# -*- coding: utf-8 -*-
"""
比对 JSON 样本字段 vs API 参考文档(.md)字段。
找出 JSON 中存在但 .md 文档"四、响应字段详解"中缺失的字段。

特殊处理：
- settlement_records / recharge_settlements: 从 settleList 内层提取字段
  siteProfile 子字段不提取（ODS 中存为 siteprofile jsonb 列）
- stock_goods_category_tree: 从 goodsCategoryList 内层提取字段
- 嵌套对象（siteProfile, tableProfile）作为整体字段名
"""
import json
import os
import re
import sys

SAMPLES_DIR = os.path.join("docs", "api-reference", "samples")
DOCS_DIR = os.path.join("docs", "api-reference")

# 结构包装器字段（不应出现在比对中）
WRAPPER_FIELDS = {"settleList", "siteProfile", "tableProfile",
                  "goodsCategoryList", "data", "code", "msg",
                  "settlelist", "siteprofile", "tableprofile",
                  "goodscategorylist"}

# 表头关键字（跳过）— 注意 "type" 不能放这里，因为有些表有 type 业务字段
CROSS_REF_HEADERS = {"字段名", "类型", "示例值", "说明", "field", "example", "description"}


def extract_json_fields(table_name: str) -> set:
    """从 JSON 样本提取所有字段名（小写）"""
    path = os.path.join(SAMPLES_DIR, f"{table_name}.json")
    if not os.path.exists(path):
        return set()

    with open(path, "r", encoding="utf-8") as f:
        data = json.load(f)

    # settlement_records / recharge_settlements: settleList 内层
    if table_name in ("settlement_records", "recharge_settlements"):
        settle = data.get("settleList", {})
        if isinstance(settle, list):
            settle = settle[0] if settle else {}
        fields = set()
        for k in settle.keys():
            kl = k.lower()
            if kl in {"siteprofile"}:
                fields.add(kl)  # 作为整体 jsonb 列
                continue
            fields.add(kl)
        return fields

    # stock_goods_category_tree: goodsCategoryList 内层
    if table_name == "stock_goods_category_tree":
        cat_list = data.get("goodsCategoryList", [])
        if cat_list:
            return {k.lower() for k in cat_list[0].keys()
                    if k.lower() not in WRAPPER_FIELDS}
        return set()

    # role_area_association: roleAreaRelations 内层
    if table_name == "role_area_association":
        rel_list = data.get("roleAreaRelations", [])
        if rel_list:
            return {k.lower() for k in rel_list[0].keys()
                    if k.lower() not in WRAPPER_FIELDS}
        return set()

    # 通用：顶层字段
    fields = set()
    for k in data.keys():
        kl = k.lower()
        if kl in WRAPPER_FIELDS:
            # 嵌套对象作为整体
            if kl in ("siteprofile", "tableprofile"):
                fields.add(kl)
            continue
        fields.add(kl)
    return fields


def extract_md_fields(table_name: str) -> set:
    """从 .md 文档的"四、响应字段详解"章节提取字段名（小写）"""
    md_path = os.path.join(DOCS_DIR, f"{table_name}.md")
    if not os.path.exists(md_path):
        return set()

    with open(md_path, "r", encoding="utf-8") as f:
        lines = f.readlines()

    fields = set()
    in_section = False
    in_siteprofile = False
    field_pattern = re.compile(r'^\|\s*`([^`]+)`\s*\|')
    siteprofile_header = re.compile(r'^###.*siteProfile', re.IGNORECASE)

    for line in lines:
        s = line.strip()

        if s.startswith("## 四、") and "响应字段" in s:
            in_section = True
            in_siteprofile = False
            continue

        if in_section and s.startswith("## ") and not s.startswith("## 四"):
            break

        if not in_section:
            continue

        # siteProfile 子章节处理
        if table_name in ("settlement_records", "recharge_settlements"):
            if siteprofile_header.search(s):
                in_siteprofile = True
                continue
            if s.startswith("### ") and in_siteprofile:
                if not siteprofile_header.search(s):
                    in_siteprofile = False

        m = field_pattern.match(s)
        if m:
            raw = m.group(1).strip()
            if raw.lower() in {h.lower() for h in CROSS_REF_HEADERS}:
                continue
            if table_name in ("settlement_records", "recharge_settlements"):
                if in_siteprofile:
                    continue
                if raw.startswith("siteProfile."):
                    continue
            if raw.lower() in WRAPPER_FIELDS and raw.lower() not in ("siteprofile", "tableprofile"):
                continue
            fields.add(raw.lower())

    return fields


def main():
    samples = sorted([
        f.replace(".json", "")
        for f in os.listdir(SAMPLES_DIR)
        if f.endswith(".json")
    ])

    results = []
    for table in samples:
        json_fields = extract_json_fields(table)
        md_fields = extract_md_fields(table)

        # JSON 中有但 .md 中没有的
        json_only = json_fields - md_fields
        # .md 中有但 JSON 中没有的（可能是条件性字段，仅供参考）
        md_only = md_fields - json_fields

        results.append({
            "table": table,
            "json_count": len(json_fields),
            "md_count": len(md_fields),
            "json_only": sorted(json_only),
            "md_only": sorted(md_only),
        })

    # 输出
    print("=" * 80)
    print("JSON 样本 vs .md 文档 字段比对报告")
    print("=" * 80)

    issues = 0
    for r in results:
        if r["json_only"]:
            issues += 1
            print(f"\n❌ {r['table']} — JSON={r['json_count']}, MD={r['md_count']}")
            print(f"   JSON 中有但 .md 缺失 ({len(r['json_only'])} 个):")
            for f in r["json_only"]:
                print(f"     - {f}")
            if r["md_only"]:
                print(f"   .md 中有但 JSON 无 ({len(r['md_only'])} 个，可能是条件性字段):")
                for f in r["md_only"]:
                    print(f"     - {f}")
        else:
            status = "✅" if not r["md_only"] else "⚠️"
            extra = ""
            if r["md_only"]:
                extra = f" (.md 多 {len(r['md_only'])} 个条件性字段)"
            print(f"\n{status} {r['table']} — JSON={r['json_count']}, MD={r['md_count']}{extra}")

    print(f"\n{'=' * 80}")
    print(f"总计: {len(results)} 个表, {issues} 个有 JSON→MD 缺失")

    # 输出 JSON 格式供后续处理
    out_path = os.path.join("docs", "reports", "json_vs_md_gaps.json")
    with open(out_path, "w", encoding="utf-8") as f:
        json.dump(results, f, ensure_ascii=False, indent=2)
    print(f"\n详细结果已写入: {out_path}")


if __name__ == "__main__":
    main()

# AI_CHANGELOG:
# - 日期: 2026-02-14
# - Prompt: P20260214-044500 — "md文档和json数据不对应！全面排查"
# - 直接原因: 用户要求全面排查 JSON 样本与 .md 文档的字段一致性
# - 变更摘要: 新建脚本，从 JSON 样本提取字段与 .md 文档"响应字段详解"章节比对；
#   修复 3 个 bug（type 过滤、siteProfile/tableProfile 例外、roleAreaRelations 包装器）
# - 风险与验证: 纯分析脚本，无运行时影响；运行 `python scripts/check_json_vs_md.py` 验证输出