fix(sources): query reference edge by in/out in retry endpoint (#899)

POST /sources/{id}/retry looked up a source's notebooks with
`SELECT notebook FROM reference WHERE source = $source_id`, but `reference`
is a graph edge (RELATE source->reference->notebook) with only `in`/`out`
columns. The query matched nothing, so `notebook_ids` was always empty and
the endpoint returned 400 for every source. Mirror the working query used in
the source-list path: `SELECT VALUE out FROM reference WHERE in = $source_id`.

Adds regression tests asserting retry re-queues a linked source and only 400s
when a source is genuinely unlinked.

CHANGELOG.md

@@ -7,6 +7,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Fixed
+- `POST /sources/{id}/retry` no longer returns `400 "Source is not associated with any notebooks"` for every source; it now queries the `reference` graph edge by its `in`/`out` columns instead of a non-existent `source` column (#861)
+
 ## [1.9.0] - 2026-06-02
 
 ### Added

api/routers/sources.py

@@ -842,10 +842,15 @@ async def retry_source_processing(source_id: str):
                 )
                 # Continue with retry if we can't check status
 
-        # Get notebooks that this source belongs to
-        query = "SELECT notebook FROM reference WHERE source = $source_id"
-        references = await repo_query(query, {"source_id": source_id})
-        notebook_ids = [str(ref["notebook"]) for ref in references]
+        # Get notebooks that this source belongs to. `reference` is a graph edge
+        # (RELATE source->reference->notebook), so it only has `in`/`out` columns —
+        # there is no `source`/`notebook` column. Mirror the working query at the
+        # source-list path above. See issue #861.
+        references = await repo_query(
+            "SELECT VALUE out FROM reference WHERE in = $source_id",
+            {"source_id": ensure_record_id(source.id or source_id)},
+        )
+        notebook_ids = [str(nb_id) for nb_id in references] if references else []
 
         if not notebook_ids:
             raise HTTPException(

tests/test_sources_api.py

@@ -138,5 +138,57 @@ class TestAsyncSourceAssetPersistence:
         assert source.asset is None
 
 
+class TestRetrySourceProcessing:
+    """POST /sources/{id}/retry must find a source's notebooks via the reference
+    edge's in/out columns, not a non-existent `source` column (#861)."""
+
+    @pytest.mark.asyncio
+    @patch("api.routers.sources.CommandService.submit_command_job", new_callable=AsyncMock)
+    @patch("api.routers.sources.repo_query", new_callable=AsyncMock)
+    @patch("api.routers.sources.Source.get", new_callable=AsyncMock)
+    async def test_retry_finds_notebooks_and_requeues(
+        self, mock_get, mock_repo_query, mock_submit, client
+    ):
+        source = MagicMock()
+        source.id = "source:1"
+        source.command = None
+        source.title = "My source"
+        source.topics = []
+        source.full_text = None
+        source.asset = MagicMock(file_path=None, url="https://example.com/post")
+        source.save = AsyncMock()
+        source.get_embedded_chunks = AsyncMock(return_value=0)
+        mock_get.return_value = source
+
+        # The corrected query returns the linked notebook(s)
+        mock_repo_query.return_value = ["notebook:1"]
+        mock_submit.return_value = "123"
+
+        response = client.post("/api/sources/source:1/retry")
+
+        assert response.status_code == 200
+        # Regression guard: must query the reference edge by its `in` column
+        called_query = mock_repo_query.await_args.args[0]
+        assert "WHERE in = $source_id" in called_query
+        assert "SELECT VALUE out FROM reference" in called_query
+
+    @pytest.mark.asyncio
+    @patch("api.routers.sources.repo_query", new_callable=AsyncMock)
+    @patch("api.routers.sources.Source.get", new_callable=AsyncMock)
+    async def test_retry_400_only_when_truly_unlinked(
+        self, mock_get, mock_repo_query, client
+    ):
+        source = MagicMock()
+        source.id = "source:1"
+        source.command = None
+        mock_get.return_value = source
+        mock_repo_query.return_value = []  # genuinely no notebooks
+
+        response = client.post("/api/sources/source:1/retry")
+
+        assert response.status_code == 400
+        assert "not associated with any notebooks" in response.json()["detail"]
+
+
 if __name__ == "__main__":
     pytest.main([__file__, "-v"])