty" style="min-height:420px; overflow:auto;">
-            <div class="empty-graph-copy">
-              <h4>No papers mapped yet</h4>
-              <p>Search papers, select candidates, or upload a PDF to grow the graph in an interactive 3D view.</p>
-            </div>
-          </div>
-        </div>
-        """
-
-    graph_data = _prepare_3d_graph_data(nodes, edges, title)
-    payload_json = json.dumps(graph_data, ensure_ascii=False)
-
-    iframe_html = f"""
-<!doctype html>
-<html>
-<head>
-<meta charset="utf-8" />
-<meta name="viewport" content="width=device-width, initial-scale=1" />
-<style>
-  html, body {{ margin:0; height:100%; background:#0b1020; color:#eef2ff; font-family: Inter, ui-sans-serif, system-ui, sans-serif; overflow:hidden; }}
-  #wrap {{ position:relative; width:100%; height:100%; background:radial-gradient(circle at top, #18213c 0%, #0b1020 60%, #060910 100%); }}
-  #graph {{ position:absolute; inset:0; }}
-  .overlay {{
-    position:absolute; left:16px; top:16px; z-index:10; max-width:min(460px, calc(100% - 32px));
-    padding:14px 16px; border:1px solid rgba(255,255,255,.12); border-radius:16px;
-    background:rgba(10,14,28,.72); backdrop-filter: blur(14px); box-shadow:0 12px 28px rgba(0,0,0,.28);
-  }}
-  .overlay h3 {{ margin:0 0 6px; font-size:18px; line-height:1.2; }}
-  .overlay p {{ margin:0; font-size:13px; color:#cbd5e1; line-height:1.5; }}
-  .legend {{ display:flex; flex-wrap:wrap; gap:8px 10px; margin-top:10px; }}
-  .chip {{
-    display:inline-flex; align-items:center; gap:8px; padding:6px 10px; border-radius:999px; font-size:12px;
-    color:#e2e8f0; background:rgba(255,255,255,.06); border:1px solid rgba(255,255,255,.08);
-  }}
-  .dot {{ width:10px; height:10px; border-radius:999px; display:inline-block; }}
-  .panel {{
-    position:absolute; right:16px; top:16px; z-index:10; width:min(360px, calc(100% - 32px));
-    max-height:calc(100% - 32px); overflow:auto; padding:14px 16px; border:1px solid rgba(255,255,255,.12);
-    border-radius:16px; background:rgba(10,14,28,.72); backdrop-filter: blur(14px); box-shadow:0 12px 28px rgba(0,0,0,.28);
-  }}
-  .panel h4 {{ margin:0 0 8px; font-size:14px; color:#f8fafc; }}
-  .panel p {{ margin:0; font-size:12px; color:#cbd5e1; line-height:1.5; }}
-  .panel dl {{ margin:12px 0 0; display:grid; grid-template-columns:auto 1fr; gap:6px 10px; font-size:12px; }}
-  .panel dt {{ color:#93c5fd; }}
-  .panel dd {{ margin:0; color:#e2e8f0; word-break:break-word; }}
-  .stats {{ margin-top:10px; display:grid; grid-template-columns:repeat(3,1fr); gap:8px; }}
-  .stat {{ padding:8px 10px; border-radius:12px; background:rgba(255,255,255,.06); border:1px solid rgba(255,255,255,.08); }}
-  .stat strong {{ display:block; font-size:15px; color:#fff; }}
-  .stat span {{ font-size:11px; color:#cbd5e1; }}
-  .hint {{
-    position:absolute; left:16px; bottom:16px; z-index:10; padding:10px 12px; border-radius:12px;
-    font-size:12px; color:#dbeafe; background:rgba(15,23,42,.75); border:1px solid rgba(255,255,255,.08);
-  }}
-</style>
-<script src="https://unpkg.com/three@0.160.0/build/three.min.js"></script>
-<script src="https://unpkg.com/3d-force-graph"></script>
-</head>
-<body>
-<div id="wrap">
-  <div id="graph"></div>
-  <div class="overlay">
-    <h3></h3>
-    <p>Drag the background to orbit, scroll to zoom, right-drag to pan, and drag a node to move or pin it in 3D space.</p>
-    <div class="legend" id="legend"></div>
-    <div class="stats" id="stats"></div>
-  </div>
-  <div class="panel" id="panel">
-    <h4>Node details</h4>
-    <p>Click any node to inspect its label, type, venue, DOI, year, and source.</p>
-  </div>
-  <div class="hint">Interactive 3D graph: orbit, zoom, pan, drag nodes.</div>
-</div>
-<script>
-const payload = {payload_json};
-document.querySelector('.overlay h3').textContent = payload.title || 'Self-Learning Knowledge Graph';
-
-const legendEl = document.getElementById('legend');
-const statEl = document.getElementById('stats');
-const panelEl = document.getElementById('panel');
-
-const kindMap = {{
-  query: ['Research topic', '#1f8ef1'],
-  paper: ['Paper', '#00c49a'],
-  upload: ['Uploaded PDF', '#ff9f43'],
-  concept: ['Concept', '#a66cff'],
-  author: ['Author', '#f368e0'],
-  claim: ['Claim', '#ff6b6b'],
-  reference: ['Reference', '#6c757d'],
-  frontier: ['Frontier candidate', '#ffd166']
-}};
-
-Object.entries(kindMap).forEach(([key, value]) => {{
-  const chip = document.createElement('div');
-  chip.className = 'chip';
-  chip.innerHTML = `<span class="dot" style="background:${{value[1]}}"></span>${{value[0]}}`;
-  legendEl.appendChild(chip);
-}});
-
-const summary = payload.summary || {{nodes: payload.nodes.length, edges: payload.links.length, counts: {{}}}};
-[['Nodes', summary.nodes], ['Edges', summary.edges], ['Types', Object.keys(summary.counts || {{}}).length]].forEach(item => {{
-  const box = document.createElement('div');
-  box.className = 'stat';
-  box.innerHTML = `<strong>${{item[1]}}</strong><span>${{item[0]}}</span>`;
-  statEl.appendChild(box);
-}});
-
-function roundRect(ctx, x, y, width, height, radius) {{
-  ctx.beginPath();
-  ctx.moveTo(x + radius, y);
-  ctx.lineTo(x + width - radius, y);
-  ctx.quadraticCurveTo(x + width, y, x + width, y + radius);
-  ctx.lineTo(x + width, y + height - radius);
-  ctx.quadraticCurveTo(x + width, y + height, x + width - radius, y + height);
-  ctx.lineTo(x + radius, y + height);
-  ctx.quadraticCurveTo(x, y + height, x, y + height - radius);
-  ctx.lineTo(x, y + radius);
-  ctx.quadraticCurveTo(x, y, x + radius, y);
-  ctx.closePath();
-}}
-
-function makeTextSprite(text) {{
-  const canvas = document.createElement('canvas');
-  const ctx = canvas.getContext('2d');
-  const fontSize = 28;
-  ctx.font = `600 ${{fontSize}}px Inter, Arial, sans-serif`;
-  const width = Math.min(900, Math.max(180, ctx.measureText(text).width + 28));
-  canvas.width = width;
-  canvas.height = 52;
-  ctx.font = `600 ${{fontSize}}px Inter, Arial, sans-serif`;
-  ctx.fillStyle = 'rgba(7,12,24,0.78)';
-  roundRect(ctx, 0, 0, width, 52, 18);
-  ctx.fill();
-  ctx.fillStyle = '#f8fafc';
-  ctx.textBaseline = 'middle';
-  ctx.fillText(text.slice(0, 64), 14, 26);
-  const texture = new THREE.CanvasTexture(canvas);
-  const material = new THREE.SpriteMaterial({{ map: texture, transparent: true }});
-  const sprite = new THREE.Sprite(material);
-  sprite.scale.set(width / 10, 5.2, 1);
-  sprite.position.set(0, 10, 0);
-  return sprite;
-}}
-
-function nodeObject(node) {{
-  const group = new THREE.Group();
-  const geo = new THREE.SphereGeometry(Math.max(2.8, (node.val || 6) * 0.5), 18, 18);
-  const mat = new THREE.MeshStandardMaterial({{ color: node.color || '#94a3b8', metalness: 0.15, roughness: 0.45 }});
-  const sphere = new THREE.Mesh(geo, mat);
-  group.add(sphere);
-  group.add(makeTextSprite(node.label || node.id));
-  return group;
-}}
-
-function renderPanel(node, pinned) {{
-  const d = node.detail || {{}};
-  panelEl.innerHTML = `
-    <h4>Node details</h4>
-    <dl>
-      <dt>Label</dt><dd>${{node.label || ''}}</dd>
-      <dt>Type</dt><dd>${{d.kind || node.kind || ''}}</dd>
-      <dt>Venue</dt><dd>${{d.venue || '—'}}</dd>
-      <dt>Year</dt><dd>${{d.year || '—'}}</dd>
-      <dt>DOI</dt><dd>${{d.doi || '—'}}</dd>
-      <dt>Source</dt><dd>${{d.source || '—'}}</dd>
-      <dt>Authors</dt><dd>${{d.authors_text || '—'}}</dd>
-      <dt>Text</dt><dd>${{d.text || '—'}}</dd>
-      <dt>Pinned</dt><dd>${{pinned ? 'Yes' : (node.fx != null ? 'Yes' : 'No')}}</dd>
-    </dl>`;
-}}
-
-const elem = document.getElementById('graph');
-const Graph = ForceGraph3D()(elem)
-  .backgroundColor('#00000000')
-  .graphData(payload)
-  .nodeRelSize(6)
-  .nodeOpacity(1)
-  .nodeLabel(node => `<div style="padding:6px 8px"><strong>${{node.label}}</strong><br/>${{(node.detail || {{}}).kind || node.kind}}</div>`)
-  .nodeThreeObject(nodeObject)
-  .linkColor(link => {{
-    const m = {{
-      ABOUT:'#4dabf7',
-      MENTIONS:'#b197fc',
-      WRITTEN_BY:'#f783ac',
-      ASSERTS:'#ff8787',
-      CITES:'#adb5bd',
-      FRONTIER:'#ffe066',
-      FRONTIER_CANDIDATE:'#ffe066',
-      UPLOADED_SOURCE:'#ffa94d'
-    }};
-    return m[link.type] || 'rgba(226,232,240,0.42)';
-  }})
-  .linkWidth(link => ['ABOUT','UPLOADED_SOURCE','FRONTIER','FRONTIER_CANDIDATE'].includes(link.type) ? 2.6 : 1.35)
-  .linkDirectionalParticles(link => ['ABOUT','FRONTIER','FRONTIER_CANDIDATE'].includes(link.type) ? 2 : 0)
-  .linkDirectionalParticleWidth(2.2)
-  .cooldownTicks(140)
-  .d3VelocityDecay(0.24)
-  .d3Force('charge').strength(-180)
-  .onNodeClick(node => {{
-    const distance = 90;
-    const distRatio = 1 + distance / Math.hypot(node.x || 1, node.y || 1, node.z || 1);
-    Graph.cameraPosition(
-      {{ x: (node.x || 0) * distRatio, y: (node.y || 0) * distRatio, z: (node.z || 0) * distRatio }},
-      node,
-      900
-    );
-    renderPanel(node);
-  }})
-  .onNodeDragEnd(node => {{
-    node.fx = node.x;
-    node.fy = node.y;
-    node.fz = node.z;
-    renderPanel(node, true);
-  }});
-
-Graph.scene().add(new THREE.AmbientLight(0xffffff, 1.1));
-const dirLight = new THREE.DirectionalLight(0xffffff, 0.8);
-dirLight.position.set(120, 120, 120);
-Graph.scene().add(dirLight);
-Graph.cameraPosition({{ z: 210 }});
-
-window.addEventListener('resize', () => {{
-  Graph.width(elem.clientWidth);
-  Graph.height(elem.clientHeight);
-}});
-</script>
-</body>
-</html>
-"""
-    return f"""
-    <div class="panel brain-shell" style="overflow:auto; max-width:100%;">
-      <iframe
-        title="{safe_text(title)}"
-        style="width:100%; height:{GRAPH_IFRAME_HEIGHT}px; border:0; border-radius:18px; overflow:auto; background:#0b1020;"
-        sandbox="allow-scripts allow-same-origin"
-        srcdoc="{html.escape(iframe_html, quote=True)}"
-      ></iframe>
-    </div>
-    """
-
-
-def build_learning_graph_state(query, papers, uploaded_name=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-
-    for i, paper in enumerate(papers[:6], start=1):
-        pid = f"paper_{i}"
-        nodes.append({
-            "id": pid,
-            "label": paper.get("title", f"Paper {i}"),
-            "kind": "paper",
-            "title": paper.get("title"),
-            "venue": paper.get("venue"),
-            "year": paper.get("year"),
-            "source": paper.get("source"),
-            "authors_text": paper.get("authors_text"),
-        })
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-        for concept in (paper.get("concepts") or [])[:3]:
-            cid = f"concept_{i}_{slugify(concept)[:30]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-    return nodes, edges
-
-
-def graph_from_selected(query, selected_papers, uploaded_name=None, parsed_state=None, frontier=None):
-    nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
-    edges = []
-
-    for i, paper in enumerate(selected_papers[:8], start=1):
-        pid = f"paper_{i}"
-        nodes.append({
-            "id": pid,
-            "label": paper.get("title", f"Paper {i}"),
-            "kind": "paper",
-            "title": paper.get("title"),
-            "venue": paper.get("venue"),
-            "year": paper.get("year"),
-            "doi": paper.get("doi"),
-            "source": paper.get("source"),
-            "authors_text": paper.get("authors_text"),
-        })
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-
-        for author in paper.get("authors", [])[:3]:
-            aid = f"author_{i}_{slugify(author)[:30]}"
-            nodes.append({"id": aid, "label": author, "kind": "author"})
-            edges.append({"source": pid, "target": aid, "type": "WRITTEN_BY"})
-
-        for concept in (paper.get("concepts") or [])[:4]:
-            cid = f"concept_{i}_{slugify(concept)[:30]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-
-        for claim in (paper.get("claims") or [])[:2]:
-            cid = f"claim_{i}_{slugify(claim)[:30]}"
-            nodes.append({"id": cid, "label": truncate_text(claim, 82), "kind": "claim", "text": claim})
-            edges.append({"source": pid, "target": cid, "type": "ASSERTS"})
-
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload", "title": uploaded_name})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-        if parsed_state and isinstance(parsed_state, dict):
-            for concept in (parsed_state.get("concepts") or [])[:4]:
-                cid = f"upload_concept_{slugify(concept)[:30]}"
-                nodes.append({"id": cid, "label": concept, "kind": "concept"})
-                edges.append({"source": "upload", "target": cid, "type": "MENTIONS"})
-            for claim in (parsed_state.get("claims") or [])[:3]:
-                cid = f"upload_claim_{slugify(claim)[:30]}"
-                nodes.append({"id": cid, "label": truncate_text(claim, 82), "kind": "claim", "text": claim})
-                edges.append({"source": "upload", "target": cid, "type": "ASSERTS"})
-            for ref in (parsed_state.get("references") or [])[:6]:
-                if ref.get("title"):
-                    rid = f"ref_{slugify(ref.get('title'))[:30]}"
-                    nodes.append({"id": rid, "label": truncate_text(ref.get("title"), 82), "kind": "reference", "doi": ref.get("doi", "")})
-                    edges.append({"source": "upload", "target": rid, "type": "CITES"})
-
-    for j, fp in enumerate(ensure_list(frontier)[:6], start=1):
-        fid = f"frontier_{j}"
-        nodes.append({
-            "id": fid,
-            "label": fp.get("title", f"Frontier {j}"),
-            "kind": "frontier",
-            "title": fp.get("title"),
-            "source": fp.get("source"),
-            "authors_text": fp.get("authors_text"),
-            "year": fp.get("year"),
-            "doi": fp.get("doi"),
-        })
-        edges.append({"source": "query", "target": fid, "type": "FRONTIER_CANDIDATE"})
-
-    dedup_nodes = []
-    seen = set()
-    for node in nodes:
-        if node["id"] not in seen:
-            seen.add(node["id"])
-            dedup_nodes.append(node)
-    return dedup_nodes, edges
-
-
-def extract_references_from_text(text: str) -> List[Dict[str, str]]:
-    refs = []
-    seen = set()
-    lines = [norm_text(x) for x in clean_extracted_text(text).splitlines() if norm_text(x)]
-    ref_lines = lines[-250:]
-    doi_re = re.compile(r"10\.\d{4,9}/[-._;()/:A-Z0-9]+", re.I)
-
-    for line in ref_lines:
-        doi_match = doi_re.search(line)
-        title = line
-        doi = doi_match.group(0) if doi_match else ""
-        if doi:
-            title = line.replace(doi, "").strip(" .;,-")
-        if len(title) < 12:
-            continue
-        key = (title.lower(), doi.lower())
-        if key in seen:
-            continue
-        seen.add(key)
-        refs.append({"title": truncate_text(title, 240), "doi": normalize_doi(doi)})
-        if len(refs) >= 40:
-            break
-    return refs
-
-
-def parse_pdf_with_grobid(pdf_path):
-    if not GROBID_URL:
-        raise RuntimeError("GROBID_URL is not set")
-
-    with open(pdf_path, "rb") as f:
-        files = {"input": (Path(pdf_path).name, f, "application/pdf")}
-        response = HTTP.post(
-            f"{GROBID_URL.rstrip('/')}/api/processFulltextDocument",
-            files=files,
-            data={"includeRawAffiliations": "1", "segmentSentences": "1"},
-            timeout=180,
-        )
-    response.raise_for_status()
-
-    root = ET.fromstring(response.text)
-    ns = {"tei": "http://www.tei-c.org/ns/1.0"}
-
-    title = clean_extracted_text(root.findtext(".//tei:titleStmt/tei:title", default="", namespaces=ns) or Path(pdf_path).name)
-
-    abstract_parts = []
-    for p in root.findall(".//tei:profileDesc/tei:abstract//tei:p", ns):
-        abstract_parts.append(clean_extracted_text(" ".join(list(p.itertext()))))
-    abstract = truncate_text(" ".join(abstract_parts), MAX_ABSTRACT_CHARS)
-
-    authors = []
-    for author in root.findall(".//tei:sourceDesc//tei:author", ns):
-        parts = []
-        for forename in author.findall(".//tei:forename", ns):
-            parts.append(norm_text(" ".join(forename.itertext())))
-        for surname in author.findall(".//tei:surname", ns):
-            parts.append(norm_text(" ".join(surname.itertext())))
-        name = clean_extracted_text(" ".join(parts))
-        if name:
-            authors.append(name)
-
-    sections = []
-    text_pool = []
-    for div in root.findall(".//tei:text//tei:body//tei:div", ns):
-        head = clean_extracted_text(div.findtext("./tei:head", default="", namespaces=ns) or "Section")
-        paras = []
-        for p in div.findall(".//tei:p", ns):
-            para_text = clean_extracted_text(" ".join(list(p.itertext())))
-            if para_text:
-                paras.append(para_text)
-        joined = "\n".join(paras)
-        if head or joined:
-            sections.append({"heading": head or "Section", "text": truncate_text(joined, 5000)})
-            if joined:
-                text_pool.append(joined)
-
-    references = []
-    for bibl in root.findall(".//tei:listBibl//tei:biblStruct", ns)[:80]:
-        ref_title = clean_extracted_text(bibl.findtext(".//tei:title", default="", namespaces=ns) or "")
-        ref_doi = ""
-        for idno in bibl.findall(".//tei:idno", ns):
-            if (idno.attrib.get("type") or "").lower() == "doi":
-                ref_doi = norm_text(" ".join(idno.itertext()))
-                break
-        if ref_title:
-            references.append({"title": ref_title, "doi": normalize_doi(ref_doi)})
-
-    semantic_text = truncate_text(" ".join([title, abstract] + text_pool[:6]), MAX_RAW_TEXT_CHARS)
-    return {
-        "parser": "grobid",
-        "title": title,
-        "abstract": abstract,
-        "authors": authors[:12],
-        "sections": sections[:16],
-        "references": references[:60],
-        "claims": extract_claim_like_sentences(semantic_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(semantic_text, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": "",
-        "parser_quality": "scholarly-structured",
-    }
-
-
-def parse_pdf_with_pymupdf(pdf_path):
-    if fitz is None:
-        raise RuntimeError("PyMuPDF not installed")
-
-    doc = fitz.open(pdf_path)
-    page_texts = [page.get_text("text") for page in doc]
-    raw_text = truncate_text(clean_extracted_text("\n".join(page_texts).strip()), MAX_RAW_TEXT_CHARS)
-    first_page = clean_extracted_text("\n".join(page_texts[:2]))[:5000]
-
-    title = extract_title_from_text(first_page, fallback=Path(pdf_path).name)
-
-    abstract = ""
-    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n(?:1|i)[\.\s]|\nintroduction)", raw_text, re.I | re.S)
-    if match:
-        abstract = truncate_text(clean_extracted_text(match.group(1)), 2600)
-
-    sections = []
-    blocks = re.split(r"\n(?=[A-Z][A-Za-z\s]{2,40}\n)", raw_text)
-    for block in blocks[:10]:
-        lines = [norm_text(x) for x in block.splitlines() if norm_text(x)]
-        if not lines:
-            continue
-        heading = lines[0] if len(lines[0]) < 60 else "Section"
-        body = " ".join(lines[1:] if len(lines) > 1 else lines)
-        if len(body) > 80:
-            sections.append({"heading": clean_extracted_text(heading), "text": truncate_text(body, 4200)})
-
-    return {
-        "parser": "pymupdf",
-        "title": title,
-        "abstract": abstract,
-        "authors": [],
-        "sections": sections[:12] or ([{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else []),
-        "references": extract_references_from_text(raw_text),
-        "claims": extract_claim_like_sentences(raw_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(" ".join([title, abstract, raw_text[:18000]]), max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": raw_text,
-        "parser_quality": "text-fallback-cleaned",
-    }
-
-
-def parse_pdf_with_docling(pdf_path):
-    try:
-        from docling.document_converter import DocumentConverter
-    except Exception as e:
-        raise RuntimeError(f"Docling import failed: {e}")
-
-    converter = DocumentConverter()
-    result = converter.convert(pdf_path)
-    doc = result.document
-    markdown = truncate_text(clean_extracted_text(doc.export_to_markdown()), MAX_RAW_TEXT_CHARS)
-    title = extract_title_from_text(markdown, fallback=Path(pdf_path).name)
-
-    sections = []
-    chunks = re.split(r"\n(?=# )", markdown)
-    for chunk in chunks[:12]:
-        lines = [norm_text(x.lstrip('#')) for x in chunk.splitlines() if norm_text(x)]
-        if not lines:
-            continue
-        heading = lines[0][:80]
-        body = " ".join(lines[1:])
-        if body:
-            sections.append({"heading": heading, "text": truncate_text(body, 4200)})
-
-    return {
-        "parser": "docling",
-        "title": title,
-        "abstract": "",
-        "authors": [],
-        "sections": sections[:12] or ([{"heading": "Document", "text": markdown[:12000]}] if markdown else []),
