ENH: include assign_self_weight (formerly fill_diagonal) in Graph

[martinfleis]

Closes #668

@knaaptime this includes pandas-fu version as well as roundtrip via sparse for now. Timing below.

import geopandas as gpd
from geodatasets import get_path
from libpysal import graph

df = gpd.read_file(get_path("geoda.ncovr"))
contig  = graph.Graph.build_contiguity(df)

%%timeit
contig.fill_diagonal()
5.31 ms ± 108 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

%%timeit
contig.fill_diagonal_sparse()
16.3 ms ± 92.8 µs per loop (mean ± std. dev. of 7 runs, 100 loops each)

Here, the sparse version is about 3x slower. This ratio seems to be more or less consistent with any size of the graph.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (27c1f49) 85.0% compared to head (0fe196f) 85.1%.
Report is 8 commits behind head on main.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #670   +/-   ##
=====================================
  Coverage   85.0%   85.1%           
=====================================
  Files        139     139           
  Lines      14889   14903   +14     
=====================================
+ Hits       12663   12677   +14     
  Misses      2226    2226           

Files	Coverage Δ
libpysal/graph/base.py	97.7% <100.0%> (+<0.1%)	⬆️
libpysal/graph/tests/test_base.py	100.0% <100.0%> (ø)

... and 1 file with indirect coverage changes

[jGaboardi]

Looking good.

[sjsrey]

#668 (comment)

[jGaboardi]

Will the terminology in this PR be affected by the discussion in #668?

[martinfleis]

Will the terminology in this PR be affected by the discussion in #668?

Yes. I'll update this PR as soon we have an agreement.

[martinfleis]

Updated as per #668. Check the docstring please if it makes sense.

[jGaboardi]

Shall we update the title of the issue to reflect changes in terminology?