app.py

from __future__ import annotations

import hashlib
import json
import re
from html import escape
from pathlib import Path

import gradio as gr
import pandas as pd
import gradio_client.utils as gradio_client_utils

from src.totem_workbook import (
    DEFAULT_WORKBOOK,
    LOG_COLUMNS,
    METRICS,
    export_updated_workbook,
    manuscript_tracker_table,
    recalculate_log,
    score_log,
    score_single_row,
    viability_table,
    workbook_overview,
    workbook_path,
    workstack_table,
)

from src.codex_extractor import process_upload, format_fingerprint_report
from smoke_signal_tab import smoke_signal_tab, SS_CSS

ORIGINAL_WORKBOOK_PATH = "data/order69_macmillan_totem_rebuilt.xlsx"
CODEX_CATALOGUE_PATH = Path("data/codex_catalogue.xlsx")


def _patch_gradio_schema_bool_compat() -> None:
    """
    Compatibility shim for Gradio API schema parsing where boolean JSON schema
    nodes can appear as `additionalProperties: true` in newer Pydantic output.
    """
    original_get_type = gradio_client_utils.get_type

    def _safe_get_type(schema):
        if isinstance(schema, bool):
            return "boolean"
        return original_get_type(schema)

    gradio_client_utils.get_type = _safe_get_type


_patch_gradio_schema_bool_compat()

CSS = """
:root {
  --studio-green: #0e4a1d;
  --studio-green-2: #17642a;
  --studio-gold: #e4aa1a;
  --studio-cream: #fffaf0;
  --studio-ink: #15351d;
  --studio-muted: #6d725f;
  --studio-line: #eadfbd;
  --studio-red: #cf4b3f;
}

.gradio-container {
  max-width: none !important;
  padding: 0 !important;
  background: #fffaf0 !important;
  color: var(--studio-ink) !important;
  font-family: Inter, ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif !important;
}

footer { display: none !important; }

.gradio-container [role="tablist"] {
  position: sticky !important;
  top: 0 !important;
  z-index: 60 !important;
  background: #fffaf0 !important;
  border-bottom: 1px solid var(--studio-line);
  padding: 8px 10px;
  gap: 8px;
  overflow: visible !important;
}

.gradio-container [role="tab"] {
  opacity: 1 !important;
  visibility: visible !important;
  color: var(--studio-ink) !important;
  background: #f4ecd6 !important;
  border: 1px solid var(--studio-line) !important;
  border-radius: 8px !important;
  padding: 8px 14px !important;
  font-weight: 700 !important;
}

.gradio-container [role="tab"][aria-selected="true"] {
  background: linear-gradient(90deg, #f5c93c, #e5a721) !important;
  color: white !important;
  border-color: #d99e1b !important;
}

.nav-item[role="button"] {
  cursor: pointer;
}

#hidden-export, #hidden-status {
  max-width: 1180px;
  margin: 0 auto 18px auto;
}

#studio-actions {
  max-width: 1180px;
  margin: 18px auto 16px auto;
  padding-left: 18px;
  position: relative;
  z-index: 5;
}

#studio-actions .wrap {
  max-width: 430px;
  display: grid;
  grid-template-columns: 1fr 1fr;
  gap: 12px;
}

#studio-actions button {
  border-radius: 8px !important;
  min-height: 48px !important;
  font-weight: 800 !important;
}

#path-panel {
  max-width: 1180px;
  margin: 18px auto;
  padding: 0 18px;
}

#path-panel .wrap {
  display: grid;
  grid-template-columns: minmax(360px, 1fr) 210px;
  gap: 12px;
  max-width: 760px;
}

#path-panel textarea,
#path-panel input {
  border-radius: 8px !important;
  border: 1px solid var(--studio-line) !important;
  background: white !important;
}

#path-panel button {
  border-radius: 8px !important;
  min-height: 52px !important;
  font-weight: 800 !important;
}

#score-panel {
  max-width: 1180px;
  margin: 18px auto 44px auto;
  padding: 0 18px;
}

#score-panel .score-card {
  border: 1px solid var(--studio-line);
  background: #fffef8;
  border-radius: 8px;
  padding: 18px;
}

#score-panel h3 {
  margin: 0 0 12px 0;
  font-size: 18px;
  color: var(--studio-green);
}

#score-panel button {
  border-radius: 8px !important;
  font-weight: 800 !important;
}

#score-panel .wrap {
  gap: 12px;
}

#codex-panel {
  max-width: 1180px;
  margin: 18px auto 44px auto;
  padding: 0 18px;
}

#codex-panel .codex-header {
  background: linear-gradient(135deg, #0e4a1d, #17642a);
  border-radius: 8px 8px 0 0;
  padding: 20px 24px;
  color: white;
}

#codex-panel .codex-header h3 {
  margin: 0;
  color: #f8e838;
  font-size: 20px;
  font-family: Georgia, serif;
}

#codex-panel .codex-header p {
  margin: 6px 0 0;
  color: #d9ead0;
  font-size: 13px;
}

#codex-panel .codex-body {
  border: 1px solid var(--studio-line);
  border-top: none;
  border-radius: 0 0 8px 8px;
  padding: 24px;
  background: #fffef8;
}

#codex-panel .confidence-high {
  background: #e8f5e9;
  border: 1px solid #a5d6a7;
  border-radius: 6px;
  padding: 10px 14px;
  color: #1b5e20;
  font-weight: 700;
}

#codex-panel .confidence-medium {
  background: #fff8e1;
  border: 1px solid #ffe082;
  border-radius: 6px;
  padding: 10px 14px;
  color: #e65100;
  font-weight: 700;
}

#codex-panel .confidence-low {
  background: #ffebee;
  border: 1px solid #ef9a9a;
  border-radius: 6px;
  padding: 10px 14px;
  color: #b71c1c;
  font-weight: 700;
}

.dataframe, .table-wrap, .sheet, .tabs, .tabitem {
  border-radius: 8px !important;
}

