Keplergl problem in Streamlit

[mcharles96]

I can't deploy keplergl map in Streamlit when I use geopandas and movingpandas in my dataframe. When I only use original dataframe, with latitude and longitude data, works well, but when I use these libraries to analysis stop points and trajectories, I get an error.

Obs.: I used Jupyter to test and visualize the map, but the app is .py file made in VS Code.

[chrieke]

Hey @mcharles96 thanks for opening this issue! Looking through the additional screenshots you sent me on Twitter, I'm not sure if this in issue of this streamlit-keplergl component, but rather an issue of movingpandas + keplergl working together. Were you successful in running your example in the Jupyter notebook environment?
If not, I think it would be better to ask for help in the movingpandas or keplergl communities. Maybe this article also helps to clarify the behaviour.

[chrieke]

@mcharles96 Were you able to solve the problem? Closing, please feel free to reopen if new info.

[bhupinders]

Hi @chrieke Thanks for the package. I am also getting the same error.

The following renders without any issues.
map_1 = KeplerGl(height=400)

But on adding any data, it fails.
map_1 = KeplerGl(height=600, data={ "gps": data })

Where data is a pandas dataframe with lat/lon columns. the map_1 loads in jupyter without using streamlit.

The main error is:
AttributeError: 'str' object has no attribute '_geom'

Which leads to:
TypeError: fig argument has to be a keplergl map object of type keplergl.KeplerGl!

Any idea what might be going on here? (It doesn't work on first loading of the page also and not just reloading)

Thanks!

[bhupinders]

Somehow it works If I save the data to disk and read again. Maybe modified dataframes are not working due to some reason.