Update app.py

app.py

@@ -1,9 +1,9 @@
 # ======================================================
 # Federal FOIA Intelligence Search
-# HF Reviewer–Safe / Court-Safe Reference Implementation
+# Court-Safe / HF-Reviewer-Safe Reference Implementation
 # ======================================================
 
-import os, io, zipfile, tempfile, hashlib, base64, time
+import os, io, zipfile, tempfile, hashlib, base64
 from datetime import datetime
 from urllib.parse import quote_plus
 import requests
@@ -13,33 +13,24 @@ from fastapi import FastAPI
 from fastapi.staticfiles import StaticFiles
 from fastapi.responses import JSONResponse
 
-# Optional PDF support
-PDF_THUMBNAILS_AVAILABLE = False
-PDF_TEXT_AVAILABLE = False
+# ======================================================
+# OPTIONAL PDF SUPPORT
+# ======================================================
 
+PDF_THUMBNAILS_AVAILABLE = False
 try:
     from pdf2image import convert_from_bytes
     PDF_THUMBNAILS_AVAILABLE = True
 except Exception:
     pass
 
-try:
-    from pdfminer.high_level import extract_text
-    PDF_TEXT_AVAILABLE = True
-except Exception:
-    pass
-
-from reportlab.platypus import SimpleDocTemplate, Paragraph, PageBreak
-from reportlab.lib.styles import getSampleStyleSheet
-from reportlab.lib.pagesizes import LETTER
-
 # ======================================================
-# HARD GOVERNANCE FLAGS (NON-NEGOTIABLE)
+# GOVERNANCE FLAGS (HARD)
 # ======================================================
 
-ENABLE_AI = True                       # USER-INITIATED ONLY
-ENABLE_FAISS_PHASE_4 = False           # FORMAL APPROVAL REQUIRED
-ENABLE_DOC_LEVEL_APIS = False          # CIA/FBI DO NOT CURRENTLY PROVIDE
+ENABLE_AI = True
+ENABLE_FAISS_PHASE_4 = False          # FORMAL APPROVAL REQUIRED
+ENABLE_DOC_LEVEL_APIS = False         # CIA/FBI DO NOT PROVIDE
 
 # ======================================================
 # SESSION STATE (EPHEMERAL)
@@ -49,7 +40,7 @@ LAST_RESULTS = []
 AI_APPENDIX = None
 
 # ======================================================
-# CRYPTOGRAPHIC CORE
+# CRYPTOGRAPHY
 # ======================================================
 
 def sha256_text(t: str) -> str:
@@ -57,7 +48,7 @@ def sha256_text(t: str) -> str:
 
 def provenance_block(payload: str, ai=False) -> str:
     return "\n".join([
-        "Tool-Version: 2.0.0",
+        "Tool-Version: 2.1.0",
         f"Generated-UTC: {datetime.utcnow().isoformat()}",
         f"Content-SHA256: {sha256_text(payload)}",
         "Public-Source-Only: true",
@@ -66,7 +57,7 @@ def provenance_block(payload: str, ai=False) -> str:
     ])
 
 # ======================================================
-# FOIA ADAPTERS (LINK-OUT ONLY — ACCURATE)
+# FOIA ADAPTERS (LINK-OUT ONLY)
 # ======================================================
 
 class FOIAAdapter:
@@ -81,7 +72,7 @@ class FOIAAdapter:
             "resolved_url": url,
             "timestamp": datetime.utcnow().isoformat(),
             "is_pdf": False,
-            "thumbnails": []
+            "thumbnails": [],
         }]
 
 class CIA(FOIAAdapter):
@@ -108,7 +99,7 @@ ALL_ADAPTERS = {
 }
 
 # ======================================================
-# PDF DETECTION (SAFE — NO SCRAPING)
+# PDF DETECTION (SAFE)
 # ======================================================
 
 def resolve_pdf(url):
@@ -157,7 +148,7 @@ def run_search(query, agencies):
     return rows, render_cards(), "Search complete."
 
 # ======================================================
-# ASK-AI (STRICTLY GATED)
+# ASK-AI (OPT-IN)
 # ======================================================
 
 def ask_ai(index: int):
@@ -165,25 +156,25 @@ def ask_ai(index: int):
     r = LAST_RESULTS[index]
 
     if not (ENABLE_AI and r["is_pdf"]):
-        return "AI is disabled for this result."
+        return "AI assistance is unavailable for this result."
 
