Jupyter Notebook with Plotly Express file size too big and not displaying when converted to HTML

📊 Plotly Python
1

Hello everyone,

For the past 2 year or so I have used jupyter notebook with plotly express charts to generate interactive HTML reports.
Recently I had to replace my work laptop with a new one, and following that I had to reinstall and setup everything back from scratch (Python, Jupyter, Plotly etc)

After the migration I have encountered 2 major issues:

  1. The jupyter notebook filesize with the charts is currently very big, around 400MB (it was only about 20MB before the migration)
  2. The charts are working on the jupyter notebook, but when I convert it to HTML using nbconvert, the charts don’t show up on the HTML file.

Current versions of python & libraries used:

  1. Python 3.13.1
  2. Plotly 6.0.0, plotly-express 0.4.1
  3. Jupyter notebook 7.3.2
  4. Jupyter lab 4.3.5

Snippet of code used to generate the charts:

import plotly.express as px

fig = px.box(df, x=‘year and ISO week’, y=parameter, color=‘year and ISO week’)

fig.update_layout(xaxis={‘type’:‘category’})

fig.show()

Thanks in advance for any suggestions!

2

Hey @ngom52 ! Could you test with plotly<6.0.0 to verify whether this is an issue with the latest major version? We made some changes to how Plotly charts work in notebooks so it’s possible. If that’s the case I can create a ticket on your behalf.

3

Hi, thanks for the suggestion. I downgraded to plotly 5.24.1 and now it’s working as expected

4

By downgrading plotly from 6.0.0 to 5.24.1 (conda-forge), the same notebook went from 18.5MB to 4.5MB.

Is there anything new in 6.0.0 that could be disabled or configured to keep plot sizes (MB) controlled?

5

Hi @nathandrezner, new to this forum but hoping you can help me out. I’ve been having the same problem as ngom52. I was able to downgrade my plotly to 5.24.1, which solved the nbconvert issue, but when I do this, my figures no longer render inline in jupyterlab notebooks. My workaround is to have two envs, one for working in notebooks, one for exporting notebooks, both with different versions of plotly. But I havent been able to determine what is the root cause of this issue. I’d prefer to use Plotly 6.0.0 for everything, but cannot upgrade until I figure this out.

6

Hey @jcasino – our team is actively working on this issue: You can keep track of progress in the ticket (plotly on jupyter creates HUGE notebooks file sizes · Issue #5056 · plotly/plotly.py · GitHub) and pull request (Add back the mime renderer JL extension by marthacryan · Pull Request #5096 · plotly/plotly.py · GitHub). I’ll add a note for the nbconvert issue which seems related but want to make sure is being tracked.

Once this is merged we’ll issue a patch or minor release which hopefully should unblock the folks ITT :rocket: