Files
Video-Tree-TRM5/research-wiki/plans/2026-07-16-preflight-wp4-resilience.md
T

28 KiB
Raw Blame History

WP4 韧性与持久化 Implementation Plan

For agentic workers: REQUIRED SUB-SKILL: Use subagent-driven-development to implement this plan task-by-task. Steps use checkbox (- [ ]) syntax for tracking.

Goal: 修复 LLM 治理栈与持久化层的 10 处韧性缺陷,使训练信号不被缓存重放/截断响应/断连/INFRA 故障污染,且 workspace 元数据崩溃可恢复。

Architecture: 缓存加 salt 维度让跨 epoch 评估真实重采样(同 epoch 续跑仍命中);SSE 未收 [DONE] 视为截断进重试且不写缓存;扩展瞬时错误清单覆盖断连族;熔断半开只放一个探针;gate 基线臂 INFRA 故障不污染 BaselineCache(算法 #6 保真区,逐行比对);manifest 补原子写;只读查询不改基线元数据。

Tech Stack: Python 3.11、httpx、Redis、asyncio、SQLite、pytest、pydantic-settings。

设计源research-wiki/designs/2026-07-16-preflight-fixes-design.md §8


关键锚点(实现前必读)

用途 位置
chat 缓存/重试 adapters/llm.py:271 chatget L299 / set L373 / 重试 L336);_call_streaming:506_consume_stream:533(不校验 done);_iter_sse_deltas:94done 标志 L116);_is_transient_error:173-186_SseAnomaly:40
缓存键 adapters/redis_cache.py:32 _build_key(无 salt);get:50set:73ttl 分支 L89-92
Protocol core/protocols.py:22-28 LLMProvider.chat
TTL main.py:41 redis_cache_ttl=86400:93 ttl_s = ... if >0 else None
熔断 adapters/breaker.py:24 is_open(L36-37 到期即放行,无探针锁);record_failure:39record_success:63
gate INFRA app/harness/validate.py:257 _resolve_baseline_blockput L303-304);_check_infra_guard:384(累计检查 L593);BaselineCache in gate_ladder.py:344put L384 原子 L401-403
持久化 app/harness/workspace.py 4 处非原子 write_text_scaffold:111/update_manifest:282/record_run:311/update_best:365;原子范式 checkpoint.py:257-260
基线元数据 app/harness/log.py:57 HarnessLog.__init__upsert _runs L73-82);RunLogImpl._read_table:247(只读)
dual_metric app/harness/observation.py:119 write_dual_metricversion_kind final L138);runner 调用 runner.py:1444-1448
配置 .env:65 REDIS_CACHE_TTL=0.env.example:49 =86400

核心算法保真校验

触及算法 #6(块顺序验证:基线缓存/INFRA 护栏/配对翻转)——Task 6gate 基线臂 INFRA 隔离)改 _resolve_baseline_block 与护栏时序,必须逐行比对 TRM4 /home/iomgaa/Projects/Video-Tree-TRM4/core/harness/validate.py 的 INFRA 护栏语义,确保只改"INFRA unit 不写缓存/不计 W/L + 护栏前置",不动配对翻转与基线快照复用逻辑。其余 task 不触碰核心算法。


Task 1: 缓存 cache_salt 贯穿

Files:

  • Modify: core/protocols.py:22-28

  • Modify: adapters/llm.py:271-373

  • Modify: adapters/redis_cache.py:32-92

  • Modify: VLM 适配转发(adapters/ 内实现 LLMProvider/转发 chat 的类,grep 定位)

  • Test: tests/unit/test_redis_cache.pytests/unit/test_governed_llm.py

  • Step 1: 写失败测试(salt 入键)

tests/unit/test_redis_cache.py 追加:

