Automatically set correct renderer for Plotly charts in VSCode Notebooks

Because https://github.com/microsoft/vscode-notebook-renderers/pull/199 is unmerged, users of Plotly.py using the VSCode notebooks cannot access the latest Plotly.js/Plotly.py features and no discernible error message is provided. This leads to confusion, because users are working with the latest version of Plotly.py and their charts don't work properly.

This has caused issues for adoption of new features like MapLibre and subtitles in charts, and may others. See e.g.:

• https://github.com/plotly/plotly.py/issues/4715
• https://github.com/plotly/plotly.py/issues/4775
• https://github.com/plotly/plotly.py/issues/4701
• https://github.com/plotly/plotly.py/issues/4583
• https://github.com/plotly/plotly.py/issues/4560
• https://github.com/plotly/plotly.py/issues/4475
• https://github.com/plotly/plotly.py/issues/4528
• https://github.com/plotly/plotly.py/issues/4264
• https://github.com/plotly/plotly.py/issues/4173
• https://github.com/plotly/plotly.py/issues/4094

Developers can workaround this by adding to their notebook:

import plotly.io as pio
pio.renderers.default = "notebook_connected"

While this is trivial to change, it's undiscoverable and non-intuitive. If you have the latest version of Plotly.py installed, you should be able to use the latest features in any Python environment!

Regardless of whether https://github.com/microsoft/vscode-notebook-renderers/pull/199 is merged, this dependency means VSCode Notebook users will always lag behind Plotly.py releases as we wait for VSCode to adopt and release their own updates, which means VSCode users would never have day-of updates without a workaround 🙁

Is it possible to detect that we're in a VSCode Notebook environment and automatically set the correct renderer to avoid this dependency on a VSCode update?

Aucun fichier du dépôt ni aucun test n’est nommé. Commencez par lire la configuration des renderers exposée par plotly.io, en particulier le point d’entrée pio.renderers.default, et étudiez comment les environnements de notebook sont identifiés. Le travail est considéré comme terminé lorsque les utilisateurs de VSCode Notebook reçoivent automatiquement le renderer compatible actuel sans avoir à ajouter le workaround documenté, avec une couverture du comportement de détection.