revert shapely hard dependecy

[martinfleis]

Closes #453

Since this a regression of some sort, I'd recommend issuing 4.6.1 release.

I have also included a CI environment that does not have any optional dependencies so we can catch this issue next time before merging and releasing.

[martinfleis]

I also had to do some adaptation in tests to make sure they're properly skipped if the required optional dep is missing.

[codecov]

Codecov Report

Merging #454 (81147b2) into master (0a4b079) will increase coverage by 0.6%.
The diff coverage is 100.0%.

@@           Coverage Diff            @@
##           master    #454     +/-   ##
========================================
+ Coverage    78.8%   79.4%   +0.6%     
========================================
  Files         122     122             
  Lines       12910   12957     +47     
========================================
+ Hits        10168   10286    +118     
+ Misses       2742    2671     -71     

Impacted Files	Coverage Δ
libpysal/cg/tests/test_voronoi.py	89.7% <100.0%> (+3.3%)	⬆️
libpysal/weights/tests/test_contiguity.py	97.8% <100.0%> (+1.8%)	⬆️
libpysal/weights/tests/test_distance.py	97.6% <100.0%> (+0.2%)	⬆️
libpysal/weights/tests/test_gabriel.py	100.0% <100.0%> (ø)
libpysal/weights/util.py	78.5% <100.0%> (+2.8%)	⬆️
libpysal/weights/tests/test_util.py	99.1% <0.0%> (+0.9%)	⬆️
libpysal/cg/alpha_shapes.py	62.5% <0.0%> (+0.9%)	⬆️
libpysal/weights/contiguity.py	80.4% <0.0%> (+1.2%)	⬆️
libpysal/weights/gabriel.py	45.3% <0.0%> (+1.6%)	⬆️
... and 11 more

[jGaboardi]

Shall we merge or get another review from @sjsrey?

[ljwolf]

I should have also checked this before the tag for v4.6.1. Happy to make 4.6.2 ?

[martinfleis]

I should have also checked this before the tag for v4.6.1. Happy to make 4.6.2 ?

I forgot that it is still there. Yes, I would do 4.6.2 with this.

[ljwolf]

on it