Improve plotting

[astrojuanlu]

At the moment, default plotting is using a simple Folium map which has bad performance, and more complex .plot using ipyleaflet is still limiting the maximum number of features:

https://github.com/satellogic/telluric/blob/master/telluric/collections.py#L179-L183

We have to re-evaluate if this is needed.

Also, mapboxgl-jupyter just released a new version with Choropleth support:

https://github.com/mapbox/mapboxgl-jupyter/releases/tag/0.6.0

It still does not support all geometry types (mapbox/mapboxgl-jupyter#4) but it's in a better shape than two months ago.

[astrojuanlu]

mapboxgl-jupyter already supports most geometry types (see again mapbox/mapboxgl-jupyter#4 and new docs https://mapbox-mapboxgl-jupyter.readthedocs-hosted.com/)

Some pending, interesting issues or additions:

• Clarification on how Mapbox API & Tokens work Improve documentation about how Mapbox API & Tokens work mapbox/mapboxgl-jupyter#57
• Default to mapboxgl-jupyter in Jupyter Lab [WIP] Replace leaflet with mapbox-gl jupyterlab/jupyter-renderers#122
• JavaScript errors Add element to display Javascript errors mapbox/mapboxgl-jupyter#68 and Propagate errors from Mapbox GL JS  mapbox/mapboxgl-jupyter#62 (for better debugging, and to avoid white, empty maps)

[astrojuanlu]

mapboxgl-jupyter now supports all the geometry types we need, but the rest of the issues remain.

[astrojuanlu]

ipyleaflet now supports custom projections and a fit_bounds method jupyter-widgets/ipyleaflet#737 (comment)

On the other hand, matboxgl-jupyter looks rather abandoned and none of the issues I highlighted has been fixed. mapbox-gl-js is still active though.