preprocess.impute

preprocess.impute(data, how='median')

Dataset-level imputation, applied before any clock sees the matrix.

This is the coarsest of the three imputation stages and the one to reach for last. Clock-level imputation (in :func:falconage.models.linear.align) knows which features a given clock needs and can use the values that clock’s authors published; this one only knows the column.

how="none" is a no-op, and is the right choice when you want the coverage check to fail rather than to be papered over.