tests/test_multilingual.py

import re
from indextts.utils.front import TextNormalizer
from indextts.text.hindi_phonemizer import hindi_to_phoneme


def test_hindi_not_through_english_phonemizer():
    txt = "मुझे बहुत खुशी हो रही है क्योंकि आज मेरा सपना पूरा हो गया।"
    tn = TextNormalizer()
    tn.load()
    normalized = tn.normalize(txt)
    # The normalized output should contain Devanagari characters (not Latin transliteration)
    assert re.search(r"[\u0900-\u097F]", normalized), "Normalized Hindi should contain Devanagari script"

    # It should be equal to the joined hindi_to_phoneme tokens
    tokens = hindi_to_phoneme(txt)
    joined = " ".join(tokens)
    assert normalized == joined or joined in normalized, "Hindi must be routed to hindi_to_phoneme tokens"


def test_english_and_chinese_preserved():
    tn = TextNormalizer()
    tn.load()
    en = "This is a test of the emergency broadcast system."
    zh = "我爱你。"
    norm_en = tn.normalize(en)
    norm_zh = tn.normalize(zh)
    # English should contain ASCII letters
    assert re.search(r"[A-Za-z]", norm_en)
    # Chinese should contain CJK chars
    assert re.search(r"[\u4e00-\u9fff]", norm_zh)