nltk_comtrans.py 文件源码

python
阅读 18 收藏 0 点赞 0 评论 0

项目:master-thesis 作者: AndreasMadsen 项目源码 文件源码
def __iter__(self) -> Iterator[Tuple[str, str]]:

        with NLTKEnv() as nltk_env:
            nltk_env.download('perluniprops')
            nltk_env.download('comtrans')

            from nltk.corpus import comtrans
            from nltk.tokenize.moses import MosesDetokenizer

            als = comtrans.aligned_sents(self._comtrans_string())

            source_detokenizer = MosesDetokenizer(lang=self._source_lang)
            target_detokenizer = MosesDetokenizer(lang=self._target_lang)

        for source, target in self._comtrans_maybe_swap(als):
            source = source_detokenizer.detokenize(source, return_str=True)
            target = target_detokenizer.detokenize(target, return_str=True)

            if self._length_checker(source, target):
                yield (source, target)
评论列表
文章目录


问题


面经


文章

微信
公众号

扫码关注公众号