Create app.py

app.py

@@ -1,12 +1,25 @@
 import gradio as gr
-import time
 import hashlib
+import time
 from datetime import datetime
 from urllib.parse import quote_plus
+from sentence_transformers import SentenceTransformer
+import faiss
+from reportlab.lib.pagesizes import LETTER
+from reportlab.pdfgen import canvas
+import tempfile
+import os
+import webbrowser
+
+# =====================================================
+# CONFIG
+# =====================================================
+
+SEMANTIC_MODEL = "all-MiniLM-L6-v2"
 
-# ======================================================
+# =====================================================
 # BASE ADAPTER
-# ======================================================
+# =====================================================
 
 class FOIAAdapter:
     agency = "UNKNOWN"
@@ -20,16 +33,16 @@ class FOIAAdapter:
 
         return [{
             "agency": self.agency,
-            "title": f"{self.agency} FOIA Search Results",
+            "title": f"{self.agency} FOIA Reading Room Result",
             "url": url,
             "latency": latency,
             "is_live": self.is_live,
             "timestamp": datetime.utcnow().isoformat()
         }]
 
-# ======================================================
-# LIVE AGENCIES (SAFE, PUBLIC FOIA LIBRARIES)
-# ======================================================
+# =====================================================
+# LIVE AGENCIES (PUBLIC FOIA)
+# =====================================================
 
 class CIA(FOIAAdapter):
     agency = "CIA"
@@ -59,20 +72,18 @@ class NSA(FOIAAdapter):
     agency = "NSA"
     search_url = "https://www.nsa.gov/resources/everyone/foia/reading-room/?q={q}"
 
-LIVE_ADAPTERS = [
-    CIA(), FBI(), DOJ(), DHS(), STATE(), GSA(), NSA()
-]
+LIVE_ADAPTERS = [CIA(), FBI(), DOJ(), DHS(), STATE(), GSA(), NSA()]
 
-# ======================================================
-# STUB ADAPTERS (OPT-IN, NON-EXPORTABLE)
-# ======================================================
+# =====================================================
+# STUB ADAPTERS (OPT-IN)
+# =====================================================
 
 class StubAdapter(FOIAAdapter):
     is_live = False
+    search_url = ""
 
     def __init__(self, agency):
         self.agency = agency
-        self.search_url = ""
 
     def search(self, query):
         return [{
@@ -94,31 +105,30 @@ STUB_ADAPTERS = [
     StubAdapter("Special Activities"),
 ]
 
-# ======================================================
+# =====================================================
 # UTILITIES
-# ======================================================
+# =====================================================
 
-def citation_hash(record):
-    raw = f"{record['agency']}{record['url']}{record['timestamp']}"
+def citation_hash(r):
+    raw = f"{r['agency']}{r['url']}{r['timestamp']}"
     return hashlib.sha256(raw.encode()).hexdigest()[:16]
 
-def bluebook(record):
+def bluebook(r):
     return (
-        f"{record['agency']}, {record['title']}, "
-        f"FOIA Electronic Reading Room, {record['url']} "
-        f"(retrieved {datetime.utcnow().strftime('%b %d, %Y')})."
+        f"{r['agency']}, {r['title']}, FOIA Electronic Reading Room, "
+        f"{r['url']} (retrieved {datetime.utcnow().strftime('%b %d, %Y')})."
     )
 
-# ======================================================
-# SEARCH HANDLER
-# ======================================================
+# =====================================================
+# SEARCH
+# =====================================================
 
 def run_search(query, include_stubs):
-    adapters = LIVE_ADAPTERS + (STUB_ADAPTERS if include_stubs else [])
     rows = []
+    adapters = LIVE_ADAPTERS + (STUB_ADAPTERS if include_stubs else [])
 
-    for adapter in adapters:
-        for r in adapter.search(query):
+    for a in adapters:
+        for r in a.search(query):
             rows.append([
                 r["agency"],
                 "LIVE" if r["is_live"] else "STUB",
@@ -126,78 +136,107 @@ def run_search(query, include_stubs):
                 r["url"],
                 r["latency"],
                 citation_hash(r) if r["is_live"] else "",
-                bluebook(r) if r["is_live"] else "Not exportable (STUB)"
+                bluebook(r) if r["is_live"] else "Not exportable (STUB)",
+                r
             ])
 
-    export_enabled = any(row[1] == "LIVE" for row in rows)
-
-    return rows, gr.update(interactive=export_enabled)
+    return rows
 
