Method __init__

utils/eval_util.py:13–21

Args: annotators: None or empty set (only tokenizes).

(self)

    NON_WS = r'[^\p{Z}\p{C}]'

    def __init__(self):
        """
        Args:
            annotators: None or empty set (only tokenizes).
        """
        self._regexp = regex.compile(
            '(%s)|(%s)' % (self.ALPHA_NUM, self.NON_WS),
            flags=regex.IGNORECASE + regex.UNICODE + regex.MULTILINE
        )

    def tokenize(self, text, uncased=False):
        matches = [m for m in self._regexp.finditer(text)]