Hi everyone,
I’m working with Code Interpreter and have a scenario where each user interaction happens in a single thread. I can upload and manage files within that thread, but I’ve run into a problem. The thread can have a maximum of 20 files, so I’ve been deleting old ones before uploading new ones. However, after deleting outdated files and then asking Code Interpreter to “do the same analysis with this new file,” it seems to still reference the previously deleted files. The assistant returns an error saying something like:
“It seems there is an issue accessing the third file. Could you please try re-uploading it, or confirm the data it should contain for further analysis?”
I’d like to maintain a clean workflow and ensure Code Interpreter doesn’t get stuck on previously deleted files. Has anyone else run into this issue? Are there best practices for properly “resetting” the file state in a thread, or any recommended approaches for ensuring that newly uploaded files are recognized cleanly without lingering references to old ones?
Thanks in advance for any insights!