"""B-1 PR① — KR 어댑터 순수 파서 fixture 테스트 (plan safety-library-1).

fixture = 2026-06-13 law.go.kr 라이브 박제 (OC 새니타이즈, tests/fixtures/statute_kr/).
파서는 순수 함수라 httpx/DB 불요 — 컨테이너 밖 로컬 실행.
"""

import gzip
import sys
import xml.etree.ElementTree as ET
from pathlib import Path

sys.path.insert(0, str(Path(__file__).parent.parent / "app"))

from workers.statute_adapters import ChangeEvent  # noqa: E402
from workers.statute_adapters.kr import detect_change, parse_search_hit  # noqa: E402

FIX = Path(__file__).parent / "fixtures" / "statute_kr"


def _read(name: str) -> str:
    p = FIX / name
    if name.endswith(".gz"):
        return gzip.decompress(p.read_bytes()).decode("utf-8")
    return p.read_text(encoding="utf-8")


def test_parse_search_hit_exact_match():
    hit = parse_search_hit(_read("lawsearch_sanab.xml"), "산업안전보건법")
    assert hit is not None
    assert hit["law_id"] == "001766"
    assert hit["mst"] == "283449"
    assert hit["promulgation_date"] == "20260219"
    assert hit["effective_date"] == "20260601"
    assert hit["status_code"] == "현행"


def test_parse_search_hit_rejects_partial_name():
    # totalCnt 3 인 응답에서 '산업안전보건법 시행령' 등 부분 일치는 비매칭이어야 함
    hit = parse_search_hit(_read("lawsearch_sanab.xml"), "산업안전보건")
    assert hit is None


def test_detect_change_same_watermark_is_silent():
    hit = parse_search_hit(_read("lawsearch_sanab.xml"), "산업안전보건법")
    assert detect_change(hit, "kr-law:001766", "산업안전보건법", watermark="283449") is None


def test_detect_change_new_mst_is_amend():
    hit = parse_search_hit(_read("lawsearch_sanab.xml"), "산업안전보건법")
    ev = detect_change(hit, "kr-law:001766", "산업안전보건법", watermark="283448")
    assert isinstance(ev, ChangeEvent)
    assert ev.kind == "amend"
    assert ev.new_version_key == "283449"
    assert ev.effective_date == "20260601"


def test_detect_change_empty_watermark_is_amend():
    # 첫 폴링(워터마크 부재) = 변경으로 감지 — PR② 부트스트랩 전 관찰 모드의 기대 동작
    hit = parse_search_hit(_read("lawsearch_sanab.xml"), "산업안전보건법")
    ev = detect_change(hit, "kr-law:001766", "산업안전보건법", watermark=None)
    assert ev is not None and ev.kind == "amend"


def test_detect_change_repeal_keyword():
    hit = {"mst": "9", "revision_type": "폐지", "promulgation_date": None,
           "effective_date": None, "law_id": "x", "status_code": None}
    ev = detect_change(hit, "kr-law:x", "x", watermark="1")
    assert ev is not None and ev.kind == "repeal"


def test_lawservice_snapshot_semantics_rule():
    """R7-M3 판정 박제: 전문 1콜 XML = 조문+별표 전체 스냅샷 (PR② payload 계약의 전제)."""
    root = ET.fromstring(_read("lawservice_rule.xml.gz"))
    articles = root.findall(".//조문단위")
    annexes = root.findall(".//별표단위")
    assert len(articles) >= 800, "산안기준규칙 조문 853 기대 — 전문 1콜 판정 근거"
    assert len(annexes) == 23, "별표 23 전부 본문 XML 포함 = 스냅샷 의미론"
    # R7-M3 ②: 별표 식별 = 구조화 필드 (suffix 문자열 파싱 불요)
    first = annexes[0]
    assert first.findtext("별표번호") is not None
    assert first.findtext("별표가지번호") is not None


def test_lawservice_sanab_basic_info():
    root = ET.fromstring(_read("lawservice_sanab.xml.gz"))
    assert root.findtext(".//법령ID") == "001766"
    assert len(root.findall(".//조문단위")) >= 200
    # 별표 없는 법령 = 별표단위 0 (스냅샷 의미론의 반대쪽 케이스)
    assert len(root.findall(".//별표단위")) == 0