WS2b של עיצוב-מחדש זרימת-העבודה. מאפשר ניתוח כתב-הערר לבד ואז איחוד
מסמך-עיקרי שנוסף מאוחר, בלי force-delete גורף, עם diff ליו"ר.
- SCHEMA_V49: documents.claims_extracted_at + claims_extraction_status
(אירוע-חילוץ per-מסמך, לא נגזר מ-doc_type), + אינדקס חלקי
idx_documents_claims_pending. idempotent (ADD COLUMN IF NOT EXISTS).
V48 שמור ל-#357.
- claims_extractor חותם את המסמך אחרי שמירת/אי-מציאת טענות.
- db.primary_docs_not_analyzed (is_primary V47 + claims_extracted_at IS NULL)
+ mark_document_claims_extracted; _row_to_doc חושף claims_analyzed.
- reanalyze_claims (כלי-MCP): snapshot→חילוץ-מאחד רק למסמכים חדשים/לא-נותחו
(store_claims מחליף per-source ⇒ טענות אחרות נשמרות)→aggregate force=True
(מסלול קנוני, מוחק רק legal_arguments)→_impact_diff before↔after ליו"ר.
- workflow_status חושף primary_docs_not_analyzed + next-step.
- ספ: 02-data-model §2ג (דגל לא-נותח), 04-analysis-writing §1.3.
- 5 בדיקות-יחידה ל-_impact_diff/_snapshot (פונקציות טהורות).
Invariants:
- G1 (נרמול-במקור): claims_extracted_at נחתם בנקודת-החילוץ, claims_analyzed
נגזר ממנו — לא תיקון-בקריאה.
- G2 (מסלול קנוני יחיד / merge-not-fork): reanalyze מרחיב את מסלול
claims_extractor+argument_aggregator הקיים, לא forks; האיחוד דרך
store_claims per-source; aggregate force=True מוחק רק נגזר (legal_arguments).
- INV-DM (מודל-נתונים): דגל-אירוע per-מסמך, אינדקס חלקי, מקור-אמת יחיד.
- G10 (שער-אנושי): בדיקת-ההשפעה מוצגת ליו"ר, לא מוחלת אוטומטית.
- INV-TOOL idempotency: SCHEMA_V49 idempotent; ניתוח-מחדש חוזר על מסמך
כבר-נותח הוא refresh נקי (store_methods replace per-source).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
מנגנון-הייצור (claude_session→claude -p) לא העביר model/effort: BLOCK_CONFIG
הגדיר temp+model per-בלוק שהיו מטא-דאטה-מת (block_writer.py:447 העביר רק
prompt/timeout/tools). הייצור רץ על מודל ברירת-המחדל של ה-CLI עם פרומפטים-חופשיים
→ מקור אי-העקביות בטיוטת-הביניים.
תיקון (עקביות-מבנית, WS5):
- נעיצת model="claude-opus-4-8" (GENERATION_MODEL) לכל בלוקי-ה-AI + effort per-בלוק
שמועבר בפועל דרך claude_session.query → `claude -p --model --effort`.
מפת-effort: ה=medium · ו=medium · ז=high · ח=medium · ט=high · יא=high · י=xhigh.
- ניסוח-פתיחה קבוע לבלוק-ה (תמיד "לפנינו ערר…", סוף ל-OR "עניינה של החלטה זו").
- מבנה-פנימי קבוע לבלוק-ה; סדר-בלוקים בייצוא כבר קבוע (_INTERIM_BLOCK_ORDER).
- הסרת שדות temp/model המתים מ-BLOCK_CONFIG; "temp" deprecated (Opus 4.7/4.8 דוחים
temperature→400). decision_blocks.temperature נשמר=0 לתאימות-עמודה בלבד.
- timeout נגזר מ-effort (לא מפיצול sonnet/opus שכבר לא קיים).
- עדכון endpoint reference /api/settings/mcp/blocks + blocks-tab + McpBlock לחשוף effort.
- ספ: docs/block-schema.md §3 (effort מחליף temperature), docs/spec/06-export.md
(דטרמיניזם-מבני של טיוטת-הביניים).
G2 (מקור-אמת יחיד) — אין מסלול-ייצור מקביל: write_interim_draft מתזמר את אותו
write_and_store_block→write_block של ההחלטה המלאה.
claude_session-local-only — הייצור נשאר דרך גשר-ה-CLI, לא SDK/קונטיינר.
