cma-management/backend/scripts/verify_threshold_regression_20260911.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""阈值判定变更回归（project-20260911-threshold-fix ⑥）

对比三条链路的判定结果，产出「改前改后差异样本」：
  A. 旧实现：alert_generator 顺序 绿→黄→红 + 正则 (>=|<=|>|<|=|!=)\\s*([\\d.]+)（不支持负数/区间）
  B. 新实现 red→green→yellow（统一 threshold_parser）
  C. 若按任务字面「red→yellow→green」严重度顺序（用于说明为何 green 必须先于 yellow）

输出：逐 (KPI, 期间最新值) 的 A/B 差异清单 + 预警集合差异 + 真实 alert_generator 事务内预演（不落库）
"""
import os
import re
import sys
import json
from collections import Counter

BASE = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.insert(0, BASE)

envf = os.path.join(BASE, ".env")
for line in open(envf, encoding="utf-8"):
    line = line.strip()
    if line and not line.startswith("#") and "=" in line:
        k, v = line.split("=", 1)
        os.environ.setdefault(k.strip(), v.strip().strip('"').strip("'"))

from app.database import get_session_local
from app.models import KPIDefinition, KPIValue
from app.utils.threshold_parser import calc_level_from_exprs

OLD_RE = re.compile(r"(>=|<=|>|<|=|!=)\s*([\d.]+)")


def old_level(actual, kpi) -> str:
    """旧实现：绿→黄→红，命中即止；不支持负数/区间。"""
    if actual is None:
        return "info"
    for level, field in (("green", "threshold_green"), ("yellow", "threshold_yellow"),
                         ("red", "threshold_red")):
        expr = getattr(kpi, field, None)
        if not expr:
            continue
        m = OLD_RE.match(str(expr).strip())
        if not m:
            continue
        op, val = m.group(1), float(m.group(2))
        hit = {">=": actual >= val, "<=": actual <= val, ">": actual > val,
               "<": actual < val, "=": actual == val, "!=": actual != val}[op]
        if hit:
            return level
    return "info"


def severity_order_level(actual, kpi) -> str:
    """按任务字面顺序 red→yellow→green（严重度优先）——用于对比说明。"""
    for level, field in (("red", "threshold_red"), ("yellow", "threshold_yellow"),
                         ("green", "threshold_green")):
        expr = getattr(kpi, field, None)
        if not expr:
            continue
        from app.utils.threshold_parser import eval_threshold
        if eval_threshold(actual, expr):
            return level
    return "info"


def main():
    db = get_session_local()()
    kpis = db.query(KPIDefinition).filter(KPIDefinition.status == "active").all()
    samples = []
    diffs_ab = []
    diffs_bc = []
    count_old = Counter()
    count_new = Counter()
    for kpi in kpis:
        vals = db.query(KPIValue).filter(KPIValue.kpi_id == kpi.id).all()
        # 每期间取最新一条（对齐 alert_generator 取数方式）
        latest_by_period = {}
        for v in vals:
            if v.actual_value is None:
                continue
            key = v.period
            cur = latest_by_period.get(key)
            if cur is None or (v.calculated_at and cur.calculated_at and v.calculated_at > cur.calculated_at):
                latest_by_period[key] = v
        for period, v in sorted(latest_by_period.items()):
            a = old_level(v.actual_value, kpi)
            b = calc_level_from_exprs(v.actual_value, kpi.threshold_green,
                                      kpi.threshold_yellow, kpi.threshold_red)
            c = severity_order_level(v.actual_value, kpi)
            samples.append({
                "kpi_id": kpi.id, "kpi_code": kpi.kpi_code, "period": period,
                "actual": v.actual_value, "old": a, "new": b, "severity_order": c,
                "G": kpi.threshold_green, "Y": kpi.threshold_yellow, "R": kpi.threshold_red,
            })
            count_old[a] += 1
            count_new[b] += 1
            if a != b:
                diffs_ab.append(samples[-1])
            if c != b:
                diffs_bc.append(samples[-1])

    print(f"参与回归的 (KPI, 期间) 判定点: {len(samples)}")
    print("旧实现级别分布 :", dict(count_old))
    print("新实现级别分布 :", dict(count_new))
    print(f"\n=== A(旧) vs B(新) 差异 {len(diffs_ab)} 条 ===")
    for d in diffs_ab:
        print(f"  kpi={d['kpi_code']}(id={d['kpi_id']}) {d['period']} actual={d['actual']} "
              f"old={d['old']} -> new={d['new']}  G={d['G']!r} Y={d['Y']!r} R={d['R']!r}")
    print(f"\n=== B(new=red>green>yellow) vs C(red>yellow>green) 差异 {len(diffs_bc)} 条 ===")
    for d in diffs_bc[:20]:
        print(f"  kpi={d['kpi_code']}(id={d['kpi_id']}) {d['period']} actual={d['actual']} "
              f"B={d['new']} vs C={d['severity_order']}  G={d['G']!r} Y={d['Y']!r} R={d['R']!r}")
    if len(diffs_bc) > 20:
        print(f"  ... 其余 {len(diffs_bc) - 20} 条")

    # 预警集合差异（yellow/red 才落库）
    old_alerts = [(d['kpi_id'], d['period'], d['old']) for d in samples if d['old'] in ("yellow", "red")]
    new_alerts = [(d['kpi_id'], d['period'], d['new']) for d in samples if d['new'] in ("yellow", "red")]
    print(f"\n=== 预警集合：旧 {len(old_alerts)} 条 → 新 {len(new_alerts)} 条 ===")
    so, sn = set(old_alerts), set(new_alerts)
    for x in sorted(so - sn):
        print("  仅旧有(解除):", x)
    for x in sorted(sn - so):
        print("  仅新有(新增):", x)

    # 真实 alert_generator 事务内预演（阻断 commit，事后 rollback，零落库）
    import importlib.util  # noqa: E402
    _spec = importlib.util.spec_from_file_location(
        "cma_alert_generator", os.path.join(BASE, "scripts", "alert_generator.py"))
    ag = importlib.util.module_from_spec(_spec)
    _spec.loader.exec_module(ag)
    from datetime import datetime
    period = datetime.now().strftime("%Y-%m")
    real_commit = db.commit
    db.commit = lambda: None  # 阻断落库
    try:
        new_count = ag.run_alert_check(db, period)
    finally:
        db.commit = real_commit
        db.rollback()
    print(f"\n=== alert_generator 事务内预演（period={period}，已 rollback 未落库）新增预警={new_count} ===")

    out = os.path.join(BASE, "scripts", "backups", "threshold_regression_20260911.json")
    os.makedirs(os.path.dirname(out), exist_ok=True)
    with open(out, "w", encoding="utf-8") as fh:
        json.dump({"samples": len(samples), "count_old": dict(count_old), "count_new": dict(count_new),
                   "diff_ab": diffs_ab, "diff_bc": diffs_bc,
                   "alert_only_old": sorted(so - sn), "alert_only_new": sorted(sn - so),
                   "alert_gen_dry_run_new": new_count}, fh, ensure_ascii=False, indent=2)
    print("报告 JSON:", out)
    db.close()


if __name__ == "__main__":
    main()