@pytest.mark.asyncio
async def test_salt_changes_key(fake_redis):
    from adapters.redis_cache import RedisResponseCache

    cache = RedisResponseCache(redis=fake_redis, ttl_s=None)
    k_none = cache._build_key("m", [{"role": "user", "content": "x"}], None)
    k_e1 = cache._build_key("m", [{"role": "user", "content": "x"}], "run:e1")
    k_e2 = cache._build_key("m", [{"role": "user", "content": "x"}], "run:e2")
    assert k_none != k_e1 != k_e2 and k_e1 != k_e2
  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_redis_cache.py::test_salt_changes_key -v Expected: FAIL_build_key() takes 3 positional arguments but 4 were given)。

  • Step 3: redis_cache 加 salt 维度

adapters/redis_cache.py

  • _build_key(self, model, messages, cache_salt: str | None = None)payload 字典加 "salt": cache_salt
        payload = json.dumps(
            {"model": model, "messages": messages, "salt": cache_salt},
            sort_keys=True,
            ensure_ascii=False,
        )
  • get(self, model, messages, cache_salt: str | None = None)key = self._build_key(model, messages, cache_salt)

  • set(self, model, messages, response, cache_salt: str | None = None):同样传 cache_salt

  • Step 4: Protocol + chat 透传 salt

core/protocols.py LLMProvider.chat 签名加 cache_salt: str | None = Nonekeyword-only,放 parent_call_id 后)。 adapters/llm.py chat 签名加同参;L299 改 await self._cache.get(self._model, messages, cache_salt)L373 改 await self._cache.set(self._model, messages, response, cache_salt)。 grep adapters/ 找到转发 chat 的 VLM/包装类(如有),同步加 cache_salt 透传。

  • Step 5: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_redis_cache.py tests/unit/test_governed_llm.py -q Expected: 全 PASS(默认 cache_salt=None 保持旧键,既有缓存测试不受影响)。

  • Step 6: 提交
git add core/protocols.py adapters/llm.py adapters/redis_cache.py tests/unit/test_redis_cache.py
git commit -m "feat: add cache_salt dimension to LLM response cache"

注:训练链路注入 epoch 盐(f"{run_id}:e{epoch}")在 WP3 的推理调用点完成(本 WP 只提供 salt 能力)。


Task 2: TTL 语义修正(≤0 报错)

Files:

  • Modify: main.py:93

  • Modify: .env:65.env.example:49

  • Test: tests/unit/test_governed_llm.py 或新增 tests/unit/test_infra_settings.py

  • Step 1: 写失败测试

新增 tests/unit/test_infra_settings.py

import pytest


def test_redis_cache_ttl_zero_rejected(monkeypatch):
    """REDIS_CACHE_TTL<=0 必须启动即报错,消灭'0=永不过期'隐式语义。"""
    from main import _resolve_cache_ttl  # Step 3 抽出的纯函数

    with pytest.raises(ValueError, match="REDIS_CACHE_TTL"):
        _resolve_cache_ttl(0)
    with pytest.raises(ValueError, match="REDIS_CACHE_TTL"):
        _resolve_cache_ttl(-1)
    assert _resolve_cache_ttl(86400) == 86400
  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_infra_settings.py -v Expected: FAILcannot import name '_resolve_cache_ttl')。

  • Step 3: 抽出并实现 fail-loud

main.py:新增纯函数并替换 L93 逻辑:

def _resolve_cache_ttl(ttl: int) -> int:
    """校验 Redis 缓存 TTL:必须为正整数(消灭 0=永不过期 的隐式语义)。"""
    if ttl <= 0:
        raise ValueError(
            f"REDIS_CACHE_TTL 必须为正整数秒,实际 {ttl}。"
            "训练场景建议 >= 单次训练时长(如 86400)。"
        )
    return ttl

L93 改为 ttl_s = _resolve_cache_ttl(settings.redis_cache_ttl)(删除 if >0 else None 分支;RedisResponseCachettl_s 从此恒为正)。

  • Step 4: 改 .env / .env.example

.env:65 REDIS_CACHE_TTL=0REDIS_CACHE_TTL=86400.env.example:49 确认为正整数(当前 86400,OK);补注释 # 正整数秒,禁止 0(0 会被拒绝启动)

