Merge branch 'feat/support-knowledge-metadata' into deploy/dev

api/core/rag/retrieval/dataset_retrieval.py

@@ -787,9 +787,9 @@ class DatasetRetrieval:
         user_id: str,
         metadata_filtering_mode: str,
         metadata_model_config: ModelConfig,
-        metadata_filtering_conditions: MetadataFilteringCondition,
+        metadata_filtering_conditions: Optional[MetadataFilteringCondition],
         inputs: dict,
-    ) -> dict[str, list[str]]:
+    ) -> Optional[dict[str, list[str]]]:
         document_query = db.session.query(Document).filter(
             Document.dataset_id.in_(dataset_ids),
             Document.indexing_status == "completed",
@@ -916,7 +916,7 @@ class DatasetRetrieval:
             case "≥", ">=":
                 query = query.filter(Document.doc_metadata[metadata_name] >= value)
             case _:
-                 pass
+                pass
 
     def _fetch_model_config(
         self, tenant_id: str, model: ModelConfig

api/core/workflow/nodes/knowledge_retrieval/knowledge_retrieval_node.py

@@ -30,7 +30,6 @@ from core.workflow.nodes.knowledge_retrieval.template_prompts import (
     METADATA_FILTER_USER_PROMPT_1,
     METADATA_FILTER_USER_PROMPT_3,
 )
-from core.workflow.nodes.list_operator.exc import InvalidConditionError
 from core.workflow.nodes.llm.entities import LLMNodeChatModelMessage, LLMNodeCompletionModelPromptTemplate
 from core.workflow.nodes.llm.node import LLMNode
 from core.workflow.nodes.question_classifier.template_prompts import QUESTION_CLASSIFIER_USER_PROMPT_2

api/models/dataset.py

@@ -203,7 +203,7 @@ class Dataset(db.Model):  # type: ignore[name-defined]
     def doc_metadata(self):
         dataset_metadatas = db.session.query(DatasetMetadata).filter(DatasetMetadata.dataset_id == self.id).all()
 
-        return [
+        doc_metadata = [
             {
                 "id": dataset_metadata.id,
                 "name": dataset_metadata.name,
@@ -211,6 +211,43 @@ class Dataset(db.Model):  # type: ignore[name-defined]
             }
             for dataset_metadata in dataset_metadatas
         ]
+        if self.built_in_field_enabled:
+            doc_metadata.append(
+                {
+                    "id": "built-in",
+                    "name": BuiltInField.document_name,
+                    "type": "string",
+                }
+            )
+            doc_metadata.append(
+                {
+                    "id": "built-in",
+                    "name": BuiltInField.uploader,
+                    "type": "string",
+                }
+            )
+            doc_metadata.append(
+                {
+                    "id": "built-in",
+                    "name": BuiltInField.upload_date,
+                    "type": "date",
+                }
+            )
+            doc_metadata.append(
+                {
+                    "id": "built-in",
+                    "name": BuiltInField.last_update_date,
+                    "type": "date",
+                }
+            )
+            doc_metadata.append(
+                {
+                    "id": "built-in",
+                    "name": BuiltInField.source,
+                    "type": "string",
+                }
+            )
+        return doc_metadata
 
     @staticmethod
     def gen_collection_name_by_id(dataset_id: str) -> str:

api/services/metadata_service.py

@@ -31,8 +31,7 @@ class MetadataService:
         return metadata
 
     @staticmethod
-    def \
+    def update_metadata_name(dataset_id: str, metadata_id: str, name: str) -> DatasetMetadata:
-            update_metadata_name(dataset_id: str, metadata_id: str, name: str) -> DatasetMetadata:
         lock_key = f"dataset_metadata_lock_{dataset_id}"
         try:
             MetadataService.knowledge_base_metadata_lock_check(dataset_id, None)