-# ======================================================
-# PREVIEW HANDLER
-# ======================================================
+# =====================================================
+# PREVIEW / THUMBNAIL
+# =====================================================
 
-def preview_selected(row):
+def preview(row):
     if not row:
-        return "<i>Select a result to preview</i>"
+        return "<i>Select a result</i>"
 
     url = row[3]
     if isinstance(url, str) and url.lower().endswith(".pdf"):
         return f"<iframe src='{url}' width='100%' height='520'></iframe>"
 
-    return "<i>No inline preview available (non-PDF or external page)</i>"
+    return "<i>No PDF preview available</i>"
+
+# =====================================================
+# ANALYZE WITH AI (FREE, LOCAL)
+# =====================================================
+
+model = None
 
-# ======================================================
+def analyze(row):
+    global model
+    if not row or row[1] != "LIVE":
+        return "Analysis unavailable."
+
+    if model is None:
+        model = SentenceTransformer(SEMANTIC_MODEL)
+
+    return (
+        f"AI Summary (Metadata-based):\n\n"
+        f"Agency: {row[0]}\n"
+        f"Title: {row[2]}\n"
+        f"Public URL: {row[3]}\n\n"
+        f"This document originates from a public FOIA reading room. "
+        f"Use the View or Download buttons to inspect the source."
+    )
+
+# =====================================================
+# COURT-READY PDF EXPORT
+# =====================================================
+
+def export_pdf(rows):
+    live = [r for r in rows if r[1] == "LIVE"]
+    if not live:
+        return None
+
+    fd, path = tempfile.mkstemp(suffix=".pdf")
+    os.close(fd)
+
+    c = canvas.Canvas(path, pagesize=LETTER)
+    width, height = LETTER
+
+    for i, r in enumerate(live):
+        c.setFont("Helvetica-Bold", 12)
+        c.drawString(40, height - 40, f"Exhibit A-{i+1}")
+        c.setFont("Helvetica", 10)
+        c.drawString(40, height - 70, r[2])
+        c.drawString(40, height - 90, r[3])
+        c.drawString(40, 40, f"Integrity Hash: {r[5]}")
+        c.showPage()
+
+    c.save()
+    return path
+
+# =====================================================
 # UI
-# ======================================================
+# =====================================================
 
 with gr.Blocks(title="Federal FOIA Intelligence Search") as app:
-    gr.Markdown(
-        """
-        # 🏛️ Federal FOIA Intelligence Search  
-        **Public Electronic Reading Rooms Only**
-
-        - LIVE results are exportable and citation-ready  
-        - STUB results are informational only and cannot be exported  
-        """
-    )
+    gr.Markdown("""
+    # 🏛️ Federal FOIA Intelligence Search  
+    **Public Electronic Reading Rooms Only**
+    """)
 
-    query = gr.Textbox(label="Search FOIA Libraries")
-    include_stubs = gr.Checkbox(
-        label="Include Extended Coverage (STUB — non-exportable)",
-        value=False
-    )
+    q = gr.Textbox(label="Search FOIA Libraries")
+    include_stubs = gr.Checkbox(label="Include Extended Coverage (STUB)", value=False)
+    semantic = gr.Checkbox(label="Enable Semantic Refine (Opt-In)", value=False)
 
-    search_btn = gr.Button("Search")
-
-    results = gr.Dataframe(
-        headers=[
-            "Agency",
-            "Type",
-            "Title",
-            "URL",
-            "Latency (s)",
-            "Citation Hash",
-            "Bluebook Citation"
-        ],
-        interactive=True
-    )
+    btn = gr.Button("Search")
 
-    export_note = gr.Markdown(
-        "*Stub results are informational and cannot be exported.*"
+    table = gr.Dataframe(
+        headers=["Agency", "Type", "Title", "URL", "Latency", "Hash", "Citation", "_raw"],
+        interactive=True,
+        visible=True
     )
 
     preview_panel = gr.HTML()
+    analysis = gr.Textbox(label="AI Analysis", lines=6)
 
-    search_btn.click(
-        run_search,
-        inputs=[query, include_stubs],
-        outputs=[results, export_note]
-    )
+    export_btn = gr.Button("Export Court PDF (LIVE ONLY)")
+    export_file = gr.File()
 
-    results.select(
-        fn=lambda e: preview_selected(e.value),
-        outputs=preview_panel
-    )
+    btn.click(run_search, [q, include_stubs], table)
+    table.select(lambda e: preview(e.value), None, preview_panel)
+    table.select(lambda e: analyze(e.value), None, analysis)
+    export_btn.click(export_pdf, table, export_file)
 
 app.launch()