Merge pull request #72 from TukaTek/feature/implement-GPT-5

Add the odt fix

Author: AnkitTukatek

backend/onyx/server/documents/connector.py

@@ -1988,18 +1988,13 @@ def should_process_file(file_path: str) -> bool:
             if file.content_type == "application/vnd.oasis.opendocument.text" or (
                 file.filename and file.filename.lower().endswith('.odt')
             ):
-                extracted_text = convert_odt_to_txt(file.file)  # Define this helper; see below
-                text_file_id = file_store.save_file(
-                    content=io.BytesIO(extracted_text.encode()),
-                    display_name=file.filename,
-                    file_origin=file_origin,
-                    file_type="text/plain",
-                )
+                text_file_id = convert_odt_to_txt(file, file_store)
                 deduped_file_paths.append(text_file_id)
                 deduped_file_names.append(file.filename)
                 continue
 
 
+
             # Special handling for doc files - only store the plaintext version
             file_type = mime_type_to_chat_file_type(file.content_type)
             if file_type == ChatFileType.DOC: