Pattern 

Pattern(pattern)

Class for representing a pattern (a regular expression).

Initialize a Pattern.

Parameters:

  • pattern (str | Pattern) –

    Str or Pattern object representing a regular expressions. It must be compatible with re.compile method.

filter 

filter(strings)

Return a boolean mask for strings matching the pattern.

Parameters:

  • strings (list[str]) –

    Strings to filter.

Returns:

  • matches –

    Numpy array of bool type.

match 

match(string)

Verify if the pattern matches the string (at any place).

scores 

scores(strings, labels, full_labels_list=None)

Calculate classification quality scores for pattern's matches and the labels.

Pattern's matches are considered as "predictions" when calculating quality metrics.

Parameters:

  • strings (list[str]) –

    A list of strings.

  • labels (list[int]) –

    A list of strings labels.

  • full_labels_list (list[int], default: None ) –

    A full list of unique labels that can be present among labels.

Returns:

  • response ( dict ) –

    Contains keys: 'n_strings', 'total_positive', 'n_matches', 'precision', 'recall' and 'accuracy'. If there are more than 2 classes, each key is a list corresponding to each class.