Aislo/resources/tester/spreadsheet/test_spreadsheet_golden.py

"""F — 금빛 시험: 실무 구조도 xlsx 세 벌의 수식을 새 엔진으로 풀어 엑셀 캐시값과 대조.

xlsx 는 **시험에서만** openpyxl 로 읽음(앱에는 가져오기 없음). 식 = 엑셀 글 그대로 · 값 = 칸 값 ·
바깥 통합문서 참조(`[172]관집계!…`)는 식을 못 푸니 캐시값을 값으로 굳힘(대조에서 뺌).
갈린 칸은 조용히 맞추지 않음 — 목록을 `tmp/spreadsheet_golden_report.json` 에 적고 시험을 떨굼.
엔진(A · B)이 빈 몸이면 건너뜀 — 엔진이 차는 대로 다시 돌림.
"""

import datetime as dt
import json
import re
import shutil
from decimal import Decimal, InvalidOperation
from pathlib import Path

import pytest

openpyxl = pytest.importorskip("openpyxl")
from openpyxl.utils.datetime import to_excel  # noqa: E402

from common_util import common_util_spreadsheet as spreadsheet  # noqa: E402

ROOT = Path(__file__).resolve().parents[3]
PRACTICE = ROOT / "resources" / "knowledge" / "original" / "실무문서"
NAMES = ("07-구조도-소광리.xlsx", "5. 구조도(기번3).xlsx", "구조도(변경).xlsx")
FILES = [next(PRACTICE.rglob(name)) for name in NAMES]
REPORT = ROOT / "tmp" / "spreadsheet_golden_report.json"
EXPECTED_FORMULAS = 2699
_EXTERNAL = re.compile(r"\[\d+\]")
_ERRORS = {"#NULL!", "#DIV/0!", "#VALUE!", "#REF!", "#NAME?", "#NUM!", "#N/A"}

pytestmark = pytest.mark.skipif(shutil.which("node") is None, reason="node 가 없음")


def _plain(value):
    """칸 값 → JSON(날짜 · 시각 형식 칸은 엑셀 일련수)."""
    if isinstance(value, (dt.datetime, dt.date, dt.time, dt.timedelta)):
        return to_excel(value)
    return value


def _book(path: Path):
    """xlsx → 통합문서 + 대조할 칸 `[(시트 id, A1, 캐시값)]`."""
    formulas = openpyxl.load_workbook(path, data_only=False)
    cached = openpyxl.load_workbook(path, data_only=True)
    sheets, checks = [], []
    for i, ws in enumerate(formulas.worksheets, start=1):
        sid, cells = f"s{i}", {}
        values = cached[ws.title]
        for row in ws.iter_rows():
            for c in row:
                if c.value is None:
                    continue
                keep = _plain(values[c.coordinate].value)
                if c.data_type == "f" and isinstance(c.value, str):
                    if _EXTERNAL.search(c.value):
                        if keep is not None:
                            cells[c.coordinate] = {"값": keep}
                        continue
                    cells[c.coordinate] = {"식": c.value[1:]}
                    checks.append((sid, c.coordinate, keep))
                elif c.data_type != "f":
                    cells[c.coordinate] = {"값": _plain(c.value)}
        sheets.append({"id": sid, "이름": ws.title, "칸": cells})
    book = {"종류": "통합문서", "판": 1, "열": path.stem, "서식": [{}], "시트": sheets}
    return book, checks


def _same(want, got) -> bool:
    if got is None:
        return False
    if isinstance(want, bool):
        return got.get("참") is want
    if isinstance(want, (int, float)):
        try:
            a, b = Decimal(got["수"]), Decimal(repr(float(want)))
        except (KeyError, InvalidOperation):
            return False
        return abs(a - b) <= Decimal("1e-9") * max(Decimal(1), abs(b))
    if want is None or want == "":
        return got.get("글") == "" or got.get("수") == "0"
    if isinstance(want, str) and want in _ERRORS:
        return got.get("오류") == want
    return got.get("글") == want


def test_실무_구조도_수식_엑셀과_같음():
    loaded = [(path, *_book(path)) for path in FILES]
    total = sum(len(checks) for _p, _b, checks in loaded)
    external = 6  # 소광리 `[172]관집계` — 값으로 굳혀 대조에서 뺌
    assert total + external == EXPECTED_FORMULAS
    try:
        results = spreadsheet.recalc_workbooks([book for _p, book, _c in loaded])
    except spreadsheet.RecalcError as error:
        if "아직 없음" in str(error):
            pytest.skip(f"엔진 아직 — {error}")
        raise
    mismatches = []
    for (path, book, checks), values in zip(loaded, results, strict=True):
        names = {s["id"]: s["이름"] for s in book["시트"]}
        for sid, a1, want in checks:
            got = values.get(sid, {}).get(a1)
            if not _same(want, got):
                formula = next(s for s in book["시트"] if s["id"] == sid)["칸"][a1]["식"]
                mismatches.append(
                    {
                        "파일": path.name,
                        "시트": names[sid],
                        "칸": a1,
                        "식": formula,
                        "엑셀": want,
                        "엔진": got,
                    }
                )
    REPORT.parent.mkdir(exist_ok=True)
    REPORT.write_text(
        json.dumps(
            {"대조": total, "갈림": len(mismatches), "목록": mismatches},
            ensure_ascii=False,
            indent=1,
            default=str,
        ),
        encoding="utf-8",
    )
    head = "\n".join(
        f"{m['파일']} {m['시트']}!{m['칸']} ={m['식']}  엑셀 {m['엑셀']!r} · 엔진 {m['엔진']}"
        for m in mismatches[:30]
    )
    assert not mismatches, f"{len(mismatches)}/{total} 칸 갈림 — 전체 {REPORT}\n{head}"