@media (max-width: 900px) {
  #studio-actions {
    margin-top: 0;
    padding: 14px;
  }
  #studio-actions .wrap,
  #path-panel .wrap {
    grid-template-columns: 1fr;
  }
}
"""

HEAD = """
<script>
(() => {
  const switchToCodexTab = () => {
    const tabs = Array.from(document.querySelectorAll('[role="tab"]'));
    if (!tabs.length) return false;
    const target = tabs.find((tab) => /codex extractor/i.test((tab.textContent || "").trim()));
    if (target) {
      target.click();
      return true;
    }
    if (tabs.length > 1) {
      tabs[1].click();
      return true;
    }
    return false;
  };

  const bindSidebarCodexLinks = () => {
    document.querySelectorAll('.js-open-codex').forEach((el) => {
      if (el.dataset.boundCodexNav === "1") return;
      el.dataset.boundCodexNav = "1";
      el.addEventListener('click', (e) => {
        e.preventDefault();
        switchToCodexTab();
      });
      el.addEventListener('keydown', (e) => {
        if (e.key === 'Enter' || e.key === ' ') {
          e.preventDefault();
          switchToCodexTab();
        }
      });
    });
  };

  const observer = new MutationObserver(() => bindSidebarCodexLinks());
  observer.observe(document.documentElement, { childList: true, subtree: true });

  if (document.readyState === 'loading') {
    document.addEventListener('DOMContentLoaded', bindSidebarCodexLinks);
  } else {
    bindSidebarCodexLinks();
  }
})();
</script>
"""


REVISION_ACTIONS = {
    "Clarity": "Simplify the line and sharpen the subject/action.",
    "Rhythm": "Rework beat pattern and remove drag.",
    "Read-aloud Flow": "Run a speak-test pass and cut mouth knots.",
    "Emotional Truth": "Anchor the feeling in the child-facing moment.",
    "Visual Strength": "Sharpen the drawable page beat.",
    "Commercial Publishability": "Tighten hook, age fit, and list-readiness.",
}


def _clean_path(uploaded_file) -> Path:
    return workbook_path(uploaded_file)


def _validate_workbook_path(path: Path) -> Path:
    if not path.exists():
        raise gr.Error(f"Workbook path does not exist: {path}")
    if path.suffix.lower() not in {".xlsx", ".xlsm"}:
        raise gr.Error("Upload or load an Excel workbook: .xlsx or .xlsm.")
    return path


def _score_summary(log_df: pd.DataFrame | None) -> str:
    if log_df is None or log_df.empty:
        return "No scored rows yet."
    scored = log_df[log_df["Weighted Score"].astype(str) != ""].copy()
    if scored.empty:
        return "No scored rows yet."
    scored["Weighted Score"] = pd.to_numeric(scored["Weighted Score"], errors="coerce")
    average = round(float(scored["Weighted Score"].mean()), 1)
    revisions = int((scored["Revision Flag"] == "Yes").sum())
    return f"{len(scored)} scored rows. Average weighted score {average}/10. Revision flags {revisions}."


def _metric_value(log_df: pd.DataFrame, metric: str, fallback: float) -> int:
    if log_df is not None and not log_df.empty and metric in log_df:
        values = pd.to_numeric(log_df[metric], errors="coerce").dropna()
        if not values.empty:
            return int(round(float(values.mean()) * 10))
    return int(round(fallback * 10))


def _small_spark(value: int, tone: str) -> str:
    heights = [19, 23, 16, 18, 17, 20, 22, 30, 25, 29, 34, 27]
    color = "#3f8f2f" if tone == "green" else "#dda10c" if tone == "gold" else "#d84f45"
    bars = "".join(f"<i style='height:{h}px;background:{color}'></i>" for h in heights)
    return f"<div class='spark' aria-label='score trend {value}'>{bars}</div>"


def _kpi_card(title: str, value: int, icon: str, tone: str, delta: str) -> str:
    return f"""
    <article class="kpi-card">
      <div class="kpi-top">
        <span class="kpi-icon {tone}">{icon}</span>
        <div><b>{escape(title)}</b><strong>{value}<em>/100</em></strong></div>
      </div>
      {_small_spark(value, tone)}
      <p class="{tone}">{escape(delta)}</p>
    </article>
    """


def _priority_badge(gate: str) -> tuple[str, str]:
    if gate in {"HARD FAIL", "SOFT FAIL", "READ-ALOUD BLOCK"}:
        return "High", "high"
    if gate in {"COMMERCIAL CHECK", "REVISE"}:
        return "Medium", "medium"
    return "Low", "low"


def _revision_rows(log_df: pd.DataFrame, tracker_df: pd.DataFrame) -> str:
    rows = []
    if log_df is not None and not log_df.empty:
        working = log_df.copy()
        working["Weighted Score"] = pd.to_numeric(working["Weighted Score"], errors="coerce")
        working = working.sort_values(["Revision Flag", "Weighted Score"], ascending=[False, True])
        for _, row in working.head(5).iterrows():
            metric = str(row.get("Priority Fix") or "Read-aloud Flow")
            gate = str(row.get("Gate") or "REVISE")
            priority, cls = _priority_badge(gate)
            block = str(row.get("Stanza ID") or row.get("Sequence") or "Live block")
            action = REVISION_ACTIONS.get(metric, "Revise the weakest pressure point first.")
            rows.append(
                f"""
                <div class="queue-row">
                  <span>{escape(block)}</span>
                  <span>{escape(metric)}</span>
                  <span>{escape(gate.title())}</span>
                  <b class="{cls}">{priority}</b>
                  <span>{escape(action)}</span>
                </div>
                """
            )