INV-EX1 — סדר-ייצוא דטרמיניסטי נשמר; G11 — עקרונות-הכתיבה ל-12 הבלוקים נשמרים.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Golden ratios were measured from the corpus (style_distance.measure_corpus_ratios)
but never reached the writer — it used only the hardcoded STRUCTURE_GUIDANCE text.
So the structural target was frozen, and drafts kept deviating (e.g. discussion
block 10% vs a 40-47% target) with no corrective signal.
_structure_guidance_with_corpus(outcome, practice_area) now appends the
corpus-measured section ratios for the block's outcome to the structure guidance,
at both writer entry points (write_block + get_block_context) via one shared helper
(G2). Because measure_corpus_ratios reads style_corpus — which grows with every
enrolled final (see exemplar-growth, #343) — the writer's structural target now
reflects Dafna's ACTUAL current distribution and improves as the corpus grows.
Deterministic (no LLM); best-effort.
This is the STRUCTURAL half of the voice-profile-refresh (#7). The remaining half —
LLM-regenerating the abstract voice-fingerprint PROSE from the corpus (chair-gated)
— is a distinct larger initiative, deliberately deferred rather than rushed.
Invariants: G1 (the writer gets the live measured target, not a frozen constant),
G2 (one structure-guidance builder shared by both entry points; reuses the existing
measure_corpus_ratios — no parallel measurement), INV-LRN4 (the corpus the loop
grows now feeds the writer's structural target).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The chair-feedback chain was DEAD: 27 feedback rows captured, 0 ever became a
lesson the writer reads (no weekly-analysis job ever existed). The chair's own
corrections — the most authoritative signal of all, and exactly "learn from every
returned draft" — went nowhere.
decision_lessons can't carry them: it's FK-coupled to a style_corpus row (a signed
final), which a case-in-progress doesn't have. So chair STYLE feedback instead
rides the discussion_rules channel that already reaches the writer for all blocks —
the same path /training promote uses.
- db.append_global_rule: the locked read-modify-write append, extracted from web
`_append_methodology_override` into one shared impl (G2). The web function is now
a thin wrapper that seeds defaults; chair-feedback calls it directly.
- record_chair_feedback (MCP tool): a STYLE-category correction (style/wrong_tone/
wrong_structure) with a lesson_extracted flows immediately to discussion_rules.
The chair IS the gate — no separate approval (INV-LRN1 graduated gate). SUBSTANCE
feedback (missing_content/factual_error/other) is case-specific → recorded only.
Invariants: INV-LRN1 (chair-authored style = highest authority, flows; substance
not auto-flowed), G2 (single append impl shared by promote + feedback), INV-LRN5
(style channel only).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The style-exemplar corpus (channel B — the writer's block-level retrieval of
Dafna's real prose) was FROZEN at the one-time seed backfill: new finals were
enrolled into style_corpus but never broken into exemplars, so the richest style
channel never grew (8137/8126/8174 had 0 exemplars). The writer kept retrieving
only March–April seed paragraphs no matter how many finals were signed.
Extract the per-decision exemplar logic (section→paragraph→Voyage-embed→replace)
into a shared service `legal_mcp.services.style_exemplars.extract_and_store` —
the SINGLE implementation now used by BOTH the one-time backfill and the live
enrollment path (G2; no parallel extractor). `_enroll_final_in_library` calls it
on every final upload (source='internal_committee', the same source the writer's
search_style_exemplars reads). Voyage embeds over REST → container-safe;
best-effort, surfaced in the upload response, never fails the upload.
Effect: every signed final now grows the exemplar corpus, so the writer's
block-level style retrieval improves with each decision — the core "learn from
every decision" fix for channel B. Path A (style_distance_history) will track
whether the larger exemplar pool reduces style-distance over time.
Invariants: G2 (one extraction path shared by backfill + enroll), INV-LRN5
(style/structure prose only — substance routes elsewhere), INV-LRN4 (the
draft↔final loop now feeds the exemplar channel, not just the lesson channel).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The learning loop measured anti-patterns but never corrected them. style_distance
detects markdown headers / bullet lists / mid-paragraph mini-lists (from the
canonical lessons.ANTI_PATTERNS), but the writer only received anti-pattern
guidance if a chair `anti_patterns` override existed in appeal_type_rules — and
none does. So _build_style_context's `if ov:` branch injected nothing, the writer
was never told to avoid them, and drafts kept emitting them (8137: 28 hits, the
worst, newest — anti-patterns were trending UP, not down).
