Update app.py

app.py

@@ -5,13 +5,12 @@ import io
 import json
 import zipfile
 import base64
-from datetime import datetime
+from datetime import datetime, timedelta
 from urllib.parse import quote_plus
 import requests
+import os
 
-from reportlab.platypus import (
-    SimpleDocTemplate, Paragraph, Spacer, PageBreak
-)
+from reportlab.platypus import SimpleDocTemplate, Paragraph, Spacer, PageBreak
 from reportlab.lib.styles import getSampleStyleSheet
 
 from citations import bluebook_exhibit, table_of_authorities
@@ -34,7 +33,6 @@ PDF_THUMBNAIL_AVAILABLE = False
 
 try:
     from pdfminer.high_level import extract_text
-    from pdfminer.pdfpage import PDFPage
     PDF_TEXT_AVAILABLE = True
 except Exception:
     pass
@@ -53,17 +51,54 @@ LAST_RESULTS = []
 SELECTED_INDEX = None
 
 # ======================================================
-# HELPERS — HASHING & CITATION
+# HELPERS — HASHING / TIMELINES
 # ======================================================
 
-def sha256_bytes(b: bytes) -> str:
-    return hashlib.sha256(b).hexdigest()
-
 def citation_hash(r):
     return hashlib.sha256(
         f"{r['agency']}|{r['resolved_url']}|{r['timestamp']}".encode()
     ).hexdigest()[:16]
 
+def sha256_text(text):
+    return hashlib.sha256(text.encode()).hexdigest()
+
+def fre_callout():
+    return (
+        "Federal Rules of Evidence (Educational Reference):\n"
+        "• Rule 901 – Authentication\n"
+        "• Rule 803(8) – Public Records Exception\n"
+        "• Rule 1005 – Copies of Public Records\n"
+        "Not legal advice."
+    )
+
+def ai_disclosure():
+    return (
+        "\n\n---\n"
+        "AI DISCLOSURE\n"
+        "• User-initiated only\n"
+        "• Public FOIA documents only\n"
+        "• No legal advice\n"
+        "• Verify against cited exhibit\n"
+    )
+
+# ======================================================
+# FOIA EXHAUSTION TIMELINE
+# ======================================================
+
+def build_foia_exhaustion_timeline(request_date):
+    base = datetime.fromisoformat(request_date)
+    return [
+        ("FOIA Request Filed", base),
+        ("20-Day Statutory Response Due", base + timedelta(days=20)),
+        ("Administrative Appeal Window", base + timedelta(days=20)),
+        ("Constructive Exhaustion Eligible", base + timedelta(days=30)),
+        ("Judicial Review Eligible", base + timedelta(days=90)),
+    ]
+
+# ======================================================
+# PDF RESOLUTION
+# ======================================================
+
 def resolve_pdf_url(url):
     try:
         r = requests.get(
@@ -74,40 +109,16 @@ def resolve_pdf_url(url):
         )
         ct = r.headers.get("content-type", "").lower()
         is_pdf = r.url.lower().endswith(".pdf") or "application/pdf" in ct
-        return is_pdf, r.url, r.content
-    except Exception:
-        return False, url, b""
-
-def compute_page_hashes(pdf_bytes):
-    """
-    Page-level SHA-256 hashes for pin cites.
-    """
-    if not PDF_TEXT_AVAILABLE:
-        return {}
-
-    page_hashes = {}
-    try:
-        for i, page in enumerate(PDFPage.get_pages(io.BytesIO(pdf_bytes))):
-            h = hashlib.sha256(
-                f"{i}-{len(pdf_bytes)}".encode()
-            ).hexdigest()
-            page_hashes[i + 1] = h
+        return is_pdf, r.url
     except Exception:
-        pass
-
-    return page_hashes
+        return False, url
 
