"""OPEN #19 regression snapshot: table records and expected-side proposals of the reference piers.

usage: python tools/open19_snapshot.py <tag>   -> 검증/open19_snapshot_<tag>.json
Run before and after the OPEN #19 changes and compare (tools/test_open19.py reads both).
Sources are read only: Ungok P1 / P2 data folders, Ulsan texts from the recycle-bin copy + protected template.
"""
import json
import re
import sys
from pathlib import Path

APP = Path(__file__).resolve().parents[1]
sys.path.insert(0, str(APP))
sys.path.insert(0, str(APP / "tools"))
from engine import reader  # noqa: E402
from testlib import pier_dir  # noqa: E402

PIERS = {
    "운곡교P1": (APP / "data/운곡교P1/drawings/texts.json", APP / "data/운곡교P1"),
    "운곡교p2": (pier_dir("운곡교p2") / "drawings/texts.json", pier_dir("운곡교p2")),
    "산외교울산P1": (APP / "data/_휴지통/산외교울산P1_20261013_131140/drawings/texts.json",
                 APP / "data/_템플릿/산외교울산P1형"),
}


def snapshot():
    out = {}
    for name, (tx, d) in PIERS.items():
        texts = json.loads(tx.read_text(encoding="utf-8"))
        P = json.loads((d / "params.json").read_text(encoding="utf-8"))
        X = json.loads((d / "expected.json").read_text(encoding="utf-8"))
        tables = reader.table_candidates(texts)
        recs = {}
        for t in tables:
            if not t["rows"]:
                continue
            pred = (lambda s: s.isdigit()) if not t["key"].startswith("schedule") else (lambda s: bool(reader.MARK_RE.match(s)))
            recs[f"{t['anchor']['dwg']}|{t['key']}|{t['anchor']['y']}"] = [r for _, r in reader.table_records(t["rows"], pred)]
        rows = reader.proposals(reader.field_candidates(tables), P, X)
        out[name] = {"records": recs,
                     "proposals": [{k: r[k] for k in ("file", "path", "current", "candidate", "same")} for r in rows]}
    return out


if __name__ == "__main__":
    tag = sys.argv[1]
    p = APP / "검증" / f"open19_snapshot_{tag}.json"
    p.write_text(json.dumps(snapshot(), ensure_ascii=False, indent=1), encoding="utf-8")
    s = json.loads(p.read_text(encoding="utf-8"))
    for k, v in s.items():
        props = v["proposals"]
        print(k, "tables", len(v["records"]), "records", sum(len(x) for x in v["records"].values()),
              "proposals", len(props), "differ", sum(not r["same"] for r in props),
              "params rows", sum(r["file"] == "params" for r in props))