Anti-patterns are structural invariants of Dafna's voice (continuous legal
narrative — no markdown, no bullets), not overridable preferences. So inject the
canonical ANTI_PATTERNS notes ALWAYS, from the same list style_distance measures
against (single source of truth), with any chair additions layered on top. This
closes the measure-but-don't-correct gap: the next draft should show the markdown/
bullet anti-patterns drop, and Path A (style_distance_history) will confirm it.
Invariants: G1 (correct at source — the writer, not a post-hoc stripper), G2
(one canonical anti-pattern list shared by detection and instruction — no parallel
list), INV-LRN4 (closes the feedback half of the draft↔final loop).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
תיקוני-מבנה ללולאת-הלמידה (TaskMaster #157):
1. כרטיס ה-curator (/training טאב "אוצֵר") ספר decision_lessons WHERE source='curator'
— ערך שאף מסלול-קוד לא כתב → תמיד 0. עוצב-מחדש (דרך שער-העיצוב Claude Design, אושר)
להציג ביושר את שלושת ערוצי-ההזנה לכותב: דיסטילציה→appeal_type_rules (180, זורם),
פאנל→decision_lessons (81, ממתין), אוצֵר→source='curator'. get_curator_stats שוכתב.
2. drift ספ↔סוכן: 07-learning.md §1.1 + INV-LRN3 קבעו שהאוצֵר רושם ממצאים כ-decision_lesson
source='curator', אך הסוכן כתב comments בלבד — הממצאים האיכותיים אבדו. נוסף כלי-MCP
record_curator_findings + §A.5b ב-hermes-curator.md (read-only נשמר; הצעה מגודרת-שער).
3. get_recent_decision_lessons(limit=15) חתך בשקט — נוסף WARN על מה שנחתך (חוקה §6);
הפתרון האמיתי = סינתזת-לקחים (TaskMaster #158).
Invariants: מקיים INV-LRN1/G10 (שער-יו"ר), INV-LRN3 (לכידה מובנית), INV-IA2/IA5 (מקור-אמת
יחיד), G12 (leak-guard עובר), G2 (אין מסלול מקביל — אותם מאגרים). פער-מימוש פתוח מתועד:
§A לא רץ אוטומטית על mark-final (pipeline-wake exits) → נדחה לתכנון נפרד.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A clean held-out test of voice learning was no longer runnable: every
final-uploaded case already has its lessons folded, and lessons are stored
universal/untagged so leave-one-out is impossible. Path A makes the test
prospective instead — capture the generalization datapoint at the one moment
it's clean.
On final upload, after the draft↔final pair is created but BEFORE this
case's lessons are folded (folding is a separate manual /training step), we
snapshot style_distance (anti_pattern_total, golden-ratio max-deviation,
change_percent) alongside the current voice-lesson pool size. Because the
draft was written with only the PRIOR pool, each row is a clean "with N
accumulated lessons, our draft on this unseen case scored X" datapoint. As
the pool grows over cases, a downward trend = learning generalizes.
- db: SCHEMA_V45 style_distance_history (append-only) + helpers
voice_lesson_pool_sizes / record_style_distance_snapshot /
get_style_distance_history.
- app: best-effort capture in api_upload_final_decision (never fails the
upload); GET /api/learning/style-distance-history for the trend.
Reuses the existing style_distance service + appeal_type_rules pool — no
parallel metric path. The 8 existing cases are already folded, so the table
starts empty and fills from the next final (their clean window is past).
Invariants: G2 (reuse style_distance/appeal_type_rules — one path),
INV-LRN4 (measure the draft↔final gap; this is its trend surface). LLM-free
(style_distance is deterministic) so it runs in the container.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
When a chair's signed final decision is uploaded it is enrolled into the
precedent library, but _enroll_final_in_library only set deterministic
fields (citation, proceeding_type, date) and copied subject_tags from the
case's subject_categories — which is usually empty. The Gemini metadata
pass (subject_tags / summary / headnote / key_quote) was never triggered,
so the row sat at metadata_extraction_status='pending' with no subject
tags until a drain happened to pick it up. In practice a freshly-enrolled
final showed an empty "תגיות נושא" in the precedent edit UI.
