Vector store is rejecting documented 'supported files'

API Bugs
1

Hi! I was experimenting with creating a vector store with code. I have a very minimal .sh file that prevented the store from being created:

#!/bin/bash

# Start the SSH service
service ssh start

# Start the application
flask run --host=0.0.0.0 --port=5000

and a sample of the implementation code:

		file_paths = self._get_supported_files(my_files)
		file_streams = [open(path, "rb") for path in file_paths]

		vector_store = self.openai.beta.vector_stores.create(name="my_vector_store")

		file_batch = self.openai.beta.vector_stores.file_batches.upload_and_poll(
			vector_store_id=vector_store.id, files=file_streams
		)

Error message

{‘error’: {‘message’: ‘Invalid file format text/x-shellscript. Supported formats: “c”, “cpp”, “css”, “csv”, “docx”, “gif”, “html”, “java”, “jpeg”, “jpg”, “js”, “json”, “md”, “pdf”, “php”, “png”, “pptx”, “py”, “rb”, “tar”, “tex”, “ts”, “txt”, “webp”, “xlsx”, “xml”, “zip”’, ‘type’: ‘invalid_request_error’, ‘param’: None, ‘code’: None}}

Doc that says .sh is supported:
https://platform.openai.com/docs/assistants/tools/file-search/supported-files

Seems like the doc and what’s returned from the server are different.

2

tar works though. The mime type inspector doesn’t get inside the archive, if you can incorporate expanding into your code interpreter workflow.

Uploading to the vector store, this would be very poorly retrieved anyway. You could do what is required to have it work for file search based on AI query - make it into a document that has example question and answer about how to initialize flask.