fix: prevent future article timestamps from pinning feeds

docker/crawler/tests/test_article_publish_time.py

@@ -0,0 +1,48 @@
+import importlib.util
+import pathlib
+import unittest
+from datetime import datetime
+
+
+ROOT = pathlib.Path(__file__).resolve().parents[1]
+MODULE_PATH = ROOT / "src" / "parser" / "publish_time.py"
+
+
+def load_module():
+    if not MODULE_PATH.exists():
+        raise AssertionError("缺少文章发布时间规范化模块")
+    spec = importlib.util.spec_from_file_location("publish_time", MODULE_PATH)
+    module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    return module
+
+
+class ArticlePublishTimeTest(unittest.TestCase):
+    def setUp(self):
+        self.module = load_module()
+        self.now = datetime(2026, 8, 3, 15, 0, 0)
+
+    def test_keeps_normal_publish_time_within_timezone_tolerance(self):
+        value = self.module.normalize_published_at(
+            "2026-08-03 23:01:04",
+            now=self.now,
+        )
+
+        self.assertEqual(value, "2026-08-03 23:01:04")
+
+    def test_rejects_clearly_future_publish_time(self):
+        value = self.module.normalize_published_at(
+            "2031-07-21 04:50:09",
+            now=self.now,
+        )
+
+        self.assertEqual(value, "")
+
+    def test_rejects_unparseable_publish_time(self):
+        value = self.module.normalize_published_at("not-a-date", now=self.now)
+
+        self.assertEqual(value, "")
+
+
+if __name__ == "__main__":
+    unittest.main()