@jeffkaufman, I'm jotting down a thought so I don't forget, but this is not urgent.
Our current interface for getting predictors has evolved to be somewhat convoluted. I think it could be nice to refactor so that the return type of the estimate_*() functions was dict[TaxID, list[Predictor]]. I think it makes sense that the predictors always be grouped by TaxID, since there's not really a uses for them ungrouped. I also don't think we need the keys to be frozenset anymore because it seems like we should be avoiding pathogens that are not clades.
@jeffkaufman, I'm jotting down a thought so I don't forget, but this is not urgent.
Our current interface for getting predictors has evolved to be somewhat convoluted. I think it could be nice to refactor so that the return type of the
estimate_*()
functions wasdict[TaxID, list[Predictor]]
. I think it makes sense that the predictors always be grouped byTaxID
, since there's not really a uses for them ungrouped. I also don't think we need the keys to befrozenset
anymore because it seems like we should be avoiding pathogens that are not clades.