    if len(rows) < 5 and tracker_df is not None and not tracker_df.empty:
        for _, row in tracker_df.head(5 - len(rows)).iterrows():
            block = str(row.get("Block", "Block"))
            metric = str(row.get("TOTEM priority", "Read-aloud Flow"))
            action = str(row.get("Next action", REVISION_ACTIONS.get(metric, "Continue next pass.")))
            rows.append(
                f"""
                <div class="queue-row">
                  <span>{escape(block)}</span>
                  <span>{escape(metric)}</span>
                  <span>Development Gate</span>
                  <b class="medium">Medium</b>
                  <span>{escape(action)}</span>
                </div>
                """
            )

    return "\n".join(rows) or "<p class='empty'>No revision rows found.</p>"


def _risk_cards(log_df: pd.DataFrame) -> str:
    if log_df is None or log_df.empty:
        risks = [("Workbook Intake", "No scored rows detected yet.", "Medium Risk", "medium", "▁▂▃▂▁▂")]
    else:
        counts: dict[str, int] = {}
        for metric in METRICS:
            values = pd.to_numeric(log_df[metric], errors="coerce").dropna()
            weak = int((values < 7).sum())
            if weak:
                counts[metric] = weak
        if not counts:
            counts = {"Commercial Publishability": 1}
        ordered = sorted(counts.items(), key=lambda item: item[1], reverse=True)[:3]
        risks = []
        for metric, count in ordered:
            tone = "high" if count >= 2 else "medium"
            label = "High Risk" if tone == "high" else "Medium Risk"
            risks.append((metric, f"Detected under target in {count} scored block(s).", label, tone, "▂▅▃▇▁▆▂▅"))

    cards = []
    icons = {
        "Read-aloud Flow": "≋", "Rhythm": "≋", "Visual Strength": "◉",
        "Emotional Truth": "♡", "Commercial Publishability": "↗",
    }
    for metric, detail, label, tone, bars in risks:
        cards.append(
            f"""
            <div class="risk-row">
              <span class="risk-icon {tone}">{icons.get(metric, "!")}</span>
              <div><b>{escape(metric)}</b><small>{escape(detail)}</small></div>
              <em class="{tone}">{escape(label)}</em>
              <code>{escape(bars)}</code>
            </div>
            """
        )
    return "\n".join(cards)


def _recent_workbooks(path: Path, overall: int) -> str:
    name = path.stem.replace("_", " ")
    return f"""
    <div class="book-card">
      <div class="book-cover">TOTEM</div>
      <div>
        <b>{escape(name[:38])}</b>
        <small>Current workbook · Loaded now</small>
        <div class="progress"><i style="width:{max(8, min(overall, 100))}%"></i></div>
      </div>
      <strong>{overall}%</strong>
    </div>
    """


def dashboard_html(path: Path, notice: str = "") -> str:
    path = _validate_workbook_path(path)
    overview = workbook_overview(path)
    log_df = score_log(path)
    viability_df, viability_summary = viability_table(path)
    tracker_df = manuscript_tracker_table(path)
    workstack_df = workstack_table(path)

    avg_viability = float(viability_df["Score"].mean()) if viability_df is not None and not viability_df.empty else 0
    overall = int(round(avg_viability * 10)) if avg_viability else 0
    read_flow = _metric_value(log_df, "Read-aloud Flow", 6.4)
    emotional = _metric_value(log_df, "Emotional Truth", 7.8)
    visual = _metric_value(log_df, "Visual Strength", 6.9)
    commercial = _metric_value(log_df, "Commercial Publishability", avg_viability or 7.1)
    weakest = "Read-aloud Flow"
    strongest = "Emotional Truth"
    if log_df is not None and not log_df.empty:
        metric_means = {
            metric: pd.to_numeric(log_df[metric], errors="coerce").dropna().mean()
            for metric in METRICS
        }
        metric_means = {metric: value for metric, value in metric_means.items() if pd.notna(value)}
        if metric_means:
            weakest = min(metric_means, key=metric_means.get)
            strongest = max(metric_means, key=metric_means.get)

    next_item = ""
    if workstack_df is not None and not workstack_df.empty and "Status" in workstack_df:
        active = workstack_df[workstack_df["Status"].isin(["Active", "Queued"])]
        if not active.empty:
            next_item = str(active.iloc[0].get("Next item", "Run next pass"))
    next_item = next_item or "Run the next TOTEM pass"

    notice_block = f"<div class='notice'>{escape(notice)}</div>" if notice else ""

