padding doesn't work with map tiles

[rsignell]

I want to plot a single station location (lon,lat point) on a map, but set a larger lon/lat extent than default so I can more easily see the "big picture" for where the point is located.

@ahuang11 suggested I post an issue here, as padding should work, but seems to fail when map tiles are specified.

For example, this code:

import hvplot.pandas
import pandas as pd

# Create a DataFrame with the single lon, lat point
data = {'lon': [-70.4], 'lat': [42.35]}  # A point in Massachusetts Bay
df = pd.DataFrame(data)
df.hvplot.points('lon', 'lat', geo=True, size=10, color='red', tiles='ESRI', padding=2.0)

results in plot where it seems the padding was just ignored:

My workaround is to calculate and specify the xlim and ylim, but it would be nice if padding just worked!

I'm running WSL on Windows 11 with:

(pangeo) rsignell@OSC:~$ conda list | grep -E "bokeh|pyviz|widgets|hvplot|geoviews|holoviews|panel|datashader"
bokeh                     3.6.2              pyhd8ed1ab_1    conda-forge
datashader                0.16.3             pyhd8ed1ab_1    conda-forge
geoviews                  1.14.0               hd8ed1ab_0    conda-forge
geoviews-core             1.14.0             pyha770c72_0    conda-forge
holoviews                 1.20.0             pyhd8ed1ab_1    conda-forge
hvplot                    0.11.2             pyhd8ed1ab_0    conda-forge
ipywidgets                8.1.5              pyhd8ed1ab_1    conda-forge
jupyterlab_widgets        3.0.13             pyhd8ed1ab_1    conda-forge
panel                     1.5.4              pyhd8ed1ab_0    conda-forge
pyviz_comms               3.0.3              pyhd8ed1ab_0    conda-forge
widgetsnbextension        4.0.13             pyhd8ed1ab_1    conda-forge

[holovizbot]

This issue has been mentioned on HoloViz Discourse. There might be relevant details there:

https://discourse.holoviz.org/t/setting-padding-in-hvplot-when-using-tiles/8521/10

[philippjfr]

Thanks @rsignell, I think it's way past time we revisited the map tiles renderer implementation in Bokeh which is at fault for a lot of the oddities (like this one) that happen if you add a tile to your plot.

[maximlt]

@philippjfr if you have more concrete ideas on the changes required in Bokeh to improve this, I encourage you to capture them in a Bokeh issue 🙃

[rsignell]

Thanks @philippjfr and @maximlt for jumping in on this!