-        "references": extract_references_from_text(markdown),
-        "claims": extract_claim_like_sentences(markdown, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(markdown, max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": markdown,
-        "parser_quality": "layout-aware-cleaned",
-    }
-
-
-def parse_uploaded_pdf(file_obj, parser_order):
-    if not file_obj:
-        return "### PDF parse status\n\nNo PDF uploaded yet.", {}
-
-    path = getattr(file_obj, "name", None) or str(file_obj)
-    parser_order = ensure_list(parser_order) or PDF_PARSERS
-    errors = []
-
-    for parser_name in parser_order:
-        try:
-            if parser_name == "grobid":
-                result = parse_pdf_with_grobid(path)
-            elif parser_name == "docling":
-                result = parse_pdf_with_docling(path)
-            elif parser_name == "pymupdf":
-                result = parse_pdf_with_pymupdf(path)
-            else:
-                continue
-
-            summary = (
-                f"### PDF parse status\n\n"
-                f"- Parser used: {result['parser']}\n"
-                f"- Parser quality: {result.get('parser_quality', 'unknown')}\n"
-                f"- Title: {result.get('title') or 'Unknown'}\n"
-                f"- Authors: {', '.join(result.get('authors')[:6]) if result.get('authors') else 'Unknown'}\n"
-                f"- Abstract found: {'Yes' if result.get('abstract') else 'No'}\n"
-                f"- Sections extracted: {len(result.get('sections') or [])}\n"
-                f"- References extracted: {len(result.get('references') or [])}\n"
-                f"- Concepts extracted: {len(result.get('concepts') or [])}\n"
-                f"- Claims extracted: {len(result.get('claims') or [])}\n"
-            )
-            return summary, result
-        except Exception as e:
-            errors.append(f"{parser_name}: {e}")
-
-    fail_summary = "### PDF parse status\n\n" + "\n".join([f"- {x}" for x in errors])
-    return fail_summary, {"parser": None, "errors": errors}
-
-
-def render_parse_result(parsed):
-    if not parsed or not isinstance(parsed, dict) or (not parsed.get("title") and not parsed.get("sections")):
-        return '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
-
-    sections_html = []
-    for section in parsed.get("sections", [])[:8]:
-        sections_html.append(
-            f"""
-            <details class="agent-step">
-              <summary class="agent-summary">
-                <div class="agent-index">§</div>
-                <div class="agent-head">
-                  <h4>{safe_text(section.get('heading', 'Section'))}</h4>
-                  <span>section</span>
-                </div>
-              </summary>
-              <div class="agent-copy">
-                <p>{safe_text(section.get('text', '')[:2200])}</p>
-              </div>
-            </details>
-            """
-        )
-
-    refs = parsed.get("references", [])[:14]
-    refs_html = "".join(
-        f"<li>{safe_text(r.get('title') or 'Untitled')} {'· DOI ' + safe_text(r.get('doi')) if r.get('doi') else ''}</li>"
-        for r in refs
-    ) or "<li>No references extracted.</li>"
-
-    concepts = parsed.get("concepts", [])[:12]
-    claims = parsed.get("claims", [])[:8]
-    concepts_html = "".join(f"<li>{safe_text(x)}</li>" for x in concepts) or "<li>No concepts extracted.</li>"
-    claims_html = "".join(f"<li>{safe_text(x)}</li>" for x in claims) or "<li>No claims extracted.</li>"
-
-    title = safe_text(parsed.get("title") or "Parsed document")
-    abstract = safe_text((parsed.get("abstract") or "")[:2600]) or "No abstract extracted."
-    parser_name = safe_text(parsed.get("parser") or "unknown")
-    parser_quality = safe_text(parsed.get("parser_quality") or "unknown")
-
-    return f"""
-    <div class="panel" style="padding:18px">
-      <div class="brain-header">
-        <div>
-          <p class="eyebrow">PDF Parse</p>
-          <h3>{title}</h3>
-        </div>
-        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name} · {parser_quality}</span></div>
-      </div>
-      <div class="parse-grid">
-        <div class="parse-card">
-          <h4>Abstract</h4>
-          <p>{abstract}</p>
-        </div>
-        <div class="parse-card">
-          <h4>References</h4>
-          <ul class="ref-list">{refs_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Concepts</h4>
-          <ul class="ref-list">{concepts_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Claims</h4>
-          <ul class="ref-list">{claims_html}</ul>
-        </div>
-      </div>
-      <div class="timeline" style="margin-top:14px; max-height:560px; overflow:auto;">
-        {''.join(sections_html) if sections_html else '<div class="panel" style="padding:16px;"><p>No sections extracted.</p></div>'}
-      </div>
-    </div>
-    """
-
-
-def add_node(nodes_by_id: Dict[str, Dict], node_id: str, node_type: str, label: str = "", **attrs):
-    if not node_id:
-        return
-    current = nodes_by_id.get(node_id, {})
-    merged = {"id": node_id, "type": node_type, "label": label or current.get("label", node_id)}
-    merged.update(current)
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            merged[key] = value
-    nodes_by_id[node_id] = merged
-
-
-def add_edge(edges: List[Dict], source: str, target: str, edge_type: str, **attrs):
-    if not source or not target or source == target:
-        return
-    edge = {"source": source, "target": target, "type": edge_type}
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            edge[key] = value
-    edges.append(edge)
-
-
-def build_ingest_payload(query, selected_papers, parsed_pdf=None, frontier=None):
-    nodes_by_id = {}
-    edges = []
-
-    topic_id = "topic:query"
-    add_node(nodes_by_id, topic_id, "Topic", label=query or "Research topic", query=query or "")
-
-    for i, paper in enumerate(selected_papers, start=1):
-        paper_id = normalize_doi(paper.get("doi")) or (paper.get("external_ids") or {}).get("arxiv") or f"paper:{i}:{slugify(paper.get('title', 'paper'))[:40]}"
-        add_node(
-            nodes_by_id,
-            paper_id,
-            "Paper",
-            label=paper.get("title") or f"Paper {i}",
-            title=paper.get("title"),
-            year=paper.get("year"),
-            venue=paper.get("venue"),
-            doi=normalize_doi(paper.get("doi")),
-            source=paper.get("source"),
-            url=paper.get("url"),
-            pdf=paper.get("pdf"),
-            score=paper.get("score"),
-            learned_score=paper.get("learned_score", paper.get("score")),
-            open_access=paper.get("open_access"),
-            authors_text=paper.get("authors_text"),
-        )
-        add_edge(edges, topic_id, paper_id, "ABOUT", weight=paper.get("learned_score", paper.get("score", 0)))
-
-        for author in paper.get("authors", [])[:6]:
-            author_id = f"author:{slugify(author)[:64]}"
-            add_node(nodes_by_id, author_id, "Author", label=author, name=author)
-            add_edge(edges, paper_id, author_id, "WRITTEN_BY")
-
-        for concept in (paper.get("concepts") or [])[:8]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, paper_id, concept_id, "MENTIONS")
-
-        for claim in (paper.get("claims") or [])[:4]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:140], text=claim)
-            add_edge(edges, paper_id, claim_id, "ASSERTS")
-
-    if parsed_pdf and isinstance(parsed_pdf, dict) and parsed_pdf.get("title"):
-        doc_id = "upload:pdf"
-        add_node(nodes_by_id, doc_id, "UploadedPDF", label=parsed_pdf.get("title"), title=parsed_pdf.get("title"), parser=parsed_pdf.get("parser"))
-        add_edge(edges, topic_id, doc_id, "UPLOADED_SOURCE")
-
-        for concept in (parsed_pdf.get("concepts") or [])[:8]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, doc_id, concept_id, "MENTIONS")
-
-        for claim in (parsed_pdf.get("claims") or [])[:6]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:140], text=claim)
-            add_edge(edges, doc_id, claim_id, "ASSERTS")
-
-        for idx, ref in enumerate(parsed_pdf.get("references", [])[:20], start=1):
-            ref_title = ref.get("title") or f"Reference {idx}"
-            ref_doi = normalize_doi(ref.get("doi") or "")
-            ref_id = ref_doi or f"ref:{idx}:{slugify(ref_title)[:40]}"
-            add_node(nodes_by_id, ref_id, "Reference", label=ref_title, title=ref_title, doi=ref_doi)
-            add_edge(edges, doc_id, ref_id, "CITES")
-
-    for idx, item in enumerate(ensure_list(frontier)[:18], start=1):
-        fid = normalize_doi(item.get("doi")) or f"frontier:{idx}:{slugify(item.get('title', 'paper'))[:40]}"
-        add_node(
-            nodes_by_id,
-            fid,
-            "FrontierPaper",
-            label=item.get("title") or f"Frontier {idx}",
-            title=item.get("title"),
-            frontier_score=item.get("frontier_score"),
-            url=item.get("url"),
-            source=item.get("source"),
-            authors_text=item.get("authors_text"),
-            year=item.get("year"),
-            doi=item.get("doi"),
-        )
-        add_edge(edges, topic_id, fid, "FRONTIER_CANDIDATE", weight=item.get("frontier_score", item.get("learned_score", item.get("score", 0))))
-
-    return {"status": "ok", "nodes": list(nodes_by_id.values())[:GRAPH_MAX_NODES], "edges": edges[:GRAPH_MAX_EDGES]}
-
-
-def learn_from_payload(payload: Dict, query: str = "") -> Dict:
-    if not payload:
-        return GRAPH_MEMORY
-
-    GRAPH_MEMORY["queries"].append(query or "")
-    GRAPH_MEMORY["events"].append({
-        "ts": time.time(),
-        "query": query or "",
-        "nodes": len(payload.get("nodes", [])),
-        "edges": len(payload.get("edges", [])),
-    })
-    GRAPH_MEMORY["payloads"].append(payload)
-
-    for node in payload.get("nodes", []):
-        node_id = node.get("id")
-        if not node_id:
-            continue
-        GRAPH_MEMORY["nodes"][node_id] = node
-        node_type = (node.get("type") or "").lower()
-        if node_type in {"paper", "frontierpaper"}:
-            GRAPH_MEMORY["papers"][node_id] = node
-        if node_type == "concept" and node.get("label"):
-            GRAPH_MEMORY["concept_counts"][node["label"].lower()] += 1
-        if node_type == "claim" and node.get("label"):
-            GRAPH_MEMORY["claim_counts"][node["label"].lower()] += 1
-
-    GRAPH_MEMORY["edges"].extend(payload.get("edges", []))
-    GRAPH_MEMORY["edges"] = GRAPH_MEMORY["edges"][:GRAPH_MAX_EDGES]
-    return GRAPH_MEMORY
-
-
-def export_learning_state() -> str:
-    snapshot = {
-        "papers": list(GRAPH_MEMORY["papers"].values())[:60],
-        "nodes": list(GRAPH_MEMORY["nodes"].values())[:250],
-        "edges": GRAPH_MEMORY["edges"][:500],
-        "top_concepts": GRAPH_MEMORY["concept_counts"].most_common(24),
-        "top_claims": GRAPH_MEMORY["claim_counts"].most_common(24),
-        "queries": GRAPH_MEMORY["queries"][-20:],
-        "events": GRAPH_MEMORY["events"][-20:],
-        "frontier": GRAPH_MEMORY["frontier"][:24],
-    }
-    return json.dumps(snapshot, indent=2, ensure_ascii=False)
-
-
-def resolve_selected_papers(selected_indices, papers_state):
-    papers = ensure_list(papers_state)
-    selected_indices = ensure_list(selected_indices)
-    selected = []
-    if not selected_indices:
-        return selected
-
-    value_map = {paper_choice_value(i, paper): paper for i, paper in enumerate(papers)}
-    label_map = {paper_choice_label(i, paper): paper for i, paper in enumerate(papers)}
-
-    for idx in selected_indices:
-        try:
-            if isinstance(idx, int):
-                if 0 <= idx < len(papers):
-                    selected.append(papers[idx])
-                continue
-            idx_str = str(idx)
-            if idx_str in value_map:
-                selected.append(value_map[idx_str])
-                continue
-            if idx_str.isdigit():
-                num = int(idx_str)
-                if 0 <= num < len(papers):
-                    selected.append(papers[num])
-                    continue
-            if "|" in idx_str:
-                left = idx_str.split("|", 1)[0]
-                if left.isdigit():
-                    num = int(left)
-                    if 0 <= num < len(papers):
-                        selected.append(papers[num])
-                        continue
-            if idx_str in label_map:
-                selected.append(label_map[idx_str])
-                continue
-        except Exception:
-            continue
-
-    out = []
-    seen = set()
-    for paper in selected:
-        key = paper_identity_key(paper)
-        if key not in seen:
-            seen.add(key)
-            out.append(paper)
-    return out
-
-
-def summarize_learning_state(query_text, papers, selected_sources):
-    concept_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:4])
-    top_concepts = [c for c, _ in Counter([c.lower() for c in concept_pool]).most_common(8)]
-    return (
-        "### Discovery results\n\n"
-        f"- Query: {query_text}\n"
-        f"- Sources: {', '.join(selected_sources)}\n"
-        f"- Candidates found: {len(papers)}\n"
-        f"- Top learned concepts: {', '.join(top_concepts) if top_concepts else 'None'}\n"
-        "- Select papers below, or leave selection empty to auto-ingest the top papers.\n"
-    )
-
-
-def build_ingest_status_markdown(query_text: str, payload: Dict, selected: List[Dict], parsed_state: Optional[Dict], frontier: List[Dict]) -> str:
-    payload = payload or {"nodes": [], "edges": []}
-    nodes = payload.get("nodes", [])
-    edges = payload.get("edges", [])
-
-    counts = Counter((node.get("type") or "Unknown") for node in nodes)
-    node_lines = []
-    for idx, node in enumerate(nodes[:24], start=1):
-        label = node.get("label") or node.get("title") or node.get("id")
-        node_lines.append(f"- {idx}. [{node.get('type', 'Node')}] {label}")
-
-    edge_lines = []
-    for idx, edge in enumerate(edges[:30], start=1):
-        edge_lines.append(f"- {idx}. {edge.get('source')} --{edge.get('type', 'RELATES_TO')}--> {edge.get('target')}")
-
-    return "\n".join([
-        "### Graph ingest status",
-        "",
-        f"- Topic: {query_text}",
-        f"- Selected papers ingested: {len(selected)}",
-        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
-        f"- Frontier candidates added: {len(frontier)}",
-        f"- Total nodes created: {len(nodes)}",
-        f"- Total edges created: {len(edges)}",
-        f"- Node breakdown: {', '.join([f'{k}={v}' for k, v in counts.items()]) if counts else 'None'}",
-        "",
-        "### Nodes",
-        *(node_lines or ["- None"]),
-        "",
-        "### Edges",
-        *(edge_lines or ["- None"]),
-    ])
-
-
-def run_paper_discovery(query, search_mode, sources, pdf_file):
-    query_text = norm_text(query or "")
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
-
-    if not query_text and not pdf_file:
-        empty_graph = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        return (
-            empty_graph,
-            '<div class="panel papers-panel" style="padding:18px"><p>Enter a topic, title, DOI, link, or upload a PDF to start learning.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            "No PDF uploaded yet.",
-            gr.update(choices=[], value=[]),
-            [],
-            "### No discovery results yet.",
-        )
-
-    if not query_text and pdf_file:
-        uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name
-        graph_nodes, graph_edges = build_learning_graph_state("", [], uploaded_name)
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges, "Uploaded PDF Waiting for Parse"),
-            '<div class="panel papers-panel" style="padding:18px"><p>No query yet. Parse the uploaded PDF or enter a research topic to begin discovery.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            "### Upload detected.\n\n- Parse the PDF to extract structure.\n- Or enter a topic to start discovery.",
-        )
-
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    try:
-        papers = discover_papers(query_text, search_mode, selected_sources, max_results=GRAPH_MAX_RESULTS)
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:6], uploaded_name)
-        graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Self-Learning Knowledge Graph")
-        papers_html = format_papers_html(papers)
-        journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
-        pdf_summary = uploaded_pdf_summary(pdf_file)
-        choices = format_selection_choices(papers)
-        status_md = summarize_learning_state(query_text, papers, selected_sources)
-        return (
-            graph_html,
-            papers_html,
-            journals_html,
-            pdf_summary,
-            gr.update(choices=choices, value=[]),
-            papers,
-            status_md,
-        )
-    except Exception as e:
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, [], uploaded_name)
-        error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges),
-            error_html,
-            build_journal_html(query_text or "biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
-        )
-
-
-def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
-    papers = ensure_list(papers_state)
-    selected = resolve_selected_papers(selected_indices, papers)
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    if not selected and papers:
-        selected = papers[:3]
-
-    if not selected and parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title"):
-        selected = []
-
-    if not selected and not (parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title")):
-        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        empty_payload = {"status": "empty", "nodes": [], "edges": []}
-        return (
-            graph_html,
-            "### Graph ingest status\n\n- No papers were selected and no parsed PDF is available.\n- Select papers or parse an uploaded PDF first.",
-            empty_payload,
-        )
-
-    query_text = norm_text(query or "")
-    if not query_text and isinstance(parsed_state, dict):
-        query_text = parsed_state.get("title") or "Research topic"
-    if not query_text:
-        query_text = "Research topic"
-
-    selected = [enrich_paper_semantics(query_text, paper) for paper in selected]
-    frontier = frontier_expand(query_text, DEFAULT_SOURCES, selected or papers[:3], parsed_state=parsed_state if isinstance(parsed_state, dict) else None, per_query=3)
-
-    graph_nodes, graph_edges = graph_from_selected(
-        query_text,
-        selected,
-        uploaded_name,
-        parsed_state if isinstance(parsed_state, dict) else None,
-        frontier=frontier,
-    )
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
-
-    payload = build_ingest_payload(
-        query_text,
-        selected,
-        parsed_state if isinstance(parsed_state, dict) else None,
-        frontier=frontier,
-    )
-    learn_from_payload(payload, query=query_text)
-
-    status_md = build_ingest_status_markdown(
-        query_text,
-        payload,
-        selected,
-        parsed_state if isinstance(parsed_state, dict) else None,
-        frontier,
-    )
-
-    return graph_html, status_md, payload
-
-
-def autonomous_expand_into_markdown(query, payload, parsed_state=None):
-    frontier = GRAPH_MEMORY.get("frontier") or []
-    lines = [
-        "### Autonomous expansion plan",
-        "",
-        f"- Seed query: {query or 'Research topic'}",
-        f"- Current nodes: {len(payload.get('nodes', [])) if isinstance(payload, dict) else 0}",
-        f"- Current edges: {len(payload.get('edges', [])) if isinstance(payload, dict) else 0}",
-        f"- Frontier candidates: {len(frontier)}",
-    ]
-
-    proposed = propose_expansion_queries(query or "", list(GRAPH_MEMORY.get("papers", {}).values())[:8], parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    if proposed:
-        lines.extend(["", "#### Proposed next queries", ""])
-        lines.extend([f"- {q}" for q in proposed])
-
-    if frontier:
-        lines.extend(["", "#### Top frontier papers", ""])
-        for item in frontier[:8]:
-            lines.append(
-                f"- {item.get('title', 'Untitled')} ({item.get('source', 'unknown')}) — frontier score {item.get('frontier_score', item.get('learned_score', item.get('score', 0)))}"
-            )
-    return "\n".join(lines)
-
-
-__all__ = [
-    "SEARCH_MODES",
-    "SOURCE_OPTIONS",
-    "DEFAULT_SOURCES",
-    "PDF_PARSERS",
-    "GRAPH_MEMORY",
-    "discover_papers",
-    "run_paper_discovery",
-    "parse_uploaded_pdf",
-    "render_parse_result",
-    "ingest_selected_papers",
-    "build_ingest_payload",
-    "learn_from_payload",
-    "frontier_expand",
-    "autonomous_expand_into_markdown",
-    "export_learning_state",
-    "format_frontier_html",
-]

