The callable accepts a list of numeric and return a bool. Alternatively
it also accepts the following strings to filter along the specified axis:
’abundance’: calls is_abundant(), filter by abundance;
’prevalence’: calls is_prevalent(), filter by prevalence;
’freq_ratio’: calls freq_ratio(), filter if there is a dominant unique value;
’unique_cut’: calls unique_cut(), filter by how diversified the values.
axis (0, 1, 's', or 'f', optional) – Apply predicate on each row (ie samples) (0, ‘s’) or each column (ie features) (1, ‘f’)
field (str or None, optional) – The column in the sample_metadata (or feature_metadata,
depending on axis). If it is None, the predicate
operates on the whole data set; if it is not None, the data
set is divided into groups according to the sample_metadata
(feature_metadata) column and the predicate operates on each
partition of data - only the features (or samples) that fail
to pass every partition will be filtered away.
negate (bool) – negate the predicate for selection
kwargs (dict) – keyword argument passing to predicate function.