    return f"""
    <style>
      .studio-shell {{
        min-height: 900px;
        display: grid;
        grid-template-columns: 250px 1fr;
        background: var(--studio-cream);
      }}
      .studio-sidebar {{
        background: linear-gradient(180deg, #145b23 0%, #063f18 100%);
        color: white;
        padding: 22px 13px;
      }}
      .brand {{
        padding: 4px 10px 22px 10px;
      }}
      .brand h1 {{
        margin: 0;
        color: #f8e838;
        font-size: 28px;
        line-height: .82;
        text-shadow: 0 2px 0 #2b8c23;
        letter-spacing: 0;
      }}
      .brand small {{
        display: block;
        margin-top: 13px;
        letter-spacing: 4px;
        font-size: 11px;
      }}
      .nav-item {{
        display: flex;
        gap: 12px;
        align-items: center;
        padding: 14px 13px;
        margin: 6px 0;
        border-radius: 8px;
        color: #eef6e8;
        font-weight: 700;
      }}
      .nav-item.active {{
        background: linear-gradient(90deg, #f5c93c, #e5a721);
        color: white;
      }}
      .sidebar-foot {{
        margin-top: 170px;
        border-top: 1px solid rgba(255,255,255,.12);
        padding: 24px 10px;
        color: #d9ead0;
        font-size: 13px;
      }}
      .studio-main {{
        background: linear-gradient(180deg, #fffdf7 0%, #fff7de 42%, #fffaf0 100%);
      }}
      .topbar {{
        height: 72px;
        display: flex;
        align-items: center;
        justify-content: space-between;
        gap: 24px;
        padding: 0 28px;
        border-bottom: 1px solid var(--studio-line);
      }}
      .search {{
        flex: 1;
        max-width: 640px;
        border: 1px solid var(--studio-line);
        border-radius: 999px;
        padding: 13px 18px;
        background: rgba(255,255,255,.78);
        color: #8d8d78;
      }}
      .profile {{
        display: flex;
        align-items: center;
        gap: 14px;
        font-weight: 800;
      }}
      .avatar {{
        width: 46px;
        height: 46px;
        border-radius: 50%;
        background: linear-gradient(135deg, #c8915e, #f3d3ad);
        display: grid;
        place-items: center;
        color: #5b2b13;
      }}
      .hero {{
        position: relative;
        padding: 34px 36px 28px;
        min-height: 205px;
        overflow: hidden;
      }}
      .hero::before {{
        content: "";
        position: absolute;
        inset: 44px 0 0 0;
        background: radial-gradient(circle at 72% 32%, rgba(246,207,80,.28), transparent 18%),
                    radial-gradient(circle at 86% 65%, rgba(89,145,49,.12), transparent 20%),
                    linear-gradient(160deg, transparent 0 35%, rgba(238,192,66,.18) 36% 52%, transparent 53%);
      }}
      .hero h2 {{
        position: relative;
        margin: 0;
        font-family: Georgia, serif;
        font-size: 44px;
        color: #0e4a1d;
        letter-spacing: 0;
      }}
      .hero p {{
        position: relative;
        margin: 8px 0 22px;
        font-size: 20px;
        color: #355b35;
      }}
      .mascot {{
        position: absolute;
        right: 88px;
        top: 10px;
        width: 156px;
        height: 156px;
        border-radius: 50%;
        background: radial-gradient(circle at 48% 43%, #fff7db 0 18%, #e7b96c 19% 48%, #3f8f2f 49% 61%, transparent 62%),
                    radial-gradient(circle at 32% 11%, #f6d08f 0 13%, transparent 14%),
                    radial-gradient(circle at 70% 8%, #f6d08f 0 13%, transparent 14%);
        box-shadow: 0 15px 35px rgba(120,88,24,.18);
      }}
      .content {{
        padding: 0 28px 34px;
      }}
      .kpi-grid {{
        display: grid;
        grid-template-columns: repeat(5, minmax(150px, 1fr));
        gap: 14px;
      }}
      .kpi-card, .panel {{
        background: rgba(255,255,255,.88);
        border: 1px solid var(--studio-line);
        border-radius: 8px;
        box-shadow: 0 10px 28px rgba(58,44,14,.06);
      }}
      .kpi-card {{
        padding: 18px;
      }}
      .kpi-top {{
        display: flex;
        align-items: center;
        gap: 14px;
      }}
      .kpi-icon {{
        width: 48px;
        height: 48px;
        border-radius: 50%;
        display: grid;
        place-items: center;
        color: white;
        font-weight: 900;
        font-size: 24px;
      }}
      .green {{ color: #348426; }}
      .gold {{ color: #d69200; }}
      .red {{ color: #d84f45; }}
      .kpi-icon.green {{ background: #4e982f; color: white; }}
      .kpi-icon.gold {{ background: #e6a500; color: white; }}
      .kpi-icon.red {{ background: #d84f45; color: white; }}
      .kpi-card b {{
        display: block;
        font-size: 13px;
        color: var(--studio-ink);
      }}
      .kpi-card strong {{
        display: block;
        color: var(--studio-ink);
        font-size: 34px;
        line-height: 1;
      }}
      .kpi-card em {{
        font-size: 13px;
        font-style: normal;
        color: var(--studio-muted);
      }}
      .spark {{
        height: 42px;
        display: flex;
        align-items: end;
        gap: 8px;
        margin: 12px 0 5px;
      }}
      .spark i {{
        width: 4px;
        border-radius: 4px;
      }}
      .kpi-card p {{
        margin: 0;
        font-size: 12px;
        background: transparent !important;
      }}
      .dashboard-grid {{
        display: grid;
        grid-template-columns: 1.55fr 1fr;
        gap: 14px;
        margin-top: 16px;
      }}
      .panel h3 {{
        margin: 0;
        padding: 16px 18px;
        border-bottom: 1px solid var(--studio-line);
        color: var(--studio-ink);
      }}
      .queue-row {{
        display: grid;
        grid-template-columns: 1.15fr .9fr .85fr .55fr 1.35fr;
        gap: 10px;
        align-items: center;
        padding: 12px 18px;
        border-bottom: 1px solid #f0e7cf;
        font-size: 13px;
      }}
      .queue-row b, .risk-row em {{
        text-align: center;
        border-radius: 999px;
        padding: 5px 9px;
        font-style: normal;
      }}
      .high {{ background: #ffe3da; color: var(--studio-red); }}
      .medium {{ background: #fff0c5; color: #c28600; }}
      .low {{ background: #e9f5dc; color: #4b8e2e; }}
      .risk-row {{
        display: grid;
