Update text_to_speech.py

[SouljaVR]

adds improved natural splitting for non-english languages + newline counts as a chunk split too. has additional measures to ensure chunks can never exceed 300 bytes for the tiktok api.

[eugenepozniak]

@mark-rez hallo! du kannst versichert das audio teil ist generiert durch:
# Function to generate audio for each text chunk
def generate_audio_chunk(index: int, text_chunk: str):
counter = 0
while counter < tries:
counter += 1