fix(tts): text_to_speech.py __main__ block passes a max_accelerate_audio kwarg that no longer exists

[Anai-Guo]

Problem

Running soni_translate/text_to_speech.py directly fails immediately:

if __name__ == "__main__":
    from segments import result_diarize

    audio_segmentation_to_voice(
        result_diarize,
        TRANSLATE_AUDIO_TO="en",
        max_accelerate_audio=2.1,        # <-- not a parameter of this function
        is_gui=True,
        ...
    )

TypeError: audio_segmentation_to_voice() got an unexpected keyword argument 'max_accelerate_audio'

Why the kwarg is the stale side

max_accelerate_audio is a parameter of accelerate_segments() (text_to_speech.py:1089-1091,
used at :1162-1163 to clamp acc_percentage), not of audio_segmentation_to_voice()
(:969, whose parameters are result_diarize, TRANSLATE_AUDIO_TO, is_gui, tts_voice00..11, dereverb_automatic, model_id_bark, model_id_coqui, delete_previous_automatic).

The pipeline in app_rvc.py reflects that split: it passes max_accelerate_audio to
accelerate_segments (app_rvc.py:989-994) and calls audio_segmentation_to_voice
positionally with only the 15 arguments it actually declares (app_rvc.py:955-970). The
__main__ block is the only call site that still hands acceleration to the TTS function.

Fix

Delete that one argument. Nothing else in the block changes, and no non-__main__ code path
is touched.

Verification

Extracted the real signature with ast and replayed both call sites with
inspect.Signature.bind:

signature: (result_diarize, TRANSLATE_AUDIO_TO, is_gui, tts_voice00, tts_voice01=None, ...,
            dereverb_automatic=None, model_id_bark=None, model_id_coqui=None,
            delete_previous_automatic=None)

--- __main__ block as written (text_to_speech.py:1566) ---
TypeError: got an unexpected keyword argument 'max_accelerate_audio'

--- __main__ block patched ---
binds OK

--- control group: app_rvc.py:955 (15 positional) ---
binds OK

🤖 Generated with Claude Code