Running particular notebook breaks UI
- Dominant language
- Jupyter Notebook
- Stars
- 13.3k
- Forks
- 5.8k
- Avg merge
- 6d 11h
- Merged PRs (30d)
- 7
Description
When running one of my previously functional notebooks I noticed that after hitting "run all", the Jupyter UI permanently breaks for that notebook, even after shutting down and restarting Jupyter. This does only happen for this particular notebook and the problems occur with both Firefox and Chrome.
I'm using Jupyter notebook 5.7.8 and I've uploaded the problematic notebook [here](https://gist.github.com/Time0o/fd314edc5527dbb2c59e878b71eee4d1). Since the notebook contains no code that interacts with the way it should be displayed it seems to me that this is some underlying problem with Jupyter (the problem persists even after clearing all cell output so I don't think its one of the displayed figures).
GitHub won't let me attach an image to this post so I have uploaded an image of what this looks like for me [here](https://imgur.com/a/6V5VUax).
EDIT: this does seem to happen for other notebooks as well, so maybe this is a problem with the version of Jupyter I have installed, it always seems to occur after executing the first cell that creates a Matplotlib figure.
ANOTHER EDIT: after further experimenting this only seems to happen if I modify `matplotlib.rcParams` in any way before plotting, e.g. `matplotlib.rcParams['font.size'] = 10`. This happens across several versions of the `notebook` and `matplotlib` pip packages.
Contributor guide
Research direction
Reproduce with the linked problematic notebook, then run the first cell that creates a Matplotlib figure after modifying matplotlib.rcParams, such as font.size. Check whether the Jupyter UI breaks in Firefox and Chrome and whether the behavior persists after clearing output or restarting. Done means identifying the cause and preventing the UI failure across the reported notebook and package versions.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- jupyter-notebook, matplotlib
- Domain
- data-visualization, frontend
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100