  • Step 5: 运行测试确认通过

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_infra_settings.py -v Expected: PASS。

  • Step 6: 提交
git add main.py .env.example tests/unit/test_infra_settings.py
git commit -m "fix: reject REDIS_CACHE_TTL<=0 (kill implicit never-expire)"

注:.env 不提交(gitignore);改动需手动同步到运行环境,runbook 会提示。


Task 3: SSE 截断检测(未收 [DONE] 即重试,不写缓存)

Files:

  • Modify: adapters/llm.py:533-578_consume_stream

  • Test: tests/unit/test_governed_llm.py

  • Step 1: 写失败测试

tests/unit/test_governed_llm.py 追加:

@pytest.mark.asyncio
async def test_truncated_stream_without_done_raises(_build_client):
    """SSE 流耗尽但未收 [DONE] → _SseAnomaly(进重试,不当成功)。"""
    from adapters.llm import _SseAnomaly, GovernedLLMClient

    async def _lines():
        yield 'data: {"choices":[{"delta":{"content":"半"}}]}'
        # 无 data: [DONE] —— 模拟服务端截断

    client = _build_client()  # 依现有 fixture
    with pytest.raises(_SseAnomaly):
        await client._consume_stream(_lines())

test_governed_llm.py 现有 _build_clientL50/_FakeRedisCache 构造;实现前读该文件对齐 client 构造签名。

  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_governed_llm.py -k truncated_stream -v Expected: FAIL(当前耗尽即正常返回,不抛)。

  • Step 3: _consume_stream 校验 done

adapters/llm.py _consume_streamreturn content, thinking, ttft_ms, max_inter_token_ms, usageL578)之前加:

        if not usage_sink.get("done"):
            raise _SseAnomaly("truncated_no_done")

_iter_sse_deltas 收到 [DONE] 时置 usage_sink["done"]=TrueL116;未置说明流被截断。_SseAnomaly 已在 _is_transient_error L186 归为可重试,故自动进重试梯且不走成功路径、不写缓存。)

  • Step 4: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_governed_llm.py tests/unit/test_streaming.py -q Expected: 全 PASS(正常流带 [DONE] 的既有测试仍通过;若既有 streaming 测试的假流未含 [DONE],同步补 data: [DONE] 终帧)。

  • Step 5: 提交
git add adapters/llm.py tests/unit/test_governed_llm.py
git commit -m "fix: treat SSE stream without [DONE] as truncated (retry, no cache)"

Task 4: 瞬时错误清单扩展(断连族)

Files:

  • Modify: adapters/llm.py:182-186

  • Test: tests/unit/test_governed_llm.py

  • Step 1: 写失败测试

tests/unit/test_governed_llm.py 追加:

def test_transient_covers_disconnect_family():
    import httpx
    from adapters.llm import _is_transient_error

    assert _is_transient_error(httpx.RemoteProtocolError("peer reset"))
    assert _is_transient_error(httpx.ReadError("read"))
    assert _is_transient_error(httpx.ConnectTimeout("ct"))
    assert _is_transient_error(httpx.PoolTimeout("pt"))
  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_governed_llm.py -k disconnect_family -v Expected: FAILRemoteProtocolError/PoolTimeout 不在当前清单)。

  • Step 3: 扩展 _is_transient_error

adapters/llm.py L182-183 的:

    if isinstance(exc, (httpx.ConnectError, httpx.ReadTimeout, httpx.WriteTimeout)):
        return True

改为(两族基类覆盖 RemoteProtocolError/ReadError/ConnectTimeout/PoolTimeout 等):

    if isinstance(exc, (httpx.TimeoutException, httpx.TransportError)):
        return True
  • Step 4: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_governed_llm.py -q Expected: 全 PASShttpx.HTTPStatusError 非 TransportError 子类,401/403 致命分支不受影响)。

  • Step 5: 提交
git add adapters/llm.py tests/unit/test_governed_llm.py
git commit -m "fix: cover httpx disconnect family in transient error set"

