Now that we've better nailed down the data structures we'll be dealing with.
Completely removes the FrequencyTable thing and now household_weights works with just DataFrames and Series as inputs, things that should be more readily available from earlier parts of the pipeline.
Coverage increased (+0.23%) when pulling 4984cf58d4fe0583eba06f3534b340d91f25f49e on ipu-rehash into 3949be049ddb0c62c06a343588d7ebc98d9c5fc0 on master.
Now that we've better nailed down the data structures we'll be dealing with.
Completely removes the
FrequencyTable
thing and nowhousehold_weights
works with justDataFrames
andSeries
as inputs, things that should be more readily available from earlier parts of the pipeline.