Non deterministic zoom level when exporting to Jupyter slides

[astrojuanlu]

Please add a code sample or a nbviewer link, copy-pastable if possible

http://nbviewer.jupyter.org/format/slides/gist/Juanlu001/7feeffbada57114423d68fe0bf491352#/0/1

Problem description

[this should explain why the current behavior is a problem and what is the expected output.]

When opening the link above, sometimes I get a zoomed map and sometimes I get a world map. It depends on F5 versus Ctrl+F5 and has a random result. Tested on Linux Mint 18.3 with both Firefox and Chromium.

Expected Output

Always a zoomed map.

Output of folium.__version__

$ python -c "import folium as f;print(f.__version__)"
0.5.0

[astrojuanlu]

I attach a screencast of the observed behavior:

[ocefpaf]

I believe that ipyleaft may be a better fit for this case. We cannot keep up with the constant changes in how jupyter/jupyterlab deals with HTML/JS display :-(

[astrojuanlu]

Thanks for your quick answer @ocefpaf. It's a pity because ipyleaflet being a widget instead of an HTML content makes it trickier to share in some use cases, but we understand the issue and in fact we have been doing experiments with both.

[ocefpaf]

being a widget instead of an HTML content makes it trickier to share in some use cases

I know, that is the only reason why we are still keeping folium afloat.

but we understand the issue and in fact we have been doing experiments with both.

The changes in how they handle HTML/JS made me loose my interest in this sort of thing. For example, to get folium to work with JupyterLab I have to actually write JS but the whole reason behind folium is to not write JS 😆

[ocefpaf]

@Juanlu001 I looked into this and I could not find the culprit. Something with revealjs does not play well with how we are loading all the folium JS deps. Closing this as "unsolvable" but I'll keep looking into a better way ti isolate things.

PS: note that the m.fit_bounds(m.get_bounds()) may help you there.