        grid-template-columns: 54px 1fr 96px 86px;
        gap: 12px;
        align-items: center;
        padding: 14px 18px;
        border-bottom: 1px solid #f0e7cf;
      }}
      .risk-icon {{
        width: 46px;
        height: 46px;
        display: grid;
        place-items: center;
        border-radius: 50%;
        color: white;
        font-weight: 900;
        background: #e6a500;
      }}
      .risk-icon.high {{ background: #d84f45; color: white; }}
      .risk-icon.medium {{ background: #e6a500; color: white; }}
      .risk-row small {{
        display: block;
        color: var(--studio-muted);
        margin-top: 4px;
      }}
      .risk-row code {{
        color: #d84f45;
        font-size: 24px;
        letter-spacing: 1px;
      }}
      .bottom-grid {{
        display: grid;
        grid-template-columns: 1fr 1fr;
        gap: 14px;
        margin-top: 16px;
      }}
      .book-card {{
        display: grid;
        grid-template-columns: 58px 1fr 48px;
        align-items: center;
        gap: 14px;
        padding: 14px 18px;
      }}
      .book-cover {{
        width: 54px;
        height: 76px;
        display: grid;
        place-items: center;
        border-radius: 6px;
        background: linear-gradient(160deg, #072d18, #135b25);
        color: #f2d444;
        font-weight: 900;
        font-size: 11px;
      }}
      .book-card small {{
        display: block;
        color: var(--studio-muted);
        margin: 5px 0 12px;
      }}
      .progress {{
        height: 7px;
        border-radius: 999px;
        background: #eadfbd;
      }}
      .progress i {{
        display: block;
        height: 7px;
        border-radius: 999px;
        background: #4e982f;
      }}
      .snapshot {{
        display: grid;
        grid-template-columns: repeat(3, 1fr);
        gap: 8px;
        padding: 18px;
      }}
      .snapshot div {{
        border-right: 1px solid var(--studio-line);
        min-height: 82px;
      }}
      .snapshot div:last-child {{
        border-right: none;
      }}
      .snapshot small {{
        color: var(--studio-muted);
        display: block;
      }}
      .snapshot b {{
        display: block;
        margin-top: 9px;
      }}
      .snapshot strong {{
        font-size: 28px;
      }}
      .notice {{
        margin: 0 28px 16px;
        padding: 12px 14px;
        background: #e9f5dc;
        border: 1px solid #c5ddb5;
        border-radius: 8px;
        font-weight: 700;
      }}
      @media (max-width: 1100px) {{
        .studio-shell {{ grid-template-columns: 1fr; }}
        .studio-sidebar {{ display: none; }}
        .kpi-grid, .dashboard-grid, .bottom-grid {{ grid-template-columns: 1fr; }}
        .mascot {{ display: none; }}
        .queue-row {{ grid-template-columns: 1fr; }}
      }}
    </style>
    <div class="studio-shell">
      <aside class="studio-sidebar">
        <div class="brand">
          <h1>TOTEM<br>Studio</h1>
          <small>PUBLISHING INTERFACE</small>
        </div>
        <div class="nav-item active">⌂ Home</div>
        <div class="nav-item">▣ Projects</div>
        <div class="nav-item">▤ Workbook Upload</div>
        <div class="nav-item">▥ TOTEM Analytics</div>
        <div class="nav-item">☷ Revision Queue</div>
        <div class="nav-item">⬡ Risk Clusters</div>
        <div class="nav-item js-open-codex" role="button" tabindex="0" aria-label="Open Codex Extractor tab">◈ Codex Extractor</div>
        <div class="nav-item">⇩ Export</div>
        <div class="sidebar-foot">Knowledge. Structure.<br>Story. Performance.</div>
      </aside>
      <main class="studio-main">
        <div class="topbar">
          <div class="search">Search projects, workbooks, blocks...</div>
          <div class="profile"><span>Editorial Workspace</span><span>🔔</span><span class="avatar">T</span><span>TOTEM<br><small>Studio</small></span></div>
        </div>
        <section class="hero">
          <h2>TOTEM Studio.</h2>
          <p>Data-driven insight for stronger stories.</p>
          <div class="mascot" aria-label="TOTEM Studio mascot"></div>
        </section>
        {notice_block}
        <section class="content">
          <div class="kpi-grid">
            {_kpi_card("Overall Publishability", overall, "✦", "green", "Source: viability lens")}
            {_kpi_card("Read-Aloud Flow", read_flow, "≋", "red" if read_flow < 65 else "gold", "Weakest live pressure")}
            {_kpi_card("Emotional Truth", emotional, "♡", "green", "Strongest story signal")}
            {_kpi_card("Visual Strength", visual, "◉", "gold" if visual < 75 else "green", "Drawable page value")}
            {_kpi_card("Commercial Viability", commercial, "↗", "green", "Publisher-facing lens")}
          </div>
          <div class="dashboard-grid">
            <section class="panel">
              <h3>Revision Priority Queue <small>{len(log_df) if log_df is not None else 0} live item(s)</small></h3>
              <div class="queue-row header"><b>Block</b><b>Weakest Dimension</b><b>Gate</b><b>Priority</b><b>Recommended Action</b></div>
              {_revision_rows(log_df, tracker_df)}
            </section>
            <section class="panel">
              <h3>Risk Clusters</h3>
              {_risk_cards(log_df)}
            </section>
          </div>
          <div class="bottom-grid">
            <section class="panel">
              <h3>Recent Workbook</h3>
              {_recent_workbooks(path, overall)}
            </section>
            <section class="panel">
              <h3>TOTEM Snapshot <small>Based on latest run</small></h3>
              <div class="snapshot">
                <div><small>Weakest Dimension</small><b>{escape(weakest)}</b><strong>{read_flow}</strong><small>/100</small></div>
                <div><small>Strongest Dimension</small><b>{escape(strongest)}</b><strong>{max(emotional, visual)}</strong><small>/100</small></div>
                <div><small>Next Work</small><b>{escape(next_item[:42])}</b><small>{escape(viability_summary)}</small></div>
              </div>
            </section>
          </div>
          <p style="color:#6d725f;font-size:12px;margin:18px 0 0;">Loaded {escape(path.name)} · {overview['sheet_count']} sheets read privately · matrix hidden from the product surface.</p>
        </section>
      </main>
    </div>
    """


