feat: check source links during scheduled fetches

apps/api/tests/test_community_importer.py

@@ -9,7 +9,7 @@ from sqlalchemy.orm import Session
 
 from app.community_importer import CommunityImportError, stage_observations
 from app.community_review import ExternalReviewError, map_observation, publish_observation, suggest_aliases
-from app.source_lifecycle import record_source_check
+from app.source_lifecycle import record_scheduled_source_check, record_source_check
 from app.database import Base
 from app.models import CatchReport, DataSource, ExternalEntityAlias, ExternalObservation, Fish, Waterbody
 from rf4_research.community_sources import parse_rf4db_catches, parse_rf4map_point, parse_rf4posts_spot
@@ -186,6 +186,35 @@ def test_non_authoritative_source_failures_do_not_withdraw(db: Session, status:
     assert item.source_check_status == status
 
 
+def test_scheduled_failure_only_affects_exact_source_url(db: Session) -> None:
+    stage_observations(db, [
+        record(external_id="matching"),
+        record(external_id="other") | {"source_url": "https://rf4db.com/catches/other"},
+    ])
+
+    affected = record_scheduled_source_check(
+        db,
+        source_system="rf4db",
+        source_url="https://rf4db.com/ru/catches/matching",
+        status="missing",
+    )
+    items = {item.source_external_id: item for item in db.scalars(select(ExternalObservation))}
+
+    assert affected == 1
+    assert items["matching"].status == "withdrawn"
+    assert items["other"].status != "withdrawn"
+    assert items["other"].source_check_status == "available"
+
+    record_scheduled_source_check(
+        db,
+        source_system="rf4db",
+        source_url="https://rf4db.com/ru/catches/matching",
+        status="available",
+    )
+    assert items["matching"].source_check_status == "available"
+    assert items["matching"].status == "withdrawn"
+
+
 def test_auto_publication_requires_enabled_source(db: Session) -> None:
     source = DataSource(key="rf4db", name="RF4DB", base_url="https://rf4db.com", default_confidence=70, enabled=False)
     fish = Fish(slug="pike", name_ru="Щука")

apps/api/tests/test_community_scheduler.py

@@ -1,10 +1,12 @@
 import pytest
 from pydantic import ValidationError
+from urllib.error import HTTPError
 
 from datetime import datetime, timedelta, timezone
 
 from app.community_scheduler import MAX_BACKOFF_SECONDS, configured_sources, _static_registry, oldest_site_source, retry_delay
 from app.config import Settings
+from app.source_lifecycle import classify_source_failure
 
 
 def test_all_authorized_sources_are_scheduled() -> None:
@@ -23,6 +25,16 @@ def test_failed_runs_back_off_but_success_resets_delay() -> None:
     assert retry_delay(["success", "failed"]) == 1800
 
 
+@pytest.mark.parametrize(("code", "expected"), [(404, "missing"), (410, "missing"), (403, "blocked"), (429, "blocked"), (500, "temporary_error")])
+def test_source_http_failure_classification(code: int, expected: str) -> None:
+    error = HTTPError("https://rf4.example/source", code, "failure", {}, None)
+    assert classify_source_failure(error) == expected
+
+
+def test_non_http_source_failure_is_temporary() -> None:
+    assert classify_source_failure(TimeoutError("timeout")) == "temporary_error"
+
+
 def test_same_site_endpoints_rotate_by_oldest_attempt() -> None:
     now = datetime.now(timezone.utc)
     all_keys = {"rf4db", "rf4stat-fishing", "rf4stat-post", "rf4map", "rf4posts-spot"}