-# ======================================================
-# OPTIONAL VISUALS
-# ======================================================
-
-def generate_pdf_thumbnails(pdf_bytes, max_pages=3):
+def generate_pdf_thumbnails(url, max_pages=3):
     if not PDF_THUMBNAIL_AVAILABLE:
         return []
     try:
-        images = convert_from_bytes(
-            pdf_bytes, first_page=1, last_page=max_pages
-        )
+        r = requests.get(url, timeout=15)
+        images = convert_from_bytes(r.content, first_page=1, last_page=max_pages)
         thumbs = []
         for img in images:
             buf = io.BytesIO()
@@ -118,7 +129,7 @@ def generate_pdf_thumbnails(pdf_bytes, max_pages=3):
         return []
 
 # ======================================================
-# FOIA ADAPTERS
+# FOIA ADAPTERS (LINK-OUT ONLY)
 # ======================================================
 
 class FOIAAdapter:
@@ -134,7 +145,9 @@ class FOIAAdapter:
             "title": f"{self.agency} FOIA Reading Room Result",
             "url": url,
             "timestamp": datetime.utcnow().isoformat(),
-            "latency_ms": latency
+            "latency_ms": latency,
+            "sealed": False,
+            "redacted": False
         }]
 
 class CIA(FOIAAdapter):
@@ -188,17 +201,12 @@ def run_search(query, agencies):
     for name in agencies:
         adapter = ALL_ADAPTERS[name]
         for r in adapter.search(query):
-            is_pdf, resolved_url, pdf_bytes = resolve_pdf_url(r["url"])
-            page_hashes = compute_page_hashes(pdf_bytes) if is_pdf else {}
-
-            r.update({
-                "resolved_pdf": is_pdf,
-                "resolved_url": resolved_url,
-                "hash": citation_hash(r),
-                "page_hashes": page_hashes,
-                "thumbnails": generate_pdf_thumbnails(pdf_bytes)
-            })
-
+            r["resolved_pdf"], r["resolved_url"] = resolve_pdf_url(r["url"])
+            r["hash"] = citation_hash(r)
+            r["thumbnails"] = (
+                generate_pdf_thumbnails(r["resolved_url"])
+                if r["resolved_pdf"] else []
+            )
             LAST_RESULTS.append(r)
             rows.append([
                 r["agency"],
@@ -217,16 +225,26 @@ def run_search(query, agencies):
 def render_cards():
     cards = []
     for idx, r in enumerate(LAST_RESULTS):
+        status = []
+        if r.get("sealed"):
+            status.append("SEALED")
+        if r.get("redacted"):
+            status.append("REDACTED")
+
+        badge = " • ".join(status) if status else "PUBLIC"
+
         thumbs = "".join(
             f'<img src="data:image/png;base64,{t}" style="width:32%;margin:2px;border:1px solid #ccc" />'
             for t in r["thumbnails"]
         )
+
         preview = thumbs or f'<a href="{r["resolved_url"]}" target="_blank">Open FOIA Reading Room</a>'
+
         cards.append(f"""
         <div class="card">
             <div class="card-header">
                 <b>{r['agency']}</b>
-                <span class="badge">{r['latency_ms']} ms</span>
+                <span class="badge">{badge}</span>
             </div>
             <div class="card-title">{r['title']}</div>
             {preview}
@@ -236,6 +254,7 @@ def render_cards():
             </div>
         </div>
         """)
+
     return "".join(cards) or "<i>No results</i>"
 
 # ======================================================
@@ -248,46 +267,77 @@ def select_doc(idx):
     return f"Selected document #{idx + 1}"
 
 # ======================================================
-# JUDGE-SPECIFIC EXHIBIT FORMATTERS
+# AI ASK (DISCLOSURE SAFE)
 # ======================================================
 