# ── CODEX EXTRACTOR FUNCTIONS ─────────────────────────────────────────────────

def _extract_uploaded_path(file_obj) -> str | None:
    """Handle Gradio file payload variants and return a filesystem path."""
    if file_obj is None:
        return None
    if isinstance(file_obj, str):
        return file_obj
    if isinstance(file_obj, dict):
        return file_obj.get("path") or file_obj.get("name")
    if hasattr(file_obj, "name"):
        return file_obj.name
    return None


def _normalise_lookup_key(value: str) -> str:
    text = str(value or "").lower()
    text = re.sub(r"[_\-]+", " ", text)
    text = re.sub(r"[^a-z0-9 ]+", " ", text)
    return re.sub(r"\s+", " ", text).strip()


def _generate_codex_author_id(author_name: str) -> str:
    """
    Deterministic fallback author ID when missing in catalogue.
    Format: CA-<3 letters>-<3 digits>
    """
    cleaned = re.sub(r"[^A-Za-z]", "", author_name or "").upper()
    prefix = (cleaned[:3] or "AUT").ljust(3, "X")
    digest = hashlib.md5((author_name or "").strip().lower().encode("utf-8")).hexdigest()
    suffix = int(digest[:4], 16) % 1000
    return f"CA-{prefix}-{suffix:03d}"


def _empty_catalogue_df() -> pd.DataFrame:
    return pd.DataFrame(columns=["authour_id", "author_name", "title"])


def _load_codex_catalogue(path: Path = CODEX_CATALOGUE_PATH) -> pd.DataFrame:
    """
    Load catalogue workbook with required columns:
    `authour_id`, `author_name`, `title`
    """
    if not path.exists():
        return _empty_catalogue_df()
    try:
        raw = pd.read_excel(path)
    except Exception:
        return _empty_catalogue_df()
    if raw is None or raw.empty:
        return _empty_catalogue_df()

    col_lookup = {str(col).strip().lower(): col for col in raw.columns}
    id_col = col_lookup.get("authour_id") or col_lookup.get("author_id")
    name_col = col_lookup.get("author_name")
    title_col = col_lookup.get("title")

    if name_col is None or title_col is None:
        return _empty_catalogue_df()

    if id_col is None:
        raw["__authour_id"] = ""
        id_col = "__authour_id"

    cat = raw[[id_col, name_col, title_col]].copy()
    cat.columns = ["authour_id", "author_name", "title"]

    for col in ["authour_id", "author_name", "title"]:
        cat[col] = cat[col].fillna("").astype(str).str.strip()

    cat = cat[(cat["author_name"] != "") & (cat["title"] != "")]
    return cat


def _match_catalogue_row(file_path: str, catalogue: pd.DataFrame) -> pd.Series | None:
    if catalogue.empty:
        return None
    stem_key = _normalise_lookup_key(Path(file_path).stem)
    if not stem_key:
        return None

    title_keys = catalogue["title"].map(_normalise_lookup_key)

    exact = catalogue[title_keys == stem_key]
    if not exact.empty:
        return exact.iloc[0]

    contains = catalogue[
        title_keys.apply(lambda t: bool(t) and (t in stem_key or stem_key in t))
    ]
    if not contains.empty:
        return contains.assign(_key_len=contains["title"].map(lambda t: len(_normalise_lookup_key(t)))) \
                       .sort_values("_key_len", ascending=False) \
                       .iloc[0]
    return None


def autofill_codex_details(
    file_obj,
    current_author_name: str,
    current_author_id: str,
    current_works: str,
) -> tuple[str, str, str, str]:
    """
    Auto-populate author fields from data/codex_catalogue.xlsx on file upload.
    Expected columns: authour_id, author_name, title.
    """
    file_path = _extract_uploaded_path(file_obj)
    if not file_path:
        return (
            current_author_name,
            current_author_id or "CA-XXX",
            current_works,
            "Ready.",
        )

    catalogue = _load_codex_catalogue()
    if catalogue.empty:
        return (
            current_author_name,
            current_author_id or "CA-XXX",
            current_works,
            "No catalogue match: add rows to data/codex_catalogue.xlsx with authour_id, author_name, title.",
        )

    row = _match_catalogue_row(file_path, catalogue)
    if row is None:
        return (
            current_author_name,
            current_author_id or "CA-XXX",
            current_works,
            "No title match found in catalogue for this filename. You can still fill fields manually.",
        )

    author_name = str(row["author_name"]).strip()
    author_id = str(row["authour_id"]).strip() or _generate_codex_author_id(author_name)

    # Populate only the matched title to avoid confusion during extraction.
    works_sampled = str(row["title"]).strip()

    return (
        author_name,
        author_id,
        works_sampled,
        f"Auto-filled from catalogue: {author_name} ({author_id}).",
    )


def run_codex_extraction(
    file_obj,
    author_name: str,
    author_id: str,
    works_sampled: str,
) -> tuple[str, str, str]:
    """
    Gradio handler for the Codex Extraction tab.
    Returns (report_text, json_output) tuple.
    """
    if file_obj is None:
        return (
            "No file uploaded. Please upload a .txt or .pdf file.",
            "",
            "ERROR: No file uploaded.",
        )

    if not author_name.strip():
        return (
            "Please enter the author's full name before extracting.",
            "",
            "ERROR: Author name is required.",
        )

    # Gradio may pass a filepath string or a file-like payload depending on runtime.
    file_path = _extract_uploaded_path(file_obj)

    if not file_path:
        return (
            "Unable to read uploaded file path. Please re-upload and try again.",
            "",
            "ERROR: File payload missing path.",
        )

    try:
        report, fp_dict = process_upload(
            file_path=file_path,
            author_name=author_name.strip(),
            author_id=author_id.strip() or "CA-XXX",
            works_sampled=works_sampled.strip(),
        )

        if not fp_dict:
            return report, "", "ERROR: Extraction failed. See report for details."

        # Format JSON output for workbook entry
        json_out = json.dumps(fp_dict, indent=2)
        status = f"SUCCESS: Fingerprint extracted ({fp_dict.get('Sample_Words', 0)} words analysed)."
        return report, json_out, status

    except Exception as e:
        return (
            f"Extraction error: {type(e).__name__}: {str(e)}",
            "",
            f"ERROR: {type(e).__name__}",
        )


def clear_codex_form() -> tuple[None, str, str, str, str, str, str]:
    """Reset the Codex extraction form."""
    return None, "", "CA-XXX", "", "", "", "Ready."


