ENH: add Graph.build_h3

[martinfleis]

Closes #667

I am open to API suggestions, especially regarding weight. Example usage:

import geopandas as gpd
from geodatasets import get_path
from tobler.util import h3fy

from libpysal import graph

gdf = gpd.read_file(get_path("geoda guerry"))
h3 = h3fy(gdf, resolution=4)
G = graph.Graph.build_h3(h3.index)

In theory, you can use this to do spatial stats without having any geometry, just the indices.

@ljwolf I am aware of you suggestion to support other hierarchical indices as well but I am not sure if the overall goal with that would be to have a single builder method (or of it is wise to have only one) or individual, so I went ahead with H3-specific one.

[knaaptime]

would it be better to move h3fy to lib?

[martinfleis]

would it be better to move h3fy to lib?

I got no opinion there, either way is fine with me.

[codecov]

Codecov Report

Attention: Patch coverage is 94.33962% with 3 lines in your changes are missing coverage. Please review.

Project coverage is 85.0%. Comparing base (bcabdbc) to head (88c3448).
Report is 2 commits behind head on main.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #694   +/-   ##
=====================================
  Coverage   85.0%   85.0%           
=====================================
  Files        141     143    +2     
  Lines      15203   15259   +56     
=====================================
+ Hits       12924   12974   +50     
- Misses      2279    2285    +6     

Files	Coverage Δ
libpysal/graph/_indices.py	100.0% <100.0%> (ø)
libpysal/graph/tests/test_indices.py	100.0% <100.0%> (ø)
libpysal/graph/base.py	97.0% <75.0%> (-0.9%)	⬇️

... and 1 file with indirect coverage changes