Fix: after enrollment + citation, call the existing reextract_metadata
path inline (G2 — one path, full status lifecycle). It runs on Gemini
Flash over REST (GOOGLE_GEMINI_API_KEY, already in Coolify), so it is
container-safe — unlike the halacha path (claude CLI, host-only).
apply_to_record fills only empty fields, so the deterministic seeds and
any chair-curated subject_categories are preserved. Result surfaced in the
upload response (out["metadata"]); failures logged, upload still succeeds.
Also corrects two now-stale comments: the "Gemini returns no_metadata for
internal decisions" note in the enroll loop, and the "MCP-tool-only path"
docstring on reextract_metadata (true for halacha, not for Gemini REST).
Invariants: G1 (fill at source on enroll, not a read-time workaround),
G2 (reuse reextract_metadata — no parallel extraction path). LLM call is
Gemini REST, not claude_session, so the container LLM-call constraint holds.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Inline # comments in gitignore are not supported — they were silently
breaking three patterns (data/checkpoints/, data/adapter-migration-state.json,
.claude/agents/.generated/). Moved comments to their own lines and added
missing entries for runtime dirs (data/audit/, data/logs/, etc.) and
temp files (.interaction_tmp.json, .design-build/, .taskmaster bak files).
Also tracks previously untracked legitimate files: scripts, tests, docs,
skills references, .env.example, taskmaster templates.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Switches the scanned-PDF fallback from Google Cloud Vision to
Mistral OCR (mistral-ocr-latest) for better Hebrew accuracy and
robustness against broken embedded OCR layers (e.g. case 1044-03-26
which returned English garbage through Vision).
Routing strategy (document-level, not per-page):
- PyMuPDF extracts all pages; pages that pass _text_quality_ok()
use PyMuPDF output directly (free, ~50ms).
- If ANY page fails quality → Mistral OCR called once for the whole
PDF, returning per-page Markdown for all pages (consistent format,
no plain-text/Markdown mix within a document).
Markdown output preserved: Mistral returns ## headers and |tables|;
chunker updated to recognise ATX Markdown headers (##/###) as section
boundaries in _split_into_sections().
Config: GOOGLE_CLOUD_VISION_API_KEY → MISTRAL_API_KEY; allowlist
updated vision.googleapis.com → api.mistral.ai.
MISTRAL_API_KEY added to Coolify container env.
Invariants: G1 (single OCR fallback path, not parallel), G2 (no
duplicate extractor route), INV-AH (Mistral handles gershayim
natively; quote-fix only on PyMuPDF path).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
aggregate_claims_to_arguments crashed with "current transaction is aborted,
commands ignored until end of transaction block" on large cases (confirmed on
1027-04-26, 195 claims; reported via CMP-186).
Root cause: the proposition INSERT (legal_argument_propositions) was wrapped in
a broad except Exception with no savepoint. When the LLM echoes a
syntactically-valid-but-nonexistent claim_id, the FK violation
(legal_argument_propositions_claim_id_fkey) puts the asyncpg transaction into
aborted state. The except caught only that INSERT's error but never issued
ROLLBACK TO SAVEPOINT, so the next statement (the following argument's INSERT
into legal_arguments) raised InFailedSQLTransactionError uncaught and crashed
the whole call. With many claims the bad-UUID probability is high -> consistent
failure.
Fix:
- Validate each claim_id against the known set of claim ids fetched for the
case before INSERT, so a hallucinated id never reaches the DB (G1: fix at
source). Malformed UUIDs are already dropped in _normalize_argument; this
catches the valid-but-nonexistent ones.
- Wrap the INSERT in a per-row savepoint (async with conn.transaction()) as
defense-in-depth, so any unexpected constraint failure rolls back to the
savepoint instead of poisoning the surrounding transaction.
Invariants: G1 (normalize at source, not symptom-catch on read). No parallel
path (G2). No silent swallow — skips are logged.
TaskMaster: legal-ai #156
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Backend half of the citation-verification tab (frontend follows in a separate PR):
- Schema V44: case_precedents gains argument_id (the legal argument it supports),
case_law_id (the corpus ruling — for cited_by + dedup), verified (the INV-AH gate
the writer respects) + verified_at. All nullable; chair_note already existed.
