Repair official collection and retire exhausted material tasks

backend/tests/test_collection_repair.py

@@ -0,0 +1,30 @@
+"""Repair initialization never renews an already initialized retry budget."""
+import asyncio
+import copy
+from types import SimpleNamespace
+from unittest.mock import AsyncMock
+
+
+def test_preview_and_repeated_execution_preserve_retry_evidence(trading, monkeypatch):
+    from alphapulse.crawler import repair_collection as module
+    original = {'document_retries': {'https://example.invalid/doc': {
+        'document': {'url': 'https://example.invalid/doc'}, 'attempts': 3,
+        'next_at': 1, 'error': 'HTTP 503'}}}
+    source = SimpleNamespace(code='doj', history_cursor=copy.deepcopy(original))
+    session = AsyncMock()
+    session.scalars.return_value = [source]
+    session.__aenter__.return_value = session
+    monkeypatch.setattr(module, 'SessionLocal', lambda: session)
+    monkeypatch.setattr(module, 'engine', SimpleNamespace(dispose=AsyncMock()))
+    async def check():
+        await module.run()
+        assert source.history_cursor == original
+        session.commit.assert_not_awaited()
+        await module.run(True)
+        entry = source.history_cursor['document_retries']['https://example.invalid/doc']
+        assert entry['attempts'] == 0 and entry['repair_previous']['attempts'] == 3
+        entry['attempts'] = 2
+        result = await module.run(True)
+        assert result['sources'][0]['already_initialized']
+        assert entry['attempts'] == 2
+    asyncio.run(check())