support strings in elements.get_node_id()

[jGaboardi]

This PR resolves #519

[jGaboardi]

Seeing now that the line isn't being hit by any tests...

[codecov]

Codecov Report

Merging #522 (5250f7a) into main (891156e) will increase coverage by 0.0%.
The diff coverage is 100.0%.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #522   +/-   ##
=====================================
  Coverage   97.5%   97.5%           
=====================================
  Files         24      24           
  Lines       4115    4123    +8     
=====================================
+ Hits        4011    4020    +9     
+ Misses       104     103    -1     

Files	Coverage Δ
momepy/elements.py	96.9% <100.0%> (+0.3%)	⬆️
momepy/tests/test_elements.py	100.0% <100.0%> (ø)

[martinfleis]

It is technically covering a corner case that should not happen when everything is as expected.

We hit it when we don't have a street network edge attached to a building footprint, which happens only if a building is really far away from any street. But we could possibly simulate that by using momepy.get_network_id with a small min_size (that API and implementation there are bad...)

[jGaboardi]

It is technically covering a corner case that should not happen when everything is as expected.

We hit it when we don't have a street network edge attached to a building footprint, which happens only if a building is really far away from any street. But we could possibly simulate that by using momepy.get_network_id with a small min_size (that API and implementation there are bad...)

Working on creating a synthetic test case now that simply converts some edge IDs in the bubenec streets to NaN.

[jGaboardi]

Passing locally and hitting the offending line.

[martinfleis]

Suggested change

	_df_buildings.loc[[0, 1], "nID"] = pd.NA
	_df_buildings["nID"] = _df_buildings["nID"].astype(str)
	_df_buildings.loc[[0, 1], "nID"] = pd.NA

Might be worth actually testing for the string ID?

[jGaboardi]

Sure. Will push up a new version.

[martinfleis]

Thanks!