dvnc_ai_v2_hf/graph_canvas_patch.py

@@ -1,977 +0,0 @@
-import html
-import json
-import re
-from typing import Any, Dict, List
-
-
-TYPE_STYLES = {
-    "topic": {"color": "#38bdf8", "size": 12, "label": "Topic"},
-    "query": {"color": "#38bdf8", "size": 12, "label": "Topic"},
-    "paper": {"color": "#34d399", "size": 9, "label": "Paper"},
-    "uploadedpdf": {"color": "#f59e0b", "size": 10, "label": "Uploaded PDF"},
-    "upload": {"color": "#f59e0b", "size": 10, "label": "Uploaded PDF"},
-    "concept": {"color": "#a78bfa", "size": 7, "label": "Concept"},
-    "author": {"color": "#f472b6", "size": 6, "label": "Author"},
-    "reference": {"color": "#94a3b8", "size": 6, "label": "Reference"},
-    "claim": {"color": "#fb7185", "size": 7, "label": "Claim"},
-    "frontierpaper": {"color": "#fde047", "size": 8, "label": "Frontier Paper"},
-    "frontier": {"color": "#fde047", "size": 8, "label": "Frontier Paper"},
-}
-
-
-def _safe_text(x: Any, default: str = "") -> str:
-    return html.escape(str(x if x is not None else default))
-
-
-def _norm_text(x: Any) -> str:
-    return re.sub(r"\s+", " ", str(x or "")).strip()
-
-
-def _truncate(text: str, limit: int = 120) -> str:
-    text = _norm_text(text)
-    return text if len(text) <= limit else text[: limit - 1].rstrip() + "…"
-
-
-def _as_dict(payload: Any) -> Dict[str, Any]:
-    if payload is None:
-        return {"status": "empty", "nodes": [], "edges": []}
-    if isinstance(payload, dict):
-        return payload
-    if isinstance(payload, str):
-        try:
-            parsed = json.loads(payload)
-            if isinstance(parsed, dict):
-                return parsed
-        except Exception:
-            pass
-    return {"status": "empty", "nodes": [], "edges": []}
-
-
-def _style_for(kind: str) -> Dict[str, Any]:
-    key = _norm_text(kind).lower().replace(" ", "")
-    return TYPE_STYLES.get(key, {"color": "#cbd5e1", "size": 6, "label": _norm_text(kind or "Node") or "Node"})
-
-
-def _normalize_payload(payload: Any, title: str) -> Dict[str, Any]:
-    raw = _as_dict(payload)
-    raw_nodes = raw.get("nodes", []) or []
-    raw_edges = raw.get("edges", raw.get("links", [])) or []
-
-    nodes: List[Dict[str, Any]] = []
-    seen = set()
-
-    for i, node in enumerate(raw_nodes):
-        if not isinstance(node, dict):
-            continue
-        node_id = _norm_text(node.get("id") or f"node_{i}")
-        if not node_id or node_id in seen:
-            continue
-        seen.add(node_id)
-
-        kind = _norm_text(node.get("kind") or node.get("type") or "paper")
-        style = _style_for(kind)
-        label = _truncate(node.get("label") or node.get("title") or node_id, 120)
-
-        nodes.append({
-            "id": node_id,
-            "label": label,
-            "kind": kind or "paper",
-            "color": style["color"],
-            "size": float(node.get("size") or style["size"]),
-            "detail": {
-                "type": style["label"],
-                "label": label,
-                "title": _norm_text(node.get("title") or node.get("label") or node_id),
-                "venue": _norm_text(node.get("venue")),
-                "year": _norm_text(node.get("year")),
-                "doi": _norm_text(node.get("doi")),
-                "source": _norm_text(node.get("source")),
-                "authors_text": _norm_text(node.get("authors_text")),
-                "text": _norm_text(node.get("text")),
-            },
-        })
-
-    links: List[Dict[str, Any]] = []
-    for edge in raw_edges:
-        if not isinstance(edge, dict):
-            continue
-        src = _norm_text(edge.get("source"))
-        tgt = _norm_text(edge.get("target"))
-        if not src or not tgt or src == tgt:
-            continue
-        links.append({
-            "source": src,
-            "target": tgt,
-            "type": _norm_text(edge.get("type") or "RELATES_TO"),
-        })
-
-    counts: Dict[str, int] = {}
-    for n in nodes:
-        label = _style_for(n["kind"])["label"]
-        counts[label] = counts.get(label, 0) + 1
-
-    return {
-        "title": _norm_text(title) or "Self-Learning Knowledge Graph",
-        "nodes": nodes,
-        "links": links,
-        "counts": counts,
-    }
-
-
-def render_graph_canvas_html(payload: Any, title: str = "Self-Learning Knowledge Graph", height: int = 780) -> str:
-    data = _normalize_payload(payload, title=title)
-    payload_json = json.dumps(data, ensure_ascii=False).replace("</script>", "<\\/script>")
-
-    html_doc = r"""
-<!doctype html>
-<html>
-<head>
-<meta charset="utf-8"/>
-<meta name="viewport" content="width=device-width, initial-scale=1"/>
-<style>
-  :root {
-    --bg0: #030712;
-    --bg1: #0b1220;
-    --bg2: #111827;
-    --panel: rgba(8, 15, 29, 0.72);
-    --panel-border: rgba(255,255,255,0.10);
-    --text: #e5eefc;
-    --muted: #a5b4cc;
-    --accent: #38bdf8;
-  }
-  * { box-sizing: border-box; }
-  html, body {
-    margin: 0;
-    width: 100%;
-    height: 100%;
-    overflow: hidden;
-    background:
-      radial-gradient(circle at 20% 20%, rgba(56,189,248,0.16), transparent 28%),
-      radial-gradient(circle at 80% 18%, rgba(167,139,250,0.12), transparent 24%),
-      linear-gradient(180deg, var(--bg2) 0%, var(--bg1) 45%, var(--bg0) 100%);
-    color: var(--text);
-    font-family: Inter, ui-sans-serif, system-ui, -apple-system, Segoe UI, Roboto, sans-serif;
-    touch-action: none;
-  }
-  #app {
-    position: relative;
-    width: 100%;
-    height: 100%;
-    overflow: hidden;
-  }
-  #canvas {
-    position: absolute;
-    inset: 0;
-    width: 100%;
-    height: 100%;
-    display: block;
-    cursor: grab;
-  }
-  #canvas.dragging { cursor: grabbing; }
-
-  .hud {
-    position: absolute;
-    z-index: 10;
-    backdrop-filter: blur(10px);
-    background: var(--panel);
-    border: 1px solid var(--panel-border);
-    box-shadow: 0 10px 30px rgba(0,0,0,0.28);
-  }
-  .topbar {
-    top: 16px;
-    left: 16px;
-    width: min(420px, calc(100% - 32px));
-    border-radius: 16px;
-    padding: 14px 16px;
-  }
-  .topbar h3 {
-    margin: 0 0 6px;
-    font-size: 18px;
-    line-height: 1.2;
-    color: #f8fbff;
-  }
-  .topbar p {
-    margin: 0;
-    font-size: 12px;
-    line-height: 1.45;
-    color: var(--muted);
-  }
-  .legend {
-    display: flex;
-    flex-wrap: wrap;
-    gap: 8px;
-    margin-top: 10px;
-  }
-  .chip {
-    display: inline-flex;
-    align-items: center;
-    gap: 8px;
-    padding: 6px 10px;
-    border-radius: 999px;
-    font-size: 11px;
-    line-height: 1;
-    color: #dbeafe;
-    background: rgba(255,255,255,0.04);
-    border: 1px solid rgba(255,255,255,0.06);
-  }
-  .dot {
-    width: 10px;
-    height: 10px;
-    border-radius: 999px;
-    flex: 0 0 10px;
-  }
-
-  .details {
-    right: 16px;
-    bottom: 16px;
-    width: min(360px, calc(100% - 32px));
-    max-height: min(48%, 360px);
-    overflow: auto;
-    border-radius: 16px;
-    padding: 14px 16px;
-  }
-  .details h4 {
-    margin: 0 0 8px;
-    font-size: 14px;
-    color: #f8fbff;
-  }
-  .details p {
-    margin: 0;
-    font-size: 12px;
-    line-height: 1.5;
-    color: var(--muted);
-  }
-  .details dl {
-    margin: 12px 0 0;
-    display: grid;
-    grid-template-columns: 78px 1fr;
-    gap: 8px 10px;
-    font-size: 12px;
-  }
-  .details dt {
-    color: #93c5fd;
-    font-weight: 600;
-  }
-  .details dd {
-    margin: 0;
-    color: #e5eefc;
-    word-break: break-word;
-  }
-
-  .statusbar {
-    position: absolute;
-    left: 16px;
-    bottom: 16px;
-    z-index: 10;
-    display: flex;
-    gap: 8px;
-    flex-wrap: wrap;
-  }
-  .status {
-    padding: 8px 10px;
-    border-radius: 12px;
-    font-size: 11px;
-    color: #dbeafe;
-    background: rgba(8, 15, 29, 0.66);
-    border: 1px solid rgba(255,255,255,0.08);
-    backdrop-filter: blur(8px);
-  }
-
-  .empty {
-    position: absolute;
-    inset: 0;
-    display: none;
-    align-items: center;
-    justify-content: center;
-    z-index: 5;
-    pointer-events: none;
-  }
-  .empty.show { display: flex; }
-  .empty-card {
-    width: min(520px, calc(100% - 32px));
-    text-align: center;
-    padding: 24px 28px;
-    border-radius: 20px;
-    background: rgba(8, 15, 29, 0.66);
-    border: 1px solid rgba(255,255,255,0.08);
-    backdrop-filter: blur(10px);
-    box-shadow: 0 16px 40px rgba(0,0,0,0.34);
-  }
-  .empty-card h3 {
-    margin: 0 0 10px;
-    font-size: 22px;
-  }
-  .empty-card p {
-    margin: 0;
-    color: var(--muted);
-    line-height: 1.55;
-  }
-
-  @media (max-width: 700px) {
-    .topbar {
-      width: calc(100% - 20px);
-      left: 10px;
-      top: 10px;
-      padding: 12px;
-    }
-    .details {
-      width: calc(100% - 20px);
-      right: 10px;
-      bottom: 10px;
-      max-height: 34%;
-      padding: 12px;
-    }
-    .statusbar {
-      left: 10px;
-      bottom: calc(34% + 22px);
-      right: 10px;
-    }
-  }
-</style>
-</head>
-<body>
-<div id="app">
-  <canvas id="canvas"></canvas>
-
-  <div class="hud topbar">
-    <h3 id="title"></h3>
-    <p>Drag background to rotate, mouse wheel or trackpad pinch to zoom, two-finger pinch on touch screens, Shift-drag or right-drag to pan, and drag a node to move it in 3D space.</p>
-    <div class="legend" id="legend"></div>
-  </div>
-
-  <div class="hud details" id="details">
-    <h4>Node details</h4>
-    <p>Tap or click a node to inspect it.</p>
-  </div>
-
-  <div class="statusbar" id="statusbar"></div>
-
-  <div class="empty" id="empty">
-    <div class="empty-card">
-      <h3>No graph data yet</h3>
-      <p>Ingest papers or pass a payload containing nodes and edges. This renderer accepts the existing graph payload shape directly.</p>
-    </div>
-  </div>
-</div>
-
-<script>
-const DATA = __DATA__;
-
-const canvas = document.getElementById("canvas");
-const ctx = canvas.getContext("2d", { alpha: true });
-const titleEl = document.getElementById("title");
-const detailsEl = document.getElementById("details");
-const legendEl = document.getElementById("legend");
-const statusbarEl = document.getElementById("statusbar");
-const emptyEl = document.getElementById("empty");
-
-titleEl.textContent = DATA.title || "Self-Learning Knowledge Graph";
-
-const legendMap = [
-  ["Topic", "#38bdf8"],
-  ["Paper", "#34d399"],
-  ["Uploaded PDF", "#f59e0b"],
-  ["Concept", "#a78bfa"],
-  ["Author", "#f472b6"],
-  ["Reference", "#94a3b8"],
-  ["Claim", "#fb7185"],
-  ["Frontier Paper", "#fde047"]
-];
-
-legendMap.forEach(([label, color]) => {
-  const chip = document.createElement("div");
-  chip.className = "chip";
-  chip.innerHTML = `<span class="dot" style="background:${color}"></span>${label}`;
-  legendEl.appendChild(chip);
-});
-
-function addStatus(text) {
-  const el = document.createElement("div");
-  el.className = "status";
-  el.textContent = text;
-  statusbarEl.appendChild(el);
-}
-
-addStatus(`Nodes: ${DATA.nodes.length}`);
-addStatus(`Edges: ${DATA.links.length}`);
-
-Object.entries(DATA.counts || {}).forEach(([k, v]) => addStatus(`${k}: ${v}`));
-
-if (!DATA.nodes.length) {
-  emptyEl.classList.add("show");
-}
-
-function hashCode(str) {
-  let h = 0;
-  for (let i = 0; i < str.length; i++) {
-    h = ((h << 5) - h) + str.charCodeAt(i);
-    h |= 0;
-  }
-  return Math.abs(h);
-}
-
-function seededFloat(key, salt) {
-  const h = hashCode(key + "|" + salt);
-  return (h % 10000) / 10000;
-}
-
-const nodes = DATA.nodes.map((n, i) => {
-  const r1 = seededFloat(n.id, "x");
-  const r2 = seededFloat(n.id, "y");
-  const r3 = seededFloat(n.id, "z");
-  const radius = 120 + (i % 7) * 18 + (n.size || 6) * 2;
-  const theta = r1 * Math.PI * 2;
-  const phi = (r2 - 0.5) * Math.PI;
-
-  return {
-    ...n,
-    x: Math.cos(theta) * Math.cos(phi) * radius,
-    y: Math.sin(phi) * radius * 0.72,
-    z: Math.sin(theta) * Math.cos(phi) * radius,
-    vx: 0,
-    vy: 0,
-    vz: 0,
-    fx: null,
-    fy: null,
-    fz: null,
-    screenX: 0,
-    screenY: 0,
-    screenR: 0,
-    depth: 0
-  };
-});
-
-const nodeById = new Map(nodes.map(n => [n.id, n]));
-
-const links = DATA.links
-  .map(l => ({
-    ...l,
-    sourceRef: nodeById.get(l.source),
-    targetRef: nodeById.get(l.target)
-  }))
-  .filter(l => l.sourceRef && l.targetRef);
-
-const state = {
-  width: 0,
-  height: 0,
-  dpr: Math.max(1, Math.min(2, window.devicePixelRatio || 1)),
-  cx: 0,
-  cy: 0,
-  yaw: 0.65,
-  pitch: 0.25,
-  zoom: 1.0,
-  panX: 0,
-  panY: 0,
-  perspective: 940,
-  cameraZ: 860,
-  hovered: null,
-  selected: null,
-  draggingNode: null,
-  draggingBackground: false,
-  panMode: false,
-  lastX: 0,
-  lastY: 0,
-  raf: null,
-  lastTouchDist: 0,
-  lastTouchCenter: null
-};
-
-function resize() {
-  const rect = canvas.getBoundingClientRect();
-  state.width = rect.width;
-  state.height = rect.height;
-  state.cx = rect.width / 2;
-  state.cy = rect.height / 2;
-  canvas.width = Math.floor(rect.width * state.dpr);
-  canvas.height = Math.floor(rect.height * state.dpr);
-  ctx.setTransform(state.dpr, 0, 0, state.dpr, 0, 0);
-}
-resize();
-window.addEventListener("resize", resize);
-
-function clamp(v, lo, hi) {
-  return Math.max(lo, Math.min(hi, v));
-}
-
-function rotatePoint(x, y, z) {
-  const cy = Math.cos(state.yaw);
-  const sy = Math.sin(state.yaw);
-  const cp = Math.cos(state.pitch);
-  const sp = Math.sin(state.pitch);
-
-  const x1 = x * cy - z * sy;
-  const z1 = x * sy + z * cy;
-
-  const y2 = y * cp - z1 * sp;
-  const z2 = y * sp + z1 * cp;
-
-  return { x: x1, y: y2, z: z2 };
-}
-
-function projectNode(n) {
-  const p = rotatePoint(n.x, n.y, n.z);
-  const depth = state.cameraZ - p.z;
-  const scale = (state.perspective / Math.max(120, depth)) * state.zoom;
-  n.depth = p.z;
-  n.screenX = state.cx + state.panX + p.x * scale;
-  n.screenY = state.cy + state.panY + p.y * scale;
-  n.screenR = clamp((n.size || 6) * scale * 0.85, 4, 22);
-  n._scale = scale;
-}
-
-function setDetails(n) {
-  if (!n) {
-    detailsEl.innerHTML = `
-      <h4>Node details</h4>
-      <p>Tap or click a node to inspect it.</p>
-    `;
-    return;
-  }
-  const d = n.detail || {};
-  detailsEl.innerHTML = `
-    <h4>Node details</h4>
-    <dl>
-      <dt>Label</dt><dd>${escapeHtml(n.label || "")}</dd>
-      <dt>Type</dt><dd>${escapeHtml(d.type || n.kind || "")}</dd>
-      <dt>Title</dt><dd>${escapeHtml(d.title || "—")}</dd>
-      <dt>Venue</dt><dd>${escapeHtml(d.venue || "—")}</dd>
-      <dt>Year</dt><dd>${escapeHtml(d.year || "—")}</dd>
-      <dt>DOI</dt><dd>${escapeHtml(d.doi || "—")}</dd>
-      <dt>Source</dt><dd>${escapeHtml(d.source || "—")}</dd>
-      <dt>Authors</dt><dd>${escapeHtml(d.authors_text || "—")}</dd>
-      <dt>Text</dt><dd>${escapeHtml(d.text || "—")}</dd>
-    </dl>
-  `;
-}
-
-function escapeHtml(str) {
-  return String(str || "")
-    .replaceAll("&", "&amp;")
-    .replaceAll("<", "&lt;")
-    .replaceAll(">", "&gt;")
-    .replaceAll('"', "&quot;")
-    .replaceAll("'", "&#039;");
-}
-
-function findNodeAt(x, y) {
-  const ordered = [...nodes].sort((a, b) => b.depth - a.depth);
-  for (const n of ordered) {
-    const dx = x - n.screenX;
-    const dy = y - n.screenY;
-    const rr = n.screenR + 6;
-    if (dx * dx + dy * dy <= rr * rr) {
-      return n;
-    }
-  }
-  return null;
-}
-
-function worldDeltaFromScreen(dx, dy, scale) {
-  const localX = dx / Math.max(0.2, scale);
-  const localY = dy / Math.max(0.2, scale);
-
-  const cp = Math.cos(-state.pitch);
-  const sp = Math.sin(-state.pitch);
-  const cy = Math.cos(-state.yaw);
-  const sy = Math.sin(-state.yaw);
-
-  let x1 = localX;
-  let y1 = localY * cp;
-  let z1 = localY * sp;
-
-  const wx = x1 * cy - z1 * sy;
-  const wz = x1 * sy + z1 * cy;
-
-  return { x: wx, y: y1, z: wz };
-}
-
-function stepPhysics() {
-  if (!nodes.length) return;
-
-  const nCount = nodes.length;
-  const pairStride = nCount > 120 ? 2 : 1;
-  const repulsion = 15000;
-  const spring = 0.0035;
-  const centering = 0.0008;
-  const damping = 0.90;
-
-  for (const n of nodes) {
-    n.ax = 0;
-    n.ay = 0;
-    n.az = 0;
-  }
-
-  for (let i = 0; i < nCount; i += pairStride) {
-    const a = nodes[i];
-    for (let j = i + 1; j < nCount; j += pairStride) {
-      const b = nodes[j];
-      const dx = a.x - b.x;
-      const dy = a.y - b.y;
-      const dz = a.z - b.z;
-      const d2 = dx * dx + dy * dy + dz * dz + 0.01;
-      const f = repulsion / d2;
-      const inv = 1 / Math.sqrt(d2);
-      const fx = dx * inv * f;
-      const fy = dy * inv * f;
-      const fz = dz * inv * f;
-      a.ax += fx; a.ay += fy; a.az += fz;
-      b.ax -= fx; b.ay -= fy; b.az -= fz;
-    }
-  }
-
-  for (const l of links) {
-    const a = l.sourceRef;
-    const b = l.targetRef;
-    const dx = b.x - a.x;
-    const dy = b.y - a.y;
-    const dz = b.z - a.z;
-    const dist = Math.sqrt(dx * dx + dy * dy + dz * dz) || 1;
-    const target = 80 + ((a.size || 6) + (b.size || 6)) * 3;
-    const f = (dist - target) * spring;
-    const fx = (dx / dist) * f;
-    const fy = (dy / dist) * f;
-    const fz = (dz / dist) * f;
-    a.ax += fx; a.ay += fy; a.az += fz;
-    b.ax -= fx; b.ay -= fy; b.az -= fz;
-  }
-
-  for (const n of nodes) {
-    if (state.draggingNode === n) {
-      n.vx = 0; n.vy = 0; n.vz = 0;
-      continue;
-    }
-    n.ax += -n.x * centering;
-    n.ay += -n.y * centering;
-    n.az += -n.z * centering;
-
-    n.vx = (n.vx + n.ax) * damping;
-    n.vy = (n.vy + n.ay) * damping;
-    n.vz = (n.vz + n.az) * damping;
-
-    n.x += n.vx;
-    n.y += n.vy;
-    n.z += n.vz;
-  }
-}
-
-function drawBackground() {
-  ctx.clearRect(0, 0, state.width, state.height);
-
-  const g = ctx.createRadialGradient(state.width * 0.5, state.height * 0.46, 40, state.width * 0.5, state.height * 0.46, state.width * 0.72);
-  g.addColorStop(0, "rgba(56,189,248,0.05)");
-  g.addColorStop(0.45, "rgba(99,102,241,0.03)");
-  g.addColorStop(1, "rgba(0,0,0,0)");
-  ctx.fillStyle = g;
-  ctx.fillRect(0, 0, state.width, state.height);
-}
-
-function drawEdge(a, b, type) {
-  const depthAlpha = clamp((a._scale + b._scale) * 0.22, 0.10, 0.45);
-  const typeColor = {
-    "ABOUT": "rgba(56,189,248,ALPHA)",
-    "MENTIONS": "rgba(167,139,250,ALPHA)",
-    "WRITTEN_BY": "rgba(244,114,182,ALPHA)",
-    "CITES": "rgba(148,163,184,ALPHA)",
-    "ASSERTS": "rgba(251,113,133,ALPHA)",
-    "FRONTIER": "rgba(253,224,71,ALPHA)",
-    "FRONTIER_CANDIDATE": "rgba(253,224,71,ALPHA)",
-    "UPLOADED_SOURCE": "rgba(245,158,11,ALPHA)"
-  };
-  const c = (typeColor[type] || "rgba(203,213,225,ALPHA)").replace("ALPHA", depthAlpha.toFixed(3));
-
-  ctx.beginPath();
-  ctx.moveTo(a.screenX, a.screenY);
-  ctx.lineTo(b.screenX, b.screenY);
-  ctx.lineWidth = type === "ABOUT" || type === "UPLOADED_SOURCE" ? 2 : 1.2;
-  ctx.strokeStyle = c;
-  ctx.stroke();
-}
-
-function drawNode(n, selected, hovered) {
-  const r = n.screenR;
-  ctx.beginPath();
-  ctx.arc(n.screenX, n.screenY, r + 8, 0, Math.PI * 2);
-  const glow = ctx.createRadialGradient(n.screenX, n.screenY, 0, n.screenX, n.screenY, r + 8);
-  glow.addColorStop(0, hexToRgba(n.color, hovered || selected ? 0.28 : 0.18));
-  glow.addColorStop(1, hexToRgba(n.color, 0));
-  ctx.fillStyle = glow;
-  ctx.fill();
-
-  ctx.beginPath();
-  ctx.arc(n.screenX, n.screenY, r, 0, Math.PI * 2);
-  ctx.fillStyle = n.color;
-  ctx.fill();
-
-  ctx.lineWidth = selected ? 2.6 : hovered ? 2.0 : 1.0;
-  ctx.strokeStyle = selected ? "rgba(255,255,255,0.95)" : hovered ? "rgba(255,255,255,0.75)" : "rgba(255,255,255,0.22)";
-  ctx.stroke();
-
-  if (selected || hovered || r > 10) {
-    drawLabel(n, selected || hovered);
-  }
-}
-
-function drawLabel(n, strong) {
-  const text = n.label || n.id;
-  ctx.font = `${strong ? 700 : 600} 12px Inter, sans-serif`;
-  const padX = 8;
-  const h = 24;
-  const w = Math.min(260, ctx.measureText(text).width + padX * 2);
-  const x = n.screenX + 10;
-  const y = n.screenY - 10 - h;
-
-  roundRect(x, y, w, h, 10);
-  ctx.fillStyle = strong ? "rgba(8,15,29,0.92)" : "rgba(8,15,29,0.72)";
-  ctx.fill();
-
-  ctx.lineWidth = 1;
-  ctx.strokeStyle = "rgba(255,255,255,0.08)";
-  ctx.stroke();
-
-  ctx.fillStyle = "#f8fbff";
-  ctx.fillText(text.length > 32 ? text.slice(0, 31) + "…" : text, x + padX, y + 16);
-}
-
-function roundRect(x, y, w, h, r) {
-  ctx.beginPath();
-  ctx.moveTo(x + r, y);
-  ctx.arcTo(x + w, y, x + w, y + h, r);
-  ctx.arcTo(x + w, y + h, x, y + h, r);
-  ctx.arcTo(x, y + h, x, y, r);
-  ctx.arcTo(x, y, x + w, y, r);
-  ctx.closePath();
-}
-
-function hexToRgba(hex, alpha) {
-  const h = hex.replace("#", "");
-  const bigint = parseInt(h.length === 3 ? h.split("").map(c => c + c).join("") : h, 16);
-  const r = (bigint >> 16) & 255;
-  const g = (bigint >> 8) & 255;
-  const b = bigint & 255;
-  return `rgba(${r}, ${g}, ${b}, ${alpha})`;
-}
-
-function render() {
-  drawBackground();
-
-  for (const n of nodes) {
-    projectNode(n);
-  }
-
-  const orderedLinks = [...links].sort((a, b) => {
-    const da = (a.sourceRef.depth + a.targetRef.depth) * 0.5;
-    const db = (b.sourceRef.depth + b.targetRef.depth) * 0.5;
-    return da - db;
-  });
-
-  for (const l of orderedLinks) {
-    drawEdge(l.sourceRef, l.targetRef, l.type);
-  }
-
-  const orderedNodes = [...nodes].sort((a, b) => a.depth - b.depth);
-  for (const n of orderedNodes) {
-    drawNode(n, state.selected === n, state.hovered === n);
-  }
-}
-
-function animate() {
-  stepPhysics();
-  render();
-  state.raf = requestAnimationFrame(animate);
-}
-animate();
-
-function pointerPos(evt) {
-  const rect = canvas.getBoundingClientRect();
-  return { x: evt.clientX - rect.left, y: evt.clientY - rect.top };
-}
-
-canvas.addEventListener("contextmenu", (e) => e.preventDefault());
-
-canvas.addEventListener("mousedown", (e) => {
-  const p = pointerPos(e);
-  state.lastX = p.x;
-  state.lastY = p.y;
-  state.hovered = findNodeAt(p.x, p.y);
-  state.panMode = e.button === 2 || e.shiftKey;
-
-  if (state.hovered && e.button !== 2) {
-    state.draggingNode = state.hovered;
-    state.selected = state.hovered;
-    setDetails(state.selected);
-  } else {
-    state.draggingBackground = true;
-  }
-
-  canvas.classList.add("dragging");
-});
-
-window.addEventListener("mousemove", (e) => {
-  const rect = canvas.getBoundingClientRect();
-  const inside = e.clientX >= rect.left && e.clientX <= rect.right && e.clientY >= rect.top && e.clientY <= rect.bottom;
-  if (!inside && !state.draggingNode && !state.draggingBackground) return;
-
-  const p = pointerPos(e);
-  const dx = p.x - state.lastX;
-  const dy = p.y - state.lastY;
-  state.lastX = p.x;
-  state.lastY = p.y;
-
-  if (state.draggingNode) {
-    const delta = worldDeltaFromScreen(dx, dy, state.draggingNode._scale || 1);
-    state.draggingNode.x += delta.x * 1.18;
-    state.draggingNode.y += delta.y * 1.18;
-    state.draggingNode.z += delta.z * 1.18;
-    state.draggingNode.vx = 0;
-    state.draggingNode.vy = 0;
-    state.draggingNode.vz = 0;
-    return;
-  }
-
-  if (state.draggingBackground) {
-    if (state.panMode) {
-      state.panX += dx;
-      state.panY += dy;
-    } else {
-      state.yaw += dx * 0.0085;
-      state.pitch = clamp(state.pitch + dy * 0.0065, -1.25, 1.25);
-    }
-    return;
-  }
-
-  state.hovered = findNodeAt(p.x, p.y);
-});
-
-window.addEventListener("mouseup", () => {
-  state.draggingNode = null;
-  state.draggingBackground = false;
-  state.panMode = false;
-  canvas.classList.remove("dragging");
-});
-
-canvas.addEventListener("wheel", (e) => {
-  e.preventDefault();
-  const delta = e.deltaY > 0 ? 0.92 : 1.09;
-  state.zoom = clamp(state.zoom * delta, 0.35, 4.0);
-}, { passive: false });
-
-canvas.addEventListener("click", (e) => {
-  const p = pointerPos(e);
-  const hit = findNodeAt(p.x, p.y);
-  if (hit) {
-    state.selected = hit;
-    setDetails(hit);
-  }
-});
-
-canvas.addEventListener("touchstart", (e) => {
-  e.preventDefault();
-  if (e.touches.length === 1) {
-    const t = e.touches[0];
-    const p = pointerPos(t);
-    state.lastX = p.x;
-    state.lastY = p.y;
-    state.hovered = findNodeAt(p.x, p.y);
-    if (state.hovered) {
-      state.draggingNode = state.hovered;
-      state.selected = state.hovered;
-      setDetails(state.selected);
-    } else {
-      state.draggingBackground = true;
-    }
-  } else if (e.touches.length === 2) {
-    state.draggingNode = null;
-    state.draggingBackground = false;
-    const a = e.touches[0];
-    const b = e.touches[1];
-    state.lastTouchDist = Math.hypot(a.clientX - b.clientX, a.clientY - b.clientY);
-    state.lastTouchCenter = {
-      x: (a.clientX + b.clientX) * 0.5,
-      y: (a.clientY + b.clientY) * 0.5
-    };
-  }
-}, { passive: false });
-
-canvas.addEventListener("touchmove", (e) => {
-  e.preventDefault();
-  if (e.touches.length === 1) {
-    const t = e.touches[0];
-    const p = pointerPos(t);
-    const dx = p.x - state.lastX;
-    const dy = p.y - state.lastY;
-    state.lastX = p.x;
-    state.lastY = p.y;
-
-    if (state.draggingNode) {
-      const delta = worldDeltaFromScreen(dx, dy, state.draggingNode._scale || 1);
-      state.draggingNode.x += delta.x * 1.18;
-      state.draggingNode.y += delta.y * 1.18;
-      state.draggingNode.z += delta.z * 1.18;
-      state.draggingNode.vx = 0;
-      state.draggingNode.vy = 0;
-      state.draggingNode.vz = 0;
-    } else if (state.draggingBackground) {
-      state.yaw += dx * 0.0085;
-      state.pitch = clamp(state.pitch + dy * 0.0065, -1.25, 1.25);
-    }
-  } else if (e.touches.length === 2) {
-    const a = e.touches[0];
-    const b = e.touches[1];
-    const dist = Math.hypot(a.clientX - b.clientX, a.clientY - b.clientY);
-    const center = {
-      x: (a.clientX + b.clientX) * 0.5,
-      y: (a.clientY + b.clientY) * 0.5
-    };
-
-    if (state.lastTouchDist) {
-      const ratio = dist / state.lastTouchDist;
-      state.zoom = clamp(state.zoom * ratio, 0.35, 4.0);
-    }
-
-    if (state.lastTouchCenter) {
-      state.panX += center.x - state.lastTouchCenter.x;
-      state.panY += center.y - state.lastTouchCenter.y;
-    }
-
-    state.lastTouchDist = dist;
-    state.lastTouchCenter = center;
-  }
-}, { passive: false });
-
-canvas.addEventListener("touchend", (e) => {
-  e.preventDefault();
-  if (e.touches.length === 0) {
-    state.draggingNode = null;
-    state.draggingBackground = false;
-    state.lastTouchDist = 0;
-    state.lastTouchCenter = null;
-  }
-}, { passive: false });
-
-setDetails(null);
-</script>
-</body>
-</html>
-"""
-
-    html_doc = html_doc.replace("__DATA__", payload_json)
-    return (
-        f'<iframe title="{_safe_text(title)}" '
-        f'style="width:100%; height:{int(height)}px; border:0; border-radius:18px; overflow:hidden; background:#030712;" '
-        f'sandbox="allow-scripts allow-same-origin" '
-        f'srcdoc="{html.escape(html_doc, quote=True)}"></iframe>'
-    )
-
-
-def render_graph_canvas_component(payload: Any, title: str = "Self-Learning Knowledge Graph", height: int = 780):
-    try:
-        import gradio as gr
-    except Exception as e:
-        raise RuntimeError(f"gradio is required for render_graph_canvas_component: {e}")
-    return gr.HTML(render_graph_canvas_html(payload, title=title, height=height))
-
-
-__all__ = [
-    "render_graph_canvas_html",
-    "render_graph_canvas_component",
-]

