值守: 新增 audit_runs.py —— 把上一轮验证过的核对办法做成工具
上一轮我照着自己写在手册里的办法手工核对, 结果 6 条全部误报"对不上"(坑在
日志结论在结束时、报告 started 在开始时, 差一个 elapsed_s)。手敲 PowerShell
容易出错, 所以做成脚本: 代码里的时间关系比人记的可靠。
用法: python scripts/audit_runs.py (退出码 0 = 全部吻合)
检查三件事:
1. 每份报告都能在日志里找到对应的"开始行", 且时间关系(开始+耗时=结论时刻)成立
2. 日志里每一轮都有对应报告
3. mock 残留: 同一秒扎堆的结论、或内容自相矛盾的结论
⛔ 写第一版时自己踩了同一类坑: 它拿"报告.started + elapsed"去找日志结论行,
能对上就说明 elapsed 一致 —— 于是"耗时不符"那个分支永远不可能触发, 是**死代码**。
看起来在查、实际查不到的东西比没有更糟(和"读不到就跳过"是同一种病)。
改成按日志里的"=== 开始: 体力换模块 ==="配对后, 时间关系和耗时才能各自独立核对。
13 个用例, 每个都是"必须能抓到某种问题"而不是"必须通过";
实测真实数据 6 轮/6 报告全部吻合, 造的三类坏数据(缺报告/耗时不符/内容矛盾)
全部被点名。全套 310/310 通过。
This commit is contained in:
@@ -0,0 +1,145 @@
|
|||||||
|
# -*- coding: utf-8 -*-
|
||||||
|
"""audit_runs 的核对逻辑。
|
||||||
|
|
||||||
|
这个工具的职责是"把日志和报告对起来", 所以测试重点不是它能通过, 而是
|
||||||
|
**它能不能抓到问题** —— 一个什么都不报的体检工具等于没有。
|
||||||
|
|
||||||
|
⛔ 第一版就踩过这个坑: 它拿"报告.started + elapsed"去找日志结论行, 能对上就说明
|
||||||
|
elapsed 一致, 于是"耗时不符"那个分支永远不可能触发 —— 是死代码。现在改成按日志里
|
||||||
|
的"=== 开始: 体力换模块 ===" 配对, 时间关系和耗时才能各自独立核对。
|
||||||
|
"""
|
||||||
|
import json
|
||||||
|
import os
|
||||||
|
import sys
|
||||||
|
import tempfile
|
||||||
|
import unittest
|
||||||
|
from pathlib import Path
|
||||||
|
|
||||||
|
sys.path.insert(0, str(Path(__file__).resolve().parents[1] / 'scripts'))
|
||||||
|
import audit_runs as ar # noqa: E402
|
||||||
|
|
||||||
|
START = '=== 开始: 体力换模块 ==='
|
||||||
|
|
||||||
|
|
||||||
|
def start_line(t):
|
||||||
|
return '[%s] %s' % (t, START)
|
||||||
|
|
||||||
|
|
||||||
|
def conclusion(t, result, secs, ok='OK'):
|
||||||
|
return '[%s] === %s (%s), 耗时 %ds ===' % (t, result, ok, secs)
|
||||||
|
|
||||||
|
|
||||||
|
def report(start, secs, result, ok=True):
|
||||||
|
return json.dumps({'started': start, 'elapsed_s': secs, 'ok': ok,
|
||||||
|
'result': result, 'steps': []}, ensure_ascii=False)
|
||||||
|
|
||||||
|
|
||||||
|
class AuditTest(unittest.TestCase):
|
||||||
|
|
||||||
|
def setUp(self):
|
||||||
|
self.d = tempfile.mkdtemp()
|
||||||
|
|
||||||
|
def write_log(self, lines):
|
||||||
|
with open(os.path.join(self.d, 'convert.log'), 'w', encoding='utf-8') as f:
|
||||||
|
f.write('\n'.join(lines) + '\n')
|
||||||
|
|
||||||
|
def write_report(self, name, content):
|
||||||
|
with open(os.path.join(self.d, name), 'w', encoding='utf-8') as f:
|
||||||
|
f.write(content)
|
||||||
|
|
||||||
|
def problems(self):
|
||||||
|
return ar.audit(self.d)['problems']
|
||||||
|
|
||||||
|
# ---------------------------------------------------------- 正常情况
|
||||||
|
|
||||||
|
def test_matching_run_is_clean(self):
|
||||||
|
self.write_log([start_line('2026-01-01 10:00:00'),
|
||||||
|
conclusion('2026-01-01 10:00:30', '合成完成: 模组 1 -> 2 (+1)', 30)])
|
||||||
|
self.write_report('convert_20260101_100000.json',
|
||||||
|
report('2026-01-01 10:00:00', 30, '合成完成: 模组 1 -> 2 (+1)'))
|
||||||
|
self.assertEqual(self.problems(), [])
|
||||||
|
|
||||||
|
def test_empty_dir_is_clean(self):
|
||||||
|
self.assertEqual(self.problems(), [])
|
||||||
|
|
||||||
|
# ---------------------------------------------------------- 抓问题
|
||||||
|
|
||||||
|
def test_report_without_run_is_flagged(self):
|
||||||
|
self.write_report('convert_20260101_110000.json',
|
||||||
|
report('2026-01-01 11:00:00', 40, '体力不足(3<20), 本轮无需合成'))
|
||||||
|
self.assertTrue(any('找不到对应的开始行' in p for p in self.problems()))
|
||||||
|
|
||||||
|
def test_run_without_report_is_flagged(self):
|
||||||
|
self.write_log([start_line('2026-01-01 12:00:00'),
|
||||||
|
