parenx & osmnx methods & data; evaluation WIP

[anastassiavybornova]

this PR contains

• updated methods notebook for osmnx: node consolidation with default param of 10m, then dropping multiedges)
• updated methods notebook for parenx incl. terminal output of cli run
• data (results for all 6 fuas) for osmnx and parenx
• added read-in functions to utils for sgeop and parenx
• adjusted parenx folder structure (to mirror rest of output folders, we now have 2 folders: parenx-voronoi and parenx-skeletonize, which each contain {fua}.parquet file)
• lastly: WIP evaluation notebook in evaluation

this resolves:

• rerun parenx simplification #130
• add osmnx data #194

& xref (wip) #200

[anastassiavybornova]

question: we now have a lot of different read-in funcs in utils, only one of those using .explode() on the gdf before returning it, would it make sense to have this as standard (cause it won't hurt?? or will it?) for all read-in funcs, or is there a specific reason why we only added it for the utils.read_manual() one?

[jGaboardi]

question: we now have a lot of different read-in funcs in utils, only one of those using .explode() on the gdf before returning it, would it make sense to have this as standard (cause it won't hurt?? or will it?) for all read-in funcs, or is there a specific reason why we only added it for the utils.read_manual() one?

Good question.

Some background:

• [proposal] uniform naming for read roads funcs in utils.py #104
• uniform naming - funcs & data #108

Can't at all remember why explode is needed in the manual one...

[jGaboardi]

would it make sense to have this as standard

I think it would surely make sense to have fewer lines of code. And probably very doable with some minor refactoring.

[jGaboardi]

To resolve the failures you'll need to update test_utils.test_read_parenex().

[anastassiavybornova]

@jGaboardi i've updated the tests, lmk what you think! the parenx_voronoi and parenx_skeletonize ones can probably be merged (DRY), i was just not sure how to pass 2 zipped and 1 non-zipped parameter to the pytest.mark.parametrize

[jGaboardi]

@jGaboardi i've updated the tests, lmk what you think! the parenx_voronoi and parenx_skeletonize ones can probably be merged (DRY), i was just not sure how to pass 2 zipped and 1 non-zipped parameter to the pytest.mark.parametrize

All good. Nice work on this!