Initial commit

tests/test_scheduler.py

@@ -0,0 +1,132 @@
+"""M7 定时任务测试。
+
+使用 mock subprocess.run,不依赖真实脚本执行。
+"""
+
+from __future__ import annotations
+
+import subprocess
+from pathlib import Path
+from unittest.mock import MagicMock, patch
+
+from scheduler import PipelineResult, StepResult, run_pipeline, run_step
+
+# --------------------------------------------------------------------------- #
+# fixtures
+# --------------------------------------------------------------------------- #
+
+def _mock_proc(returncode: int = 0, stderr: str = "") -> MagicMock:
+    p = MagicMock()
+    p.returncode = returncode
+    p.stderr = stderr
+    p.stdout = "ok"
+    return p
+
+
+# --------------------------------------------------------------------------- #
+# StepResult 模型
+# --------------------------------------------------------------------------- #
+
+def test_step_result_defaults() -> None:
+    sr = StepResult(name="crawler", success=True, elapsed_sec=12.3)
+    assert sr.name == "crawler"
+    assert sr.success is True
+    assert sr.elapsed_sec == 12.3
+    assert sr.exit_code is None
+
+
+def test_pipeline_result_all_success() -> None:
+    pr = PipelineResult(steps=[
+        StepResult(name="crawler", success=True, elapsed_sec=1),
+        StepResult(name="extractor", success=True, elapsed_sec=2),
+    ])
+    assert pr.all_success is True
+
+
+def test_pipeline_result_partial_failure() -> None:
+    pr = PipelineResult(steps=[
+        StepResult(name="crawler", success=True, elapsed_sec=1),
+        StepResult(name="extractor", success=False, elapsed_sec=0),
+    ])
+    assert pr.all_success is False
+
+
+# --------------------------------------------------------------------------- #
+# run_step
+# --------------------------------------------------------------------------- #
+
+def test_run_step_success() -> None:
+    with patch("subprocess.run", return_value=_mock_proc(returncode=0,
+                                                          stderr="INFO | 完成: 成功 20/20")):
+        sr = run_step("crawler", "20260616")
+    assert sr.success is True
+    assert sr.exit_code == 0
+
+
+def test_run_step_failure() -> None:
+    with patch("subprocess.run", return_value=_mock_proc(returncode=1)):
+        sr = run_step("extractor", "20260616")
+    assert sr.success is False
+    assert sr.exit_code == 1
+    assert "rc=1" in sr.tail_msg
+
+
+def test_run_step_timeout() -> None:
+    with patch("subprocess.run", side_effect=subprocess.TimeoutExpired(cmd=["uv"], timeout=10)):
+        sr = run_step("llm", "20260616")
+    assert sr.success is False
+    assert "超时" in sr.tail_msg
+
+
+def test_run_step_exception() -> None:
+    with patch("subprocess.run", side_effect=OSError("磁盘满")):
+        sr = run_step("embedding", "20260616")
+    assert sr.success is False
+    assert "磁盘满" in sr.tail_msg
+
+
+def test_run_step_unknown_name() -> None:
+    sr = run_step("nonexistent", "20260616")
+    assert sr.success is False
+    assert "未知" in sr.tail_msg
+
+
+# --------------------------------------------------------------------------- #
+# run_pipeline
+# --------------------------------------------------------------------------- #
+
+def test_run_pipeline_all_success() -> None:
+    with (
+        patch("subprocess.run", return_value=_mock_proc(returncode=0)),
+        patch("scheduler.reporter.generate_report", return_value=Path("/tmp/r.html")),
+    ):
+        result = run_pipeline("20260616", steps=["crawler", "extractor", "dedup"])
+    assert len(result.steps) == 3
+    assert result.all_success is True
+    assert result.started_at is not None
+    assert result.finished_at is not None
+
+
+def test_run_pipeline_continues_on_failure() -> None:
+    """中间步骤失败,后续继续执行(不阻断)。"""
+    call_count = {"n": 0}
+
+    def _side_effect(*args, **kwargs):
+        call_count["n"] += 1
+        if call_count["n"] == 2:  # extractor 失败
+            return _mock_proc(returncode=1, stderr="GNE 提取异常")
+        return _mock_proc(returncode=0, stderr="ok")
+
+    with patch("subprocess.run", side_effect=_side_effect):
+        result = run_pipeline("20260616", 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
+    assert result.steps[2].success is True
+
+
+def test_run_pipeline_custom_steps() -> None:
+    with patch("subprocess.run", return_value=_mock_proc(returncode=0, stderr="ok")):
+        result = run_pipeline("20260616", steps=["crawler", "extractor"])
+    assert len(result.steps) == 2
+    assert result.all_success is True