- Correct bug where URL can be too long due to tracking parameters ==> added clean_url function, to be called before adding an URL.

eveai_api/api/document_api.py

@@ -17,7 +17,7 @@ from common.utils.document_utils import (
     create_document_stack, process_url, start_embedding_task,
     EveAIInvalidLanguageException, EveAIDoubleURLException, EveAIUnsupportedFileType,
     get_documents_list, edit_document, refresh_document, edit_document_version,
-    refresh_document_with_info, lookup_document, refresh_document_with_content
+    refresh_document_with_info, lookup_document, refresh_document_with_content, clean_url
 )
 from common.utils.eveai_exceptions import EveAIException
 from eveai_api.api.auth import requires_service
@@ -271,11 +271,12 @@ class AddURL(Resource):
 
         try:
             args = document_ns.payload
-            file_content, filename, extension = process_url(args['url'], tenant_id)
+            cleaned_url = clean_url(args['url'])
+            file_content, filename, extension = process_url(cleaned_url, tenant_id)
 
             api_input = {
                 'catalog_id': args['catalog_id'],
-                'url': args['url'],
+                'url': cleaned_url,
                 'name': args.get('name') or filename,
                 'language': args['language'],
                 'user_context': args.get('user_context'),