reconsider non-planar edge reconnection to skeletonized artifact

[anastassiavybornova]

continue work on typology-multiples notebook on n-clusters branch.

we need to reconsider how exactly identify the edges AND their nodes which need to get reconnected to the skeleton; and whether (not implemented yet) some of these edges should be dropped entirely under certain conditions (smth like "connected to skeleton boundary on one end, and leading into the artifact cluster on the other end", eg here: https://www.openstreetmap.org/search?lat=50.6463&lon=5.4712#map=15/50.6463/5.4712 )

the current implementation solves some cases (eg below) but "over"connects others.

[anastassiavybornova]

@martinfleis i think we can close this one?

[martinfleis]

Did we ever reconsidered it? :)

[anastassiavybornova]

i don't think we came up with a solution, but i do have a vague memory of deciding not to look further into non-planar edge connection? and accept the fact that non-planar edges are cut off at artifact boundaries?

we can also label this as feature request for future work, instead of closing, but curious to hear what the current way of reconnecting non-planar edges is (if i'm mistaken in what i recall)

[martinfleis]

and accept the fact that non-planar edges are cut off at artifact boundaries?

They are not. But they maybe should be.

I would not close this issue but I remember it being a little annoying occasionally.