# ── WORKBOOK FUNCTIONS ────────────────────────────────────────────────────────

def load_workbook(uploaded_file=None, notice: str = ""):
    path = _validate_workbook_path(_clean_path(uploaded_file))
    log_df = score_log(path)
    return str(path), dashboard_html(path, notice), log_df, _score_summary(log_df)


def load_default():
    return load_workbook(None, "Bundled workbook reloaded.")


def load_uploaded(uploaded_file):
    if uploaded_file is None:
        raise gr.Error("Choose an .xlsx or .xlsm workbook first.")
    return load_workbook(uploaded_file, "Workbook uploaded and analysed.")


def load_local_path(path_text: str):
    path = _validate_workbook_path(Path(path_text or "").expanduser())
    log_df = score_log(path)
    return str(path), dashboard_html(path, "Local workbook loaded."), log_df, _score_summary(log_df)


def run_analysis(active_path: str):
    path = _validate_workbook_path(Path(active_path) if active_path else DEFAULT_WORKBOOK)
    log_df = score_log(path)
    return dashboard_html(path, "TOTEM analysis refreshed."), log_df, _score_summary(log_df)


def recalc_log(log_df, active_path: str):
    path = _validate_workbook_path(Path(active_path) if active_path else DEFAULT_WORKBOOK)
    recalculated = recalculate_log(log_df, path)
    return dashboard_html(path, "Gates recalculated."), recalculated, _score_summary(recalculated)


def export_log(log_df, active_path: str):
    path = _validate_workbook_path(Path(active_path) if active_path else DEFAULT_WORKBOOK)
    return export_updated_workbook(log_df, path)


def single_score(active_path, sequence, stanza_id, draft_pass, clarity, rhythm, flow,
                 emotional_truth, visual_strength, commercial, notes):
    path = _validate_workbook_path(Path(active_path) if active_path else DEFAULT_WORKBOOK)
    df = score_single_row(path, sequence, stanza_id, draft_pass, clarity, rhythm,
                          flow, emotional_truth, visual_strength, commercial, notes)
    return df, _score_summary(df)


def initial_dashboard_html() -> str:
    """
    Render dashboard shell at startup so sidebar/nav are visible immediately.
    Falls back gracefully if workbook read fails.
    """
    try:
        return dashboard_html(DEFAULT_WORKBOOK, "Dashboard ready. Click Run TOTEM Analysis to refresh metrics.")
    except Exception as exc:
        return (
            "<div style='padding:24px;font-family:system-ui,sans-serif'>"
            "<h3 style='margin:0 0 8px'>TOTEM Dashboard</h3>"
            "<p style='margin:0 0 12px;color:#666'>"
            "Dashboard shell is ready. Click <b>Run TOTEM Analysis</b> to load workbook data."
            "</p>"
            f"<p style='margin:0;color:#888;font-size:12px'>Startup note: {escape(str(exc))}</p>"
            "</div>"
        )


# ── GRADIO INTERFACE ──────────────────────────────────────────────────────────

with gr.Blocks(title="TOTEM Studio") as demo:
    active_path = gr.State(str(DEFAULT_WORKBOOK))
    log_state = gr.State(pd.DataFrame(columns=LOG_COLUMNS))

    with gr.Tabs():

        # ── TAB 1: DASHBOARD ─────────────────────────────────────────────────
        with gr.TabItem("Dashboard"):
            dashboard = gr.HTML(value=initial_dashboard_html())

            with gr.Row(elem_id="studio-actions"):
                with gr.Column(elem_classes=["wrap"]):
                    workbook_upload = gr.UploadButton(
                        "Upload Workbook",
                        file_types=[".xlsx", ".xlsm"],
                        type="filepath",
                        variant="primary",
                        scale=1,
                    )
                    run_button = gr.Button("Run TOTEM Analysis", variant="secondary", scale=1)

            with gr.Row(elem_id="path-panel"):
                with gr.Column(elem_classes=["wrap"]):
                    path_input = gr.Textbox(label="Local workbook path", value=ORIGINAL_WORKBOOK_PATH)
                    path_button = gr.Button("Load Local Path", variant="primary")

            with gr.Accordion("Private scoring controls", open=False, elem_id="score-panel"):
                gr.HTML("<div class='score-card'><h3>Live Score A Block</h3></div>")
                with gr.Row():
                    sequence = gr.Textbox(label="Sequence", value="Live pass")
                    stanza_id = gr.Textbox(label="Stanza ID", value="New block")
                    draft_pass = gr.Textbox(label="Draft / Pass", value="First score")
                with gr.Row():
                    clarity = gr.Slider(1, 10, value=7, step=0.5, label="Clarity")
                    rhythm = gr.Slider(1, 10, value=7, step=0.5, label="Rhythm")
                    flow = gr.Slider(1, 10, value=7, step=0.5, label="Read-aloud Flow")
                with gr.Row():
                    emotional_truth = gr.Slider(1, 10, value=7, step=0.5, label="Emotional Truth")
                    visual_strength = gr.Slider(1, 10, value=7, step=0.5, label="Visual Strength")
                    commercial = gr.Slider(1, 10, value=7, step=0.5, label="Commercial Publishability")
                notes = gr.Textbox(label="Notes", lines=2)
                with gr.Row():
                    single_button = gr.Button("Score Block", variant="primary")
                    recalc_button = gr.Button("Recalculate Gates")
                    export_button = gr.Button("Download Updated Workbook")
                single_df = gr.Dataframe(label="Latest scorecard", interactive=False, visible=False)
                score_status = gr.Markdown(elem_id="hidden-status")
                exported_file = gr.File(label="Export appears here", elem_id="hidden-export")

