ENH: street_alignment and get_nearest_street

[martinfleis]

I love how streamlined these things are now compared to the original :).

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.7%. Comparing base (4037c70) to head (02e9b54).
Report is 11 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##            main    #566     +/-   ##
=======================================
+ Coverage   97.4%   97.7%   +0.3%     
=======================================
  Files         26      32      +6     
  Lines       4328    4985    +657     
=======================================
+ Hits        4214    4869    +655     
- Misses       114     116      +2     

Files	Coverage Δ
momepy/functional/_distribution.py	100.0% <100.0%> (ø)
momepy/functional/_elements.py	96.8% <100.0%> (ø)
momepy/functional/tests/test_distribution.py	100.0% <100.0%> (ø)
momepy/functional/tests/test_elements.py	100.0% <100.0%> (ø)

[jGaboardi]

Is there is specific reasoning to using both pd and Series above, and using gpd.GeoSeries here? Should we also do from gpd import GeoSeries to conform? Or vice-versa?

[martinfleis]

No particular reason. My thought was that people will be familiar with Series but may not be with GeoSeries, so I wanted to indicate where it comes from. Do you have any preference here?

[jGaboardi]

My personal preference is "always explicit" so if I'd had my way it would be non-aliased & full usage: e.g.:

import pandas
...
pandas.Series

However, I know that my preference is not norm. If I had to choose one of the two methods already being used here, I'd probably go with non-aliased classes (e.g. from gpd import GeoSeries) for the type hinting.

[martinfleis]

I also know that you like linting rules, so have one for you - https://docs.astral.sh/ruff/rules/unconventional-import-alias/

[martinfleis]

I'll do the typing consistency as a follow-up across the codebase.