- db: create_case_precedent(+argument_id/case_law_id/verified),
set_case_precedent_verified(id, verified, chair_note), list_case_precedents(+cols).
- services/case_citation_verification.build_view(case): per legal_argument →
in-corpus suggestions (search_library per issue) each with the cited_by authority
breakdown (db.citation_authority, X11) + merged attach/verify state + per-issue
radar (case_digest_radar grouped by matched issue). Pure read/assembly; reuses the
one corpus search + one authority query + one radar (G2).
- endpoints: GET /api/cases/{n}/citation-verification (the view) and
POST .../verify (upsert verify/un-verify + chair note).
Validated on 1044-03-26: 14 arguments, all with support; 317/10 surfaces with
אומץ×11/אובחן×1; radar leads attributed per issue. Auto-approval untouched (chair
gate, INV-G10). Verify defaults to false — nothing is authoritative until the chair
marks it.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The radar's query was built from dozens of raw claims (procedural-heavy noise), so
matches were thematic but imprecise and gave no reason WHY a lead is relevant. Now:
- Prefer the analyst's distilled legal_arguments (argument_title + legal_topic — one
crisp CREAC issue per row) over raw claims.
- Search EACH issue separately and MERGE, so every lead is attributed to the case
issue(s) it answers (`matched_issues`) — the chair sees "this ruling is for your
'זכות עמידה' issue", not just a blended score.
- Fall back to the raw-claims blended query pre-aggregation; `source` reports the path.
- Shared `_radar_enrich` helper (gap status + action + matched_issues), bounded to 25
issues to cap the per-issue fan-out.
Validated: 8124-09-24 (32 args → per-issue) surfaces betterment rulings each tagged to
its issue (היעדר השבחה / זהות הנישום / סעיף 7(ב)); 1044-03-26 (0 args) falls back to
claims unchanged. No tool/endpoint signature change (new fields pass through the dict).
Invariants: G2 (reuses the one digest search + arg accessor), INV-DIG1 (radar only).
No schema change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
A digest pointing at a ruling we don't hold yet ("unlinked") was captured globally
(missing_precedents inbox) but never surfaced IN THE CONTEXT of the case being
decided — so a relevant ruling known only via a digest could fall through the cracks
at the moment it matters. Adds the case-contextual radar:
- db.search_digests_semantic: new `linked_only` filter (False = unlinked-only target set).
- digest_library.case_digest_radar(case_number): builds the case topic from title +
appeal_subtype + the analyst's claims, embeds once, matches against UNLINKED digests,
and returns leads enriched with the underlying ruling's gap status + suggested action
(new_lead / gap_open / fetched / available_link).
- MCP tool `digest_radar` + endpoint GET /api/cases/{n}/digest-radar (for the agent and
the future case-page lead).
INV-DIG1 preserved: radar only — every lead points at the underlying RULING (fetch /
upload / link), never cites the digest. Read-only.
Validated on 8124-09-24 (היטל השבחה): 5 on-topic unlinked digests, scores 0.64–0.72.
The visible case-page panel is a separate UI change → goes through the design gate.
Invariants: G2 (reuses the one digest semantic-search + gap/citation resolvers),
INV-DIG1 (no digest citation), INV-DIG3 (gap surfacing). No schema change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The internal citation graph fed only RANKING (raw in-degree), and the per-citation
TREATMENT was never classified — so a precedent distinguished N times got the same
authority boost as one followed N times (INV-COR2 violation), and the signal never
reached the agents' reasoning. Wires the full path:
Phase 1 — scripts/classify_citation_treatments.py: classify each linked edge's
treatment (followed/distinguished/…) from its match_context via
corroboration.classify_treatment (Opus 4.8 @ xhigh, local), filling
precedent_internal_citations.treatment. Idempotent.
Phase 2 — db.refresh_verified_layer: count only NON-negative treatments toward
verified/cite_count (INV-COR2/COR4). Unclassified counts as neutral-positive so
the signal degrades gracefully before classification runs.
Phase 3 — db.citation_authority(ids): per-precedent {total, positive, negative,
unclassified, by_treatment}. Surfaced as `cited_by` in search_precedent_library
hits and precedent_library_get, and `treatment` per incoming citation.
