Add min_number parameter to segmentize

[runarberg]

I am working on a neural network which operates on polylines with variable number of vertices. To fit each polyline into the ANN the simplify method works well to reduce the number of vertices of long polylines. However to increase the number of vertices for short polylines, I am out of luck (currently I just fill the matrix with zeros but I suspect that may limiting the sensitivity of the model).

I am thinking that segmentize could work similar to simplify where you either pass in max_distance or min_number (but not both), and if you pass in min_number a simple algorithm of inserting a redundant vertex in the center of the longest line iteratively until the min_number is reached.

[rafaqz]

Sounds like a nice improvement, especially if the interface matches simplify.

Can you make a PR?

[asinghvi17]

Yeah that would be cool! I also wonder if this shouldn't just be a different algorithm - perhaps that needs some setup. But the thing you do in the core loop is then a bit different, because you're trying to optimize over a global function not a local one.

[runarberg]

I can make a PR with at least a draft implementation and some tests. But I don’t know how to implement this as a different algorithm.