can you explain what weighted overlap is?
This is usually called “conservative regridding”, and indeed computes weights according to the contribution based on the area-of-overlap (at least for first-order conservative, I didn’t look at higher-order conservative). You can read more about this in Conservative Region Aggregation with Xarray, Geopandas and Sparse
Automatic vector quantity treatment would be great, since that’s somewhat complex. According to Regridding Overview | Climate Data Guide this can be done by converting to vorticity / divergence, regridding, then deriving the vectors again.