calour.filtering.filter_by_data¶

calour.filtering.filter_by_data(exp: calour.experiment.Experiment, predicate, axis=1, field=None, negate=False, inplace=False, **kwargs)[source]¶

Filter samples or features by the data matrix.

Note

This function is also available as a class method Experiment.filter_by_data()

Parameters:	exp (Experiment) – Input experiment object. predicate (str or callable) – 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.
Returns:	the filtered object
Return type:	Experiment