dvnc_ai_v2_hf/self_learning_graph.py

@@ -1,12 +1,11 @@
 import html
-import json
+import os
 import re
-import time
 import urllib.parse
 import xml.etree.ElementTree as ET
-from collections import Counter
 from pathlib import Path
-from typing import Any, Dict, List, Optional
+from typing import Dict, List, Optional
+from urllib.parse import quote
 
 import gradio as gr
 import requests
@@ -16,31 +15,11 @@ try:
 except Exception:
     fitz = None
 
-from dvnc_ai_v2_hf.graph_canvas_patch import render_graph_canvas_html
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    payload = {
-        "status": "ok" if (nodes or edges) else "empty",
-        "nodes": nodes or [],
-        "edges": edges or [],
-    }
-    return render_graph_canvas_html(payload, title=title, height=780)
-
-SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
-SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-DEFAULT_SOURCES = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
-PDF_PARSERS = ["pymupdf"]
+try:
+    from bs4 import BeautifulSoup
+except Exception:
+    BeautifulSoup = None
 
-REQUEST_TIMEOUT = 25
-GRAPH_MAX_RESULTS = 12
-GRAPH_MAX_CONCEPTS = 12
-GRAPH_MAX_CLAIMS = 8
-GRAPH_MAX_EXPANSIONS = 6
-GRAPH_MAX_NODES = 500
-GRAPH_MAX_EDGES = 1600
-GRAPH_IFRAME_HEIGHT = 760
-MAX_ABSTRACT_CHARS = 4000
-MAX_RAW_TEXT_CHARS = 60000
 
 JOURNALS = [
     {"name": "Nature", "url": "https://www.nature.com/search", "desc": "Flagship multidisciplinary research journal."},
@@ -50,34 +29,14 @@ JOURNALS = [
     {"name": "IEEE Xplore", "url": "https://ieeexplore.ieee.org/search/searchresult.jsp", "desc": "Engineering, AI, signal processing, and systems."},
 ]
 
-STOPWORDS = {
-    "a", "an", "and", "are", "as", "at", "be", "been", "being", "by", "can", "could",
-    "did", "do", "does", "for", "from", "had", "has", "have", "if", "in", "into", "is",
-    "it", "its", "may", "might", "of", "on", "or", "our", "such", "that", "the", "their",
-    "there", "these", "this", "those", "to", "using", "use", "used", "via", "was", "were",
-    "will", "with", "within", "without", "we", "they", "you", "your", "study", "paper",
-    "research", "results", "result", "method", "methods", "analysis", "approach", "based",
-    "new", "novel", "effect", "effects", "model", "models", "system", "systems", "show",
-    "shows", "shown", "however", "therefore", "also", "between", "across", "among", "et", "al",
-    "introduction", "discussion", "conclusion", "references", "figure", "table",
-}
-
-GRAPH_MEMORY: Dict[str, Any] = {
-    "papers": {},
-    "nodes": {},
-    "edges": [],
-    "queries": [],
-    "events": [],
-    "frontier": [],
-    "payloads": [],
-    "concept_counts": Counter(),
-    "claim_counts": Counter(),
-}
-
-
-# ----------------------------
-# Utility
-# ----------------------------
+SEARCH_MODES = ["topic", "title", "doi", "link", "paper_name", "autonomous_web"]
+SOURCE_OPTIONS = ["arxiv", "crossref", "openalex", "semantic_scholar", "europe_pmc"]
+DEFAULT_SOURCES = ["arxiv", "openalex", "crossref", "semantic_scholar", "europe_pmc"]
+
+SEMANTIC_SCHOLAR_API_KEY = os.getenv("SEMANTIC_SCHOLAR_API_KEY", "")
+GROBID_URL = os.getenv("GROBID_URL", "").strip()
+REQUEST_TIMEOUT = 25
+
 
 def safe_text(x, default=""):
     return html.escape(str(x if x is not None else default))
@@ -87,25 +46,6 @@ def norm_text(x: Optional[str]) -> str:
     return re.sub(r"\s+", " ", (x or "")).strip()
 
 
-def ensure_list(x):
-    return x if isinstance(x, list) else []
-
-
-def truncate_text(text: str, limit: int) -> str:
-    text = norm_text(text)
-    return text if len(text) <= limit else text[: limit - 1].rstrip() + "…"
-
-
-def slugify(text: str) -> str:
-    return re.sub(r"[^a-z0-9]+", "-", (text or "").lower()).strip("-")
-
-
-def normalize_doi(text: str) -> str:
-    text = (text or "").strip()
-    text = re.sub(r"^https?://(dx\.)?doi\.org/", "", text, flags=re.I)
-    return text.strip().rstrip("/")
-
-
 def detect_query_type(query: str) -> str:
     q = (query or "").strip()
     doi_pattern = r"^10\.\d{4,9}/[-._;()/:A-Z0-9]+$"
@@ -116,262 +56,106 @@ def detect_query_type(query: str) -> str:
     return "topic"
 
 
-def tokenize(text: str) -> List[str]:
-    return [
-        t for t in re.findall(r"[A-Za-z][A-Za-z0-9\-/+]{2,}", (text or ""))
-        if t.lower() not in STOPWORDS
-    ]
-
-
-def text_overlap_score(a: str, b: str) -> float:
-    sa = {x.lower() for x in tokenize(a)}
-    sb = {x.lower() for x in tokenize(b)}
-    if not sa or not sb:
-        return 0.0
-    return len(sa & sb) / max(1, len(sa | sb))
-
-
-def compute_recency_bonus(year: str) -> float:
-    try:
-        y = int(str(year)[:4])
-    except Exception:
-        return 0.0
-    current = time.gmtime().tm_year
-    age = max(current - y, 0)
-    return max(0.0, 0.16 - age * 0.018)
-
-
-def clean_extracted_text(text: str) -> str:
-    text = text or ""
-    replacements = {
-        "\u00ad": "",
-        "\ufb01": "fi",
-        "\ufb02": "fl",
-        "\u2010": "-",
-        "\u2011": "-",
-        "\u2012": "-",
-        "\u2013": "-",
-        "\u2014": "-",
-        "\u2212": "-",
-        "\u00a0": " ",
-    }
-    for old, new in replacements.items():
-        text = text.replace(old, new)
-
-    text = re.sub(r"(\w)-\s*\n\s*(\w)", r"\1\2", text)
-    text = re.sub(r"([a-z])\n([a-z])", r"\1 \2", text)
-    text = re.sub(r"\bnjectable\b", "injectable", text, flags=re.I)
-    text = re.sub(r"\bfhydrogel\b", "hydrogel", text, flags=re.I)
-    text = re.sub(
-        r"(?i)\b([a-z])?(hydrogel|conductive|responsive|injectable|biomaterial|scaffold|cardiac|patch|repair)\b",
-        lambda m: m.group(2),
-        text,
-    )
-    text = re.sub(r"[ \t]+", " ", text)
-    text = re.sub(r"\n{3,}", "\n\n", text)
-    return text.strip()
-
-
-def paper_identity_key(paper: Dict[str, Any]) -> str:
-    return (
-        normalize_doi(paper.get("doi") or "")
-        or ((paper.get("external_ids") or {}).get("arxiv") or "")
-        or norm_text(paper.get("title", "")).lower()
-        or str(paper.get("id", ""))
-    )
-
-
-def unique_keep_order(items: List[str]) -> List[str]:
-    seen = set()
-    out = []
-    for item in items:
-        key = norm_text(item).lower()
-        if key and key not in seen:
-            seen.add(key)
-            out.append(norm_text(item))
-    return out
-
-
-# ----------------------------
-# Concept / claim extraction
-# ----------------------------
-
-def normalize_concept_label(phrase: str) -> str:
-    phrase = norm_text(phrase)
-    mapping = {"ph": "pH", "ai": "AI", "ml": "ML", "3d": "3D", "ecg": "ECG", "mri": "MRI"}
-    return " ".join(mapping.get(p.lower(), p) for p in phrase.split())
-
-
-def looks_like_bad_phrase(phrase: str) -> bool:
-    phrase = norm_text(phrase)
-    if not phrase or len(phrase) < 4 or len(phrase) > 90:
-        return True
-    parts = phrase.split()
-    if len(parts) > 6:
-        return True
-    for p in parts:
-        t = p.strip("-.,;:()[]{}")
-        if not t:
-            return True
-        if len(t) == 1 and t.lower() not in {"p", "h"}:
-            return True
-        if re.match(r"^[bcdfghjklmnpqrstvwxyz]{5,}$", t.lower()):
-            return True
-    return False
-
-
-def extract_concepts_from_text(text: str, max_terms: int = GRAPH_MAX_CONCEPTS) -> List[str]:
-    text = clean_extracted_text(text)
-    words = re.findall(r"[A-Za-z][A-Za-z0-9\-/+]{2,}", text)
-    phrases = []
-
-    for n in (3, 2, 1):
-        for i in range(len(words) - n + 1):
-            phrase = " ".join(words[i:i + n])
-            low = phrase.lower()
-            if any(tok in STOPWORDS for tok in low.split()):
-                continue
-            if looks_like_bad_phrase(low):
-                continue
-            phrases.append(low)
+def ensure_list(x):
+    return x if isinstance(x, list) else []
 
-    counts = Counter(phrases)
-    ranked = []
-    for phrase, count in counts.most_common(max_terms * 8):
-        score = float(count)
-        score += 0.25 * len(phrase.split())
-        if any(x in phrase for x in ["hydrogel", "conductive", "responsive", "injectable", "graph", "neural", "cardiac", "biomaterial"]):
-            score += 0.5
-        ranked.append((score, phrase))
 
-    out = []
-    seen = set()
-    for _, phrase in sorted(ranked, key=lambda x: x[0], reverse=True):
-        clean = normalize_concept_label(phrase)
-        low = clean.lower()
-        if low in seen:
-            continue
-        if any((low in s or s in low) and abs(len(low) - len(s)) <= 2 for s in seen):
-            continue
-        seen.add(low)
-        out.append(clean)
-        if len(out) >= max_terms:
-            break
-    return out
+def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
+    if not nodes:
+        return """
+        <div class="panel brain-shell">
+          <div class="brain-header">
+            <div>
+              <p class="eyebrow">Learning Graph</p>
+              <h3>Self-Learning Knowledge Graph</h3>
+            </div>
+          </div>
+          <div class="brain-stage learning-empty">
+            <div class="empty-graph-copy">
+              <h4>No papers mapped yet</h4>
+              <p>Search papers, pick a topic, select candidates, or upload a PDF to grow the graph in real time.</p>
+            </div>
+          </div>
+        </div>
+        """
 
+    node_items = []
+    label_items = []
+    edge_items = []
+    coords = [(110, 110), (320, 80), (540, 130), (660, 270), (550, 410), (300, 450), (110, 340), (380, 260)]
+
+    graph_nodes = [dict(n) for n in nodes[:8]]
+    for i, node in enumerate(graph_nodes):
+        x, y = coords[i]
+        node["sx"] = x
+        node["sy"] = y
+
+    node_map = {n["id"]: n for n in graph_nodes}
+    for a, b in edges:
+        if a in node_map and b in node_map:
+            na = node_map[a]
+            nb = node_map[b]
+            edge_items.append(
+                f'<line class="learn-edge" x1="{na["sx"]}" y1="{na["sy"]}" x2="{nb["sx"]}" y2="{nb["sy"]}" />'
+            )
 
-def extract_claim_like_sentences(text: str, max_items: int = GRAPH_MAX_CLAIMS) -> List[str]:
-    text = clean_extracted_text(text)
-    sentences = [norm_text(s) for s in re.split(r"(?<=[\.\!\?])\s+", text) if norm_text(s)]
-    scored = []
-    for sentence in sentences:
-        if len(sentence) < 45 or len(sentence) > 320:
-            continue
-        lower = sentence.lower()
-        score = 0.0
-        if any(k in lower for k in ["improves", "reduces", "increases", "demonstrates", "shows", "reveals", "predicts", "achieves", "outperforms", "enables", "supports"]):
-            score += 2.0
-        if any(k in lower for k in ["significant", "associated", "correlated", "effective", "robust", "accurate", "validated", "statistically"]):
-            score += 1.0
-        score += min(len(tokenize(sentence)) / 18.0, 2.0)
-        scored.append((score, sentence))
+    for node in graph_nodes:
+        kind = node.get("kind", "paper")
+        klass = f'learn-node {kind}'
+        radius = 24 if kind == "query" else 20
+        node_items.append(
+            f'<circle class="{klass}" cx="{node["sx"]}" cy="{node["sy"]}" r="{radius}" />'
+        )
+        label_items.append(
+            f'<text class="learn-label" x="{node["sx"] + 28}" y="{node["sy"] - 10}">{safe_text(node["label"][:44])}</text>'
+        )
 
-    out = []
-    seen = set()
-    for _, sentence in sorted(scored, key=lambda x: x[0], reverse=True):
-        key = sentence.lower()
-        if key not in seen:
-            seen.add(key)
-            out.append(sentence)
-        if len(out) >= max_items:
-            break
-    return out
+    return f"""
+    <div class="panel brain-shell">
+      <div class="brain-header">
+        <div>
+          <p class="eyebrow">Learning Graph</p>
+          <h3>{safe_text(title)}</h3>
+        </div>
+        <div class="brain-legend">
+          <span><i class="dot dot-query"></i> query</span>
+          <span><i class="dot dot-paper"></i> paper</span>
+          <span><i class="dot dot-upload"></i> uploaded PDF</span>
+        </div>
+      </div>
+      <div class="brain-stage">
+        <svg viewBox="0 0 760 520" class="brain-svg" role="img" aria-label="Self-learning knowledge graph">
+          {''.join(edge_items)}
+          {''.join(node_items)}
+          {''.join(label_items)}
+        </svg>
+      </div>
+    </div>
+    """
 
 
-def extract_title_from_text(raw_text: str, fallback: str = "Uploaded PDF") -> str:
-    raw_text = clean_extracted_text(raw_text or "")
-    lines = [norm_text(x) for x in raw_text.splitlines() if norm_text(x)]
-    for line in lines[:18]:
-        if 8 <= len(line) <= 240 and not re.search(r"doi|http|www\.|@", line, flags=re.I):
-            return truncate_text(line.strip(" -:;"), 240)
-    return fallback
-
-
-def extract_references_from_text(text: str) -> List[Dict[str, str]]:
-    refs = []
-    seen = set()
-    doi_re = re.compile(r"10\.\d{4,9}/[-._;()/:A-Z0-9]+", re.I)
-
-    for line in [norm_text(x) for x in clean_extracted_text(text).splitlines() if norm_text(x)][-250:]:
-        doi_match = doi_re.search(line)
-        doi = normalize_doi(doi_match.group(0)) if doi_match else ""
-        title = line.replace(doi_match.group(0), "").strip(" .;,-") if doi_match else line
-        if len(title) < 12:
-            continue
-        key = (title.lower(), doi.lower())
-        if key in seen:
-            continue
-        seen.add(key)
-        refs.append({"title": truncate_text(title, 220), "doi": doi})
-        if len(refs) >= 40:
-            break
-    return refs
-
-
-# ----------------------------
-# Search sources
-# ----------------------------
-
-def parse_openalex_abstract(inverted_index) -> str:
-    if not inverted_index or not isinstance(inverted_index, dict):
-        return ""
-    pos_to_word = {}
-    for word, positions in inverted_index.items():
-        for pos in positions:
-            pos_to_word[pos] = word
-    return " ".join(pos_to_word[i] for i in sorted(pos_to_word)) if pos_to_word else ""
-
-
-def enrich_paper_semantics(query: str, paper: Dict[str, Any]) -> Dict[str, Any]:
-    paper = dict(paper)
-    title = clean_extracted_text(paper.get("title", ""))
-    abstract = clean_extracted_text(paper.get("abstract", "") or paper.get("summary", ""))
-    venue = clean_extracted_text(paper.get("venue", ""))
-
-    concepts = extract_concepts_from_text(" ".join([title, abstract, venue]), max_terms=GRAPH_MAX_CONCEPTS)
-    claims = extract_claim_like_sentences(abstract, max_items=GRAPH_MAX_CLAIMS)
-    rel = text_overlap_score(query, f"{title} {abstract}")
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    learned_score = float(paper.get("score", 0)) + rel * 0.52 + recency + doi_bonus + oa_bonus + min(len(concepts), 8) * 0.012
-
-    paper["title"] = title or paper.get("title", "Untitled")
-    paper["abstract"] = abstract
-    paper["summary"] = truncate_text(abstract or paper.get("summary", ""), 520)
-    paper["venue"] = venue
-    paper["concepts"] = concepts[:GRAPH_MAX_CONCEPTS]
-    paper["claims"] = claims[:GRAPH_MAX_CLAIMS]
-    paper["relevance"] = round(rel, 4)
-    paper["learned_score"] = round(learned_score, 4)
-    return paper
-
-
-def dedupe_papers(items: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
-    seen = {}
-    for item in items:
-        key = paper_identity_key(item) or f"{item.get('source', 'src')}::{item.get('title', 'paper')}"
-        cur = seen.get(key)
-        cur_score = float(cur.get("learned_score", cur.get("score", 0))) if cur else -1
-        new_score = float(item.get("learned_score", item.get("score", 0)))
-        if cur is None or new_score > cur_score:
-            seen[key] = item
-    out = list(seen.values())
-    out.sort(key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-    return out
+def build_journal_html(query):
+    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
+    rows = []
+    for journal in JOURNALS:
+        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
+        if "ieeexplore" in journal["url"]:
+            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
+        rows.append(
+            f"""
+            <a class="journal-card" href="{safe_text(url)}" target="_blank" rel="noopener noreferrer">
+              <div>
+                <h4>{safe_text(journal['name'])}</h4>
+                <p>{safe_text(journal['desc'])}</p>
+              </div>
+              <span>Open</span>
+            </a>
+            """
+        )
+    return '<div class="journal-grid">' + ''.join(rows) + '</div>'
 
