preprocess.mask_report

preprocess.mask_report(
    platform,
    *,
    clocks='scoreable',
    kind='general',
    registry=None,
)

What applying this mask would cost each clock.

One row per clock: how many of its probes the mask flags, and – the number that decides it – what share of the model’s total |coefficient| those probes carry. Two clocks losing the same probe count are not in the same position if one of them leans on the probes it is losing.

Run this before :func:apply_mask, not after.