Phase 4 — legal-researcher/analyst/writer prompts: weigh & ARGUE authority
("הלכה שאומצה ב-N החלטות ועדת-ערר"), flag distinguished/overruled, never invent
the count (INV-AH; writer is read-only of the analyst).
Auto-approval stays kill-switched off (chair gate preserved, INV-G10). No schema
change (treatment column already existed). Operational: run the classifier +
refresh_verified_layer over the 379 edges, then sync agents across companies.
Invariants: G2 (one classifier + one authority query, reused), INV-COR2/COR3/COR4
(negative never corroborates; point-specific; ≥N), INV-G10 (no auto-approval),
INV-AH (no invented numbers).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Root cause: the upload form "מראה המקום" field is stored only as case_number;
citation_formatted was left empty and filled ONLY by the async metadata
extraction, which runs in the local drainer (not the container) because it was
lumped with halacha extraction. Result: the chair saw an empty מראה-מקום during
the window and re-typed it by hand.
Fix (hybrid — Option C):
- Seed: _create_external_record / create_external_case_law store the chair's typed
citation as citation_formatted at INSERT, so the field is never blank. A
cited_only→external promotion preserves an existing non-empty value (prior edit).
- Enrich: ingest_precedent runs metadata extraction INLINE in-container
(gemini_session is direct REST, no local CLI) with force_citation=True, upgrading
the seed to the canonical derived citation (parties + reporter + date) before the
chair opens the page. Best-effort: on no key / API failure the seed remains and
the queued metadata drain stays as the fallback. Halacha extraction is untouched
(stays local — claude_session).
- apply_to_record / extract_and_apply gain force_citation: re-assemble even when
citation_formatted is non-empty, writing only when assembly SUCCEEDS (seed
preserved on abstention). The drainer keeps the default (False) so a chair's
manual edit in /precedents/[id] is never clobbered.
Ops: GOOGLE_GEMINI_API_KEY added to the legal-ai Coolify app (runtime) — the SoT
value from Infisical nautilus:/external-apis/gemini, validated against the API.
Invariants: G2 (one citation-resolution + one metadata-extraction path, reused —
no parallel logic), INV-ID2/X1§3 (citation_formatted stays a derived field; the
seed is provisional and upgraded), INV-G10 (chair edits preserved). No schema change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The precedent-detail "ציטוטים מקושרים" panel rendered only MANUAL appeal-chain
links (case_law_relations), so it stayed empty even when decisions cite the
ruling — the automatic citation graph (precedent_internal_citations) was never
surfaced. e.g. עע"מ 317/10 (שפר) has 12 incoming citations in the DB, none shown.
Fills the existing panel from the citation graph (data/logic only, no new
visual design — gate-exempt per web-ui/AGENTS.md):
- citation_extractor.list_citations_to_case_law: return source precedent_level/
court/date too (additive; reuses the one canonical incoming query — G2).
- precedent_library.get_precedent: add incoming_citations[], shaped like the
RelatedCase row. No parallel resolution path.
- web-ui: render incoming citations in the same row style, read-only (no unlink),
de-duped against manual links; manual "קשר" linking preserved alongside.
Invariants: G2 (single citation-resolution path, reused), G1 (graph self-heals
via existing relink_orphan_citations on upload). No schema change.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
extract_and_store read the WHOLE full_text, so a ruling a party cited in its
own arguments (block ז / appellant_claims / parties_claims) was recorded as the
deciding committee citing it — and surfaced as "צוטט ע״י דפנה". Example: ערר
130/10 appears only in the appellant_claims chunk of decision 1200-12-25, yet
was attributed to the chair.
Fix: build the extraction text from the discussion/ruling chunks via the
halacha extractor's _select_extractable_chunks (G2 — one definition of
"reasoning sections": legal_analysis/ruling/conclusion + discussion-anchor,
excluding facts/claims/intro). Falls back to full_text only for un-chunked rows.
A citation now reflects the deciding body's reliance, not a party's argument.
Existing data reconciled via a one-off (backup
data/audit/citation-edges-backup-*.csv): per source decision, dropped edges
whose cited number is absent from the discussion-only set — 82 of 398 removed
(incl. ערר 130/10 ↔ 1200-12-25); 316 genuine edges kept.
