feat(kb): add 'caselaw' (פסיקה) as a fourth kind

Court rulings are a distinct content type — long discussion / reasoning,
no enumerated statutory hierarchy, case identifier instead of statute id —
so they get their own kind alongside law/regulation/circular rather than
being shoehorned into one of the existing three.

Migration 003 alters the CHECK constraints on kb_source.kind and
kb_ingest_job.kind to include 'caselaw'. The chunker uses the
circulars path for caselaw too (paragraph + size split, no statutory
section regex); the case identifier is captured at upload time in
kb_source.identifier.

Updates Literal types in ingest_source, admin_kb upload, and the
kb_public SearchRequest. _KINDS in s3.py grows to four so failed/
processed/inbox listings include caselaw subdirs.

Refs Task Master #18 (espocrm-extensions/KnowledgeBase)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

api/routes/admin_kb.py

@@ -37,7 +37,7 @@ def _verify_admin(request: Request) -> None:
 @router.post("/ingest")
 async def ingest(
     request: Request,
-    kind: Literal["law", "regulation", "circular"] = Form(...),
+    kind: Literal["law", "regulation", "circular", "caselaw"] = Form(...),
     title: str = Form(...),
     identifier: str | None = Form(None),
     published_at: str | None = Form(None),
@@ -226,7 +226,7 @@ _MAX_UPLOAD_BYTES = 50 * 1024 * 1024
 @router.post("/upload")
 async def upload(
     request: Request,
-    kind: Literal["law", "regulation", "circular"] = Form(...),
+    kind: Literal["law", "regulation", "circular", "caselaw"] = Form(...),
     topic_id: int = Form(...),
     title: str | None = Form(None),
     identifier: str | None = Form(None),

api/routes/kb_public.py

@@ -36,7 +36,7 @@ def _verify_auth(request: Request) -> None:
 
 class SearchRequest(BaseModel):
     query: str = Field(..., min_length=1, max_length=500)
-    kind: Literal["any", "law", "regulation", "circular"] = "any"
+    kind: Literal["any", "law", "regulation", "circular", "caselaw"] = "any"
     top_k: int = Field(8, ge=1, le=15)
     # Multi-query expansion: ask the LLM for alternative phrasings, retrieve
     # for each, merge, then rerank against the original query. Recovers

api/services/kb/chunker.py

@@ -396,10 +396,16 @@ def chunk_circular(
     return chunks
 
 
-def chunk(text: str, kind: Literal["law", "regulation", "circular"]) -> list[dict]:
+def chunk(text: str, kind: Literal["law", "regulation", "circular", "caselaw"]) -> list[dict]:
     if kind in ("law", "regulation"):
         chunks = chunk_statute(text)
     else:
+        # circulars and caselaw share the same fallback: no enumerated
+        # statutory hierarchy, so we split by paragraph + size. The
+        # caselaw section_ref ends up empty (the chunker can't infer a
+        # case identifier from body text reliably) — the user supplies
+        # the case identifier via the upload form's `identifier` field
+        # and it's stored on the parent kb_source row instead.
         chunks = chunk_circular(text)
     # _split_oversized uses page markers still embedded in content to derive
     # per-piece page numbers; _as_dicts then strips markers and emits the

api/services/kb/ingest.py

@@ -136,7 +136,7 @@ def parse(data: bytes, filename: str) -> str:
 
 async def ingest_source(
     *,
-    kind: Literal["law", "regulation", "circular"],
+    kind: Literal["law", "regulation", "circular", "caselaw"],
     title: str,
     identifier: str | None,
     content: bytes,

api/services/kb/s3.py

@@ -1,9 +1,9 @@
 """MinIO/S3 helpers for the insurance-kb bucket.
 
 Layout in the bucket:
-    inbox/{law,regulation,circular}/<file>
-    processed/{law,regulation,circular}/<file>
-    failed/{law,regulation,circular}/<file>
+    inbox/{law,regulation,circular,caselaw}/<file>
+    processed/{law,regulation,circular,caselaw}/<file>
+    failed/{law,regulation,circular,caselaw}/<file>
 
 After a successful ingest, the file is moved from inbox/* to processed/*.
 On failure it is moved to failed/*.
@@ -20,7 +20,7 @@ from botocore.client import Config as BotoConfig
 
 logger = logging.getLogger("shira.kb.s3")
 
-_KINDS = ("law", "regulation", "circular")
+_KINDS = ("law", "regulation", "circular", "caselaw")
 
 
 def _client():

api/services/kb/search.py

@@ -34,7 +34,7 @@ _EXPANSION_TIMEOUT_S = 12.0
 
 async def _retrieve_rrf(
     query: str,
-    kind: Literal["law", "regulation", "circular", "any"],
+    kind: Literal["law", "regulation", "circular", "caselaw", "any"],
     topic_id: int | None = None,
 ) -> list[dict]:
     """Run vector + lexical retrieval, fuse with RRF. No rerank, no top_k cut.

scripts/migrations/003_caselaw_kind.sql

@@ -0,0 +1,35 @@
+-- Migration 003: add 'caselaw' (פסיקה) as a fourth content kind.
+--
+-- Task Master #18 (espocrm-extensions/KnowledgeBase). Until now the KB
+-- holds laws / regulations / circulars only. Court rulings (פסיקה) are
+-- a distinct content type — different structure (no numbered sections,
+-- long discussion / reasoning, case identifier instead of statute id) —
+-- so we surface it as a separate `kind` rather than abusing one of the
+-- existing three.
+--
+-- The CHECK constraints on kb_source.kind and kb_ingest_job.kind are
+-- recreated with the expanded set. `kb_source.kind` rules every search
+-- and ask path; `kb_ingest_job.kind` rules the upload form. Both must
+-- agree.
+--
+-- Idempotent — wrapped in a transaction; if the constraint is already
+-- in the new shape, the DROP is a no-op (IF EXISTS) and the ADD just
+-- re-asserts.
+--
+-- Run as postgres superuser:
+--   docker exec -i <pg-container> psql -U postgres -d insurance_kb \
+--     -f /tmp/003_caselaw_kind.sql
+
+BEGIN;
+
+ALTER TABLE kb_source DROP CONSTRAINT IF EXISTS kb_source_kind_check;
+ALTER TABLE kb_source
+  ADD CONSTRAINT kb_source_kind_check
+  CHECK (kind IN ('law','regulation','circular','caselaw'));
+
+ALTER TABLE kb_ingest_job DROP CONSTRAINT IF EXISTS kb_ingest_job_kind_check;
+ALTER TABLE kb_ingest_job
+  ADD CONSTRAINT kb_ingest_job_kind_check
+  CHECK (kind IN ('law','regulation','circular','caselaw'));
+
+COMMIT;