 
-def search_arxiv(query: str, max_results: int = 8) -> List[Dict[str, Any]]:
+def search_arxiv(query, max_results=8):
     encoded = urllib.parse.quote(query)
     url = (
         "http://export.arxiv.org/api/query?search_query=all:"
@@ -381,45 +165,45 @@ def search_arxiv(query: str, max_results: int = 8) -> List[Dict[str, Any]]:
     response.raise_for_status()
     root = ET.fromstring(response.text)
     ns = {"atom": "http://www.w3.org/2005/Atom"}
-
-    out = []
+    papers = []
     for entry in root.findall("atom:entry", ns):
-        title = clean_extracted_text(entry.findtext("atom:title", default="", namespaces=ns) or "")
-        summary = truncate_text(clean_extracted_text(entry.findtext("atom:summary", default="", namespaces=ns) or ""), MAX_ABSTRACT_CHARS)
-        published = entry.findtext("atom:published", default="", namespaces=ns) or ""
-        paper_id = entry.findtext("atom:id", default="", namespaces=ns) or ""
-        authors = [clean_extracted_text(a.findtext("atom:name", default="", namespaces=ns) or "") for a in entry.findall("atom:author", ns)]
+        title = " ".join((entry.findtext("atom:title", default="", namespaces=ns) or "").split())
+        summary = " ".join((entry.findtext("atom:summary", default="", namespaces=ns) or "").split())
+        published = entry.findtext("atom:published", default="", namespaces=ns)
+        paper_id = entry.findtext("atom:id", default="", namespaces=ns)
+        authors = [a.findtext("atom:name", default="", namespaces=ns) for a in entry.findall("atom:author", ns)]
         pdf_url = ""
         for link in entry.findall("atom:link", ns):
             if link.attrib.get("title") == "pdf":
                 pdf_url = link.attrib.get("href", "")
                 break
-
-        out.append({
-            "id": paper_id or title,
-            "title": title,
-            "summary": summary,
-            "abstract": summary,
-            "published": published[:10],
-            "authors": [a for a in authors[:8] if a],
-            "authors_text": ", ".join([a for a in authors[:4] if a]) or "Unknown authors",
-            "url": paper_id,
-            "pdf": pdf_url,
-            "doi": "",
-            "venue": "arXiv",
-            "year": published[:4] if published else "",
-            "source": "arxiv",
-            "score": 0.76,
-            "open_access": True,
-            "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
-        })
-    return out
+        papers.append(
+            {
+                "id": paper_id or title,
+                "title": title,
+                "summary": summary,
+                "abstract": summary,
+                "published": published[:10],
+                "authors": [a for a in authors[:8] if a],
+                "authors_text": ", ".join([a for a in authors[:4] if a]),
+                "url": paper_id,
+                "pdf": pdf_url,
+                "doi": "",
+                "venue": "arXiv",
+                "year": published[:4] if published else "",
+                "source": "arxiv",
+                "score": 0.76,
+                "open_access": True,
+                "external_ids": {"arxiv": (paper_id or "").split("/")[-1]},
+            }
+        )
+    return papers
 
 
-def search_crossref(query: str, mode: str = "topic", max_results: int = 8) -> List[Dict[str, Any]]:
-    headers = {"User-Agent": "dvnc-ai-space/1.0"}
+def search_crossref(query, mode="topic", max_results=8):
+    headers = {"User-Agent": "dvnc-ai-space/0.1"}
     if mode == "doi":
-        url = f"https://api.crossref.org/works/{urllib.parse.quote(normalize_doi(query))}"
+        url = f"https://api.crossref.org/works/{quote(query)}"
         response = requests.get(url, headers=headers, timeout=REQUEST_TIMEOUT)
         if response.status_code != 200:
             return []
@@ -440,29 +224,28 @@ def search_crossref(query: str, mode: str = "topic", max_results: int = 8) -> Li
         for a in item.get("author", []) or []:
             name = " ".join(filter(None, [a.get("given"), a.get("family")])).strip()
             if name:
-                authors.append(clean_extracted_text(name))
-
-        title = clean_extracted_text((item.get("title") or ["Untitled"])[0])
+                authors.append(name)
+        title = (item.get("title") or ["Untitled"])[0]
         year = ""
         for key in ["published-print", "published-online", "created"]:
             if item.get(key, {}).get("date-parts"):
                 year = str(item[key]["date-parts"][0][0])
                 break
-        abstract = truncate_text(clean_extracted_text(re.sub("<.*?>", " ", item.get("abstract") or "")), MAX_ABSTRACT_CHARS)
-        doi = normalize_doi(item.get("DOI", ""))
-
+        abstract = item.get("abstract") or ""
+        abstract = re.sub("<.*?>", "", abstract)
+        doi = item.get("DOI", "")
         out.append({
             "id": doi or title,
-            "title": title,
-            "summary": truncate_text(abstract, 500),
-            "abstract": abstract,
+            "title": norm_text(title),
+            "summary": norm_text(abstract)[:400],
+            "abstract": norm_text(abstract),
             "published": year,
             "authors": authors,
             "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
             "url": item.get("URL", ""),
             "pdf": "",
             "doi": doi,
-            "venue": clean_extracted_text((item.get("container-title") or [""])[0]),
+            "venue": (item.get("container-title") or [""])[0],
             "year": year,
             "source": "crossref",
             "score": 0.72,
@@ -472,42 +255,37 @@ def search_crossref(query: str, mode: str = "topic", max_results: int = 8) -> Li
     return out
 
 
-def search_openalex(query: str, mode: str = "topic", max_results: int = 8) -> List[Dict[str, Any]]:
+def search_openalex(query, mode="topic", max_results=8):
     params = {"per-page": max_results}
     if mode == "doi":
-        doi = normalize_doi(query)
+        doi = query.lower().replace("https://doi.org/", "").replace("http://doi.org/", "")
         params["filter"] = f"doi:https://doi.org/{doi}"
     else:
         params["search"] = query
-
     response = requests.get("https://api.openalex.org/works", params=params, timeout=REQUEST_TIMEOUT)
-    if response.status_code != 200:
-        return []
+    response.raise_for_status()
     items = response.json().get("results", [])
-
     out = []
     for item in items:
         authors = []
         for auth in item.get("authorships", [])[:8]:
             author = auth.get("author") or {}
             if author.get("display_name"):
-                authors.append(clean_extracted_text(author["display_name"]))
+                authors.append(author["display_name"])
         oa = item.get("open_access") or {}
-        doi = normalize_doi(item.get("doi") or "")
-        abstract = truncate_text(clean_extracted_text(parse_openalex_abstract(item.get("abstract_inverted_index"))), MAX_ABSTRACT_CHARS)
-
+        doi = (item.get("doi") or "").replace("https://doi.org/", "")
         out.append({
             "id": item.get("id") or doi or item.get("title"),
-            "title": clean_extracted_text(item.get("title") or ""),
-            "summary": truncate_text(abstract, 500),
-            "abstract": abstract,
+            "title": norm_text(item.get("title")),
+            "summary": "",
+            "abstract": "",
             "published": str(item.get("publication_year") or ""),
             "authors": authors,
             "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
-            "url": ((item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or ""),
+            "url": (item.get("primary_location") or {}).get("landing_page_url") or item.get("id") or "",
             "pdf": oa.get("oa_url") or "",
             "doi": doi,
-            "venue": clean_extracted_text((((item.get("primary_location") or {}).get("source") or {}).get("display_name") or "")),
+            "venue": ((item.get("primary_location") or {}).get("source") or {}).get("display_name") or "",
             "year": str(item.get("publication_year") or ""),
             "source": "openalex",
             "score": 0.80,
@@ -517,12 +295,15 @@ def search_openalex(query: str, mode: str = "topic", max_results: int = 8) -> Li
     return out
 
 
-def search_semantic_scholar(query: str, mode: str = "topic", max_results: int = 8) -> List[Dict[str, Any]]:
+def search_semantic_scholar(query, mode="topic", max_results=8):
+    headers = {}
+    if SEMANTIC_SCHOLAR_API_KEY:
+        headers["x-api-key"] = SEMANTIC_SCHOLAR_API_KEY
     fields = "title,authors,year,abstract,venue,externalIds,url,openAccessPdf"
     if mode == "doi":
-        doi = normalize_doi(query)
-        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{urllib.parse.quote(doi)}"
-        response = requests.get(url, params={"fields": fields}, timeout=REQUEST_TIMEOUT)
+        doi = query.lower().replace("https://doi.org/", "").replace("http://doi.org/", "")
+        url = f"https://api.semanticscholar.org/graph/v1/paper/DOI:{quote(doi)}"
+        response = requests.get(url, params={"fields": fields}, headers=headers, timeout=REQUEST_TIMEOUT)
         if response.status_code != 200:
             return []
         items = [response.json()]
@@ -530,6 +311,7 @@ def search_semantic_scholar(query: str, mode: str = "topic", max_results: int =
         response = requests.get(
             "https://api.semanticscholar.org/graph/v1/paper/search",
             params={"query": query, "limit": max_results, "fields": fields},
+            headers=headers,
             timeout=REQUEST_TIMEOUT,
         )
         if response.status_code != 200:
@@ -539,20 +321,19 @@ def search_semantic_scholar(query: str, mode: str = "topic", max_results: int =
     out = []
     for item in items:
         external = item.get("externalIds") or {}
-        authors = [clean_extracted_text(a.get("name")) for a in item.get("authors", []) if a.get("name")]
-        abstract = truncate_text(clean_extracted_text(item.get("abstract", "")), MAX_ABSTRACT_CHARS)
+        authors = [a.get("name") for a in item.get("authors", []) if a.get("name")]
         out.append({
             "id": external.get("CorpusId") or external.get("DOI") or item.get("title"),
-            "title": clean_extracted_text(item.get("title") or ""),
-            "summary": truncate_text(abstract, 500),
-            "abstract": abstract,
+            "title": norm_text(item.get("title")),
+            "summary": norm_text(item.get("abstract", ""))[:400],
+            "abstract": norm_text(item.get("abstract", "")),
             "published": str(item.get("year") or ""),
             "authors": authors,
             "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
             "url": item.get("url") or "",
-            "pdf": ((item.get("openAccessPdf") or {}).get("url") or ""),
-            "doi": normalize_doi(external.get("DOI", "")),
-            "venue": clean_extracted_text(item.get("venue") or ""),
+            "pdf": (item.get("openAccessPdf") or {}).get("url") or "",
+            "doi": external.get("DOI", ""),
+            "venue": item.get("venue") or "",
             "year": str(item.get("year") or ""),
             "source": "semantic_scholar",
             "score": 0.84,
@@ -562,34 +343,37 @@ def search_semantic_scholar(query: str, mode: str = "topic", max_results: int =
     return out
 
 
-def search_europe_pmc(query: str, mode: str = "topic", max_results: int = 8) -> List[Dict[str, Any]]:
+def search_europe_pmc(query, mode="topic", max_results=8):
     epmc_query = f'DOI:"{query}"' if mode == "doi" else query
-    params = {"query": epmc_query, "format": "json", "pageSize": max_results, "resultType": "core"}
+    params = {
+        "query": epmc_query,
+        "format": "json",
+        "pageSize": max_results,
+        "resultType": "core",
+    }
     response = requests.get("https://www.ebi.ac.uk/europepmc/webservices/rest/search", params=params, timeout=REQUEST_TIMEOUT)
     if response.status_code != 200:
         return []
     items = response.json().get("resultList", {}).get("result", [])
-
     out = []
     for item in items:
         author_string = item.get("authorString", "")
-        authors = [clean_extracted_text(x) for x in author_string.split(",")[:8] if norm_text(x)]
+        authors = [x.strip() for x in author_string.split(",")[:8] if x.strip()]
         pmcid = item.get("pmcid", "")
         pdf_url = f"https://europepmc.org/articles/{pmcid}?pdf=render" if pmcid else ""
         landing_url = f"https://europepmc.org/article/PMC/{pmcid}" if pmcid else ""
-        abstract = truncate_text(clean_extracted_text(item.get("abstractText", "")), MAX_ABSTRACT_CHARS)
         out.append({
             "id": item.get("id") or item.get("doi") or item.get("title"),
-            "title": clean_extracted_text(item.get("title") or ""),
-            "summary": truncate_text(abstract, 500),
-            "abstract": abstract,
+            "title": norm_text(item.get("title")),
+            "summary": norm_text(item.get("abstractText", ""))[:400],
+            "abstract": norm_text(item.get("abstractText", "")),
             "published": str(item.get("pubYear") or ""),
             "authors": authors,
             "authors_text": ", ".join(authors[:4]) if authors else "Unknown authors",
             "url": landing_url,
             "pdf": pdf_url,
-            "doi": normalize_doi(item.get("doi", "")),
-            "venue": clean_extracted_text(item.get("journalTitle", "")),
+            "doi": item.get("doi", ""),
+            "venue": item.get("journalTitle", ""),
             "year": str(item.get("pubYear") or ""),
             "source": "europe_pmc",
             "score": 0.78,
@@ -599,31 +383,140 @@ def search_europe_pmc(query: str, mode: str = "topic", max_results: int = 8) ->
     return out
 
 
-def resolve_link(query: str) -> List[Dict[str, Any]]:
+def resolve_link(query):
     url = (query or "").strip()
     if not url:
         return []
-    return [{
-        "id": url,
-        "title": Path(url.split("?")[0]).name or url,
-        "summary": "Direct link supplied.",
-        "abstract": "",
-        "published": "",
-        "authors": [],
-        "authors_text": "Unknown authors",
-        "url": url,
-        "pdf": url if url.lower().endswith(".pdf") else "",
-        "doi": "",
-        "venue": "Web Link",
-        "year": "",
-        "source": "link",
-        "score": 0.4,
-        "open_access": None,
-        "external_ids": {},
-    }]
-
-
-def discover_papers(query: str, mode: str, sources: List[str], max_results: int = 10) -> List[Dict[str, Any]]:
+    try:
+        response = requests.get(
+            url,
+            timeout=REQUEST_TIMEOUT,
+            allow_redirects=True,
+            headers={"User-Agent": "dvnc-ai-space/0.1"},
+        )
+        content_type = response.headers.get("content-type", "")
+        if "pdf" in content_type or url.lower().endswith(".pdf"):
+            name = Path(url.split("?")[0]).name or "linked-paper.pdf"
+            return [{
+                "id": url,
+                "title": name,
+                "summary": "Direct PDF link detected.",
+                "abstract": "",
+                "published": "",
+                "authors": [],
+                "authors_text": "Unknown authors",
+                "url": url,
+                "pdf": url,
+                "doi": "",
+                "venue": "Direct PDF",
+                "year": "",
+                "source": "link",
+                "score": 0.66,
+                "open_access": True,
+                "external_ids": {},
+            }]
+
+        if BeautifulSoup is None:
+            return [{
+                "id": url,
+                "title": url,
+                "summary": "Web page resolved. Install beautifulsoup4 for DOI extraction.",
+                "abstract": "",
+                "published": "",
+                "authors": [],
+                "authors_text": "Unknown authors",
+                "url": url,
+                "pdf": "",
+                "doi": "",
+                "venue": "Web Link",
+                "year": "",
+                "source": "link",
+                "score": 0.48,
+                "open_access": None,
+                "external_ids": {},
+            }]
+
+        soup = BeautifulSoup(response.text, "html.parser")
+        doi = ""
+        for meta_name in ["citation_doi", "dc.identifier", "dc.Identifier"]:
+            tag = soup.find("meta", attrs={"name": meta_name})
+            if tag and tag.get("content"):
+                doi = tag["content"].strip()
+                break
+
+        title = soup.title.text.strip() if soup.title else url
+        pdf_link = ""
+        for a in soup.find_all("a", href=True):
+            href = a["href"]
+            if ".pdf" in href.lower():
+                pdf_link = href if href.startswith("http") else ""
+                break
+
+        if doi:
+            results = search_crossref(doi, mode="doi", max_results=1)
+            if results:
+                if pdf_link and not results[0].get("pdf"):
+                    results[0]["pdf"] = pdf_link
+                if url and not results[0].get("url"):
+                    results[0]["url"] = url
+                return results
+
+        return [{
+            "id": url,
+            "title": title,
+            "summary": "Landing page resolved from direct link.",
+            "abstract": "",
+            "published": "",
+            "authors": [],
+            "authors_text": "Unknown authors",
+            "url": url,
+            "pdf": pdf_link,
+            "doi": doi,
+            "venue": "Web Link",
+            "year": "",
+            "source": "link",
+            "score": 0.54,
+            "open_access": bool(pdf_link),
+            "external_ids": {},
+        }]
+    except Exception as e:
+        return [{
+            "id": url,
+            "title": "Link resolution error",
+            "summary": str(e),
+            "abstract": "",
+            "published": "",
+            "authors": [],
+            "authors_text": "Unknown authors",
+            "url": url,
+            "pdf": "",
+            "doi": "",
+            "venue": "Link",
+            "year": "",
+            "source": "link",
+            "score": 0.20,
+            "open_access": None,
+            "external_ids": {},
+        }]
+
+
+def dedupe_papers(items: List[Dict]) -> List[Dict]:
+    seen = {}
+    for item in items:
+        key = (
+            (item.get("doi") or "").lower().strip()
+            or (item.get("external_ids") or {}).get("arxiv")
+            or norm_text(item.get("title", "")).lower()
+            or item.get("id")
+        )
+        if not key:
+            key = f"{item.get('source')}::{item.get('title')}"
+        if key not in seen or float(item.get("score", 0)) > float(seen[key].get("score", 0)):
+            seen[key] = item
+    return sorted(seen.values(), key=lambda x: float(x.get("score", 0)), reverse=True)
+
+
+def discover_papers(query, mode, sources, max_results=10):
     query = (query or "").strip()
     if not query:
         return []
@@ -635,93 +528,33 @@ def discover_papers(query: str, mode: str, sources: List[str], max_results: int
     if mode == "link":
         return dedupe_papers(resolve_link(query))
 
-    if "arxiv" in selected_sources and mode != "doi":
-        try:
-            results.extend(search_arxiv(query, max_results=min(max_results, 6)))
-        except Exception:
-            pass
-    if "crossref" in selected_sources:
-        try:
-            results.extend(search_crossref(query, mode=mode, max_results=min(max_results, 8)))
-        except Exception:
-            pass
-    if "openalex" in selected_sources:
-        try:
-            results.extend(search_openalex(query, mode=mode, max_results=min(max_results, 8)))
-        except Exception:
-            pass
-    if "semantic_scholar" in selected_sources:
-        try:
-            results.extend(search_semantic_scholar(query, mode=mode, max_results=min(max_results, 8)))
-        except Exception:
-            pass
-    if "europe_pmc" in selected_sources:
-        try:
-            results.extend(search_europe_pmc(query, mode=mode, max_results=min(max_results, 8)))
-        except Exception:
-            pass
-
-    papers = [enrich_paper_semantics(query, p) for p in dedupe_papers(results)]
-    papers.sort(key=lambda x: float(x.get("learned_score", x.get("score", 0))), reverse=True)
-    return papers[:max_results]
-
-
-# ----------------------------
-# Journal / paper HTML
-# ----------------------------
-
-def journal_query_links(query: str):
-    q = urllib.parse.quote_plus(query or "biomaterials cardiac repair")
-    rows = []
-    for journal in JOURNALS:
-        url = f"{journal['url']}?q={q}" if "?" not in journal["url"] else f"{journal['url']}&q={q}"
-        if "ieeexplore" in journal["url"]:
-            url = f"https://ieeexplore.ieee.org/search/searchresult.jsp?queryText={q}"
-        rows.append({"name": journal["name"], "desc": journal["desc"], "url": url})
-    return rows
-
-
-def build_journal_html(query):
-    rows = []
-    for journal in journal_query_links(query):
-        rows.append(
-            f"""
-            <a class="journal-card" href="{safe_text(journal['url'])}" target="_blank" rel="noopener noreferrer">
-              <div>
-                <h4>{safe_text(journal['name'])}</h4>
-                <p>{safe_text(journal['desc'])}</p>
-              </div>
-              <span>Open</span>
-            </a>
-            """
-        )
-    return """
-    <style>
-      .journal-grid{display:grid;grid-template-columns:repeat(auto-fit,minmax(240px,1fr));gap:12px}
-      .journal-card{display:flex;justify-content:space-between;gap:12px;padding:14px 16px;border-radius:14px;border:1px solid #dbe3f0;background:#f8fbff;text-decoration:none;color:#0f172a}
-      .journal-card h4{margin:0 0 4px;font-size:15px}
-      .journal-card p{margin:0;color:#475569;font-size:13px;line-height:1.45}
-      .journal-card span{align-self:center;font-size:12px;font-weight:700;color:#2563eb}
-    </style>
-    <div class="journal-grid">""" + "".join(rows) + "</div>"
-
-
-def paper_choice_value(index: int, paper: Dict[str, Any]) -> str:
-    doi = normalize_doi(paper.get("doi") or "")
-    title_slug = slugify(paper.get("title", ""))[:40]
-    return f"{index}|{doi}|{title_slug}"
-
-
-def paper_choice_label(index: int, paper: Dict[str, Any]) -> str:
-    score = round(float(paper.get("learned_score", paper.get("score", 0))), 3)
-    title = paper.get("title", "Untitled")
-    authors_text = paper.get("authors_text", "Unknown authors")[:90]
-    source = paper.get("source", "src")
-    return f"[{source}] {title} — {authors_text} — score {score}"
-
+    try:
+        if "arxiv" in selected_sources and mode != "doi":
+            results.extend(search_arxiv(query, max_results=max_results))
+    except Exception:
+        pass
+    try:
+        if "crossref" in selected_sources:
+            results.extend(search_crossref(query, mode=mode, max_results=max_results))
+    except Exception:
+        pass
+    try:
+        if "openalex" in selected_sources:
+            results.extend(search_openalex(query, mode=mode, max_results=max_results))
+    except Exception:
+        pass
+    try:
+        if "semantic_scholar" in selected_sources:
+            results.extend(search_semantic_scholar(query, mode=mode, max_results=max_results))
+    except Exception:
+        pass
+    try:
+        if "europe_pmc" in selected_sources:
+            results.extend(search_europe_pmc(query, mode=mode, max_results=max_results))
+    except Exception:
+        pass
 