Invariants: G1 (correct at the source) · G2 (single reasoning-section selector,
shared with halacha extraction) · INV-LRN2 (quality-at-source).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Topic hubs in /graph are built from case_law.subject_tags. The documented
extraction contract (precedent_library tool examples: קווי_בניין,
מועד_קביעת_שומה) and ~99% of the corpus store plain multi-word Hebrew tags
with underscores between words ("היטל_השבחה"). A single case (8126-03-25,
יעקב עמיאל) was tagged with spaces ("היטל השבחה"), which the graph renders as
a SECOND, distinct topic hub — a duplicate of the underscore form. The data
was normalized separately; this enforces the convention at the source so no
write path can re-introduce the split.
_normalize_subject_tags() is applied at the three (and only) case_law write
chokepoints in db.py — create_external_case_law, create_internal_committee_decision,
update_case_law — so the rule cannot be bypassed (G1: normalize at source, not
in the read/graph path). Tags carrying punctuation/digits/dashes
(e.g. "פטור מותנה — סעיף 19(ג)") are left untouched; only plain Hebrew word
phrases ([א-ת]+ separated by spaces) are converted. Also dedups post-normalize.
digests.subject_tags (TEXT[], a different graph layer) and canonical_halachot
are intentionally out of scope.
Invariants: maintains G1 (fix at source, not in the read projection),
G2 (graph_api stays a pure read projection — no parallel normalization there).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Implements the chair-approved redesign of /missing-precedents (Claude Design
card 09). The "צוטט ע״י" and "תיק" columns were empty for ~98% of rows because
the corpus-decision reliance lives in precedent_internal_citations (keyed to
case_law), a different system than missing_precedents (keyed to cases).
Backend (G2 — read-time bridge, no stored duplication):
- list_missing_precedents: new cited_by_chairs / cited_by_decisions columns,
resolved from the citation graph by normalized docket number (same
normalization the relinker uses). For an open gap cited by committee
decisions, surfaces the chair(s) (e.g. דפנה תמיר) + the deciding case numbers.
Frontend (matches approved mockup):
- "צוטט ע״י" shows the chair chip (bridge) with priority over the generic
discovery-source chips; "תיק" shows the deciding decision number(s) + "+N".
- Status filter reduced to פתוח / נסגר (removed הועלה, לא-רלוונטי, הכל).
- "פסיקה" column collapses to a single line when case_name is empty — fixes the
duplicated-citation two-row rendering.
- Lifecycle note simplified to open → closed.
Verified: bridge SQL returns chair+decisions on live data; tsc --noEmit clean;
py_compile clean.
Follow-up (data, separate): LLM backfill of legal_topic (נושא); create
missing_precedents rows for the 28 graph-cited rulings not yet tracked.
Invariants: G2 (single source of truth — bridge derived at read, no duplicated
column) · INV-AH context (closing source gaps).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
precedent_internal_citations resolves cited_case_law_id only at extraction
time. When a decision cites a ruling that isn't in the corpus yet, the edge is
stored with cited_case_law_id=NULL. The ruling is often uploaded days later —
but the orphan edge was never re-resolved, so the citation graph permanently
under-counted real connectivity (same stale-derived-value pattern as the
searchable flag). Worse, external rulings store case_number WITHOUT the court
prefix ("3213/97") while citations carry it ("ע\"א 3213/97"), so a large share
of "missing" citations were actually present-but-unlinked.
Fix:
- new citation_extractor.relink_orphan_citations(case_law_id=None): re-resolves
orphan edges via the canonical _resolve_case_law_id (no parallel matching, G2);
scoped to one row, or full sweep when None. Idempotent.
- ingest_document calls it on every upload (non-fatal) so the graph self-heals.
Existing backlog already re-linked via a one-off sweep against the live DB:
linked 128 edges → in-corpus distinct rulings 67→178, unlinked distinct
262→143. So real "cited but not uploaded" ≈ 143, not the 262 the stale graph
implied.
Invariants: G1 (re-resolve the derived link at the write) · G2 (single resolver).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The `searchable` completeness flag is recomputed only at ingest end and after
the Gemini metadata extractor. Internal-committee decisions skip Gemini, so
when their summary/subject_tags/metadata are filled after ingest the flag goes
stale and the row stays invisible to RAG despite being complete. Found 4 such
decisions (1049-06-21, 8126-03-25 [יעקב עמיאל], 8181-21 [האוניברסיטה העברית],
85074-04-25) — all complete (chunks+embeddings, extraction+halacha completed,
metadata present) yet searchable=false.
