feat(learning): מטא-דאטה מלא להחלטות-פנימיות בקליטה + חילוץ-הלכות אוטומטי

סוגר את הפער שעלה על בל"מ 8126: החלטה שנכנסה לספריית-הפסיקה הופיעה ללא
מטא-דאטה (summary/citation/date ריקים, proceeding_type שגוי) כי מחלץ-ה-Gemini
מיועד לפסיקה חיצונית ומחזיר no_metadata לפנימיות, והחילוץ-הלכות נשאר pending.

web/app.py — `_enroll_final_in_library` עכשיו ממלא **דטרמיניסטית** מהתיק (בלי LLM):
- proceeding_type (מהתיק — בל"מ/ערר, גם idempotency key נכון מהקליטה הראשונה),
  decision_date (fallback ל-hearing_date), subject_tags, summary (=subject).
- `citation_formatted` נבנה דטרמיניסטית (`_build_internal_citation`):
  'ועדת ערר ... בל"מ <num> <עורר> נ' <משיב> (יו"ר עו"ד <chair>)'.

scripts/final_halacha_pipeline.py — שלב [0] חדש: `precedent_extract_halachot`
על ההחלטה עצמה (idempotent — מדלג כש-completed/dry-run), כך שהלכות-ההחלטה
לא נשארות pending.

אומת: py_compile ✓ · ה-pipeline רץ dry-run נקי (4 שלבים). 8126 כבר תוקן ידנית;
מכאן זה אוטומטי לכל החלטה. Invariants: INV-LRN4/X11 · G1 (נרמול-במקור) ·
DM7 · feedback_silent_swallow.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

web/app.py

@@ -3387,6 +3387,33 @@ def _committee_chair_for_case(case: dict, case_number: str) -> str:
     return COMMITTEE_CHAIR_BY_PREFIX.get(case_number[:1], COMMITTEE_CHAIR_DEFAULT)
 
 
+def _party_name(parties) -> str:
+    """First party's display name from a list of {name|party_name} dicts or strings."""
+    if isinstance(parties, list) and parties:
+        p = parties[0]
+        if isinstance(p, dict):
+            return (p.get("name") or p.get("party_name") or "").strip()
+        return str(p).strip()
+    return ""
+
+
+def _build_internal_citation(
+    case: dict, case_number: str, chair_name: str, proceeding_type: str, district: str,
+) -> str:
+    """Deterministic uniform citation for OUR committee decisions — the Gemini metadata
+    extractor targets external rulings and returns nothing for internal ones, so we build
+    it from the case record instead. E.g.:
+    'ועדת ערר מחוזית לתכנון ובניה ירושלים, בל"מ 8126-03-25 פלוני נ' הוועדה המקומית (יו"ר עו"ד דפנה תמיר)'."""
+    appellant = _party_name(case.get("appellants"))
+    respondent = _party_name(case.get("respondents"))
+    parties = f" {appellant} נ' {respondent}" if (appellant and respondent) else (
+        f" {appellant}" if appellant else "")
+    proc = proceeding_type or "ערר"
+    chair_clause = f' (יו"ר עו"ד {chair_name})' if chair_name else ""
+    return (f"ועדת ערר מחוזית לתכנון ובניה {district}, {proc} {case_number}"
+            f"{parties}{chair_clause}").strip()
+
+
 async def _enroll_final_in_library(
     case: dict, case_number: str, final_text: str, chair_name: str,
 ) -> dict:
@@ -3404,12 +3431,23 @@ async def _enroll_final_in_library(
     if not final_text.strip():
         out["error"] = "no final text extracted"
         return out
+
+    # Deterministic metadata from the case record — the Gemini metadata extractor is
+    # tuned for EXTERNAL rulings and returns no_metadata for internal decisions, so we
+    # populate proceeding_type / date / tags / summary / citation ourselves (no LLM).
+    district = "ירושלים"
+    proceeding_type = (case.get("proceeding_type") or "ערר").strip()
+    decision_date = case.get("decision_date") or case.get("hearing_date")
+    subject_tags = case.get("subject_categories") or []
+    summary = (case.get("subject") or case.get("title") or "").strip()
     try:
         res = await int_svc.ingest_internal_decision(
             case_number=case_number, case_name=case.get("title", ""),
-            decision_date=case.get("decision_date"), chair_name=chair_name,
-            district="ירושלים", practice_area=case.get("practice_area", ""),
-            appeal_subtype=case.get("appeal_subtype", ""), text=final_text,
+            decision_date=decision_date, chair_name=chair_name,
+            district=district, practice_area=case.get("practice_area", ""),
+            appeal_subtype=case.get("appeal_subtype", ""),
+            proceeding_type=proceeding_type, subject_tags=subject_tags,
+            summary=summary, text=final_text,
         )
     except Exception as e:
         logger.warning("library enrollment failed for %s: %s", case_number, e)
@@ -3421,6 +3459,21 @@ async def _enroll_final_in_library(
     if not case_law_id:
         return out
 
+    # Build the uniform citation deterministically (Gemini won't, for internal rows).
+    try:
+        citation = _build_internal_citation(
+            case, case_number, chair_name, proceeding_type, district)
+        pool = await db.get_pool()
+        async with pool.acquire() as conn:
+            await conn.execute(
+                "UPDATE case_law SET citation_formatted = $2 "
+                "WHERE id = $1 AND COALESCE(citation_formatted, '') = ''",
+                UUID(case_law_id), citation,
+            )
+        out["citation"] = citation
+    except Exception as e:
+        logger.warning("citation build failed for %s: %s", case_number, e)
+
     # The precedents this decision cites → link to the library; flag the ones not found.
     try:
         await cit_tools.extract_internal_citations(case_law_id=case_law_id, limit=0)