Task 5: 熔断半开单探针锁

Files:

  • Modify: adapters/breaker.py:18-70

  • Test: tests/unit/test_breaker.py

  • Step 1: 写失败测试

tests/unit/test_breaker.py::TestCircuitBreaker 追加:

    def test_half_open_admits_single_probe(self):
        from adapters.breaker import CircuitBreaker

        b = CircuitBreaker(fail_threshold=2, cooldown_s=10.0)
        b.record_failure("p", now=0.0)
        b.record_failure("p", now=0.0)  # 开路至 t=10
        assert b.is_open("p", now=5.0) is True  # 冷却中
        # 冷却到期:只放行第一个探针
        assert b.is_open("p", now=11.0) is False  # 探针 1 放行
        assert b.is_open("p", now=11.0) is True   # 探针 2 被挡(探针在途)
        b.record_success("p")  # 探针成功 → 闭合
        assert b.is_open("p", now=12.0) is False
  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_breaker.py -k half_open_admits_single -v Expected: FAIL(当前到期后所有调用都返回 False)。

  • Step 3: 实现半开单探针锁

adapters/breaker.py

  • __init__self._half_open_inflight: dict[str, bool] = {}
  • is_open 改为:
    def is_open(self, source_name: str, now: float) -> bool:
        until = self._open_until.get(source_name)
        if until is None:
            return False
        if now < until:
            return True  # 冷却中,全挡
        # 冷却到期:half-open,只放行一个探针
        if self._half_open_inflight.get(source_name):
            return True  # 已有探针在途,继续挡
        self._half_open_inflight[source_name] = True
        return False
  • record_successself._half_open_inflight.pop(source_name, None)(探针成功 → 清在途 + 已有的清 fails/open_until
  • record_failure:探针失败会累计并可能重开路;末尾加 self._half_open_inflight.pop(source_name, None)(让下一轮 cooldown 后可再探)
  • force_open:加 self._half_open_inflight.pop(source_name, None)

is_open 在 asyncio 单线程内同步执行,"检查+标记探针"原子,无竞态。)

  • Step 4: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_breaker.py tests/unit/test_governed_llm.py -q Expected: 全 PASS(既有 circuit_open 测试若假设"到期即多次放行"需同步更新为单探针语义)。

  • Step 5: 提交
git add adapters/breaker.py tests/unit/test_breaker.py
git commit -m "fix: half-open circuit admits single probe (no thundering herd)"

Task 6: gate 基线臂 INFRA 隔离(算法 #6 保真区)

Files:

  • Modify: app/harness/validate.py:257-311_resolve_baseline_block+ 护栏时序
  • Test: tests/unit/test_harness_validate.py

保真前置:实现前完整读 app/harness/validate.py:257-311 _resolve_baseline_block_candidate_correctness_from_db、块循环(575-600),并逐行比对 TRM4 core/harness/validate.py 的 INFRA 护栏语义。目标仅为:① 基线臂 INFRA 故障(stop_reason∈{error,parse_error})的 unit 不写 BaselineCache不计入 W/L 翻转;② 护栏检查移到"写缓存之前"。不得改动配对翻转、基线快照复用、unit 折叠逻辑。

  • Step 1: 写失败测试

tests/unit/test_harness_validate.py 追加(用 fake run_inference 让某基线 unit 返回 stop_reason=error):

@pytest.mark.asyncio
async def test_baseline_infra_error_not_cached(tmp_path):
    """基线臂 INFRA error 的 unit 不写入 BaselineCache(不永久污染)。"""
    # 依现有 test_harness_validate.py 的 fake run_inference / BaselineCache fixture 构造;
    # 让 miss unit u1 的推理返回 prediction=None, stop_reason='error'
    # 调 _resolve_baseline_block 后断言 baseline_cache.get(..., u1) is None
    ...

test_harness_validate.py 现有 fixturefake run_inferenceBaselineCacheHarnessLog);实现前读该文件对齐构造,勿臆造签名。

