MMD-Coder/persian-tts-piper

1import hazm2import typing3 4normalizer = hazm.Normalizer()5sent_tokenizer = hazm.SentenceTokenizer()6word_tokenizer = hazm.WordTokenizer()7 8tagger = hazm.POSTagger(9            model=str("gyroing/PersianTextCorrection_Hazm/pos_tagger.model")10        )11 12def preprocess_text(text: str) -> typing.List[typing.List[str]]:13        """Split/normalize text into sentences/words with hazm"""14        text = normalizer.normalize(text)15        processed_sentences = []16 17        for sentence in sent_tokenizer.tokenize(text):18            words = word_tokenizer.tokenize(sentence)19            processed_words = fix_words(words)20            processed_sentences.append(" ".join(processed_words))21 22        return  " ".join(processed_sentences)23    24def fix_words(words: typing.List[str]) -> typing.List[str]:25        fixed_words = []26 27        for word, pos in tagger.tag(words):28            if pos[-1] == "Z":29                if word[-1] != "ِ":30                    if (word[-1] == "ه") and (word[-2] != "ا"):31                        word += "‌ی"32                word += "ِ"33                    34 35            fixed_words.append(word)36 37        return fixed_words