conclusion('2026-01-01 12:00:10', '跑了但没留报告', 10)])
|
||||||
|
self.assertTrue(any('没有对应报告' in p for p in self.problems()))
|
||||||
|
|
||||||
|
def test_elapsed_mismatch_is_flagged(self):
|
||||||
|
"""日志说耗时 30s, 报告说 99s —— 必须能发现。
|
||||||
|
|
||||||
|
老版本这里永远查不出来(死代码), 所以这条用例是专门守它的。
|
||||||
|
"""
|
||||||
|
self.write_log([start_line('2026-01-01 10:00:00'),
|
||||||
|
conclusion('2026-01-01 10:00:30', '合成完成', 30)])
|
||||||
|
self.write_report('convert_20260101_100000.json',
|
||||||
|
report('2026-01-01 10:00:00', 99, '合成完成'))
|
||||||
|
got = self.problems()
|
||||||
|
self.assertTrue(any('耗时不符' in p for p in got), got)
|
||||||
|
|
||||||
|
def test_end_time_mismatch_is_flagged(self):
|
||||||
|
# 开始和耗时都对得上, 但结论时刻对不上 -> 时间关系有问题
|
||||||
|
self.write_log([start_line('2026-01-01 10:00:00'),
|
||||||
|
conclusion('2026-01-01 10:05:00', '合成完成', 30)])
|
||||||
|
self.write_report('convert_20260101_100000.json',
|
||||||
|
report('2026-01-01 10:00:00', 30, '合成完成'))
|
||||||
|
self.assertTrue(any('时间对不上' in p for p in self.problems()))
|
||||||
|
|
||||||
|
def test_result_text_mismatch_is_flagged(self):
|
||||||
|
self.write_log([start_line('2026-01-01 10:00:00'),
|
||||||
|
conclusion('2026-01-01 10:00:30', '合成完成: 模组 1 -> 2 (+1)', 30)])
|
||||||
|
self.write_report('convert_20260101_100000.json',
|
||||||
|
report('2026-01-01 10:00:00', 30, '合成完成: 模组 5 -> 9 (+4)'))
|
||||||
|
self.assertTrue(any('结果文本不符' in p for p in self.problems()))
|
||||||
|
|
||||||
|
def test_same_second_cluster_is_flagged(self):
|
||||||
|
"""同一秒多行结论 = 测试批量产生的典型特征。"""
|
||||||
|
self.write_log([conclusion('2026-01-01 12:00:00', '第一条', 0),
|
||||||
|
conclusion('2026-01-01 12:00:00', '第二条', 0)])
|
||||||
|
self.assertTrue(any('同一秒' in p for p in self.problems()))
|
||||||
|
|
||||||
|
def test_impossible_content_is_flagged(self):
|
||||||
|
self.write_log([start_line('2026-01-01 12:00:00'),
|
||||||
|
conclusion('2026-01-01 12:00:05', '看门狗执行失败: powershell not found', 5)])
|
||||||
|
self.assertTrue(any('可疑内容' in p for p in self.problems()))
|
||||||
|
|
||||||
|
# ---------------------------------------------------------- 解析
|
||||||
|
|
||||||
|
def test_parse_pairs_start_with_conclusion(self):
|
||||||
|
self.write_log([
|
||||||
|
'[2026-01-01 09:00:00] === 开始: 体力换模块 ===',
|
||||||
|
'[2026-01-01 09:00:05] 游戏已在运行',
|
||||||
|
conclusion('2026-01-01 09:01:00', '合成完成', 60),
|
||||||
|
])
|
||||||
|
runs = ar.parse_runs(os.path.join(self.d, 'convert.log'))
|
||||||
|
self.assertEqual(len(runs), 1)
|
||||||
|
self.assertEqual(runs[0]['secs'], 60)
|
||||||
|
self.assertEqual(runs[0]['start'].strftime('%H:%M:%S'), '09:00:00')
|
||||||
|
|
||||||
|
def test_parse_handles_missing_file(self):
|
||||||
|
self.assertEqual(ar.parse_runs(os.path.join(self.d, 'nope.log')), [])
|
||||||
|
|
||||||
|
def test_conclusion_without_start_line_still_listed(self):
|
||||||
|
# 没有开始行的结论不该被丢掉(要能报"没有对应报告")
|
||||||
|
self.write_log([conclusion('2026-01-01 12:00:00', '独苗', 10)])
|
||||||
|
runs = ar.parse_runs(os.path.join(self.d, 'convert.log'))
|
||||||
|
self.assertEqual(len(runs), 1)
|
||||||
|
self.assertIsNone(runs[0]['start'])
|
||||||
|
|
||||||
|
def test_skip_broken_report_json(self):
|
||||||
|
with open(os.path.join(self.d, 'convert_20260101_100000.json'), 'w', encoding='utf-8') as f:
|
||||||
|
f.write('{ 半截')
|
||||||
|
self.assertEqual(ar.load_reports(self.d), [])
|
||||||
|
|
||||||
|
|
||||||
|
if __name__ == '__main__':
|
||||||
|
unittest.main()
|
||||||
Reference in New Issue
Block a user