-def format_selection_choices(papers):
-    return [(paper_choice_label(i, paper), paper_choice_value(i, paper)) for i, paper in enumerate(papers)]
+    return dedupe_papers(results)
 
 
 def format_papers_html(papers):
@@ -730,12 +563,10 @@ def format_papers_html(papers):
 
     items = []
     for i, paper in enumerate(papers, start=1):
-        summary = safe_text((paper.get("summary") or paper.get("abstract") or "")[:320])
+        summary = safe_text((paper.get("summary") or "")[:280])
         doi_line = f'<span class="paper-badge doi-badge">{safe_text(paper.get("doi"))}</span>' if paper.get("doi") else ""
         pdf_link = paper.get("pdf") or "#"
         abs_link = paper.get("url") or "#"
-        concepts_text = ", ".join((paper.get("concepts") or [])[:5])
-
         items.append(
             f"""
             <article class="paper-card">
@@ -749,8 +580,7 @@ def format_papers_html(papers):
               <div class="paper-meta-stack">
                 <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
                 <div><strong>Venue:</strong> {safe_text(paper.get('venue', 'Unknown venue'))}</div>
-                <div><strong>Learned score:</strong> {safe_text(round(float(paper.get('learned_score', paper.get('score', 0))), 3))}</div>
-                <div><strong>Concepts:</strong> {safe_text(concepts_text or 'None extracted')}</div>
+                <div><strong>Score:</strong> {safe_text(round(float(paper.get('score', 0)), 3))}</div>
               </div>
               <div class="paper-links">
                 <a href="{safe_text(abs_link)}" target="_blank" rel="noopener noreferrer">Abstract</a>
@@ -762,29 +592,12 @@ def format_papers_html(papers):
     return '<div class="papers-grid">' + ''.join(items) + '</div>'
 
 
-def format_frontier_html(frontier):
-    if not frontier:
-        return '<div class="panel papers-panel" style="padding:18px"><p>No autonomous expansion candidates yet.</p></div>'
-    cards = []
-    for i, paper in enumerate(frontier[:12], start=1):
-        cards.append(
-            f"""
-            <article class="paper-card frontier-card">
-              <div class="paper-topline">
-                <span class="paper-badge">frontier</span>
-                <span class="paper-badge alt">{safe_text(paper.get('source', 'paper'))}</span>
-              </div>
-              <h4>{i}. {safe_text(paper.get('title', 'Untitled'))}</h4>
-              <p>{safe_text((paper.get('summary') or paper.get('abstract') or '')[:280])}</p>
-              <div class="paper-meta-stack">
-                <div><strong>Frontier score:</strong> {safe_text(paper.get('frontier_score', paper.get('learned_score', paper.get('score', 0))))}</div>
-                <div><strong>Concept overlap:</strong> {safe_text(paper.get('frontier_concept_overlap', 0))}</div>
-                <div><strong>Authors:</strong> {safe_text(paper.get('authors_text', 'Unknown authors'))}</div>
-              </div>
-            </article>
-            """
-        )
-    return '<div class="papers-grid">' + ''.join(cards) + '</div>'
+def format_selection_choices(papers):
+    choices = []
+    for i, paper in enumerate(papers):
+        label = f"[{paper.get('source', 'src')}] {paper.get('title', 'Untitled')} — {paper.get('authors_text', 'Unknown authors')[:80]}"
+        choices.append((label, str(i)))
+    return choices
 
 
 def uploaded_pdf_summary(file_obj):
@@ -792,391 +605,217 @@ def uploaded_pdf_summary(file_obj):
         return "No PDF uploaded yet."
     path = getattr(file_obj, "name", None) or str(file_obj)
     p = Path(path)
-    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, references, concepts, and claims."
-
-
-# ----------------------------
-# 3D graph
-# ----------------------------
-
-def graph_kind_style(kind: str) -> Dict[str, Any]:
-    palette = {
-        "query": {"color": "#1f8ef1", "size": 14, "label": "Research topic"},
-        "paper": {"color": "#00c49a", "size": 10, "label": "Paper"},
-        "upload": {"color": "#ff9f43", "size": 11, "label": "Uploaded PDF"},
-        "concept": {"color": "#a66cff", "size": 8, "label": "Concept"},
-        "author": {"color": "#f368e0", "size": 7, "label": "Author"},
-        "claim": {"color": "#ff6b6b", "size": 8, "label": "Claim"},
-        "reference": {"color": "#6c757d", "size": 7, "label": "Reference"},
-        "frontier": {"color": "#ffd166", "size": 8, "label": "Frontier candidate"},
-    }
-    return palette.get(kind, {"color": "#9aa0a6", "size": 7, "label": kind.title()})
-
-
-def summarize_graph(nodes: List[Dict], edges: List[Dict]) -> Dict[str, Any]:
-    counts = Counter((n.get("kind") or n.get("type") or "unknown").lower() for n in nodes)
-    return {"nodes": len(nodes), "edges": len(edges), "counts": dict(counts)}
-
-
-def _prepare_3d_graph_data(nodes: List[Dict], edges: List[Dict], title: str) -> Dict[str, Any]:
-    node_out = []
-    for node in nodes:
-        kind = (node.get("kind") or node.get("type") or "paper").lower()
-        if kind == "topic":
-            kind = "query"
-        if kind == "uploadedpdf":
-            kind = "upload"
-        if kind == "frontierpaper":
-            kind = "frontier"
-        style = graph_kind_style(kind)
-        node_out.append({
-            "id": node.get("id"),
-            "label": truncate_text(node.get("label") or node.get("title") or node.get("id") or "node", 120),
-            "kind": kind,
-            "color": style["color"],
-            "val": style["size"],
-            "detail": {
-                "kind": style["label"],
-                "title": node.get("title") or node.get("label") or node.get("id"),
-                "venue": node.get("venue") or "",
-                "year": node.get("year") or "",
-                "doi": node.get("doi") or "",
-                "source": node.get("source") or "",
-                "authors_text": node.get("authors_text") or "",
-                "text": node.get("text") or "",
-            },
-        })
-    edge_out = []
-    for edge in edges:
-        edge_out.append({
-            "source": edge.get("source"),
-            "target": edge.get("target"),
-            "type": edge.get("type") or "RELATES_TO",
-            "label": edge.get("type") or "RELATES_TO",
-        })
-    return {"title": title, "nodes": node_out, "links": edge_out, "summary": summarize_graph(nodes, edges)}
-
-
-def build_learning_graph_html(nodes, edges, title="Self-Learning Knowledge Graph"):
-    if not nodes:
-        return """
-        <div class="panel brain-shell" style="overflow:auto; max-width:100%;">
-          <div class="brain-header">
-            <div>
-              <p class="eyebrow">Learning Graph</p>
-              <h3>Self-Learning Knowledge Graph</h3>
-            </div>
-          </div>
-          <div class="brain-stage learning-empty" style="min-height:420px; overflow:auto;">
-            <div class="empty-graph-copy">
-              <h4>No papers mapped yet</h4>
-              <p>Search papers, select candidates, or upload a PDF to grow the graph in an interactive 3D view.</p>
-            </div>
-          </div>
-        </div>
-        """
-
-    graph_data = _prepare_3d_graph_data(nodes, edges, title)
-    payload_json = json.dumps(graph_data, ensure_ascii=False)
-
-    iframe_html = f"""
-<!doctype html>
-<html>
-<head>
-<meta charset="utf-8" />
-<meta name="viewport" content="width=device-width, initial-scale=1" />
-<style>
-  html, body {{ margin:0; height:100%; background:#0b1020; color:#eef2ff; font-family: Inter, ui-sans-serif, system-ui, sans-serif; overflow:hidden; }}
-  #wrap {{ position:relative; width:100%; height:100%; background:radial-gradient(circle at top, #18213c 0%, #0b1020 60%, #060910 100%); }}
-  #graph {{ position:absolute; inset:0; }}
-  .overlay {{
-    position:absolute; left:16px; top:16px; z-index:10; max-width:min(460px, calc(100% - 32px));
-    padding:14px 16px; border:1px solid rgba(255,255,255,.12); border-radius:16px;
-    background:rgba(10,14,28,.72); backdrop-filter: blur(14px); box-shadow:0 12px 28px rgba(0,0,0,.28);
-  }}
-  .overlay h3 {{ margin:0 0 6px; font-size:18px; line-height:1.2; }}
-  .overlay p {{ margin:0; font-size:13px; color:#cbd5e1; line-height:1.5; }}
-  .panel {{
-    position:absolute; right:16px; top:16px; z-index:10; width:min(360px, calc(100% - 32px));
-    max-height:calc(100% - 32px); overflow:auto; padding:14px 16px; border:1px solid rgba(255,255,255,.12);
-    border-radius:16px; background:rgba(10,14,28,.72); backdrop-filter: blur(14px); box-shadow:0 12px 28px rgba(0,0,0,.28);
-  }}
-  .panel h4 {{ margin:0 0 8px; font-size:14px; color:#f8fafc; }}
-  .panel p {{ margin:0; font-size:12px; color:#cbd5e1; line-height:1.5; }}
-  .panel dl {{ margin:12px 0 0; display:grid; grid-template-columns:auto 1fr; gap:6px 10px; font-size:12px; }}
-  .panel dt {{ color:#93c5fd; }}
-  .panel dd {{ margin:0; color:#e2e8f0; word-break:break-word; }}
-  .hint {{
-    position:absolute; left:16px; bottom:16px; z-index:10; padding:10px 12px; border-radius:12px;
-    font-size:12px; color:#dbeafe; background:rgba(15,23,42,.75); border:1px solid rgba(255,255,255,.08);
-  }}
-</style>
-<script src="https://unpkg.com/three@0.160.0/build/three.min.js"></script>
-<script src="https://unpkg.com/3d-force-graph"></script>
-</head>
-<body>
-<div id="wrap">
-  <div id="graph"></div>
-  <div class="overlay">
-    <h3></h3>
-    <p>Drag the background to orbit, scroll to zoom, right-drag to pan, and drag a node to move or pin it in 3D space.</p>
-  </div>
-  <div class="panel" id="panel">
-    <h4>Node details</h4>
-    <p>Click any node to inspect its label, type, venue, DOI, year, and source.</p>
-  </div>
-  <div class="hint">Interactive 3D graph: orbit, zoom, pan, drag nodes.</div>
-</div>
-<script>
-const payload = {payload_json};
-document.querySelector('.overlay h3').textContent = payload.title || 'Self-Learning Knowledge Graph';
-const panelEl = document.getElementById('panel');
-
-const Graph = ForceGraph3D()(document.getElementById('graph'))
-  .backgroundColor('#00000000')
-  .graphData(payload)
-  .nodeRelSize(6)
-  .nodeOpacity(1)
-  .nodeLabel(node => `<div style="padding:6px 8px"><strong>${{node.label}}</strong><br/>${{(node.detail || {{}}).kind || node.kind}}</div>`)
-  .linkWidth(link => ['ABOUT','UPLOADED_SOURCE','FRONTIER_CANDIDATE'].includes(link.type) ? 2.6 : 1.35)
-  .linkDirectionalParticles(link => ['ABOUT','FRONTIER_CANDIDATE'].includes(link.type) ? 2 : 0)
-  .linkDirectionalParticleWidth(2.2)
-  .cooldownTicks(140)
-  .d3VelocityDecay(0.24)
-  .d3Force('charge').strength(-180)
-  .nodeColor(node => node.color)
-  .nodeVal(node => node.val)
-  .onNodeClick(node => {{
-    const d = node.detail || {{}};
-    panelEl.innerHTML = `
-      <h4>Node details</h4>
-      <dl>
-        <dt>Label</dt><dd>${{node.label || ''}}</dd>
-        <dt>Type</dt><dd>${{d.kind || node.kind || ''}}</dd>
-        <dt>Venue</dt><dd>${{d.venue || '—'}}</dd>
-        <dt>Year</dt><dd>${{d.year || '—'}}</dd>
-        <dt>DOI</dt><dd>${{d.doi || '—'}}</dd>
-        <dt>Source</dt><dd>${{d.source || '—'}}</dd>
-        <dt>Authors</dt><dd>${{d.authors_text || '—'}}</dd>
-        <dt>Text</dt><dd>${{d.text || '—'}}</dd>
-      </dl>`;
-  }})
-  .onNodeDragEnd(node => {{
-    node.fx = node.x;
-    node.fy = node.y;
-    node.fz = node.z;
-  }});
-
-Graph.scene().add(new THREE.AmbientLight(0xffffff, 1.1));
-const dirLight = new THREE.DirectionalLight(0xffffff, 0.8);
-dirLight.position.set(120, 120, 120);
-Graph.scene().add(dirLight);
-Graph.cameraPosition({{ z: 210 }});
-</script>
-</body>
-</html>
-"""
-    return f"""
-    <div class="panel brain-shell" style="overflow:auto; max-width:100%;">
-      <iframe
-        title="{safe_text(title)}"
-        style="width:100%; height:{GRAPH_IFRAME_HEIGHT}px; border:0; border-radius:18px; overflow:auto; background:#0b1020;"
-        sandbox="allow-scripts allow-same-origin"
-        srcdoc="{html.escape(iframe_html, quote=True)}"
-      ></iframe>
-    </div>
-    """
+    return f"Uploaded PDF ready for ingestion: {p.name}. Use Parse uploaded PDF to extract title, abstract, sections, and references."
 
 
 def build_learning_graph_state(query, papers, uploaded_name=None):
     nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
     edges = []
-
-    for i, paper in enumerate(papers[:6], start=1):
+    for i, paper in enumerate(papers[:5], start=1):
         pid = f"paper_{i}"
-        nodes.append({
-            "id": pid,
-            "label": paper.get("title", f"Paper {i}"),
-            "kind": "paper",
-            "title": paper.get("title"),
-            "venue": paper.get("venue"),
-            "year": paper.get("year"),
-            "source": paper.get("source"),
-            "authors_text": paper.get("authors_text"),
-        })
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
-        for concept in (paper.get("concepts") or [])[:3]:
-            cid = f"concept_{i}_{slugify(concept)[:30]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
-
+        nodes.append({"id": pid, "label": paper["title"], "kind": "paper"})
+        edges.append(("query", pid))
     if uploaded_name:
         nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
+        edges.append(("query", "upload"))
+        if len(nodes) > 2:
+            edges.append(("upload", "paper_1"))
     return nodes, edges
 
 
-def graph_from_selected(query, selected_papers, uploaded_name=None, parsed_state=None, frontier=None):
+def graph_from_selected(query, selected_papers, uploaded_name=None):
     nodes = [{"id": "query", "label": query or "Research topic", "kind": "query"}]
     edges = []
-
-    for i, paper in enumerate(selected_papers[:8], start=1):
+    for i, paper in enumerate(selected_papers[:6], start=1):
         pid = f"paper_{i}"
-        nodes.append({
-            "id": pid,
-            "label": paper.get("title", f"Paper {i}"),
-            "kind": "paper",
-            "title": paper.get("title"),
-            "venue": paper.get("venue"),
-            "year": paper.get("year"),
-            "doi": paper.get("doi"),
-            "source": paper.get("source"),
-            "authors_text": paper.get("authors_text"),
-        })
-        edges.append({"source": "query", "target": pid, "type": "ABOUT"})
+        nodes.append({"id": pid, "label": paper["title"], "kind": "paper"})
+        edges.append(("query", pid))
+        if paper.get("doi"):
+            doi_id = f"doi_{i}"
+            nodes.append({"id": doi_id, "label": f"DOI {paper['doi']}", "kind": "paper"})
+            edges.append((pid, doi_id))
+    if uploaded_name:
+        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload"})
+        edges.append(("query", "upload"))
+        if len(selected_papers) > 0:
+            edges.append(("upload", "paper_1"))
+    return nodes, edges
 
-        for author in paper.get("authors", [])[:3]:
-            aid = f"author_{i}_{slugify(author)[:30]}"
-            nodes.append({"id": aid, "label": author, "kind": "author"})
-            edges.append({"source": pid, "target": aid, "type": "WRITTEN_BY"})
 
-        for concept in (paper.get("concepts") or [])[:4]:
-            cid = f"concept_{i}_{slugify(concept)[:30]}"
-            nodes.append({"id": cid, "label": concept, "kind": "concept"})
-            edges.append({"source": pid, "target": cid, "type": "MENTIONS"})
+def parse_pdf_with_grobid(pdf_path):
+    if not GROBID_URL:
+        raise RuntimeError("GROBID_URL is not set")
 
-        for claim in (paper.get("claims") or [])[:2]:
-            cid = f"claim_{i}_{slugify(claim)[:30]}"
-            nodes.append({"id": cid, "label": truncate_text(claim, 82), "kind": "claim", "text": claim})
-            edges.append({"source": pid, "target": cid, "type": "ASSERTS"})
+    with open(pdf_path, "rb") as f:
+        files = {"input": (Path(pdf_path).name, f, "application/pdf")}
+        response = requests.post(
+            f"{GROBID_URL.rstrip('/')}/api/processFulltextDocument",
+            files=files,
+            data={"includeRawAffiliations": "1", "segmentSentences": "1"},
+            timeout=120,
+        )
 
-    if uploaded_name:
-        nodes.append({"id": "upload", "label": uploaded_name, "kind": "upload", "title": uploaded_name})
-        edges.append({"source": "query", "target": "upload", "type": "UPLOADED_SOURCE"})
-        if parsed_state and isinstance(parsed_state, dict):
-            for concept in (parsed_state.get("concepts") or [])[:4]:
-                cid = f"upload_concept_{slugify(concept)[:30]}"
-                nodes.append({"id": cid, "label": concept, "kind": "concept"})
-                edges.append({"source": "upload", "target": cid, "type": "MENTIONS"})
-            for claim in (parsed_state.get("claims") or [])[:3]:
-                cid = f"upload_claim_{slugify(claim)[:30]}"
-                nodes.append({"id": cid, "label": truncate_text(claim, 82), "kind": "claim", "text": claim})
-                edges.append({"source": "upload", "target": cid, "type": "ASSERTS"})
-            for ref in (parsed_state.get("references") or [])[:6]:
-                if ref.get("title"):
-                    rid = f"ref_{slugify(ref.get('title'))[:30]}"
-                    nodes.append({"id": rid, "label": truncate_text(ref.get("title"), 82), "kind": "reference", "doi": ref.get("doi", "")})
-                    edges.append({"source": "upload", "target": rid, "type": "CITES"})
-
-    for j, fp in enumerate(ensure_list(frontier)[:6], start=1):
-        fid = f"frontier_{j}"
-        nodes.append({
-            "id": fid,
-            "label": fp.get("title", f"Frontier {j}"),
-            "kind": "frontier",
-            "title": fp.get("title"),
-            "source": fp.get("source"),
-            "authors_text": fp.get("authors_text"),
-            "year": fp.get("year"),
-            "doi": fp.get("doi"),
-        })
-        edges.append({"source": "query", "target": fid, "type": "FRONTIER_CANDIDATE"})
+    response.raise_for_status()
+    tei_xml = response.text
+
+    if BeautifulSoup is None:
+        return {
+            "parser": "grobid",
+            "title": Path(pdf_path).name,
+            "abstract": "",
+            "authors": [],
+            "sections": [],
+            "references": [],
+            "raw_text": "",
+            "tei_xml": tei_xml[:30000],
+        }
+
+    soup = BeautifulSoup(tei_xml, "xml")
+    title_stmt = soup.find("titleStmt")
+    title_tag = title_stmt.find("title") if title_stmt else soup.find("title")
+    abstract_tag = soup.find("abstract")
+
+    authors = []
+    for author in soup.find_all("author"):
+        pers = author.find("persName")
+        if pers:
+            name = " ".join(
+                x.get_text(" ", strip=True)
+                for x in pers.find_all(["forename", "surname"])
+            ).strip()
+            if name:
+                authors.append(name)
 
-    dedup_nodes = []
-    seen = set()
-    for node in nodes:
-        if node["id"] not in seen:
-            seen.add(node["id"])
-            dedup_nodes.append(node)
-    return dedup_nodes, edges
+    sections = []
+    for div in soup.find_all("div"):
+        head = div.find("head")
+        paras = [p.get_text(" ", strip=True) for p in div.find_all("p")]
+        text = "\n".join([p for p in paras if p.strip()])
+        if head and text.strip():
+            sections.append({"heading": head.get_text(" ", strip=True), "text": text[:4000]})
+
+    references = []
+    for bibl in soup.find_all("biblStruct")[:30]:
+        ref_title = ""
+        ref_doi = ""
+        title_node = bibl.find("title")
+        if title_node:
+            ref_title = title_node.get_text(" ", strip=True)
+        doi_node = bibl.find("idno", attrs={"type": "DOI"})
+        if doi_node:
+            ref_doi = doi_node.get_text(" ", strip=True)
+        references.append({"title": ref_title, "doi": ref_doi})
 
+    return {
+        "parser": "grobid",
+        "title": title_tag.get_text(" ", strip=True) if title_tag else Path(pdf_path).name,
+        "abstract": abstract_tag.get_text(" ", strip=True) if abstract_tag else "",
+        "authors": authors,
+        "sections": sections,
+        "references": references,
+        "raw_text": "",
+        "tei_xml": tei_xml[:60000],
+    }
 
-# ----------------------------
-# PDF parsing
-# ----------------------------
 
-def parse_pdf_with_pymupdf(pdf_path: str) -> Dict[str, Any]:
+def parse_pdf_with_pymupdf(pdf_path):
     if fitz is None:
         raise RuntimeError("PyMuPDF not installed")
 
     doc = fitz.open(pdf_path)
-    page_texts = [page.get_text("text") for page in doc[: min(len(doc), 10)]]
-    raw_text = truncate_text(clean_extracted_text("\n".join(page_texts).strip()), MAX_RAW_TEXT_CHARS)
-    first_page = clean_extracted_text("\n".join(page_texts[:2]))[:5000]
-
-    title = extract_title_from_text(first_page, fallback=Path(pdf_path).name)
+    pages = [page.get_text("text") for page in doc]
+    raw_text = "\n".join(pages).strip()
+    first_page = raw_text[:4000]
+    lines = [x.strip() for x in first_page.splitlines() if x.strip()]
+    title = lines[0][:300] if lines else Path(pdf_path).name
 
     abstract = ""
-    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n(?:1|i)[\.\s]|\nintroduction)", raw_text, re.I | re.S)
+    match = re.search(r"abstract\s*(.+?)(?:\n\s*\n|\n1[\.\s]|introduction)", raw_text, re.I | re.S)
     if match:
-        abstract = truncate_text(clean_extracted_text(match.group(1)), 2600)
-
-    sections = []
-    blocks = re.split(r"\n(?=[A-Z][A-Za-z\s]{2,40}\n)", raw_text)
-    for block in blocks[:10]:
-        lines = [norm_text(x) for x in block.splitlines() if norm_text(x)]
-        if not lines:
-            continue
-        heading = lines[0] if len(lines[0]) < 60 else "Section"
-        body = " ".join(lines[1:] if len(lines) > 1 else lines)
-        if len(body) > 80:
-            sections.append({"heading": clean_extracted_text(heading), "text": truncate_text(body, 4200)})
+        abstract = norm_text(match.group(1))[:2000]
 
     return {
         "parser": "pymupdf",
         "title": title,
         "abstract": abstract,
         "authors": [],
-        "sections": sections[:12] or ([{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else []),
-        "references": extract_references_from_text(raw_text),
-        "claims": extract_claim_like_sentences(raw_text, max_items=GRAPH_MAX_CLAIMS),
-        "concepts": extract_concepts_from_text(" ".join([title, abstract, raw_text[:18000]]), max_terms=GRAPH_MAX_CONCEPTS),
-        "raw_text": raw_text,
-        "parser_quality": "text-fallback-cleaned",
+        "sections": [{"heading": "Full Text", "text": raw_text[:12000]}] if raw_text else [],
+        "references": [],
+        "raw_text": raw_text[:50000],
+        "tei_xml": "",
+    }
+
+
+def parse_pdf_with_docling(pdf_path):
+    try:
+        from docling.document_converter import DocumentConverter
+    except Exception as e:
+        raise RuntimeError(f"Docling import failed: {e}")
+
+    converter = DocumentConverter()
+    result = converter.convert(pdf_path)
+    doc = result.document
+    markdown = doc.export_to_markdown()
+
+    title = Path(pdf_path).name
+    first_nonempty = next((line.strip("# ").strip() for line in markdown.splitlines() if line.strip()), "")
+    if first_nonempty:
+        title = first_nonempty[:300]
+
+    return {
+        "parser": "docling",
+        "title": title,
+        "abstract": "",
+        "authors": [],
+        "sections": [{"heading": "Document", "text": markdown[:12000]}] if markdown else [],
+        "references": [],
+        "raw_text": markdown[:50000],
+        "tei_xml": "",
     }
 