-    summary = (
+    text = (
         "AI Assistive Summary (Non-Authoritative)\n\n"
         f"Agency: {r['agency']}\n"
         f"Source: {r['resolved_url']}\n\n"
-        "This summary assists review of a public FOIA document only."
+        "This assists review of a public FOIA document only."
     )
 
     AI_APPENDIX = {
-        "text": summary,
-        "hash": sha256_text(summary),
-        "prov": provenance_block(summary, ai=True)
+        "text": text,
+        "hash": sha256_text(text),
+        "prov": provenance_block(text, ai=True),
     }
 
-    return summary + "\n\n" + AI_APPENDIX["prov"]
+    return text + "\n\n" + AI_APPENDIX["prov"]
 
 # ======================================================
-# COURT BUNDLE (CM/ECF-READY)
+# COURT BUNDLE
 # ======================================================
 
 def generate_court_bundle():
@@ -202,30 +193,18 @@ def generate_court_bundle():
                 z.writestr("Exhibit_AI_Appendix.txt", AI_APPENDIX["text"])
                 z.writestr("Exhibit_AI_Appendix.provenance.txt", AI_APPENDIX["prov"])
 
-            z.writestr("Judicial_Notice.txt", JUDICIAL_NOTICE)
-            z.writestr("HF_Reviewer_Cover_Letter.txt", HF_REVIEWER_COVER_LETTER)
-
         return path
 
 # ======================================================
-# STATIC GOVERNANCE TEXT
+# GOVERNANCE SITE LOADING
 # ======================================================
 
-JUDICIAL_NOTICE = """
-This system provides navigation to public FOIA reading rooms only.
-It does not host, certify, authenticate, or modify records.
-Authoritative documents remain with issuing agencies.
-"""
-
-HF_REVIEWER_COVER_LETTER = """
-This Hugging Face Space is a governance-first reference implementation.
-
-• Link-out only
-• Public FOIA sources only
-• AI is opt-in, hashed, and user-initiated
-• No document scraping or indexing
-• Court-safe by design
-"""
+def load_governance_md():
+    path = "governance-site/index.md"
+    if os.path.exists(path):
+        with open(path, "r", encoding="utf-8") as f:
+            return f.read()
+    return "_Governance documentation not found._"
 
 # ======================================================
 # FASTAPI
@@ -245,33 +224,36 @@ if os.path.exists("governance-site"):
 # ======================================================
 
 CSS = """
-button { border-radius:999px !important; }
-.tab-nav { position:sticky; top:0; background:#fff; z-index:999; }
+.tab-nav { position: sticky; top: 0; z-index: 999; background: #0f0f0f; }
+@media (max-width: 768px) {
+  .desktop-only { display: none !important; }
+}
+button { border-radius: 999px !important; }
 """
 
 def render_cards():
     cards = []
     for i, r in enumerate(LAST_RESULTS):
-        thumbs = "".join(
-            f'<img src="data:image/png;base64,{t}" style="width:120px;border-radius:8px;margin-right:6px;" />'
-            for t in r["thumbnails"]
+        download_html = (
+            f'<a href="{r["resolved_url"]}" download>Download</a>'
+            if r["is_pdf"]
+            else '<span style="opacity:.4">Download</span>'
         )
-        disabled = "" if r["is_pdf"] else "opacity:0.4;pointer-events:none;"
+
         cards.append(f"""
-        <div style="border:1px solid #ddd;border-radius:16px;padding:16px;margin-bottom:16px;">
+        <div style="border:1px solid #333;border-radius:18px;padding:16px;margin-bottom:16px;">
           <b>{r['agency']}</b><br/>
           {r['title']}<br/>
-          {thumbs}
           <div style="margin-top:10px;">
             <a href="{r['resolved_url']}" target="_blank">View</a> |
-            <a href="{r['resolved_url']}" download style="{disabled}">Download</a> |
+            {download_html} |
             <a href="{r['resolved_url']}" target="_blank">Share</a>
-            <button style="background:#1e88e5;color:white;padding:4px 12px;margin-left:10px;border:none;"
+            <button style="background:#1e88e5;color:white;padding:4px 14px;margin-left:10px;border:none;"
               onclick="fetch('/ask_ai?index={i}')">
               Ask AI
             </button>
           </div>
-          <div style="font-size:0.75em;color:#666;margin-top:6px;">
+          <div style="font-size:.75em;color:#777;margin-top:6px;">
             Why am I seeing this? This is a public FOIA reading-room result.
           </div>
         </div>
@@ -284,18 +266,19 @@ with gr.Blocks(css=CSS) as demo:
     with gr.Tab("Search"):
         agencies = gr.CheckboxGroup(list(ALL_ADAPTERS.keys()), value=list(ALL_ADAPTERS.keys()))
         query = gr.Textbox(placeholder="Search FOIA reading rooms")
-        table = gr.Dataframe(headers=["Agency", "Title", "URL", "Hash"])
+        table = gr.Dataframe(
+            headers=["Agency", "Title", "URL", "Hash"],
+            elem_classes=["desktop-only"],
+        )
         cards = gr.HTML()
         status = gr.Textbox()
-        gr.Button("Search", elem_classes=["primary"]).click(
-            run_search, [query, agencies], [table, cards, status]
-        )
+        gr.Button("Search").click(run_search, [query, agencies], [table, cards, status])
 
     with gr.Tab("Court / Clerk"):
         gr.Button("Generate Court Bundle").click(lambda: generate_court_bundle(), None, gr.File())
 
     with gr.Tab("Governance"):
-        gr.Markdown(HF_REVIEWER_COVER_LETTER)
+        gr.Markdown(load_governance_md())
 
 demo.queue()
 demo.launch(server_name="0.0.0.0", server_port=7860)