Python: vectorized search (many queries in one go)
First of all, thanks for creating geo-index! It's so much faster than rtree! Indexing 100k rectangles alone is an order of magnitude faster 🚀
I'm curious if you have a sense how complicated is would be to implement vectorized search? I'd be happy to help. Not sure if you're aiming for it but I found rtree's intersection_v() return type to be elegant (a continuous array if intersecting indices plus an array of counts to figure out which index is associated to which bounding box)
I don't think it would be particularly complex; the Rust side would be just a for loop.
The only slightly complex part would be designing the right user-facing API. Presumably the result would be an Arrow ListArray, because there'll be a different number of results for each input