#!/usr/bin/env python3
# -*- coding: utf-8 -*-
r"""coverage_enum.py — 本体枚举器（游戏内容↔测试验证对账，B 面板·义务层双源之机器源）

两个文本扫描器＋一次比对，全程不依赖跑游戏（交互命中维度除外——由探针覆盖登记另行承载）：

  扫描器 1（游戏本体枚举）——「游戏有什么」：
    game/data/tables/*.tres       行集主键（`"xxx_id": "yyy"` 形态，逐表逐行）
    game/data/schemas/*.gd        词表常量与字段名（const 块内的引号 token＋var 字段名）
    game/presentation/**/*.tscn   BaseButton 族节点（scene＋节点名）
    game/presentation/**/*.gd     _gui_input/_input 自绘命中类（脚本名）
    game/presentation/views/*.gd  屏与模态清单（*_screen/*_overlay/*_panel 视图类）

  扫描器 2（验证关联）——「验证了什么」：复用 tools/extract_checks.py 的断言静态提取
  （unit/**＋integration/**＋system_test.gd 全量），实体 id 在「断言标签＋cond＋源文件＋函数」
  文料中出现即记一次关联。

  比对＝差集：本体存在、验证零关联＝缺口清单，按类别分组、标首见批次（本工具生成时间）。

诚实边界（契约 §7 在案）：grep 关联有假阳假阴——断言测了实体但文本没带 id＝假缺口（人工核销
兜）；文本带 id 但断言弱＝假覆盖（oracle 改造堵）。推荐约定「断言文本含被测实体 id」提高精度，
不立规约。交互点（按钮/物件被点过没有）不在本工具判据内——那归探针真输入覆盖登记。

用法:  python tools/coverage_enum.py [--out <路径>]
产出:  game/tests/.gates/coverage/ontology_gaps.js（JSONP，B 面板消费）
"""
from __future__ import annotations

import argparse
import datetime
import json
import re
import sys
from pathlib import Path

ROOT = Path(__file__).resolve().parent.parent
GAME = ROOT / "game"
DEFAULT_OUT = GAME / "tests" / ".gates" / "coverage" / "ontology_gaps.js"

TRES_ID_RE = re.compile(r'"(\w+_id)":\s*"([^"]+)"')
SCHEMA_TOKEN_RE = re.compile(r'"([a-z][a-z0-9_]{2,40})"')
SCHEMA_VAR_RE = re.compile(r"^@?export\s+var\s+(\w+)\s*:", re.M)
TSCN_NODE_RE = re.compile(r'\[node name="([^"]+)"[^>]*type="(\w*Button\w*)"')
GUI_INPUT_RE = re.compile(r"^func (_gui_input|_input)\(", re.M)
VIEW_KIND_RE = re.compile(r"(screen|overlay|panel)$")


def scan_tables() -> list[dict]:
    out: list[dict] = []
    tables = GAME / "data" / "tables"
    for f in sorted(tables.glob("*.tres")):
        text = f.read_text(encoding="utf-8", errors="replace")
        seen: set[tuple[str, str]] = set()
        for m in TRES_ID_RE.finditer(text):
            key = (m.group(1), m.group(2))
            if key not in seen:
                seen.add(key)
                out.append({"category": "表行:%s" % f.stem, "id": m.group(2), "id_field": m.group(1)})
    return out


def scan_schemas() -> list[dict]:
    out: list[dict] = []
    schemas = GAME / "data" / "schemas"
    for f in sorted(schemas.glob("*.gd")):
        text = f.read_text(encoding="utf-8", errors="replace")
        tokens = set(SCHEMA_TOKEN_RE.findall(text))
        fields = set(SCHEMA_VAR_RE.findall(text))
        for t in sorted(tokens):
            out.append({"category": "词表/枚值:%s" % f.stem, "id": t, "id_field": "token"})
        for fl in sorted(fields):
            out.append({"category": "schema字段:%s" % f.stem, "id": fl, "id_field": "field"})
    return out