 
-def parse_uploaded_pdf(file_obj, parser_order=None):
+def parse_uploaded_pdf(file_obj, parser_order):
     if not file_obj:
-        return "### PDF parse status\n\nNo PDF uploaded yet.", {}
+        return "No PDF uploaded yet.", {}
 
     path = getattr(file_obj, "name", None) or str(file_obj)
-    parser_order = ensure_list(parser_order) or PDF_PARSERS
+    parser_order = ensure_list(parser_order) or ["grobid", "docling", "pymupdf"]
     errors = []
 
     for parser_name in parser_order:
         try:
-            if parser_name == "pymupdf":
+            if parser_name == "grobid":
+                result = parse_pdf_with_grobid(path)
+            elif parser_name == "docling":
+                result = parse_pdf_with_docling(path)
+            elif parser_name == "pymupdf":
                 result = parse_pdf_with_pymupdf(path)
             else:
                 continue
 
             summary = (
-                f"### PDF parse status\n\n"
+                f"### Parsed PDF\n\n"
                 f"- Parser used: {result['parser']}\n"
-                f"- Parser quality: {result.get('parser_quality', 'unknown')}\n"
                 f"- Title: {result.get('title') or 'Unknown'}\n"
                 f"- Authors: {', '.join(result.get('authors')[:6]) if result.get('authors') else 'Unknown'}\n"
                 f"- Abstract found: {'Yes' if result.get('abstract') else 'No'}\n"
                 f"- Sections extracted: {len(result.get('sections') or [])}\n"
                 f"- References extracted: {len(result.get('references') or [])}\n"
-                f"- Concepts extracted: {len(result.get('concepts') or [])}\n"
-                f"- Claims extracted: {len(result.get('claims') or [])}\n"
             )
             return summary, result
         except Exception as e:
             errors.append(f"{parser_name}: {e}")
 
-    fail_summary = "### PDF parse status\n\n" + "\n".join([f"- {x}" for x in errors])
+    fail_summary = "### PDF parsing failed\n\n" + "\n".join([f"- {x}" for x in errors])
     return fail_summary, {"parser": None, "errors": errors}
 
 
 def render_parse_result(parsed):
-    if not parsed or not isinstance(parsed, dict) or (not parsed.get("title") and not parsed.get("sections")):
+    if not parsed:
         return '<div class="panel" style="padding:18px"><p>No parsed document yet.</p></div>'
 
     sections_html = []
-    for section in parsed.get("sections", [])[:8]:
+    for section in parsed.get("sections", [])[:6]:
         sections_html.append(
             f"""
             <details class="agent-step">
@@ -1188,27 +827,21 @@ def render_parse_result(parsed):
                 </div>
               </summary>
               <div class="agent-copy">
-                <p>{safe_text(section.get('text', '')[:2200])}</p>
+                <p>{safe_text(section.get('text', '')[:1800])}</p>
               </div>
             </details>
             """
         )
 
-    refs = parsed.get("references", [])[:14]
+    refs = parsed.get("references", [])[:12]
     refs_html = "".join(
         f"<li>{safe_text(r.get('title') or 'Untitled')} {'· DOI ' + safe_text(r.get('doi')) if r.get('doi') else ''}</li>"
         for r in refs
     ) or "<li>No references extracted.</li>"
 
-    concepts = parsed.get("concepts", [])[:12]
-    claims = parsed.get("claims", [])[:8]
-    concepts_html = "".join(f"<li>{safe_text(x)}</li>" for x in concepts) or "<li>No concepts extracted.</li>"
-    claims_html = "".join(f"<li>{safe_text(x)}</li>" for x in claims) or "<li>No claims extracted.</li>"
-
     title = safe_text(parsed.get("title") or "Parsed document")
-    abstract = safe_text((parsed.get("abstract") or "")[:2600]) or "No abstract extracted."
+    abstract = safe_text((parsed.get("abstract") or "")[:2400]) or "No abstract extracted."
     parser_name = safe_text(parsed.get("parser") or "unknown")
-    parser_quality = safe_text(parsed.get("parser_quality") or "unknown")
 
     return f"""
     <div class="panel" style="padding:18px">
@@ -1217,7 +850,7 @@ def render_parse_result(parsed):
           <p class="eyebrow">PDF Parse</p>
           <h3>{title}</h3>
         </div>
-        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name} · {parser_quality}</span></div>
+        <div class="brain-legend"><span><i class="dot dot-upload"></i> {parser_name}</span></div>
       </div>
       <div class="parse-grid">
         <div class="parse-card">
@@ -1228,433 +861,61 @@ def render_parse_result(parsed):
           <h4>References</h4>
           <ul class="ref-list">{refs_html}</ul>
         </div>
-        <div class="parse-card">
-          <h4>Concepts</h4>
-          <ul class="ref-list">{concepts_html}</ul>
-        </div>
-        <div class="parse-card">
-          <h4>Claims</h4>
-          <ul class="ref-list">{claims_html}</ul>
-        </div>
       </div>
-      <div class="timeline" style="margin-top:14px; max-height:560px; overflow:auto;">
+      <div class="timeline" style="margin-top:14px;">
         {''.join(sections_html) if sections_html else '<div class="panel" style="padding:16px;"><p>No sections extracted.</p></div>'}
       </div>
     </div>
     """
 
 
-# ----------------------------
-# Ingest payload / learning
-# ----------------------------
-
-def add_node(nodes_by_id: Dict[str, Dict], node_id: str, node_type: str, label: str = "", **attrs):
-    if not node_id:
-        return
-    current = nodes_by_id.get(node_id, {})
-    merged = {"id": node_id, "type": node_type, "label": label or current.get("label", node_id)}
-    merged.update(current)
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            merged[key] = value
-    nodes_by_id[node_id] = merged
-
-
-def add_edge(edges: List[Dict], source: str, target: str, edge_type: str, **attrs):
-    if not source or not target or source == target:
-        return
-    edge = {"source": source, "target": target, "type": edge_type}
-    for key, value in attrs.items():
-        if value not in [None, ""]:
-            edge[key] = value
-    edges.append(edge)
-
-
-def build_ingest_payload(query, selected_papers, parsed_pdf=None, frontier=None):
-    nodes_by_id = {}
+def build_ingest_payload(query, selected_papers, parsed_pdf=None):
+    nodes = [{"id": "topic:query", "type": "Topic", "label": query or "Research topic"}]
     edges = []
 
-    topic_id = "topic:query"
-    add_node(nodes_by_id, topic_id, "Topic", label=query or "Research topic", query=query or "")
-
-    for i, paper in enumerate(selected_papers, start=1):
-        paper_id = normalize_doi(paper.get("doi")) or ((paper.get("external_ids") or {}).get("arxiv")) or f"paper:{i}:{slugify(paper.get('title', 'paper'))[:40]}"
-        add_node(
-            nodes_by_id,
-            paper_id,
-            "Paper",
-            label=paper.get("title") or f"Paper {i}",
-            title=paper.get("title"),
-            year=paper.get("year"),
-            venue=paper.get("venue"),
-            doi=normalize_doi(paper.get("doi")),
-            source=paper.get("source"),
-            url=paper.get("url"),
-            pdf=paper.get("pdf"),
-            score=paper.get("score"),
-            learned_score=paper.get("learned_score", paper.get("score")),
-            open_access=paper.get("open_access"),
-            authors_text=paper.get("authors_text"),
-        )
-        add_edge(edges, topic_id, paper_id, "ABOUT", weight=paper.get("learned_score", paper.get("score", 0)))
-
-        for author in paper.get("authors", [])[:6]:
-            author_id = f"author:{slugify(author)[:64]}"
-            add_node(nodes_by_id, author_id, "Author", label=author, name=author)
-            add_edge(edges, paper_id, author_id, "WRITTEN_BY")
-
-        for concept in (paper.get("concepts") or [])[:8]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, paper_id, concept_id, "MENTIONS")
+    for i, p in enumerate(selected_papers, start=1):
+        paper_id = p.get("doi") or (p.get("external_ids") or {}).get("arxiv") or f"paper:{i}"
+        nodes.append({
+            "id": paper_id,
+            "type": "Paper",
+            "title": p.get("title"),
+            "year": p.get("year"),
+            "venue": p.get("venue"),
+            "doi": p.get("doi"),
+            "source": p.get("source"),
+            "url": p.get("url"),
+            "pdf": p.get("pdf"),
+        })
+        edges.append({"source": "topic:query", "target": paper_id, "type": "ABOUT", "weight": p.get("score", 0)})
 
-        for claim in (paper.get("claims") or [])[:4]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:140], text=claim)
-            add_edge(edges, paper_id, claim_id, "ASSERTS")
+        for author in p.get("authors", [])[:8]:
+            author_id = f"author:{author.lower()}"
+            nodes.append({"id": author_id, "type": "Author", "label": author})
+            edges.append({"source": paper_id, "target": author_id, "type": "WRITTEN_BY"})
 
-    if parsed_pdf and isinstance(parsed_pdf, dict) and parsed_pdf.get("title"):
+    if parsed_pdf and parsed_pdf.get("title"):
         doc_id = "upload:pdf"
-        add_node(nodes_by_id, doc_id, "UploadedPDF", label=parsed_pdf.get("title"), title=parsed_pdf.get("title"), parser=parsed_pdf.get("parser"))
-        add_edge(edges, topic_id, doc_id, "UPLOADED_SOURCE")
-
-        for concept in (parsed_pdf.get("concepts") or [])[:8]:
-            concept_id = f"concept:{slugify(concept)[:72]}"
-            add_node(nodes_by_id, concept_id, "Concept", label=concept, name=concept)
-            add_edge(edges, doc_id, concept_id, "MENTIONS")
-
-        for claim in (parsed_pdf.get("claims") or [])[:6]:
-            claim_id = f"claim:{slugify(claim)[:72]}"
-            add_node(nodes_by_id, claim_id, "Claim", label=claim[:140], text=claim)
-            add_edge(edges, doc_id, claim_id, "ASSERTS")
-
-        for idx, ref in enumerate(parsed_pdf.get("references", [])[:20], start=1):
-            ref_title = ref.get("title") or f"Reference {idx}"
-            ref_doi = normalize_doi(ref.get("doi") or "")
-            ref_id = ref_doi or f"ref:{idx}:{slugify(ref_title)[:40]}"
-            add_node(nodes_by_id, ref_id, "Reference", label=ref_title, title=ref_title, doi=ref_doi)
-            add_edge(edges, doc_id, ref_id, "CITES")
-
-    for idx, item in enumerate(ensure_list(frontier)[:18], start=1):
-        fid = normalize_doi(item.get("doi")) or f"frontier:{idx}:{slugify(item.get('title', 'paper'))[:40]}"
-        add_node(
-            nodes_by_id,
-            fid,
-            "FrontierPaper",
-            label=item.get("title") or f"Frontier {idx}",
-            title=item.get("title"),
-            frontier_score=item.get("frontier_score"),
-            url=item.get("url"),
-            source=item.get("source"),
-            authors_text=item.get("authors_text"),
-            year=item.get("year"),
-            doi=item.get("doi"),
-        )
-        add_edge(edges, topic_id, fid, "FRONTIER_CANDIDATE", weight=item.get("frontier_score", item.get("learned_score", item.get("score", 0))))
-
-    return {"status": "ok", "nodes": list(nodes_by_id.values())[:GRAPH_MAX_NODES], "edges": edges[:GRAPH_MAX_EDGES]}
-
-
-def learn_from_payload(payload: Dict, query: str = "") -> Dict:
-    if not payload:
-        return GRAPH_MEMORY
-
-    GRAPH_MEMORY["queries"].append(query or "")
-    GRAPH_MEMORY["events"].append({
-        "ts": time.time(),
-        "query": query or "",
-        "nodes": len(payload.get("nodes", [])),
-        "edges": len(payload.get("edges", [])),
-    })
-    GRAPH_MEMORY["payloads"].append(payload)
-
-    for node in payload.get("nodes", []):
-        node_id = node.get("id")
-        if not node_id:
-            continue
-        GRAPH_MEMORY["nodes"][node_id] = node
-        node_type = (node.get("type") or "").lower()
-        if node_type in {"paper", "frontierpaper"}:
-            GRAPH_MEMORY["papers"][node_id] = node
-        if node_type == "concept" and node.get("label"):
-            GRAPH_MEMORY["concept_counts"][node["label"].lower()] += 1
-        if node_type == "claim" and node.get("label"):
-            GRAPH_MEMORY["claim_counts"][node["label"].lower()] += 1
-
-    GRAPH_MEMORY["edges"].extend(payload.get("edges", []))
-    GRAPH_MEMORY["edges"] = GRAPH_MEMORY["edges"][:GRAPH_MAX_EDGES]
-    return GRAPH_MEMORY
-
-
-def export_learning_state() -> str:
-    snapshot = {
-        "papers": list(GRAPH_MEMORY["papers"].values())[:60],
-        "nodes": list(GRAPH_MEMORY["nodes"].values())[:250],
-        "edges": GRAPH_MEMORY["edges"][:500],
-        "top_concepts": GRAPH_MEMORY["concept_counts"].most_common(24),
-        "top_claims": GRAPH_MEMORY["claim_counts"].most_common(24),
-        "queries": GRAPH_MEMORY["queries"][-20:],
-        "events": GRAPH_MEMORY["events"][-20:],
-        "frontier": GRAPH_MEMORY["frontier"][:24],
-    }
-    return json.dumps(snapshot, indent=2, ensure_ascii=False)
-
-
-# ----------------------------
-# Frontier expansion
-# ----------------------------
-
-def score_frontier_candidate(query: str, seed_concepts: List[str], paper: Dict[str, Any]) -> Dict[str, Any]:
-    title = paper.get("title", "")
-    abstract = paper.get("abstract", "") or paper.get("summary", "")
-    venue = paper.get("venue", "")
-    base_text = " ".join([title, abstract, venue])
-    rel = text_overlap_score(query, base_text)
-    concept_overlap = text_overlap_score(" ".join(seed_concepts), " ".join(paper.get("concepts") or [])) if seed_concepts else 0.0
-    recency = compute_recency_bonus(paper.get("year"))
-    doi_bonus = 0.02 if paper.get("doi") else 0.0
-    oa_bonus = 0.03 if paper.get("open_access") else 0.0
-    score = float(paper.get("learned_score", paper.get("score", 0))) + rel * 0.45 + concept_overlap * 0.22 + recency + doi_bonus + oa_bonus
-    paper["frontier_score"] = round(score, 4)
-    paper["frontier_relevance"] = round(rel, 4)
-    paper["frontier_concept_overlap"] = round(concept_overlap, 4)
-    return paper
-
-
-def propose_expansion_queries(query: str, papers: List[Dict], parsed_state: Optional[Dict] = None, limit: int = GRAPH_MAX_EXPANSIONS) -> List[str]:
-    concept_pool = []
-    venue_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:4])
-        if paper.get("venue"):
-            venue_pool.append(paper["venue"])
-    if parsed_state and isinstance(parsed_state, dict):
-        concept_pool.extend((parsed_state.get("concepts") or [])[:6])
-
-    ranked_concepts = [c for c, _ in Counter([norm_text(c).lower() for c in concept_pool if c]).most_common(limit * 2)]
-    expansions = [norm_text(query)] if query else []
-    for concept in ranked_concepts:
-        if concept:
-            expansions.append(f"{query} {concept}".strip())
-    for venue in unique_keep_order(venue_pool)[:2]:
-        if venue:
-            expansions.append(f"{query} {venue}".strip())
-    return unique_keep_order(expansions)[:limit]
-
-
-def frontier_expand(query: str, sources: List[str], selected_papers: List[Dict], parsed_state: Optional[Dict] = None, per_query: int = 4) -> List[Dict]:
-    seed_concepts = []
-    for p in selected_papers[:6]:
-        seed_concepts.extend((p.get("concepts") or [])[:4])
-    if parsed_state and isinstance(parsed_state, dict):
-        seed_concepts.extend((parsed_state.get("concepts") or [])[:6])
-
-    expansion_queries = propose_expansion_queries(query, selected_papers, parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    frontier = []
-    for eq in expansion_queries:
-        try:
-            items = discover_papers(eq, "topic", sources, max_results=per_query)
-            for item in items:
-                frontier.append(score_frontier_candidate(query or eq, seed_concepts, item))
-        except Exception:
-            continue
-
-    frontier = dedupe_papers(frontier)
-    frontier.sort(key=lambda x: float(x.get("frontier_score", x.get("learned_score", x.get("score", 0)))), reverse=True)
-    GRAPH_MEMORY["frontier"] = frontier[: GRAPH_MAX_EXPANSIONS * per_query]
-    return GRAPH_MEMORY["frontier"]
-
-
-def autonomous_expand_into_markdown(query, payload, parsed_state=None):
-    frontier = GRAPH_MEMORY.get("frontier") or []
-    lines = [
-        "### Autonomous expansion plan",
-        "",
-        f"- Seed query: {query or 'Research topic'}",
-        f"- Current nodes: {len(payload.get('nodes', [])) if isinstance(payload, dict) else 0}",
-        f"- Current edges: {len(payload.get('edges', [])) if isinstance(payload, dict) else 0}",
-        f"- Frontier candidates: {len(frontier)}",
-    ]
-
-    proposed = propose_expansion_queries(query or "", list(GRAPH_MEMORY.get("papers", {}).values())[:8], parsed_state=parsed_state, limit=GRAPH_MAX_EXPANSIONS)
-    if proposed:
-        lines.extend(["", "#### Proposed next queries", ""])
-        lines.extend([f"- {q}" for q in proposed])
-
-    if frontier:
-        lines.extend(["", "#### Top frontier papers", ""])
-        for item in frontier[:8]:
-            lines.append(
-                f"- {item.get('title', 'Untitled')} ({item.get('source', 'unknown')}) — frontier score {item.get('frontier_score', item.get('learned_score', item.get('score', 0)))}"
-            )
-    return "\n".join(lines)
-
-
-# ----------------------------
-# Selection / ingest
-# ----------------------------
-
-def resolve_selected_papers(selected_indices, papers_state):
-    papers = ensure_list(papers_state)
-    selected_indices = ensure_list(selected_indices)
-    selected = []
-    if not selected_indices:
-        return selected
-
-    value_map = {paper_choice_value(i, paper): paper for i, paper in enumerate(papers)}
-    label_map = {paper_choice_label(i, paper): paper for i, paper in enumerate(papers)}
-
-    for idx in selected_indices:
-        try:
-            if isinstance(idx, int):
-                if 0 <= idx < len(papers):
-                    selected.append(papers[idx])
-                continue
-            idx_str = str(idx)
-            if idx_str in value_map:
-                selected.append(value_map[idx_str])
-                continue
-            if idx_str.isdigit():
-                num = int(idx_str)
-                if 0 <= num < len(papers):
-                    selected.append(papers[num])
-                    continue
-            if "|" in idx_str:
-                left = idx_str.split("|", 1)[0]
-                if left.isdigit():
-                    num = int(left)
-                    if 0 <= num < len(papers):
-                        selected.append(papers[num])
-                        continue
-            if idx_str in label_map:
-                selected.append(label_map[idx_str])
-                continue
-        except Exception:
-            continue
-
-    out = []
-    seen = set()
-    for paper in selected:
-        key = paper_identity_key(paper)
-        if key not in seen:
-            seen.add(key)
-            out.append(paper)
-    return out
-
-
-def build_ingest_status_markdown(query_text: str, payload: Dict, selected: List[Dict], parsed_state: Optional[Dict], frontier: List[Dict]) -> str:
-    payload = payload or {"nodes": [], "edges": []}
-    nodes = payload.get("nodes", [])
-    edges = payload.get("edges", [])
-    counts = Counter((node.get("type") or "Unknown") for node in nodes)
-
-    node_lines = []
-    for idx, node in enumerate(nodes[:24], start=1):
-        label = node.get("label") or node.get("title") or node.get("id")
-        node_lines.append(f"- {idx}. [{node.get('type', 'Node')}] {label}")
-
-    edge_lines = []
-    for idx, edge in enumerate(edges[:30], start=1):
-        edge_lines.append(f"- {idx}. {edge.get('source')} --{edge.get('type', 'RELATES_TO')}--> {edge.get('target')}")
-
-    return "\n".join([
-        "### Graph ingest status",
-        "",
-        f"- Topic: {query_text}",
-        f"- Selected papers ingested: {len(selected)}",
-        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
-        f"- Frontier candidates added: {len(frontier)}",
-        f"- Total nodes created: {len(nodes)}",
-        f"- Total edges created: {len(edges)}",
-        f"- Node breakdown: {', '.join([f'{k}={v}' for k, v in counts.items()]) if counts else 'None'}",
-        "",
-        "### Nodes",
-        *(node_lines or ["- None"]),
-        "",
-        "### Edges",
-        *(edge_lines or ["- None"]),
-    ])
-
-
-def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
-    papers = ensure_list(papers_state)
-    selected = resolve_selected_papers(selected_indices, papers)
-    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
-
-    if not selected and papers:
-        selected = papers[:3]
-
-    if not selected and parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title"):
-        selected = []
-
-    if not selected and not (parsed_state and isinstance(parsed_state, dict) and parsed_state.get("title")):
-        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
-        empty_payload = {"status": "empty", "nodes": [], "edges": []}
-        return (
-            graph_html,
-            "### Graph ingest status\n\n- No papers were selected and no parsed PDF is available.\n- Select papers or parse an uploaded PDF first.",
-            empty_payload,
-        )
-
-    query_text = norm_text(query or "")
-    if not query_text and isinstance(parsed_state, dict):
-        query_text = parsed_state.get("title") or "Research topic"
-    if not query_text:
-        query_text = "Research topic"
-
-    selected = [enrich_paper_semantics(query_text, paper) for paper in selected]
-    frontier = frontier_expand(query_text, DEFAULT_SOURCES, selected or papers[:3], parsed_state=parsed_state if isinstance(parsed_state, dict) else None, per_query=3)
-
-    graph_nodes, graph_edges = graph_from_selected(
-        query_text,
-        selected,
-        uploaded_name,
-        parsed_state if isinstance(parsed_state, dict) else None,
-        frontier=frontier,
-    )
-    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
-
-    payload = build_ingest_payload(
-        query_text,
-        selected,
-        parsed_state if isinstance(parsed_state, dict) else None,
-        frontier=frontier,
-    )
-    learn_from_payload(payload, query=query_text)
-
-    status_md = build_ingest_status_markdown(
-        query_text,
-        payload,
-        selected,
-        parsed_state if isinstance(parsed_state, dict) else None,
-        frontier,
-    )
-
-    return graph_html, status_md, payload
-
-
-# ----------------------------
-# Discovery
-# ----------------------------
+        nodes.append({
+            "id": doc_id,
+            "type": "UploadedPDF",
+            "title": parsed_pdf.get("title"),
+            "parser": parsed_pdf.get("parser"),
+        })
+        edges.append({"source": "topic:query", "target": doc_id, "type": "UPLOADED_SOURCE"})
+        for idx, section in enumerate(parsed_pdf.get("sections", [])[:8], start=1):
+            sec_id = f"{doc_id}:section:{idx}"
+            nodes.append({"id": sec_id, "type": "Section", "label": section.get("heading") or f"Section {idx}"})
+            edges.append({"source": doc_id, "target": sec_id, "type": "HAS_SECTION"})
+        for idx, ref in enumerate(parsed_pdf.get("references", [])[:12], start=1):
+            ref_id = f"{doc_id}:ref:{idx}"
+            nodes.append({"id": ref_id, "type": "Reference", "label": ref.get("title") or f"Reference {idx}", "doi": ref.get("doi")})
+            edges.append({"source": doc_id, "target": ref_id, "type": "CITES"})
 
