# -*- coding: utf-8 -*-
"""验证字符串+结构指纹方案的**跨构建一致性** —— 只基于 har/online/ 全新样本。

案例 §5.3 的验收标准: "3 个 tdc.js 版本、37 个模块、全部正确识别，无一 unknown"。
本脚本再进一步：不仅要求"无 unknown"，还要求
  ① 每份构建内部 37 项都有名字
  ② 各构建之间的**语义名集合**一致（说明识别的是同一批采集器）
  ③ 语义名 -> 每项在 cd 里的**相对位置**可交叉核对
"""
import glob
import sys
from collections import Counter
from pathlib import Path

ROOT = Path(__file__).resolve().parents[2]
sys.path.insert(0, str(ROOT / "tools"))
import tdc_modules as tm  # noqa: E402

BATCH = sorted(glob.glob(str(ROOT / "har" / "online" / "*" / "tdc_*.js")))
print("全新在线样本 %d 份（只使用这批数据）" % len(BATCH))
print()

results = {}
for f in BATCH:
    rows = tm.analyze(f)
    names = [r["name"] for r in rows]
    results[f] = rows
    unk = sum(1 for n in names if n == "unknown")
    via_struct = sum(1 for r in rows if r.get("via_struct"))
    via_deleg = sum(1 for r in rows if r.get("via_delegate"))
    print("  %-24s 项%2d  unknown %d  | 字符串直判 %2d  委派链 %d  结构兜底 %d"
          % (f.split("/")[-1], len(rows), unk,
             len(rows) - via_struct - via_deleg, via_deleg, via_struct))

print()
keys = list(results)
base = Counter(r["name"] for r in results[keys[0]])
print("以 %s 为基准：%d 个语义名（去重 %d）"
      % (keys[0].split("/")[-1], len(base), len(set(base))))
print()
for f in keys[1:]:
    c = Counter(r["name"] for r in results[f])
    same = sum((base & c).values())
    only_base = sum((base - c).values())
    only_it = sum((c - base).values())
    print("  vs %-22s 相同 %2d | 仅基准有 %2d | 仅有它有 %2d"
          % (f.split("/")[-1], same, only_base, only_it))

print()
common = set(base)
for f in keys[1:]:
    common &= set(Counter(r["name"] for r in results[f]))
allnames = set()
for f in keys:
    allnames |= set(r["name"] for r in results[f])
print(">>> 所有构建**共同**识别出的语义名: %d 个" % len(common))
print(">>> 出现过的语义名总数          : %d 个" % len(allnames))
print(">>> 差异名:", sorted(allnames - common) if allnames - common else "无（完全一致）")
print()

# 唯一性检查：单构建内是否有重名（重名 = 锚点有歧义）
print("单构建内重名检查（重名说明该锚点有歧义）:")
for f in keys:
    c = Counter(r["name"] for r in results[f])
    dup = {k: v for k, v in c.items() if v > 1}
    print("  %-24s 重名 %d 组 %s"
          % (f.split("/")[-1], len(dup), dict(sorted(dup.items()))))