adding delete_by_filter and update_by_filter

Author: davidsbatista

integrations/supabase/src/haystack_integrations/document_stores/supabase/groonga_document_store.py

@@ -234,6 +234,48 @@ def write_documents(
 
         return written
 
+    def delete_by_filter(self, filters: dict[str, Any]) -> int:
+        """
+        Deletes documents matching the given filters.
+
+        :param filters: Filters to select documents for deletion.
+        :returns: Number of documents deleted.
+        """
+        docs = self.filter_documents(filters=filters)
+        if not docs:
+            return 0
+        self.delete_documents([doc.id for doc in docs])
+        return len(docs)
+
+    def update_by_filter(self, filters: dict[str, Any], meta: dict[str, Any]) -> int:
+        """
+        Updates the metadata of documents matching the given filters.
+
+        Provided meta fields are merged into the existing document metadata.
+
+        :param filters: Filters to select documents to update.
+        :param meta: Metadata fields to set on matching documents.
+        :returns: Number of documents updated.
+        """
+        if self._client is None:
+            msg = "Call warm_up() before using the document store."
+            raise RuntimeError(msg)
+
+        docs = self.filter_documents(filters=filters)
+        if not docs:
+            return 0
+
+        for doc in docs:
+            row = {
+                "id": doc.id,
+                "content": doc.content or "",
+                "meta": {**doc.meta, **meta},
+                "score": None,
+            }
+            self._client.table(self.table_name).upsert(row).execute()
+
+        return len(docs)
+
     def delete_all_documents(self) -> None:
         """
         Deletes all documents from the store.

integrations/supabase/tests/test_groonga_document_store.py

@@ -140,6 +140,51 @@ def test_write_documents_fail_on_duplicate(self, groonga_store, mock_supabase_cl
         with pytest.raises(DuplicateDocumentError):
             groonga_store.write_documents([Document(content="duplicate doc")], policy=DuplicatePolicy.FAIL)
 
+    def test_delete_by_filter(self, groonga_store, mock_supabase_client):
+        mock_table = mock_supabase_client.table.return_value
+        mock_table.select.return_value.execute.return_value = MagicMock(
+            data=[{"id": "1", "content": "doc one", "meta": {"lang": "en"}, "score": None}]
+        )
+        mock_table.delete.return_value.in_.return_value.execute.return_value = MagicMock(data=[])
+
+        deleted = groonga_store.delete_by_filter(
+            filters={"conditions": [{"field": "meta.lang", "operator": "==", "value": "en"}]}
+        )
+        assert deleted == 1
+
+    def test_delete_by_filter_no_matches(self, groonga_store, mock_supabase_client):
+        mock_supabase_client.table.return_value.select.return_value.execute.return_value = MagicMock(data=[])
+
+        deleted = groonga_store.delete_by_filter(
+            filters={"conditions": [{"field": "meta.lang", "operator": "==", "value": "fr"}]}
+        )
+        assert deleted == 0
+
+    def test_update_by_filter(self, groonga_store, mock_supabase_client):
+        mock_table = mock_supabase_client.table.return_value
+        mock_table.select.return_value.execute.return_value = MagicMock(
+            data=[{"id": "1", "content": "doc one", "meta": {"lang": "en"}, "score": None}]
+        )
+        mock_table.upsert.return_value.execute.return_value = MagicMock(data=[{}])
+
+        updated = groonga_store.update_by_filter(
+            filters={"conditions": [{"field": "meta.lang", "operator": "==", "value": "en"}]},
+            meta={"reviewed": True},
+        )
+        assert updated == 1
+        mock_table.upsert.assert_called_once()
+        upserted_row = mock_table.upsert.call_args[0][0]
+        assert upserted_row["meta"] == {"lang": "en", "reviewed": True}
+
+    def test_update_by_filter_no_matches(self, groonga_store, mock_supabase_client):
+        mock_supabase_client.table.return_value.select.return_value.execute.return_value = MagicMock(data=[])
+
+        updated = groonga_store.update_by_filter(
+            filters={"conditions": [{"field": "meta.lang", "operator": "==", "value": "fr"}]},
+            meta={"reviewed": True},
+        )
+        assert updated == 0
+
     def test_delete_all_documents(self, groonga_store, mock_supabase_client):
         mock_table = mock_supabase_client.table.return_value
         mock_table.delete.return_value.neq.return_value.execute.return_value = MagicMock(data=[])

integrations/supabase/tests/test_groonga_integration.py

@@ -10,7 +10,10 @@
 from haystack.document_stores.types import DuplicatePolicy
 from haystack.testing.document_store import (
     DeleteAllTest,
+    DeleteByFilterTest,
     DocumentStoreBaseTests,
+    FilterableDocsFixtureMixin,
+    UpdateByFilterTest,
 )
 from haystack.utils import Secret
 
@@ -43,7 +46,13 @@ def document_store(request):
 
 
 @pytest.mark.integration
-class TestSupabaseGroongaDocumentStoreIntegration(DocumentStoreBaseTests, DeleteAllTest):
+class TestSupabaseGroongaDocumentStoreIntegration(
+    DocumentStoreBaseTests,
+    DeleteAllTest,
+    DeleteByFilterTest,
+    FilterableDocsFixtureMixin,
+    UpdateByFilterTest,
+):
     pass