PR #358 adds core scores, and downstream workflows need to pass relative-frequency-style values as metadata input so features can be filtered by core score.
filter-features currently only accepts FeatureTable[Frequency] and uses min_frequency / max_frequency integer thresholds. This should also support FeatureTable[RelativeFrequency], where the corresponding thresholds are fractions from 0 to 1.
Proposed change:
- Allow
filter-features to accept both FeatureTable[Frequency] and FeatureTable[RelativeFrequency].
- Rename
min_frequency / max_frequency to min_abundance / max_abundance.
- Use a TypeMap so frequency tables allow whole-number thresholds from 0 to infinity, while relative-frequency tables allow fractional thresholds from 0 to 1.
- Update the private filtering helper names to reflect abundance/fraction semantics.
- Add a regression test covering fractional filtering on a relative-frequency table.
PR #358 adds core scores, and downstream workflows need to pass relative-frequency-style values as metadata input so features can be filtered by core score.
filter-featurescurrently only acceptsFeatureTable[Frequency]and usesmin_frequency/max_frequencyinteger thresholds. This should also supportFeatureTable[RelativeFrequency], where the corresponding thresholds are fractions from 0 to 1.Proposed change:
filter-featuresto accept bothFeatureTable[Frequency]andFeatureTable[RelativeFrequency].min_frequency/max_frequencytomin_abundance/max_abundance.