Add tiles_opts parameter to customize the tiles

[maximlt]

Addresses #1293

An alternative implementation might have been overloading tiles to accept a dictionary like {"tiles": "ESRI", "alpha": 0.5}. Interested in what other maintainers think about the two different approaches.

[ahuang11]

tiles_opts follows convention of what other keywords do:
legend_opts colorbar_opts

[maximlt]

Yes although these two keywords aren't hvPlot parameters but HoloViews plot options. The features parameter of hvPlot can be overloaded:

features (default=None): dict or list
    A list of features or a dictionary of features and the scale
    at which to render it. Available features include 'borders',
    'coastline', 'lakes', 'land', 'ocean', 'rivers' and 'states'.
    Available scales include '10m'/'50m'/'110m'.

It's because of this parameter I thought that maybe tiles could also be overloaded, instead of adding tiles_opts, as hvPlot suffers more from new parameters added to its flat signature than HoloViews. Perhaps with something like:

• tiles={"alpha":0.5} (== tiles=True, tiles_opts={"alpha": 0.5})
• tiles={"tiles": "ESRI", "alpha": 0.5} (== tiles="ESRI", tiles_opts={"alpha": 0.5})

I think in the end I have a slight preference for tiles_opts. Opinions @philippjfr @hoxbro ?

[ahuang11]

tiles={"tiles": "ESRI", "alpha": 0.5} is funky (the key matches the outer key); if dict, I think
tiles={"name": "ESRI", "alpha": 0.5} would be more suitable or even tiles=("ESRI", {"alpha": 0.5}), but yes I lean towards tiles_opts or tiles_kwargs

[hoxbro]

I like tiles_opts too.

[philippjfr]

+1 on tiles_opts.