Fix: recompute_searchable() at the end of update_case_law — the single
chokepoint for metadata patches — so the derived flag never drifts from the
content (G1). Existing stale rows already corrected via a one-off canonical
recompute_searchable(None) run (corpus: 328 -> 332 searchable).
Invariants: INV-DM1 (completeness contract) · G1 (normalize derived value at
the write, no read-time patch).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Haifa (מחוז חיפה) and the North (מחוז הצפון) are separate Israeli planning
districts, each with its own appeals committee. Two divergent bugs collapsed
Haifa into צפון:
- services/internal_decisions.py _VALID_DISTRICTS omitted "חיפה" (while the
tools-layer VALID_DISTRICTS and db.py citation-abbrev map both include it —
a G2 single-source-of-truth divergence)
- _COURT_TO_DISTRICT mapped "חיפה" -> "צפון", so Haifa courts derived צפון
Result: 2 Haifa committee decisions (1074-08-23, 8508-03-24) were mis-filed
under צפון, and "חיפה" never appeared as a district.
Changes:
- add "חיפה" to service _VALID_DISTRICTS
- _COURT_TO_DISTRICT: ("חיפה","צפון") -> ("חיפה","חיפה")
- SCHEMA_V43: reclassify legacy internal rows whose court mentions חיפה from
צפון -> חיפה (normalize at source, G1; idempotent)
Verified against live DB (rollback txn): 2 rows move צפון(6)->צפון(4)+חיפה(2).
Invariants: G1 (normalize at source) · G2 (single source of truth for the
valid-district set — follow-up: consolidate the 3 divergent lists).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
ועדת-ערר decisions (source_kind=internal_committee / source_type=
appeals_committee) are persuasive authority — they do not bind another
committee, nor the committee itself. The stored is_binding column wrongly
defaulted to True across the FastAPI form + service/db layers, so 46 of 92
committee rows were marked binding and got the BINDING halacha-extraction
prompt. Authority is structural for this source (INV-DM7) — normalize at the
source (G1), not trust the input.
Changes:
- db.create_internal_committee_decision: coerce is_binding=False (structural)
- db.create_external_case_law: coerce False when source_type=appeals_committee
- db.update_case_law: coerce False when a patch relabels source_type=
appeals_committee (Gemini reclassification path)
- internal_decisions.migrate_from_external_corpus: set is_binding=FALSE on
the external→internal reclassification UPDATE
- service + FastAPI-form defaults: True → False for the internal path
- SCHEMA_V42: backfill legacy committee rows (is_binding True→False) +
CHECK constraint case_law_committee_not_binding_check so a binding
committee row can never be written again ("so it doesn't recur")
- spec: X8-field-provenance aligned to INV-DM7
Verified against live DB in a rollback transaction: 46 rows flip to 0;
court_ruling (239) + cited_only (31) unaffected; binding-committee INSERT
rejected by the constraint.
Invariants: INV-DM7 (authority ⊥ rule-type) · G1 (normalize at source) ·
G2 (single source of truth, no parallel path).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
chaim 2026-06-20: a unanimous 3-vote principle must outrank a 2-vote one
regardless of score (cross-model agreement is the more reliable keep signal).
apply_cap now sorts survivors by (votes, score), matching cluster_candidates.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Phase C — scripts/cull_principles.py: re-adjudicates every existing 'original'
principle with the SAME panel regime (panel_keep_score → classify → apply_cap),
reversible (CSV backup + rejected canonical recoverable), usage-throttled.
panel_extraction.panel_keep_score + apply_cap (shared, G2). Dry-run on 3
decisions: 37→15 survive.
Phase D — services/principles.py: source-derived label הלכה (binding court) /
כלל פרשני (committee) / עיקרון (persuasive); umbrella עקרונות משפטיים. Wired into
canonical_halacha_get/list (principle_class+principle_label). UI string changes
deferred to the Claude Design gate. spec INV-LRN7; SCRIPTS.md; 7 new tests; 428 green.
Phase E needs no new code — synthesis already targets pending_synthesis, which the
cull leaves only on survivors (rejected canonicals → 'rejected').
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