  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_validate.py -k baseline_infra_error_not_cached -v Expected: FAIL(当前 INFRA unit 的对错被 put 进缓存)。

  • Step 3: 实现 INFRA 隔离

_resolve_baseline_blockmiss 跑完后,识别 INFRA 题(从 db 读 stop_reason∈{error,parse_error} 或经 run_inference 结果),其所属 unit:不 baseline_cache.put、不计入返回的 b_units(从块 unit 集合排除,使配对与 W/L 不含它)。护栏所需的 errors_inc 在 put 之前累计并检查(护栏前置)。具体实现按 Step 前"保真前置"要求对齐现场与 TRM4。

  • Step 4: 运行测试确认通过 + 保真回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_validate.py tests/unit/test_gate_block_unit.py tests/unit/test_gates.py -q Expected: 全 PASS(配对翻转/e-process 既有测试不受影响 = 保真达成)。

  • Step 5: 提交
git add app/harness/validate.py tests/unit/test_harness_validate.py
git commit -m "fix: isolate gate baseline-arm INFRA errors from BaselineCache (algo #6)"

Task 7: manifest 原子写

Files:

  • Modify: app/harness/workspace.py_scaffold:111/update_manifest:282/record_run:311/update_best:365

  • Test: tests/unit/test_harness_workspace.py

  • Step 1: 写失败测试

tests/unit/test_harness_workspace.py 追加:

def test_update_manifest_is_atomic(workspace_dir, monkeypatch):
    """写 manifest 途中崩溃不产生半截 JSON(原子写:tmp 存在即失败也不损原文件)。"""
    import json
    from app.harness import workspace as ws

    # 先建合法 manifest
    ... # 依现有 fixture 初始化 workspace
    original = (workspace_dir / "manifest.json").read_text()

    # monkeypatch os.replace 抛异常,模拟替换阶段崩溃
    def _boom(src, dst):
        raise OSError("crash during replace")
    monkeypatch.setattr(ws.os, "replace", _boom)
    with pytest.raises(OSError):
        ws.update_manifest(workspace_dir, skills="skills/v2")
    # 原 manifest 未被破坏
    assert (workspace_dir / "manifest.json").read_text() == original
    assert json.loads((workspace_dir / "manifest.json").read_text())

test_harness_workspace.py 现有 workspace 初始化 fixture;实现前读对齐。

  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_workspace.py -k update_manifest_is_atomic -v Expected: FAIL(当前 write_text 非原子,崩溃留半截)。

  • Step 3: 加原子写 helper 并替换 4 处

app/harness/workspace.py 顶部确认 import os。新增模块级 helper

def _atomic_write_json(path: Path, data: dict) -> None:
    """原子写 JSONtmp + os.replace(对齐 checkpoint.py 范式,防半截损坏)。"""
    tmp = path.with_name(path.name + ".tmp")
    tmp.write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")
    os.replace(tmp, path)

_scaffold_workspaceL111)、update_manifestL282)、record_runL311)、update_bestL365)四处的 (...).write_text(json.dumps(...)) 替换为 _atomic_write_json(<path>, <data>)

  • Step 4: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_workspace.py -q Expected: 全 PASS。

  • Step 5: 提交
git add app/harness/workspace.py tests/unit/test_harness_workspace.py
git commit -m "fix: atomic writes for manifest/record_run/update_best (tmp+replace)"

Task 8: 只读查询不改基线元数据

Files:

  • Modify: app/harness/log.py:57-83HarnessLog.init 加 register_run 开关)

  • Modify: 只读查询基线的调用点(app/harness/runner.py:923 _init_gateapp/harness/pools.py:765,818