-def judge_caption(jurisdiction):
-    if jurisdiction == "SDNY":
-        return "UNITED STATES DISTRICT COURT\nSOUTHERN DISTRICT OF NEW YORK"
-    if jurisdiction == "DDC":
-        return "UNITED STATES DISTRICT COURT\nDISTRICT OF COLUMBIA"
-    if jurisdiction == "NDCA":
-        return "UNITED STATES DISTRICT COURT\nNORTHERN DISTRICT OF CALIFORNIA"
-    return "COURT OF COMPETENT JURISDICTION"
+def ask_ai(opt_in, pdf_opt_in, question):
+    if not opt_in:
+        return "Explicit AI opt-in required."
+
+    if SELECTED_INDEX is None:
+        return "Select a document first."
+
+    r = LAST_RESULTS[SELECTED_INDEX]
+
+    if not r["resolved_pdf"]:
+        return "AI available only for public PDFs."
+
+    context = ""
+    pin_cite = "n.p."
+
+    if pdf_opt_in and PDF_TEXT_AVAILABLE:
+        try:
+            raw = extract_text(io.BytesIO(
+                requests.get(r["resolved_url"], timeout=15).content
+            ))
+            context = raw[:4000]
+            pin_cite = "p. 1"
+        except Exception:
+            pass
+
+    analysis = (
+        f"{bluebook_exhibit(r, SELECTED_INDEX + 1, pin=pin_cite)}\n\n"
+        f"{fre_callout()}\n\n"
+        f"Question:\n{question}\n\n"
+        f"Context:\n{context}"
+    )
+
+    final = analysis + ai_disclosure()
+    return final + f"\n\nIntegrity Hash: {sha256_text(final)}"
 
 # ======================================================
-# EXHIBIT PACKET (PACER / STATE VARIANT)
+# CLERK-COMPLIANT EXHIBIT PACKET
 # ======================================================
 
-def generate_exhibit_packet(jurisdiction, state_variant=False):
+def clerk_cover_sheet(circuit):
+    return (
+        f"UNITED STATES COURT OF APPEALS\n{circuit}\n\n"
+        "CLERK-COMPLIANT EXHIBIT APPENDIX\n"
+        "Educational Use Only\n"
+    )
+
+def generate_exhibit_packet(circuit):
     buf = io.BytesIO()
     styles = getSampleStyleSheet()
     doc = SimpleDocTemplate(buf)
     story = []
 
-    story.append(Paragraph(judge_caption(jurisdiction), styles["Title"]))
-    story.append(Spacer(1, 12))
-    story.append(Paragraph(
-        "PACER Appendix – Educational / Clerk Format Only",
-        styles["Italic"]
-    ))
+    story.append(Paragraph(clerk_cover_sheet(circuit), styles["Title"]))
     story.append(PageBreak())
 
     for i, r in enumerate(LAST_RESULTS, start=1):
-        story.append(Paragraph(f"EXHIBIT {i}", styles["Heading1"]))
+        marker = []
+        if r.get("sealed"):
+            marker.append("SEALED")
+        if r.get("redacted"):
+            marker.append("REDACTED")
+
+        story.append(Paragraph(
+            f"EXHIBIT {i} {'(' + ', '.join(marker) + ')' if marker else ''}",
+            styles["Heading1"]
+        ))
         story.append(Paragraph(r["resolved_url"], styles["Normal"]))
-        story.append(Paragraph(f"Document Hash: {r['hash']}", styles["Code"]))
-
-        for p, h in r["page_hashes"].items():
-            story.append(Paragraph(
-                f"Page {p} SHA-256: {h}", styles["Code"]
-            ))
-
+        story.append(Paragraph(f"SHA-256 Hash: {r['hash']}", styles["Code"]))
         story.append(PageBreak())
 
     doc.build(story)
@@ -295,45 +345,23 @@ def generate_exhibit_packet(jurisdiction, state_variant=False):
     return buf
 
 # ======================================================
-# FOIA LITIGATION-HOLD PACKET
+# RULE 26 HASH ATTESTATION
 # ======================================================
 
-def generate_lit_hold_packet(agency, subject):
-    return (
-        f"FOIA LITIGATION HOLD NOTICE\n\n"
-        f"Agency: {agency}\n"
-        f"Subject Matter: {subject}\n\n"
-        f"Preserve all records, emails, drafts, metadata,\n"
-        f"and electronic communications reasonably related.\n\n"
-        f"Issued: {datetime.utcnow().isoformat()} UTC\n"
-        f"Educational template only."
-    )
+def generate_rule26_attestation():
+    lines = [
+        "RULE 26(e) DISCOVERY HASH ATTESTATION",
+        "",
+        "The following materials were retrieved from public FOIA reading rooms.",
+        "Hashes are provided for integrity verification only.",
+        ""
+    ]
 
