def substitute()

in augly/text/augmenters/typo.py [0:0]

• 36 lines of code
• 8 McCabe index (conditional complexity)

    def substitute(self, data: str) -> str:
        """
        Returns text where random words are typos

        @param data: the text where the word substitutions will occur
        """
        results = []
        tokens = self.tokenizer(data)
        aug_word_cnt = self._generate_aug_cnt(
            len(tokens), self.aug_min, self.aug_max, self.aug_p
        )
        filtered_word_idxes = self.skip_aug(self.pre_skip_aug(tokens), tokens)
        aug_word_idxes = set(
            get_aug_idxes(self, tokens, filtered_word_idxes, aug_word_cnt, Method.WORD)
        )

        for t_i in range(1, self.max_typo_length + 1):
            i = 0
            while i <= len(tokens) - t_i:
                if i not in aug_word_idxes:
                    results.append(tokens[i])
                    i += 1
                    continue

                misspellings = (
                    self.model.replace(" ".join(tokens[i : i + t_i]))
                    if self.model
                    else None
                )

                if misspellings:
                    misspelling = self.sample(misspellings, 1)[0]
                    results.append(self.align_capitalization(tokens[i], misspelling))
                    i += t_i - 1
                elif len(self.augmenters) > 0:
                    aug = self.sample(self.augmenters, 1)[0]
                    new_token = aug.augment(tokens[i])
                    results.append(self.align_capitalization(tokens[i], new_token))
                else:
                    # If no misspelling is found in the dict & no other typo types are being
                    # used, don't change the token
                    results.append(tokens[i])
                i += 1

            if t_i > 1:
                results.extend(tokens[-t_i + 1 :])

        return detokenize(results)