  • Test: tests/unit/test_harness_log.py

  • Step 1: 写失败测试

tests/unit/test_harness_log.py::TestHarnessLogUpsert 追加:

    def test_register_run_false_skips_upsert(self, tmp_path):
        """register_run=False 时只读打开不改写已有 _runs 行(started_at/status 不变)。"""
        db = str(tmp_path / "h.db")
        from app.harness.log import HarnessLog

        with HarnessLog(db, "r1") as log:
            log  # 初次注册
        row0 = _read_run_row(db, "r1")  # 依现有 helper 读 started_at/status

        with HarnessLog(db, "r1", register_run=False) as log:
            log.query("SELECT 1")  # 只读
        row1 = _read_run_row(db, "r1")
        assert row1["started_at"] == row0["started_at"]
        assert row1["status"] == row0["status"]

test_harness_log.py 现有读行 helper;实现前读 TestHarnessLogUpsertL196)对齐。

  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_log.py -k register_run_false -v Expected: FAILunexpected keyword argument 'register_run')。

  • Step 3: HarnessLog 加 register_run 开关

app/harness/log.py __init__ 签名加 register_run: bool = Truekeyword)。把 L73-82 的 _runs upsert 包进 if register_run:register_run=False 时跳过 upsert(仅 _init_fixed_tables 建表 + 连接,供只读查询)。__exit__/close 的 status 更新同样在 register_run 为 True 时才执行(避免只读关闭把 status 改 completed)。

  • Step 4: 只读调用点传 register_run=False

app/harness/runner.py:923_init_gate 用 baseline_run_id 只读查 predictions)、app/harness/pools.py:765,818build_or_load_pools 只读查 predictions)三处 HarnessLog(...) 调用加 register_run=False

  • Step 5: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_log.py tests/unit/test_harness_pools.py -q Expected: 全 PASS。

  • Step 6: 提交
git add app/harness/log.py app/harness/runner.py app/harness/pools.py tests/unit/test_harness_log.py
git commit -m "fix: read-only baseline queries skip _runs upsert (register_run flag)"

Task 9: dual_metric version_kind 口径修正

Files:

  • Modify: app/harness/runner.py:1444-1448(第二次 write_dual_metric

  • Test: tests/unit/test_harness_observation.pytest_harness_runner.py

  • Step 1: 写失败测试

tests/unit/test_harness_observation.py 追加(验证 slow_candidate 与 final 可区分):

def test_slow_candidate_kind_distinct_from_final(tmp_path):
    from app.harness.observation import write_dual_metric, read_dual_metric

    db = str(tmp_path / "h.db")
    write_dual_metric(db, run_id="r", epoch=1, version_kind="final",
                      skills_version="v1", prompts_version="v1", pool="val",
                      hard_acc=0.7, soft_score=None, mixed_score=None)
    write_dual_metric(db, run_id="r", epoch=1, version_kind="slow_candidate",
                      skills_version="v2", prompts_version="v2", pool="val",
                      hard_acc=0.6, soft_score=None, mixed_score=None)
    rows = read_dual_metric(db, run_id="r", epoch=1)
    kinds = {r["version_kind"] for r in rows}
    assert "final" in kinds and "slow_candidate" in kinds

test_harness_observation.py 现有 read_dual_metric 签名;实现前读对齐。

  • Step 2: 运行确认失败/通过基线

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_observation.py -k slow_candidate -v Expected: 若 read_dual_metric 已支持则 PASS 基线;重点是 Step 3 改 runner 调用点。

  • Step 3: 改 runner 慢更新第二次写为 slow_candidate

app/harness/runner.py L1444-1448 的 write_dual_metric(..., version_kind="final", ...)_slow_update_cycle Phase 8 的 R2 行)改为 version_kind="slow_candidate",使被 revert 的慢更新候选不再占用 final 语义。确认 Phase 2 的第一次(L1399-1403)仍为 finalepoch 终值唯一)。

  • Step 4: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_harness_observation.py tests/unit/test_harness_runner.py -q Expected: 全 PASS。

  • Step 5: 提交
git add app/harness/runner.py tests/unit/test_harness_observation.py
git commit -m "fix: slow-update R2 dual_metric uses slow_candidate kind (not final)"

Task 10: 离线 --retry-uncertain + docstring 修正

Files:

