Ideally, the divergence cache would allow for storing arbitrary subsets in arbitrary order based on names. #24 might help with this.
Probably hdf5 isn't really the right format for this, but pulling in a SciDB or something dependency is probably unreasonable....
Ideally, the divergence cache would allow for storing arbitrary subsets in arbitrary order based on names. #24 might help with this.
Probably hdf5 isn't really the right format for this, but pulling in a SciDB or something dependency is probably unreasonable....