# -*- coding: utf-8 -*-
"""统计引擎全量回归驱动：顺序串跑全量测试（Temp/claude 的 e2e_* 探针 + backend/scripts 的 tc 套件）。

用法：
    python run_regression.py                       # 全量（e2e 探针 + 正式 tc 套件）
    python run_regression.py gwas ocs              # 只跑名字含 gwas / ocs 的测试
    python run_regression.py --keep-logs           # 保留上次日志（默认每次清空）

覆盖：
  - Temp/claude/e2e_*.py（排除 e2e_probe* 探查工具）——61 个统计引擎/业务链探针
  - backend/scripts/test_bre_analysis_*_tc.py——正式 tc 套件（TestClient 走真实 API，
    ssGBLUP/OCS/MABC/孟德尔 golden 数值断言，新模块必测）

说明：
  - 每个测试完整输出写 Temp/claude/regression_logs/<name>.log，控制台只回显尾部 + EXIT。
  - 测试均为独立进程（不需后端 uvicorn 服务），自带 fixture 清理，顺序串跑最稳
    （曾有过并发机器争用 RC=124 超时的历史，故不并发）。
  - 最终按 EXIT 汇总：任一失败整体 exit != 0（供交付流程判断）。
  - 解释器默认取 sys.executable（须用 dpb env python 运行本脚本）；可用环境变量
    DPB_PYTHON 覆盖（如 DPB_PYTHON=C:/ai/miniconda3/envs/dpb/python.exe）。
"""
import glob
import os
import subprocess
import sys

sys.stdout.reconfigure(encoding="utf-8")
sys.stderr.reconfigure(encoding="utf-8")

ROOT = os.path.dirname(os.path.abspath(__file__))
PROBE_DIR = os.path.join(ROOT, "Temp", "claude")
TC_DIR = os.path.join(ROOT, "backend", "scripts")
LOG_DIR = os.path.join(PROBE_DIR, "regression_logs")
PY = os.environ.get("DPB_PYTHON", sys.executable)
KEEP_LOGS = "--keep-logs" in sys.argv


def main() -> int:
    filters = [a for a in sys.argv[1:] if not a.startswith("--")]
    probes = sorted(glob.glob(os.path.join(PROBE_DIR, "e2e_*.py")))
    probes = [p for p in probes if "e2e_probe" not in os.path.basename(p)]
    probes += sorted(glob.glob(os.path.join(TC_DIR, "test_bre_analysis_*_tc.py")))
    if filters:
        probes = [p for p in probes if any(f in os.path.basename(p) for f in filters)]
    if not probes:
        print("未找到匹配探针")
        return 1

    if not KEEP_LOGS and os.path.isdir(LOG_DIR):
        for old in glob.glob(os.path.join(LOG_DIR, "*.log")):
            os.remove(old)
    os.makedirs(LOG_DIR, exist_ok=True)

    env = dict(os.environ)
    env.setdefault("ENVIRONMENT", "dev")
    env["PYTHONPATH"] = os.pathsep.join([os.path.join(ROOT, "backend"), PROBE_DIR])
    env.setdefault("PYTHONIOENCODING", "utf-8")

    passed, failed, timed_out = [], [], []
    for i, p in enumerate(probes, 1):
        name = os.path.basename(p)
        log_path = os.path.join(LOG_DIR, name + ".log")
        print(f"[{i}/{len(probes)}] {name}")
        try:
            with open(log_path, "w", encoding="utf-8") as log:
                rc = subprocess.call(
                    [PY, "-X", "utf8", p], cwd=os.path.dirname(p), env=env,
                    stdout=log, stderr=subprocess.STDOUT, timeout=1800)
        except subprocess.TimeoutExpired:
            rc = 124
            with open(log_path, "a", encoding="utf-8") as log:
                log.write("\n[回归驱动] 探针超时（>30min）被杀\n")
        with open(log_path, "r", encoding="utf-8") as log:
            tail = log.readlines()[-12:]
        for line in tail:
            print("  " + line.rstrip())
        print(f"  -> EXIT={rc}")
        if rc == 0:
            passed.append(name)
        elif rc == 124:
            timed_out.append(name)
            failed.append(name)
        else:
            failed.append(name)

    print("\n" + "=" * 40)
    print(f"全量回归：pass={len(passed)} fail={len(failed)}"
          + (f"（其中超时 {len(timed_out)}）" if timed_out else ""))
    if failed:
        print("FAILED:")
        for f in failed:
            print(f"  {f}")
        print(f"完整输出见 {LOG_DIR}/")
        return 1
    print("全部通过")
    return 0


if __name__ == "__main__":
    sys.exit(main())