Code interpreter with Jupyter Notebook does not work if generated code is executed again via "Run" Button

[InventoCasa]

Bug Report

Installation Method

Docker

Environment

• Open WebUI Version: 0.5.11

• Ollama (if applicable): -

• Operating System: Ubuntu 22.04

• Browser (if applicable): Chrome

Confirmation:

• [x] I have read and followed all the instructions provided in the README.md.
• [x] I am on the latest version of both Open WebUI and Ollama.
• [x] I have included the browser console logs.
• [x] I have included the Docker container logs.
• [x] I have provided the exact steps to reproduce the bug in the "Steps to Reproduce" section below.

Expected Behavior:

When clicking on "Run" in the code block of the answer of the LLM, the code should be executed again with Jupyter notebook

Actual Behavior:

It seems like the code is executed through the standard Pyoide Code interpreter. Although I am not sure, because in my case the pyoide interpreter does not work at all since v0.5 - but this leads to "Execution timeout exceeded" when running the generated code again, so I assume the Pyoide interpreter is chosen instead of Jupyter. Also, I see no call in the jupyter logs when executing the code via "Run" again. On the initial run which is done right when the answer is generated by the LLM, is see in the jupyter logs that it is working

Description

Bug Summary: Code interpreter with Jupyter Notebook does not work if generated code is executed again via "Run" Button

Reproduction Details

Steps to Reproduce:

1. Setup Jupyter Code Interpreter
2. Select Code Interpreter in chat and prompt e.g. "Create a script which sums up 3 random numbers"
3. Observe that the answer correctly calls jupyter notebook and returns the execution output of the generated code
4. Click on "Run" again in the code block, and observe jupyter logs: Nothing will be logged. As mentioned in my case, Pyoide does not work since v0.5 (despite setting up websocket modifications correctly in nginx), that's why I noticed this bug.