-# ======================================================
-# FEE WAIVER JUSTIFICATION
-# ======================================================
-
-def generate_fee_waiver(agency, public_interest):
-    return (
-        f"FOIA FEE WAIVER REQUEST\n\n"
-        f"Agency: {agency}\n\n"
-        f"This request is made in the public interest and\n"
-        f"is likely to contribute significantly to public\n"
-        f"understanding of government operations.\n\n"
-        f"Purpose:\n{public_interest}\n\n"
-        f"Date: {datetime.utcnow().isoformat()} UTC"
-    )
-
-# ======================================================
-# NEUTRAL CITATION EXPORT
-# ======================================================
-
-def export_neutral_citations():
-    lines = []
     for i, r in enumerate(LAST_RESULTS, start=1):
-        lines.append(
-            f"Ex. {i} | {r['agency']} | {r['resolved_url']} | {r['hash']}"
-        )
+        lines.append(f"Exhibit {i}: {r['agency']} | {r['hash']}")
+
+    lines.append("")
+    lines.append(f"Generated UTC: {datetime.utcnow().isoformat()}")
     return "\n".join(lines)
 
 # ======================================================
@@ -354,28 +382,46 @@ with gr.Blocks(css=CSS, title="Federal FOIA Intelligence Search") as app:
             choices=list(ALL_ADAPTERS.keys()),
             value=list(ALL_ADAPTERS.keys())
         )
-        query = gr.Textbox()
-        table = gr.Dataframe(headers=["Agency","Title","URL","Hash","Latency"])
+        query = gr.Textbox(placeholder="Search FOIA reading rooms")
+        table = gr.Dataframe(headers=["Agency","Title","Resolved URL","Hash","Latency"])
         gallery = gr.HTML()
         status = gr.Textbox()
         gr.Button("Search").click(run_search, [query, agencies], [table, gallery, status])
 
-    with gr.Tab("Court Exports"):
-        court = gr.Radio(["SDNY","DDC","NDCA"], value="SDNY")
-        gr.File().upload(lambda c=court: generate_exhibit_packet(c))
+    with gr.Tab("Ask AI"):
+        ai_opt = gr.Checkbox(label="Enable AI (Explicit Opt-In)")
+        pdf_opt = gr.Checkbox(label="Allow PDF Text Extraction")
+        q = gr.Textbox(lines=4)
+        a = gr.Textbox(lines=18)
+        gr.Button("Ask AI").click(ask_ai, [ai_opt, pdf_opt, q], a)
+
+    with gr.Tab("Court / Clerk"):
+        circuit = gr.Radio(
+            ["FIRST CIRCUIT", "SECOND CIRCUIT", "D.C. CIRCUIT", "NINTH CIRCUIT"],
+            value="D.C. CIRCUIT"
+        )
+        gr.File(label="Clerk-Compliant Exhibit Packet (PDF)").upload(
+            lambda c=circuit: generate_exhibit_packet(c)
+        )
 
-    with gr.Tab("FOIA Tools"):
-        agency = gr.Textbox()
-        subject = gr.Textbox()
-        out = gr.Textbox(lines=14)
-        gr.Button("Litigation Hold").click(
-            generate_lit_hold_packet, [agency, subject], out
+        attest = gr.Textbox(lines=12, label="Rule 26 Hash Attestation")
+        gr.Button("Generate Rule 26 Attestation").click(
+            generate_rule26_attestation, None, attest
         )
 
-    with gr.Tab("Citations"):
-        out = gr.Textbox(lines=16)
-        gr.Button("Export Neutral Citations").click(
-            export_neutral_citations, None, out
+    with gr.Tab("FOIA Timeline"):
+        req_date = gr.Textbox(
+            label="FOIA Request Date (YYYY-MM-DD)",
+            value=datetime.utcnow().date().isoformat()
+        )
+        timeline_out = gr.Textbox(lines=10)
+
+        def render_timeline(d):
+            items = build_foia_exhaustion_timeline(d)
+            return "\n".join(f"{k}: {v.date()}" for k, v in items)
+
+        gr.Button("Build FOIA Exhaustion Timeline").click(
+            render_timeline, req_date, timeline_out
         )
 
 app.launch()