-def summarize_learning_state(query_text, papers, selected_sources):
-    concept_pool = []
-    for paper in papers[:8]:
-        concept_pool.extend((paper.get("concepts") or [])[:4])
-    top_concepts = [c for c, _ in Counter([c.lower() for c in concept_pool]).most_common(8)]
-    return (
-        "### Discovery results\n\n"
-        f"- Query: {query_text}\n"
-        f"- Sources: {', '.join(selected_sources)}\n"
-        f"- Candidates found: {len(papers)}\n"
-        f"- Top learned concepts: {', '.join(top_concepts) if top_concepts else 'None'}\n"
-        "- Select papers below, or leave selection empty to auto-ingest the top papers.\n"
-    )
+    return {"status": "ok", "nodes": nodes, "edges": edges}
 
 
 def run_paper_discovery(query, search_mode, sources, pdf_file):
-    query_text = norm_text(query or "")
-    selected_sources = ensure_list(sources) or DEFAULT_SOURCES
+    query_text = (query or "").strip()
 
     if not query_text and not pdf_file:
         empty_graph = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
@@ -1668,71 +929,73 @@ def run_paper_discovery(query, search_mode, sources, pdf_file):
             "### No discovery results yet.",
         )
 
-    if not query_text and pdf_file:
-        uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name
-        graph_nodes, graph_edges = build_learning_graph_state("", [], uploaded_name)
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges, "Uploaded PDF Waiting for Parse"),
-            '<div class="panel papers-panel" style="padding:18px"><p>No query yet. Parse the uploaded PDF or enter a research topic to begin discovery.</p></div>',
-            build_journal_html("biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            "### Upload detected.\n\n- Parse the PDF to extract structure.\n- Or enter a topic to start discovery.",
-        )
+    papers = []
+    if query_text:
+        try:
+            papers = discover_papers(query_text, search_mode, sources, max_results=10)
+        except Exception as e:
+            graph_nodes, graph_edges = build_learning_graph_state(
+                query_text,
+                [],
+                Path(getattr(pdf_file, "name", "uploaded.pdf")).name if pdf_file else None,
+            )
+            error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
+            return (
+                build_learning_graph_html(graph_nodes, graph_edges),
+                error_html,
+                build_journal_html(query_text or "biomaterials cardiac repair"),
+                uploaded_pdf_summary(pdf_file),
+                gr.update(choices=[], value=[]),
+                [],
+                f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
+            )
 
     uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
+    graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:5], uploaded_name)
+    graph_html = build_learning_graph_html(graph_nodes, graph_edges)
+    papers_html = format_papers_html(papers)
+    journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
+    pdf_summary = uploaded_pdf_summary(pdf_file)
+    choices = format_selection_choices(papers)
+
+    status_md = (
+        f"### Discovery results\n\n"
+        f"- Search mode: {search_mode}\n"
+        f"- Sources: {', '.join(ensure_list(sources) or DEFAULT_SOURCES)}\n"
+        f"- Candidates found: {len(papers)}\n"
+        f"- Select papers below, then click **Ingest selected into graph**.\n"
+    )
+    return graph_html, papers_html, journals_html, pdf_summary, gr.update(choices=choices, value=[]), papers, status_md
 
-    try:
-        papers = discover_papers(query_text, search_mode, selected_sources, max_results=GRAPH_MAX_RESULTS)
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, papers[:6], uploaded_name)
-        graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Self-Learning Knowledge Graph")
-        papers_html = format_papers_html(papers)
-        journals_html = build_journal_html(query_text or "biomaterials cardiac repair")
-        pdf_summary = uploaded_pdf_summary(pdf_file)
-        choices = format_selection_choices(papers)
-        status_md = summarize_learning_state(query_text, papers, selected_sources)
-        return (
-            graph_html,
-            papers_html,
-            journals_html,
-            pdf_summary,
-            gr.update(choices=choices, value=[]),
-            papers,
-            status_md,
-        )
-    except Exception as e:
-        graph_nodes, graph_edges = build_learning_graph_state(query_text, [], uploaded_name)
-        error_html = f'<div class="panel papers-panel" style="padding:18px"><p>Paper search failed: {safe_text(str(e))}</p></div>'
-        return (
-            build_learning_graph_html(graph_nodes, graph_edges),
-            error_html,
-            build_journal_html(query_text or "biomaterials cardiac repair"),
-            uploaded_pdf_summary(pdf_file),
-            gr.update(choices=[], value=[]),
-            [],
-            f"### Discovery failed.\n\n- Error: {safe_text(str(e))}",
-        )
 
+def ingest_selected_papers(query, selected_indices, papers_state, pdf_file, parsed_state):
+    papers = ensure_list(papers_state)
+    selected_indices = ensure_list(selected_indices)
+
+    selected = []
+    for idx in selected_indices:
+        try:
+            selected.append(papers[int(idx)])
+        except Exception:
+            pass
+
+    uploaded_name = Path(getattr(pdf_file, "name", str(pdf_file))).name if pdf_file else None
+
+    if not selected and not parsed_state:
+        graph_html = build_learning_graph_html([], [], "Self-Learning Knowledge Graph")
+        return graph_html, "### Nothing ingested yet.\n\nSelect papers or parse an uploaded PDF first.", {"status": "empty", "nodes": [], "edges": []}
+
+    graph_nodes, graph_edges = graph_from_selected(query, selected, uploaded_name)
+    graph_html = build_learning_graph_html(graph_nodes, graph_edges, "Selected Research Graph")
+    payload = build_ingest_payload(query, selected, parsed_state if isinstance(parsed_state, dict) else None)
 
-__all__ = [
-    "SEARCH_MODES",
-    "SOURCE_OPTIONS",
-    "DEFAULT_SOURCES",
-    "PDF_PARSERS",
-    "GRAPH_MEMORY",
-    "discover_papers",
-    "run_paper_discovery",
-    "parse_uploaded_pdf",
-    "render_parse_result",
-    "ingest_selected_papers",
-    "build_ingest_payload",
-    "learn_from_payload",
-    "frontier_expand",
-    "autonomous_expand_into_markdown",
-    "export_learning_state",
-    "format_frontier_html",
-    "build_learning_graph_html",
-    "build_journal_html",
-    "safe_text",
-]
+    summary_lines = [
+        "### Graph ingest ready",
+        "",
+        f"- Topic: {query or 'Research topic'}",
+        f"- Selected papers: {len(selected)}",
+        f"- Uploaded PDF parsed: {'Yes' if parsed_state and isinstance(parsed_state, dict) and parsed_state.get('title') else 'No'}",
+        f"- Nodes created: {len(payload['nodes'])}",
+        f"- Edges created: {len(payload['edges'])}",
+    ]
+    return graph_html, "\n".join(summary_lines), payload

dvnc_flip_insight_patch.py

@@ -1,203 +0,0 @@
-from __future__ import annotations
-
-import re
-from typing import Any
-
-
-PATCH_SCRIPT = r"""
-<script>
-(function () {
-  if (window.__dvncFlipPatchLoaded) return;
-  window.__dvncFlipPatchLoaded = true;
-
-  function findByFragment(fragment) {
-    return document.getElementById(fragment) || document.querySelector('[id*="' + fragment + '"]');
-  }
-
-  function getInteractiveNode(root) {
-    if (!root) return null;
-    return root.matches?.('textarea,input,button') ? root : root.querySelector?.('textarea,input,button');
-  }
-
-  function triggerRouteSwapPatched(idx) {
-    try {
-      const payloadRoot = findByFragment('route_swap_payload');
-      const payload = getInteractiveNode(payloadRoot);
-      if (!payload) {
-        console.warn('[DVNC patch] route_swap_payload not found');
-        return;
-      }
-
-      payload.focus?.();
-      payload.value = String(idx);
-
-      ['input', 'change'].forEach(function (name) {
-        payload.dispatchEvent(new Event(name, { bubbles: true }));
-      });
-
-      window.setTimeout(function () {
-        const applyRoot = findByFragment('route_swap_apply');
-        const applyBtn = getInteractiveNode(applyRoot) || applyRoot;
-        if (!applyBtn) {
-          console.warn('[DVNC patch] route_swap_apply not found');
-          return;
-        }
-        applyBtn.click?.();
-      }, 180);
-    } catch (err) {
-      console.error('[DVNC patch] triggerRouteSwap failed', err);
-    }
-  }
-
-  window.triggerRouteSwap = triggerRouteSwapPatched;
-
-  document.addEventListener('click', function (e) {
-    const mini = e.target.closest('.candidate-back .mini');
-    if (mini) return;
-
-    const card = e.target.closest('.candidate-card');
-    if (!card) return;
-
-    card.classList.toggle('flipped');
-  }, true);
-
-  document.addEventListener('keydown', function (e) {
-    const target = e.target;
-    if (!target || !target.closest) return;
-
-    const card = target.closest('.candidate-card');
-    if (!card) return;
-
-    if (e.key === 'Enter' || e.key === ' ') {
-      if (target.closest('.candidate-back .mini')) return;
-      e.preventDefault();
-      card.classList.toggle('flipped');
-    }
-  }, true);
-})();
-</script>
-"""
-
-PATCH_STYLE = r"""
-<style>
-.candidate-card { cursor: pointer; }
-.candidate-card.flipped .candidate-card-inner { transform: rotateY(180deg) !important; }
-.candidate-card:hover .candidate-card-inner,
-.candidate-card:focus .candidate-card-inner,
-.candidate-card:focus-within .candidate-card-inner { transform: none !important; }
-.candidate-back .mini { position: relative; z-index: 5; }
-</style>
-"""
-
-
-def _inject_assets(head: str) -> str:
-    if "__dvncFlipPatchLoaded" in head:
-        return head
-    return head + "\n" + PATCH_STYLE + "\n" + PATCH_SCRIPT
-
-
-def _patch_head(module: Any) -> None:
-    for attr in ("HEAD", "head", "CUSTOM_HEAD"):
-        if hasattr(module, attr):
-            value = getattr(module, attr)
-            if isinstance(value, str):
-                setattr(module, attr, _inject_assets(value))
-                return
-
-
-def _patch_cards_builder(module: Any) -> None:
-    if not hasattr(module, "build_cards_html"):
-        return
-
-    original = module.build_cards_html
-
-    def wrapped(*args, **kwargs):
-        out = original(*args, **kwargs)
-        if not isinstance(out, str):
-            return out
-        if "candidate-card" not in out:
-            return out
-
-        out = re.sub(
-            r'(<div\s+class="candidate-card"\b)',
-            r'\1 tabindex="0" role="button" aria-label="Flip insight card"',
-            out,
-        )
-        return out
-
-    module.build_cards_html = wrapped
-
-
-def _build_timeline_from_state(module: Any, route_state: Any):
-    if route_state is None:
-        return None
-
-    builder = getattr(module, "build_agent_route_cards_html", None)
-    if not callable(builder):
-        return None
-
-    try:
-        variants = route_state.get("variants") if isinstance(route_state, dict) else None
-        active_idx = route_state.get("active_variant", 0) if isinstance(route_state, dict) else 0
-        if not variants or active_idx >= len(variants):
-            return None
-
-        variant = variants[active_idx] or {}
-        steps = variant.get("steps") or variant.get("route") or []
-        if not steps:
-            return None
-
-        normalized = []
-        for i, step in enumerate(steps):
-            if isinstance(step, dict):
-                normalized.append({
-                    "step": step.get("step", i + 1),
-                    "agent": step.get("agent", f"Step {i+1}"),
-                    "tag": step.get("tag", step.get("mode", "")),
-                    "summary": step.get("summary", step.get("description", "")),
-                })
-            else:
-                normalized.append({
-                    "step": i + 1,
-                    "agent": f"Step {i+1}",
-                    "tag": "",
-                    "summary": str(step),
-                })
-
-        return builder(normalized)
-    except Exception:
-        return None
-
-
-def _patch_apply_route_swap(module: Any) -> None:
-    if not hasattr(module, "apply_route_swap"):
-        return
-
-    original = module.apply_route_swap
-
-    def wrapped(*args, **kwargs):
-        result = original(*args, **kwargs)
-        if not isinstance(result, tuple):
-            return result
-
-        if len(result) == 5:
-            chat_html, connectome_html, timeline_value, hypothesis_md, route_state = result
-            rebuilt = _build_timeline_from_state(module, route_state)
-            if rebuilt is not None:
-                gr = getattr(module, "gr", None)
-                if gr is not None and hasattr(gr, "update"):
-                    timeline_value = gr.update(value=rebuilt)
-                else:
-                    timeline_value = rebuilt
-            return chat_html, connectome_html, timeline_value, hypothesis_md, route_state
-
-        return result
-
-    module.apply_route_swap = wrapped
-
-
-def apply_patch(module: Any) -> Any:
-    _patch_head(module)
-    _patch_cards_builder(module)
-    _patch_apply_route_swap(module)
-    return module

dvnc_flip_insight_patch_old.py

@@ -1,227 +0,0 @@
-"""
-Standalone non-destructive patch for DVNC.AI Hugging Face Space.
-
-Purpose
-- Fix Bug 1: hover/focus flip interaction makes the Flip Insight button unreliable.
-- Fix Bug 2: brittle Gradio element ID lookup for route swap payload/apply controls.
-- Fix Bug 3: route swap leaves timeline stale/blank by rebuilding/preserving the timeline output.
-
-Design
-- Leaves the existing codebase untouched.
-- Intended to be imported and applied from a separate runner or notebook.
-- Uses monkey-patching where possible.
-
-Usage example
-    import app as dvnc_app
-    import dvnc_flip_insight_patch as patch
-    patch.apply_patch(dvnc_app)
-
-If you want to keep the current app.py completely untouched, create a tiny launcher file
-that imports the original app module and then calls apply_patch(dvnc_app).
-"""
-
-from __future__ import annotations
-
-import re
-from typing import Any, Callable
-
-
-PATCH_SCRIPT = r"""
-<script>
-(function () {
-  if (window.__dvncFlipPatchLoaded) return;
-  window.__dvncFlipPatchLoaded = true;
-
-  function findByFragment(fragment) {
-    return document.getElementById(fragment) || document.querySelector('[id*="' + fragment + '"]');
-  }
-
-  function getInteractiveNode(root) {
-    if (!root) return null;
-    return root.matches?.('textarea,input,button') ? root : root.querySelector?.('textarea,input,button');
-  }
-
-  function triggerRouteSwapPatched(idx) {
-    try {
-      const payloadRoot = findByFragment('route_swap_payload');
-      const payload = getInteractiveNode(payloadRoot);
-      if (!payload) {
-        console.warn('[DVNC patch] route_swap_payload not found');
-        return;
-      }
-      payload.focus?.();
-      payload.value = String(idx);
-      ['input', 'change'].forEach(function (name) {
-        payload.dispatchEvent(new Event(name, { bubbles: true }));
-      });
-
-      window.setTimeout(function () {
-        const applyRoot = findByFragment('route_swap_apply');
-        const applyBtn = getInteractiveNode(applyRoot) || applyRoot;
-        if (!applyBtn) {
-          console.warn('[DVNC patch] route_swap_apply not found');
-          return;
-        }
-        applyBtn.click?.();
-      }, 180);
-    } catch (err) {
-      console.error('[DVNC patch] triggerRouteSwap failed', err);
-    }
-  }
-
-  window.triggerRouteSwap = triggerRouteSwapPatched;
-
-  document.addEventListener('click', function (e) {
-    const mini = e.target.closest('.candidate-back .mini');
-    if (mini) return;
-
-    const card = e.target.closest('.candidate-card');
-    if (!card) return;
-
-    card.classList.toggle('flipped');
-  }, true);
-
-  document.addEventListener('keydown', function (e) {
-    const card = e.target.closest && e.target.closest('.candidate-card');
-    if (!card) return;
-    if (e.key === 'Enter' || e.key === ' ') {
-      if (e.target.closest('.candidate-back .mini')) return;
-      e.preventDefault();
-      card.classList.toggle('flipped');
-    }
-  }, true);
-})();
-</script>
-"""
-
-PATCH_STYLE = r"""
-<style>
-.candidate-card { cursor: pointer; }
-.candidate-card.flipped .candidate-card-inner { transform: rotateY(180deg) !important; }
-.candidate-card:hover .candidate-card-inner,
-.candidate-card:focus .candidate-card-inner,
-.candidate-card:focus-within .candidate-card-inner { transform: none !important; }
-.candidate-back .mini { position: relative; z-index: 5; }
-</style>
-"""
-
-
-def _inject_assets(head: str) -> str:
-    if "__dvncFlipPatchLoaded" in head:
-        return head
-    return head + "\n" + PATCH_STYLE + "\n" + PATCH_SCRIPT
-
-
-def _patch_head(module: Any) -> None:
-    for attr in ("HEAD", "head", "CUSTOM_HEAD"):
-        if hasattr(module, attr):
-            value = getattr(module, attr)
-            if isinstance(value, str):
-                setattr(module, attr, _inject_assets(value))
-                return
-
-
-def _patch_cards_builder(module: Any) -> None:
-    if not hasattr(module, "build_cards_html"):
-        return
-    original = module.build_cards_html
-
-    def wrapped(*args, **kwargs):
-        out = original(*args, **kwargs)
-        if not isinstance(out, str):
-            return out
-        if "candidate-card" not in out:
-            return out
-
-        out = re.sub(
-            r'(<div\s+class="candidate-card"\b)',
-            r'\1 tabindex="0" role="button" aria-label="Flip insight card"',
-            out,
-            count=0,
-        )
-        return out
-
-    module.build_cards_html = wrapped
-
-
-def _build_timeline_from_state(module: Any, route_state: Any) -> Any:
-    if route_state is None:
-        return None
-
-    builder = getattr(module, "build_agent_route_cards_html", None)
-    if not callable(builder):
-        return None
-
-    try:
-        variants = route_state.get("variants") if isinstance(route_state, dict) else None
-        active_idx = route_state.get("active_variant", 0) if isinstance(route_state, dict) else 0
-        if not variants or active_idx >= len(variants):
-            return None
-        variant = variants[active_idx] or {}
-        steps = variant.get("steps") or variant.get("route") or []
-        if not steps:
-            return None
-
-        normalized = []
-        for i, step in enumerate(steps):
-            if isinstance(step, dict):
-                normalized.append(
-                    {
-                        "step": step.get("step", i + 1),
-                        "agent": step.get("agent", f"Step {i+1}"),
-                        "tag": step.get("tag", step.get("mode", "")),
-                        "summary": step.get("summary", step.get("description", "")),
-                    }
-                )
-            else:
-                normalized.append(
-                    {
-                        "step": i + 1,
-                        "agent": f"Step {i+1}",
-                        "tag": "",
-                        "summary": str(step),
-                    }
-                )
-        return builder(normalized)
-    except Exception:
-        return None
-
-
-def _patch_apply_route_swap(module: Any) -> None:
-    if not hasattr(module, "apply_route_swap"):
-        return
-
-    original = module.apply_route_swap
-
-    def wrapped(*args, **kwargs):
-        result = original(*args, **kwargs)
-        if not isinstance(result, tuple):
-            return result
-
-        if len(result) == 5:
-            chat_html, connectome_html, timeline_value, hypothesis_md, route_state = result
-            timeline_rebuilt = _build_timeline_from_state(module, route_state)
-            if timeline_rebuilt is not None:
-                gr = getattr(module, "gr", None)
-                if gr is not None and hasattr(gr, "update"):
-                    timeline_value = gr.update(value=timeline_rebuilt)
-                else:
-                    timeline_value = timeline_rebuilt
-            return chat_html, connectome_html, timeline_value, hypothesis_md, route_state
-        return result
-
-    module.apply_route_swap = wrapped
-
-
-def _patch_rendered_html(module: Any) -> None:
-    for attr in ("DISCOVERY_CARD_CSS",):
-        if hasattr(module, attr) and isinstance(getattr(module, attr), str):
-            setattr(module, attr, getattr(module, attr) + "\n" + PATCH_STYLE)
-
-
-def apply_patch(module: Any) -> Any:
-    _patch_head(module)
-    _patch_rendered_html(module)
-    _patch_cards_builder(module)
-    _patch_apply_route_swap(module)
-    return module

dvnc_patch_loader_example.py

@@ -1,27 +0,0 @@
-"""
-Example launcher that keeps the original code untouched.
-
-Place this file alongside the existing app module and adjust the import path if needed.
-Run this launcher instead of the original entry point.
-"""
-
-import importlib.util
-import sys
-from pathlib import Path
-
-ORIGINAL_APP_PATH = Path("dvnc_ai_v2_hf/app.py")
-PATCH_PATH = Path("dvnc_flip_insight_patch.py")
-
-spec = importlib.util.spec_from_file_location("dvnc_original_app", ORIGINAL_APP_PATH)
-mod = importlib.util.module_from_spec(spec)
-sys.modules["dvnc_original_app"] = mod
-spec.loader.exec_module(mod)
-
-patch_spec = importlib.util.spec_from_file_location("dvnc_flip_insight_patch", PATCH_PATH)
-patch = importlib.util.module_from_spec(patch_spec)
-sys.modules["dvnc_flip_insight_patch"] = patch
-patch_spec.loader.exec_module(patch)
-patch.apply_patch(mod)
-
-# If the original app exposes demo/launch objects, they remain on mod.
-# Example: mod.demo.launch()