Merge branch 'main' into feat/agent-node-v2

api/core/rag/extractor/entity/extract_setting.py

@@ -10,7 +10,7 @@ class NotionInfo(BaseModel):
     """
 
     credential_id: str | None = None
-    notion_workspace_id: str
+    notion_workspace_id: str | None = ""
     notion_obj_id: str
     notion_page_type: str
     document: Document | None = None

api/core/rag/extractor/extract_processor.py

@@ -166,7 +166,7 @@ class ExtractProcessor:
         elif extract_setting.datasource_type == DatasourceType.NOTION:
             assert extract_setting.notion_info is not None, "notion_info is required"
             extractor = NotionExtractor(
-                notion_workspace_id=extract_setting.notion_info.notion_workspace_id,
+                notion_workspace_id=extract_setting.notion_info.notion_workspace_id or "",
                 notion_obj_id=extract_setting.notion_info.notion_obj_id,
                 notion_page_type=extract_setting.notion_info.notion_page_type,
                 document_model=extract_setting.notion_info.document,

api/core/rag/extractor/helpers.py

@@ -45,6 +45,6 @@ def detect_file_encodings(file_path: str, timeout: int = 5, sample_size: int = 1
         except concurrent.futures.TimeoutError:
             raise TimeoutError(f"Timeout reached while detecting encoding for {file_path}")
 
-    if all(encoding["encoding"] is None for encoding in encodings):
+    if all(encoding.encoding is None for encoding in encodings):
         raise RuntimeError(f"Could not detect encoding for {file_path}")
-    return [FileEncoding(**enc) for enc in encodings if enc["encoding"] is not None]
+    return [enc for enc in encodings if enc.encoding is not None]