fix: 修复补跑历史日期的静默空转与日报日期错位 (P0-3/P0-4)

- run_step: 补跑保护——历史日期跳过 crawler(首页只含当天内容)并告警,复用已有 raw 数据
- run_step: report 步骤改用传入的 date_str,不再硬编码 date.today()
- 新增 4 个补跑相关测试,修正 3 个既有测试适配新行为
- 全量 259 passed(3 个 crawler 基线失败与本次无关)

tests/test_incremental.py

@@ -266,3 +266,83 @@ def test_once_rejects_resume_with_steps(monkeypatch: pytest.MonkeyPatch) -> None
                            date="20260616")
     rc = rs._once(args)
     assert rc == 2
+
+
+# --------------------------------------------------------------------------- #
+# P0-3: 补跑模式跳过抓取 / P0-4: report 使用传入日期
+# --------------------------------------------------------------------------- #
+
+def test_crawler_backfill_skips_past_date(monkeypatch: pytest.MonkeyPatch) -> None:
+    """补跑历史日期时跳过 crawler(首页只含当天内容),不执行子进程。"""
+    from scheduler import pipeline
+
+    def _fail_run(cmd, timeout=None):  # noqa: ARG001
+        raise AssertionError("补跑模式不应执行子进程")
+
+    monkeypatch.setattr(pipeline.subprocess, "run", _fail_run)
+    sr = pipeline.run_step("crawler", "20260101")
+    assert sr.success is True
+    assert "跳过" in sr.tail_msg
+    assert sr.elapsed_sec == 0.0
+
+
+def test_crawler_today_runs_normally(monkeypatch: pytest.MonkeyPatch) -> None:
+    """当天日期正常执行 crawler(不受补跑保护影响)。"""
+    from datetime import date
+
+    from scheduler import pipeline
+
+    called: list[str] = []
+
+    def _fake_run(cmd, timeout=None):  # noqa: ARG001
+        called.append(" ".join(cmd))
+        return SimpleNamespace(returncode=0)
+
+    monkeypatch.setattr(pipeline.subprocess, "run", _fake_run)
+    sr = pipeline.run_step("crawler", date.today().strftime("%Y%m%d"))
+    assert sr.success is True
+    assert called and "run_crawler" in called[0]
+
+
+def test_report_step_uses_date_str(monkeypatch: pytest.MonkeyPatch) -> None:
+    """report 步骤必须使用传入的 date_str,而非 date.today()(P0-4)。"""
+    from scheduler import pipeline
+    from scheduler import reporter
+
+    received: list[str] = []
+
+    def _fake_generate(day_str, upload=True):  # noqa: ARG001
+        received.append(day_str)
+        return 12345
+
+    monkeypatch.setattr(reporter, "generate_report", _fake_generate)
+    sr = pipeline.run_step("report", "20260101")
+    assert sr.success is True
+    assert received == ["20260101"]
+
+
+def test_pipeline_backfill_skips_crawler_keeps_rest(
+    tmp_path: Path, monkeypatch: pytest.MonkeyPatch,
+) -> None:
+    """补跑历史日期:全链路中仅 crawler 跳过,其余步骤照常执行。"""
+    from scheduler import pipeline
+
+    calls: list[str] = []
+
+    def _fake_run(cmd, timeout=None):  # noqa: ARG001
+        name = next(c.split(".")[-1] for c in cmd if "scripts.run_" in c)
+        calls.append(name)
+        return SimpleNamespace(returncode=0)
+
+    monkeypatch.setattr(pipeline.subprocess, "run", _fake_run)
+    state_path = tmp_path / "state.json"
+    result = pipeline.run_pipeline(
+        "20260101",
+        steps=["crawler", "extractor", "dedup"],
+        state_path=state_path,
+    )
+    # crawler 跳过(未执行子进程),extractor/dedup 正常
+    assert calls == ["run_extractor", "run_dedup"]
+    assert all(s.success for s in result.steps)
+    state = pipeline._load_pipeline_state(state_path)
+    assert state["20260101"]["crawler"]["status"] == "ok"

tests/test_scheduler.py

@@ -56,9 +56,12 @@ def test_pipeline_result_partial_failure() -> None:
 # --------------------------------------------------------------------------- #
 
 def test_run_step_success() -> None:
+    # crawler 补跑保护:仅当天日期才执行子进程,故用今天的日期
+    from datetime import date
+    today = date.today().strftime("%Y%m%d")
     with patch("subprocess.run", return_value=_mock_proc(returncode=0,
                                                           stderr="INFO | 完成: 成功 20/20")):
-        sr = run_step("crawler", "20260616")
+        sr = run_step("crawler", today)
     assert sr.success is True
     assert sr.exit_code == 0
 
@@ -109,6 +112,8 @@ def test_run_pipeline_all_success() -> None:
 
 def test_run_pipeline_continues_on_failure() -> None:
     """中间步骤失败,后续继续执行(不阻断)。"""
+    from datetime import date
+    today = date.today().strftime("%Y%m%d")  # crawler 补跑保护:需当天日期才执行
     call_count = {"n": 0}
 
     def _side_effect(*args, **kwargs):
@@ -118,7 +123,7 @@ def test_run_pipeline_continues_on_failure() -> None:
         return _mock_proc(returncode=0, stderr="ok")
 
     with patch("subprocess.run", side_effect=_side_effect):
-        result = run_pipeline("20260616", steps=["crawler", "extractor", "dedup", "llm"])
+        result = run_pipeline(today, steps=["crawler", "extractor", "dedup", "llm"])
     assert len(result.steps) == 4
     # extractor 失败,但后续仍执行
     assert result.steps[1].success is False
@@ -126,7 +131,9 @@ def test_run_pipeline_continues_on_failure() -> None:
 
 
 def test_run_pipeline_custom_steps() -> None:
+    from datetime import date
+    today = date.today().strftime("%Y%m%d")
     with patch("subprocess.run", return_value=_mock_proc(returncode=0, stderr="ok")):
-        result = run_pipeline("20260616", steps=["crawler", "extractor"])
+        result = run_pipeline(today, steps=["crawler", "extractor"])
     assert len(result.steps) == 2
     assert result.all_success is True