        # ── TAB 2: CODEX EXTRACTOR ───────────────────────────────────────────
        with gr.TabItem("◈ Codex Extractor"):
            gr.HTML("""
            <div style="
                background: linear-gradient(135deg, #0e4a1d, #17642a);
                border-radius: 8px;
                padding: 20px 24px;
                margin-bottom: 16px;
            ">
                <h3 style="margin:0; color:#f8e838; font-family:Georgia,serif; font-size:20px;">
                    Codex Fingerprint Extractor
                </h3>
                <p style="margin:8px 0 0; color:#d9ead0; font-size:13px;">
                    Upload an author's text or PDF. The extractor computes 17 Tier 1 voice metrics
                    (VM-001 to VM-013, VM-024 to VM-028) mathematically from the text.
                    Copy the output into CODEX_03_FINGERPRINTS in the workbook.
                    Use Codex Build Prompt 2 (ChatGPT/Gemini) for the 10 Tier 2 qualitative metrics.
                </p>
            </div>
            """)

            with gr.Row():
                with gr.Column(scale=1):
                    gr.Markdown("### Author Details")
                    codex_author_name = gr.Textbox(
                        label="Author Full Name",
                        placeholder="e.g. Julia Donaldson",
                    )
                    codex_author_id = gr.Textbox(
                        label="Codex Author ID",
                        placeholder="e.g. CA-001",
                        value="CA-XXX",
                    )
                    codex_works = gr.Textbox(
                        label="Works Sampled (comma-separated)",
                        placeholder="e.g. The Gruffalo, Room on the Broom, Zog",
                        lines=2,
                    )
                    codex_file = gr.File(
                        label="Upload Text or PDF",
                        file_types=[".txt", ".pdf"],
                        type="filepath",
                    )
                    with gr.Row():
                        codex_extract_btn = gr.Button(
                            "Extract Fingerprint",
                            variant="primary",
                            scale=2,
                        )
                        codex_clear_btn = gr.Button(
                            "Clear",
                            variant="secondary",
                            scale=1,
                        )
                    codex_status = gr.Textbox(
                        label="Extractor Status",
                        lines=2,
                        interactive=False,
                        value="Ready.",
                        placeholder="Status and extraction errors will appear here.",
                    )

                    gr.HTML("""
                    <div style="
                        background: #fff8e1;
                        border: 1px solid #ffe082;
                        border-radius: 6px;
                        padding: 12px 14px;
                        margin-top: 12px;
                        font-size: 13px;
                        color: #5d4037;
                    ">
                        <b>File requirements:</b><br>
                        • Selectable text preferred; scanned PDFs are OCR-processed automatically<br>
                        • Minimum 1,000 words for HIGH confidence fingerprint<br>
                        • Combine multiple works in one file to increase sample size<br>
                        • Visual-primary books (Van Allsburg, Jeffers) will flag LOW confidence
                    </div>
                    """)

                with gr.Column(scale=2):
                    gr.Markdown("### Extraction Report — Tier 1 Metrics")
                    codex_report = gr.Textbox(
                        label="",
                        lines=32,
                        interactive=False,
                        placeholder="Upload a file and click Extract Fingerprint to see results here...",
                        elem_id="codex-report",
                    )

            gr.Markdown("### Raw Output — Copy into CODEX_03_FINGERPRINTS")
            codex_json = gr.Textbox(
                label="",
                lines=20,
                interactive=False,
                placeholder="JSON values appear here after extraction. Copy individual metric values into the workbook row.",
                elem_id="codex-json",
            )

            gr.HTML("""
            <div style="
                background: #e8f5e9;
                border: 1px solid #a5d6a7;
                border-radius: 6px;
                padding: 14px 18px;
                margin-top: 8px;
                font-size: 13px;
            ">
                <b style="color:#1b5e20;">Tier 2 reminder:</b>
                <span style="color:#2e7d32;">
                VM-014 (Narrative person) through VM-023 (Animal/nature imagery ratio) require
                qualitative judgment. Use Codex Build Prompt 2 from the Codex Build Prompts document
                with the same text in ChatGPT or Gemini to complete the remaining 10 metrics.
                </span>
            </div>
            """)

        # ── SMOKE SIGNAL TAB ─────────────────────────────────────────────────
        smoke_signal_tab()

    # ── EVENT WIRING ─────────────────────────────────────────────────────────

    # Dashboard tab
    # demo.load removed: user triggers load via Run TOTEM Analysis button
    workbook_upload.upload(load_uploaded, inputs=[workbook_upload],
                           outputs=[active_path, dashboard, log_state, score_status])
    run_button.click(run_analysis, inputs=[active_path],
                     outputs=[dashboard, log_state, score_status])
    path_button.click(load_local_path, inputs=[path_input],
                      outputs=[active_path, dashboard, log_state, score_status])
    recalc_button.click(recalc_log, inputs=[log_state, active_path],
                        outputs=[dashboard, log_state, score_status])
    export_button.click(export_log, inputs=[log_state, active_path], outputs=[exported_file])
    single_button.click(
        single_score,
        inputs=[active_path, sequence, stanza_id, draft_pass, clarity, rhythm,
                flow, emotional_truth, visual_strength, commercial, notes],
        outputs=[single_df, score_status],
    )

    # Codex Extractor tab
    codex_extract_btn.click(
        run_codex_extraction,
        inputs=[codex_file, codex_author_name, codex_author_id, codex_works],
        outputs=[codex_report, codex_json, codex_status],
        trigger_mode="multiple",
    )
    # Auto-fill author metadata from catalogue on file upload.
    codex_file.upload(
        autofill_codex_details,
        inputs=[codex_file, codex_author_name, codex_author_id, codex_works],
        outputs=[codex_author_name, codex_author_id, codex_works, codex_status],
        trigger_mode="always_last",
    )
    codex_clear_btn.click(
        clear_codex_form,
        outputs=[codex_file, codex_author_name, codex_author_id, codex_works,
                 codex_report, codex_json, codex_status],
    )


if __name__ == "__main__":
    demo.launch(ssr_mode=False, css=CSS + SS_CSS, head=HEAD)