fix multi thread app context (#868)

Co-authored-by: jyong <jyong@dify.ai>

api/core/indexing_runner.py

@@ -7,6 +7,7 @@ import time
 import uuid
 from typing import Optional, List, cast
 
+from flask import current_app, Flask
 from flask_login import current_user
 from langchain.schema import Document
 from langchain.text_splitter import RecursiveCharacterTextSplitter, TextSplitter
@@ -522,7 +523,8 @@ class IndexingRunner:
                 sub_documents = all_documents[i:i + 10]
                 for doc in sub_documents:
                     document_format_thread = threading.Thread(target=self.format_qa_document, kwargs={
-                        'tenant_id': tenant_id, 'document_node': doc, 'all_qa_documents': all_qa_documents})
+                        'flask_app': current_app._get_current_object(), 'tenant_id': tenant_id, 'document_node': doc,
+                        'all_qa_documents': all_qa_documents})
                     threads.append(document_format_thread)
                     document_format_thread.start()
                 for thread in threads:
@@ -530,10 +532,11 @@ class IndexingRunner:
             return all_qa_documents
         return all_documents
 
-    def format_qa_document(self, tenant_id: str, document_node, all_qa_documents):
+    def format_qa_document(self, flask_app: Flask, tenant_id: str, document_node, all_qa_documents):
         format_documents = []
         if document_node.page_content is None or not document_node.page_content.strip():
             return
+        with flask_app.app_context():
             try:
                 # qa model document
                 response = LLMGenerator.generate_qa_document(tenant_id, document_node.page_content)