[Bug][Jupyter Widget] Google Colab Dependency Conflict Prevents Rendering

[inventingian]

Describe the bug
Using Google Colab and running ! pip install keplergl , a dependency conflict arises with jupyterlabs requirement for ipykernal version of >6.5.0 and google-colab requires version 5.5.6. This jupyerlabs dependency seems to prevent the widget from rendering using the provided keplergl documentation. This worked prior to the latest releases between December 2024 and January 2025

To Reproduce
Steps to reproduce the behavior:

1. Open new Google Colab doc
2. pip install keplergl
3. map_1 = KeplerGl()
   3a. map_1

Expected behavior
Should display the kepler widget

Screenshots
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. google-colab 1.0.0 requires ipykernel==5.5.6, but you have ipykernel 6.29.5 which is incompatible.

Environment (please complete the following information):

• Python version: python3
• keplergl Widget version: 0.3.6

Additional context
The jupyerlabs v4.1.6 requires ipykernal 6.5.0

[lixun910]

We just updated keplergl-jupyter to v0.3.6 so it can work with latest Jupyter notebook with ipywidgets > 8 and notebook > 6. Can you try if the older version works on Google colab? Thanks!

!pip install keplergl==0.3.4a3

Ref: #2588

[inventingian]

Interesting, it seems a few other things changed too.

The data has to be passed in as data={'df':df}, used to work with data=df. Easy enough fix.

However it seems my real issue using a custom config. If I create a custom config and try to use it, it doesnt show the map. (in my own notebook)

I tried troubleshooting with your sample ^ and if I run w1.config, it returns an empty dictionary before and after changing layer settings.

This occurs in 0.3.2 as well, but didn't used to.

[danielsparing]

• 0.3.7 doesn't show map
• 0.3.6 doesn't show map
• 0.3.5 doesn't show map
• 0.3.4 doesn't show map
• 0.3.4a4 doesn't show map
• 0.3.4a3 shows map even though it also throws the error google-colab 1.0.0 requires ipykernel==5.5.6, but you have ipykernel 6.29.5 which is incompatible.
• 0.3.2 shows map without error.

(all in Colab)

[danielsparing]

One more level of detail, on Colab:

tldr: show() still works, but with its limitations, like no auto-zoom/pan of viewport to data.

!pip install keplergl --quiet

import pandas as pd
from keplergl import KeplerGl

import keplergl
keplergl.__version__
# Returns '0.3.7'

from google.colab import output
output.enable_custom_widget_manager()

df = pd.DataFrame(
    {'City': ['Buenos Aires', 'Brasilia', 'Santiago', 'Bogota', 'Caracas'],
     'Latitude': [-34.58, -15.78, -33.45, 4.60, 10.48],
     'Longitude': [-58.66, -47.91, -70.66, -74.08, -66.86]})

m = KeplerGl()
m.add_data(data=df, name='cities')
m
# does not show map

m.show()
# shows map with data, but no auto-zoom/pan of viewport to data

compare this with 0.3.2, where it was possible to just call (without show(), and with auto-zoom to data):

!pip install keplergl==0.3.2 --quiet

import pandas as pd
from keplergl import KeplerGl

from google.colab import output
output.enable_custom_widget_manager()

df = pd.DataFrame(
    {'City': ['Buenos Aires', 'Brasilia', 'Santiago', 'Bogota', 'Caracas'],
     'Latitude': [-34.58, -15.78, -33.45, 4.60, 10.48],
     'Longitude': [-58.66, -47.91, -70.66, -74.08, -66.86]})

m = KeplerGl()
m.add_data(data=df, name='cities')
m
# shows map zoomed to data