Jupyter interactive plots are very slow / laggy when opened in separate tab (e.g seaborn pairplot)

[sws144]

microsoft/vscode#101238 (comment)

[IanMatthewHuff]

@sws144 Thanks for reporting this, and sorry about the issue. Do you have a code example that you could share that exhibits this behavior? I do know that we have issues with generating very large file sizes with dense plots since we generate both SVG and PNG versions of the plot. The SVG versions can get very large causing some performance issues. You can skip generating SVG versions with a setting, but turning off that setting turns of the plot viewer since that uses SVGs for display.

https://github.com/microsoft/vscode-python/issues/12388

[sws144]

Here are the two files to run (up until row 135 for the sns.pairplot). This is a fairly big pairplot, but renders plot viewer unusable.
https://github.com/sws144/titanic-model-app/blob/master/titanic-logistic-backup.py

Source data file to create plots:
https://github.com/sws144/titanic-model-app/blob/master/titanic_train.csv

Perhaps we can allow PNG versions of the plot viewer as well? Thank you!

[meyerzinn]

I am encountering this issue as well on RedHat Linux 7, even with a fairly beefy computer (albeit moderate CPU clock speeds).

[EnriqueSampaio]

Same here

[NoCh-Git]

Same here. Mac, Big Sur

[mcpeixoto]

Same here, Ubuntu 20.04 LTS

[IanMatthewHuff]

Defaulting to PNG should help resolve this, so closing.