fix: knowledge hit-testing render failed. (#36106)

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

api/tests/unit_tests/controllers/console/datasets/test_hit_testing_base.py

@@ -120,7 +120,7 @@ class TestParseArgs:
 class TestPerformHitTesting:
     def test_success(self, dataset):
         response = {
-            "query": "hello",
+            "query": {"content": "hello"},
             "records": [],
         }
 
@@ -134,7 +134,7 @@ class TestPerformHitTesting:
         assert result["query"] == "hello"
         assert result["records"] == []
 
-    def test_success_normalizes_legacy_query_and_nullable_list_fields(self, dataset):
+    def test_success_prepares_nullable_list_fields(self, dataset):
         response = {
             "query": {"content": "hello"},
             "records": [
@@ -170,6 +170,18 @@ class TestPerformHitTesting:
             }
         ]
 
+    def test_invalid_query_response_raises_value_error(self):
+        with pytest.raises(ValueError, match="Invalid hit testing query response"):
+            DatasetsHitTestingBase._extract_hit_testing_query("hello")
+
+    def test_invalid_records_response_raises_value_error(self):
+        with pytest.raises(ValueError, match="Invalid hit testing records response"):
+            DatasetsHitTestingBase._prepare_hit_testing_records({"records": []})
+
+    def test_invalid_record_response_raises_value_error(self):
+        with pytest.raises(ValueError, match="Invalid hit testing record response"):
+            DatasetsHitTestingBase._prepare_hit_testing_records(["record"])
+
     def test_index_not_initialized(self, dataset):
         with patch.object(
             HitTestingService,

api/tests/unit_tests/controllers/service_api/dataset/test_hit_testing.py

@@ -103,7 +103,7 @@ class TestHitTestingApiPost:
         mock_dataset_svc.get_dataset.return_value = mock_dataset
         mock_dataset_svc.check_dataset_permission.return_value = None
 
-        mock_hit_svc.retrieve.return_value = {"query": "test query", "records": []}
+        mock_hit_svc.retrieve.return_value = {"query": {"content": "test query"}, "records": []}
         mock_hit_svc.hit_testing_args_check.return_value = None
         mock_marshal.return_value = []
 
@@ -149,7 +149,7 @@ class TestHitTestingApiPost:
             "score_threshold": 0.8,
         }
 
-        mock_hit_svc.retrieve.return_value = {"query": "complex query", "records": []}
+        mock_hit_svc.retrieve.return_value = {"query": {"content": "complex query"}, "records": []}
         mock_hit_svc.hit_testing_args_check.return_value = None
         mock_marshal.return_value = []
 
@@ -194,7 +194,7 @@ class TestHitTestingApiPost:
 
         mock_dataset_svc.get_dataset.return_value = mock_dataset
         mock_dataset_svc.check_dataset_permission.return_value = None
-        mock_hit_svc.retrieve.return_value = {"query": "filtered query", "records": []}
+        mock_hit_svc.retrieve.return_value = {"query": {"content": "filtered query"}, "records": []}
         mock_hit_svc.hit_testing_args_check.return_value = None
         mock_marshal.return_value = []
 
@@ -232,7 +232,7 @@ class TestHitTestingApiPost:
     @patch("controllers.console.datasets.hit_testing_base.HitTestingService")
     @patch("controllers.console.datasets.hit_testing_base.DatasetService")
     @patch("controllers.console.datasets.hit_testing_base.current_user", new_callable=lambda: Mock(spec=Account))
-    def test_post_normalizes_legacy_query_and_nullable_list_fields(
+    def test_post_prepares_nullable_list_fields(
         self,
         mock_current_user,
         mock_dataset_svc,
@@ -241,7 +241,7 @@ class TestHitTestingApiPost:
         mock_ns,
         app,
     ):
-        """Test service API normalizes legacy query shape and nullable list fields."""
+        """Test service API prepares nullable list fields from marshalled records."""
         dataset_id = str(uuid.uuid4())
         tenant_id = str(uuid.uuid4())
 

api/tests/unit_tests/services/test_hit_testing_service_dump_records.py

@@ -0,0 +1,88 @@
+from unittest.mock import Mock, patch
+
+from services.hit_testing_service import HitTestingService
+
+
+def _retrieval_record(payload: dict):
+    record = Mock()
+    record.model_dump.return_value = payload
+    return record
+
+
+def _dataset_document(
+    document_id: str = "document-1",
+    name: str = "guide.md",
+    data_source_type: str = "upload_file",
+    doc_type: str | None = None,
+    doc_metadata: dict | None = None,
+):
+    document = Mock()
+    document.id = document_id
+    document.name = name
+    document.data_source_type = data_source_type
+    document.doc_type = doc_type
+    document.doc_metadata = doc_metadata
+    return document
+
+
+class TestHitTestingServiceDumpRecords:
+    def test_dump_dataset_document_returns_frontend_required_fields(self):
+        document = _dataset_document(doc_metadata={"source": "manual"})
+
+        assert HitTestingService._dump_dataset_document(document) == {
+            "id": "document-1",
+            "data_source_type": "upload_file",
+            "name": "guide.md",
+            "doc_type": None,
+            "doc_metadata": {"source": "manual"},
+        }
+
+    def test_dump_retrieval_records_returns_dumped_records_without_document_ids(self):
+        record = _retrieval_record({"segment": None, "score": 0.95})
+
+        assert HitTestingService._dump_retrieval_records([record]) == [{"segment": None, "score": 0.95}]
+
+    def test_dump_retrieval_records_injects_documents_and_keeps_non_segment_records(self):
+        record_without_segment = _retrieval_record({"segment": None, "score": 0.95})
+        record_with_document = _retrieval_record(
+            {
+                "segment": {
+                    "id": "segment-1",
+                    "document_id": "document-1",
+                },
+                "score": 0.9,
+            }
+        )
+        scalars_result = Mock()
+        scalars_result.all.return_value = [_dataset_document()]
+
+        with patch("services.hit_testing_service.db.session.scalars", return_value=scalars_result):
+            result = HitTestingService._dump_retrieval_records([record_without_segment, record_with_document])
+
+        assert result[0] == {"segment": None, "score": 0.95}
+        assert result[1]["segment"]["document"] == {
+            "id": "document-1",
+            "data_source_type": "upload_file",
+            "name": "guide.md",
+            "doc_type": None,
+            "doc_metadata": None,
+        }
+
+    def test_dump_retrieval_records_skips_records_with_missing_documents(self, caplog):
+        record = _retrieval_record(
+            {
+                "segment": {
+                    "id": "segment-1",
+                    "document_id": "missing-document",
+                },
+                "score": 0.95,
+            }
+        )
+        scalars_result = Mock()
+        scalars_result.all.return_value = []
+
+        with patch("services.hit_testing_service.db.session.scalars", return_value=scalars_result):
+            result = HitTestingService._dump_retrieval_records([record])
+
+        assert result == []
+        assert "Skipping hit-testing records with missing documents" in caplog.text