cool. The way i'd handled that before was to just cheat and inculude extra args like tiles in the kwargs for one of the earlier maps (like stick them in the edge_kws). This is definitely preferable

A side note - I also realised that we can directly plot weight. Neat.

oh dope!

The way i'd handled that before was to just cheat and inculude extra args like tiles in the kwargs for one of the earlier maps

Yeah but that is very unintuitive and requires you to know the implementation.

Yeah but that is very unintuitive and requires you to know the implementation.

💯

A side note - I also realised that we can directly plot weight. Neat.

does that mean that works for the static plots too?

does that mean that works for the static plots too?

No, because we don't reuse GeoDataFrame.plot there but build collections directly. Going via GeoPandas would cause significant performance hit. We could potentially map colours ourselves though...

Codecov Report

Merging #627 (9c17c03) into main (4290c11) will increase coverage by 0.0%.
The diff coverage is 100.0%.

@@          Coverage Diff          @@
##            main    #627   +/-   ##
=====================================
  Coverage   84.5%   84.5%           
=====================================
  Files        139     139           
  Lines      14963   14967    +4     
=====================================
+ Hits       12638   12642    +4     
  Misses      2325    2325           

@knaaptime would you like to cut 4.9.1? :)

@knaaptime I suggest we first figure out what caused this conda-forge error https://dev.azure.com/conda-forge/feedstock-builds/_build/results?buildId=817319&view=logs&j=656edd35-690f-5c53-9ba3-09c10d0bea97&t=986b1512-c876-5f92-0d81-ba851554a0a3

cc @jGaboardi

It is midnight here so that is not a task for me today :)

so, i dunno why thats failing. The error seems to come from xarray not being available when weights is imported, but afaict xarray is never imported in the weights module except inside a function-level 'try' block.

...that said, it looks like we do define xarray as a hard dependency in pyproject.toml, but not in the conda forge recipe. A quick fix would be to include xarray in conda, but do we need it? Also not clear why the error is actually materializing here