def scan_tscn_buttons() -> list[dict]:
    out: list[dict] = []
    for f in sorted(GAME.glob("presentation/**/*.tscn")):
        text = f.read_text(encoding="utf-8", errors="replace")
        for m in TSCN_NODE_RE.finditer(text):
            out.append({"category": "场景按钮", "id": "%s:%s" % (f.name, m.group(1)), "id_field": m.group(2)})
    return out


def scan_gui_input_classes() -> list[dict]:
    out: list[dict] = []
    for f in sorted(GAME.glob("presentation/**/*.gd")):
        text = f.read_text(encoding="utf-8", errors="replace")
        kinds = set(GUI_INPUT_RE.findall(text))
        for k in kinds:
            out.append({"category": "自绘命中类", "id": f.stem, "id_field": k})
            break   # 一类一行（一个脚本同具两面也只记一次）
    return out


def scan_screens() -> list[dict]:
    out: list[dict] = []
    for f in sorted((GAME / "presentation" / "views").glob("*.gd")):
        if VIEW_KIND_RE.search(f.stem):
            out.append({"category": "屏与模态", "id": f.stem, "id_field": "view"})
    return out


def assertion_corpus() -> tuple[list[dict], str]:
    sys.path.insert(0, str(ROOT / "tools"))
    import extract_checks as ec   # noqa: PLC0415 —— 复用现役提取机制（契约 §7 指定扩面）
    rows, _problems = ec.extract_all(ec.default_sources())
    parts: list[str] = []
    for r in rows:
        parts.append("%s %s %s %s" % (r.get("src", ""), r.get("func", ""),
                                      r.get("label_literal", ""), r.get("cond", "")))
    return rows, "\n".join(parts)


def main() -> int:
    if hasattr(sys.stdout, "reconfigure"):
        sys.stdout.reconfigure(encoding="utf-8", errors="replace")
    ap = argparse.ArgumentParser(description="本体枚举器（游戏内容↔测试验证对账）")
    ap.add_argument("--out", default=str(DEFAULT_OUT))
    args = ap.parse_args()
    entities = scan_tables() + scan_schemas() + scan_tscn_buttons() + scan_gui_input_classes() + scan_screens()
    _rows, corpus = assertion_corpus()
    categories: dict[str, dict] = {}
    for ent in entities:
        cat = categories.setdefault(ent["category"], {"total": 0, "matched": 0, "gaps": []})
        cat["total"] += 1
        if ent["id"] and ent["id"] in corpus:
            cat["matched"] += 1
        else:
            cat["gaps"].append({"id": ent["id"], "id_field": ent.get("id_field", "")})
    payload = {
        "meta": {
            "generated_at": datetime.datetime.now().isoformat(timespec="seconds"),
            "entities_total": len(entities),
            "corpus_rows": len(_rows),
            "honest_boundary": "grep 关联有假阳假阴——文本没带 id 的假缺口人工核销；交互命中归探针真输入登记",
        },
        "categories": [
            {"category": k, "total": v["total"], "matched": v["matched"], "gaps": v["gaps"]}
            for k, v in sorted(categories.items())
        ],
    }
    out = Path(args.out)
    out.parent.mkdir(parents=True, exist_ok=True)
    js = "/* 本体对账清单——coverage_enum.py 随跑重生成，人工核销走豁免清单（coverage_exemptions.js），不落本文件。 */\n"
    js += "window.__QA_COVERAGE__ = window.__QA_COVERAGE__ || {};\n"
    js += "window.__QA_COVERAGE__.ontology = %s;\n" % json.dumps(payload, ensure_ascii=False, indent=2)
    out.write_text(js, encoding="utf-8", newline="\n")
    gap_total = sum(len(c["gaps"]) for c in payload["categories"])
    print("[ontology] 本体实体 %d / 断言文料行 %d / 缺口 %d（%d 类）→ %s" % (
        len(entities), len(_rows), gap_total, len(payload["categories"]), out))
    for c in payload["categories"]:
        print("  %s：%d 实体 / %d 缺口" % (c["category"], c["total"], len(c["gaps"])))
    return 0


if __name__ == "__main__":
    raise SystemExit(main())