Merge feat/safety-library-a1 (B-4 licensed_restricted + watch 폴더 license) into ds-board-merged

B-4 PR①②: licensed_restricted 단일 술어(retrieval 3-leg/digest/briefing/study 풀이 공유)
+ file_watcher Books/Manuals/Papers_Purchased license 주입. prod 통합 브랜치 배포용.
This commit is contained in:
hyungi
2026-06-13 14:53:34 +09:00
7 changed files with 185 additions and 17 deletions
+80
View File
@@ -0,0 +1,80 @@
"""B-4 — licensed_restricted 차단 술어 + watch 타깃 (material/jurisdiction/license) 매핑 순수 테스트.
차단 술어(_license_sql)는 retrieval 3-leg + digest 가 공유하는 단일 술어. 실제 제외 동작은
GPU 라이브(합성 restricted doc 검색 제외)로 검증 — 여기선 술어 형태 + 매핑 표 계약만.
[[feedback_external_api_fixture_first]] / [[feedback_structural_integrity_over_path_discipline]]
"""
import sys
from pathlib import Path
sys.path.insert(0, str(Path(__file__).parent.parent / "app"))
from services.search.license_filter import ( # noqa: E402
restricted_exclude_orm,
restricted_exclude_sql,
)
from services.search.retrieval_service import _license_sql # noqa: E402
from workers.file_watcher import _TARGET_AXIS # noqa: E402
def test_shared_predicate_single_source():
# retrieval/digest/briefing 가 같은 술어 정의를 공유 — drift 방지(단일 source 계약)
assert _license_sql("d") == " AND " + restricted_exclude_sql("d")
assert _license_sql("") == " AND " + restricted_exclude_sql("")
assert restricted_exclude_sql("d").startswith("COALESCE(d.extract_meta")
def test_restricted_exclude_orm_constructs():
# study 풀이(explanation_rag)용 ORM 표현 — 컴파일 SQL 이 raw 술어와 동일 구조인지
from sqlalchemy.dialects import postgresql
clause = restricted_exclude_orm()
sql = str(clause.compile(dialect=postgresql.dialect(),
compile_kwargs={"literal_binds": True}))
assert "extract_meta" in sql
assert "'license'" in sql and "'restricted'" in sql # JSONB 경로 키
assert "'false'" in sql and "'true'" in sql # COALESCE 기본 + 비교값
def test_license_sql_shape_with_alias():
sql = _license_sql("d")
assert sql.startswith(" AND ") # 항상 ' AND ...' (WHERE 합성용)
assert "COALESCE(d.extract_meta -> 'license' ->> 'restricted', 'false')" in sql
assert "<> 'true'" in sql # restricted=true 만 제외
def test_license_sql_shape_no_alias():
# alias='' = 단일 FROM documents (컬럼 직접 참조)
sql = _license_sql("")
assert "COALESCE(extract_meta -> 'license' ->> 'restricted', 'false')" in sql
assert ".extract_meta" not in sql # 점 없는 컬럼 직접
def test_axis_books_papers_are_restricted():
for folder, mt in (("Books", "book"), ("Papers_Purchased", "paper")):
material, jur, lic = _TARGET_AXIS[folder]
assert material == mt
assert jur is None # 책/논문 = 관할 없음(A-2 paper NULL 강제와 정합)
assert lic["scheme"] == "proprietary"
assert lic["restricted"] is True # RAG/digest 차단 대상
assert lic["redistribute"] is False
def test_axis_manuals_proprietary_but_not_restricted():
material, jur, lic = _TARGET_AXIS["Manuals"]
assert material == "manual"
assert lic["scheme"] == "proprietary"
assert lic["restricted"] is False # 사용자 결정 2026-06-13 (검색·RAG 활용)
def test_axis_kgs_law_kr_public_not_restricted():
material, jur, lic = _TARGET_AXIS["KGS_Code"]
assert (material, jur) == ("law", "KR")
assert lic["scheme"] == "kogl"
assert lic["restricted"] is False # 법정 위임 공공 → 차단 아님
def test_axis_non_target_folder_yields_none():
# Inbox/Recordings 등 비대상 = (None, None, None) → material/license 미주입
assert _TARGET_AXIS.get("Inbox", (None, None, None)) == (None, None, None)