Hybrid RAG: BM25+Dense (sqlite-vec/BGE-M3) + cross-encoder reranker (bge-reranker-v2-m3)

src/kpaa/cases/index.py

@@ -263,6 +263,41 @@ def search(
     ]
 
 
+def get_cases(
+    ntt_ids: Iterable[str], *, db_path: Path | None = None
+) -> dict[str, Case]:
+    """ntt_id 리스트 → {ntt_id: Case} dict.
+
+    Hybrid retrieval 의 dense 검색 결과(chunk_id='case_<ntt_id>')를 Case 로 복원할 때 사용.
+    """
+    ids = [str(i) for i in ntt_ids]
+    if not ids:
+        return {}
+    path = db_path or default_db_path()
+    if not path.exists():
+        return {}
+    placeholders = ",".join("?" * len(ids))
+    conn = _connect(path)
+    try:
+        rows = conn.execute(
+            f"SELECT * FROM cases WHERE ntt_id IN ({placeholders})", ids
+        ).fetchall()
+    finally:
+        conn.close()
+    return {
+        r["ntt_id"]: Case(
+            ntt_id=r["ntt_id"], ntt_no=r["ntt_no"],
+            title=r["title"], summary=r["summary"], body=r["body"],
+            type_code=r["type_code"], type_label=r["type_label"],
+            category1=r["category1"], category2=r["category2"], category3=r["category3"],
+            reg_dt=r["reg_dt"], case_year=r["case_year"],
+            source_note=r["source_note"], detail_url=r["detail_url"],
+            chunk_context=(r["chunk_context"] if "chunk_context" in r.keys() else "") or "",
+        )
+        for r in rows
+    }
+
+
 def count(db_path: Path | None = None) -> int:
     path = db_path or default_db_path()
     if not path.exists():