Esri / arcgis-python-api

Documentation and samples for ArcGIS API for Python
https://developers.arcgis.com/python/
Apache License 2.0
1.81k stars 1.08k forks source link

Jupyterlab mapping ipywidget incompatibility with ArcGIS 3.x cloned environments #1487

Closed ChimskyChim closed 1 year ago

ChimskyChim commented 1 year ago

Beginning in the fall of 2022 in a version of arcgis pro v3.x, the procedures for getting jupyterlab fully configured in a cloned environment (https://developers.arcgis.com/python/guide/using-the-jupyter-lab-environment/ [nam02.safelinks.protection.outlook.com] ) seem to no longer allow for mapping in juputerlab.

I’m currently using arcgis pro v3.1.0 with arcgis 2.1.0.1. The environment can be successfully cloned, albeit very large >4gb. Nodejs can be installed. I noticed that the jupyterlab-manager is now preinstalled by esri. However, the arcgis-map-ipywidget installation either fails due to compatibility issues with jupyterlab (e.g. arcgis-map-ipywidget@2.1.0) or results in errors (e.g., java script errors or just failure to produce a map) when trying to display maps in jupyterlab. When trying to install other versions of jupyterlab and/or other versions of arcgis and arcgis-map-ipywidget, such that the versions align as indicated in the instructions, compatibility issues still arise. Online maps in jupyter notebooks and in notebooks within arcgis pro still work just fine. Are there any new instructions that should be followed to get jupyterlab configured to display maps in a cloned environment?

I filed an esri case about this: Esri Case #03278504. esri support services said they were too able to reproduce this problem on their end and that I should report a bug.

nanaeaubry commented 1 year ago

Hello, we are aware of the issues concerning the map widget in jupyter lab and at this time we cannot support it.

We will work on this for future releases and take the esri case into account.

Thank you

bofethe commented 1 year ago

This also affects VS Code's Jupyter implementation.

nanaeaubry commented 1 year ago

@bofethe Yes, we cannot guarantee compatibility in there. That is maintained by VS Code.