Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (58017dc) 77.8% compared to head (cc163de) 77.8%.

@@          Coverage Diff          @@
##            main    #439   +/-   ##
=====================================
  Coverage   77.8%   77.8%           
=====================================
  Files         27      27           
  Lines       2638    2638           
=====================================
  Hits        2053    2053           
  Misses       585     585           

I had thought I had a nice solution here, but the new method still is being caught. So after checking, we are pulling geopandas==0.14.0+55.gfb3dd81 as dev from GitHub. But, of course, we can get geopandas==0.14.2 from PyPI.

@martinfleis I am wondering if you think we should change something on our end or if this is a tagging issue in the GeoPandas repo?

I am wondering if you think we should change something on our end or if this is a tagging issue in the GeoPandas repo?

How do you mean this would be an issue in GeoPandas?

I am wondering if you think we should change something on our end or if this is a tagging issue in the GeoPandas repo?

How do you mean this would be an issue in GeoPandas?

This is not exactly related to the union_all functionality. It seems like maybe the tag isn't exactly catching when installing via GitHub:

$ mamba create -n geopandas_from_github python pip git
$ mamba activate geopandas_from_github
$ pip install git+https://github.com/geopandas/geopandas.git
$ python -c "import geopandas; print(geopandas.__version__)"
0.14.0+55.gfb3dd81

And we can see the same version from within our CI. Should this not be something like 0.14.2+##.###? Perhaps I am missing something?

This is expected since we cut minor patches from a branch with backported commits.

@gegen07 & @martinfleis

The fixture I wrote combines all network instantiation stuff for testing and dramatically reduces warning thrown.