  • Modify: app/harness/video_split_cli.pybuild_arg_parser + run_pipeline

  • Modify: app/harness/baseline_diagnosis.pydone 集排除 uncertain + docstring

  • Test: tests/unit/test_video_split_cli.pytest_baseline_diagnosis

  • Step 1: 写失败测试

定位 baseline_diagnosis 的 done_question_ids 计算函数(app/harness/baseline_diagnosis.pygrep done_question_ids),追加测试:retry_uncertain=True 时 tier=uncertain 的已落库题不计入 done(会被重诊)。

def test_retry_uncertain_excludes_uncertain_from_done(...):
    # 构造 baseline_diagnosis 表含 tier=T2 与 tier=uncertain 行;
    # done_ids(retry_uncertain=False) 含 uncertain 题;
    # done_ids(retry_uncertain=True) 不含 uncertain 题
    ...

实现前读 baseline_diagnosis.py 的 done 集查询函数签名对齐。

  • Step 2: 运行确认失败

Run: conda run -n Video-Tree-TRM python -m pytest -k retry_uncertain -v Expected: FAIL(无该参数)。

  • Step 3: 实现 retry_uncertain

app/harness/baseline_diagnosis.pydone 集查询函数加 retry_uncertain: bool = FalseTrue 时 WHERE ... AND tier != 'uncertain'(uncertain 题不算完成,会被重诊)。修正模块 docstring 把"崩溃最多丢正在写的一行/逐行落库"表述改为实际的"run 末批量落库(Phase 3),崩溃丢本次 run 全部未落库结果,靠 Redis 缓存缓解重烧"。 app/harness/video_split_cli.pybuild_arg_parser--retry-uncertainstore_true);run_pipeline 透传到 run_baseline_diagnosis 的 done 集计算。

  • Step 4: 运行测试确认通过 + 回归

Run: conda run -n Video-Tree-TRM python -m pytest tests/unit/test_video_split_cli.py -q Expected: 全 PASS。

  • Step 5: 提交
git add app/harness/video_split_cli.py app/harness/baseline_diagnosis.py tests/unit/
git commit -m "feat: --retry-uncertain re-diagnoses uncertain rows; fix docstring"

Self-Review(作者自查,执行者复核)

  • cache_salt 四处(protocols/llm/redis_cache/VLM 转发)贯通,默认 None 保持旧键。
  • TTL≤0 fail-loud + .env/.env.example 同步(.env 手动,runbook 提示)。
  • SSE 未收 [DONE] → _SseAnomaly(已在瞬时清单 → 重试 + 不写缓存)。
  • 瞬时清单用 TimeoutException+TransportError 两族基类覆盖断连族,不误纳 HTTPStatusError。
  • 熔断半开单探针:is_open 检查+标记原子(asyncio 单线程)。
  • Task 6 触及算法 #6,已设保真前置(逐行比对 TRM4,只改 INFRA 不写缓存 + 护栏前置)。
  • manifest 4 处原子写;只读查询 register_run=False 不改基线元数据。

核心算法保真校验结论

Task 6 触及算法 #6(块顺序验证),已在该 Task 设"保真前置"要求逐行比对 TRM4 validate.py,仅改 INFRA 隔离与护栏时序,不动配对翻转/基线快照/unit 折叠。其余 Task 均为治理栈/持久化层,不涉及核心算法。

验收标准

  1. pytest tests/unit/test_redis_cache.py tests/unit/test_governed_llm.py tests/unit/test_breaker.py tests/unit/test_streaming.py tests/unit/test_harness_validate.py tests/unit/test_harness_workspace.py tests/unit/test_harness_log.py tests/unit/test_harness_observation.py tests/unit/test_infra_settings.py 全绿。
  2. 缓存加 salt 后跨 epoch 键不同、同 epoch 键相同。
  3. 熔断半开只放一个探针;SSE 截断进重试不写缓存。
  4. manifest 原子写;基线 _runs 元数据只读查询不被改写。