openvoice2 / openvoice_app.py
poemsforaphrodite's picture
Update openvoice_app.py
7233731 verified
raw
history blame
5.02 kB
import os
import torch
import argparse
import gradio as gr
import requests
from openvoice import se_extractor
from openvoice.api import BaseSpeakerTTS, ToneColorConverter
from dotenv import load_dotenv
from openai import OpenAI
from elevenlabs.client import ElevenLabs
from elevenlabs import play, save
from flask import Flask
from flask_limiter import Limiter
from flask_limiter.util import get_remote_address
# Load environment variables
load_dotenv()
# Initialize Flask app
app = Flask(__name__)
# Setup Limiter for rate limiting and quota management based on IP address
limiter = Limiter(get_remote_address, app=app, default_limits=["5 per minute"])
# Argument parsing
parser = argparse.ArgumentParser()
parser.add_argument("--share", action='store_true', default=False, help="make link public")
args = parser.parse_args()
# Initialize ElevenLabs client
client = ElevenLabs(api_key=os.environ.get("ELEVENLABS_API_KEY"))
device = 'cuda' if torch.cuda.is_available() else 'cpu'
output_dir = 'outputs'
os.makedirs(output_dir, exist_ok=True)
api_key = os.environ.get("ELEVENLABS_API_KEY")
supported_languages = ['zh', 'en']
# Function to get all voices
def get_voices(api_key):
url = "https://api.elevenlabs.io/v1/voices"
headers = {"xi-api-key": api_key}
response = requests.request("GET", url, headers=headers)
return response.json()
# Function to delete a voice by ID
def delete_voice(api_key, voice_id):
url = f"https://api.elevenlabs.io/v1/voices/{voice_id}"
headers = {"xi-api-key": api_key}
response = requests.request("DELETE", url, headers=headers)
return response.status_code, response.text
# Predict function with rate limiting based on IP address
#@limiter.limit("100 per minute")
def predict(prompt, style, audio_file_pth, voice_name):
text_hint = ''
if len(prompt) < 2:
text_hint += "[ERROR] Please provide a longer prompt text.\n"
return text_hint, None, None
if len(prompt) > 200:
text_hint += "[ERROR] Text length limited to 200 characters. Please try shorter text.\n"
return text_hint, None, None
print(audio_file_pth)
voice = client.clone(
name=voice_name,
description="A trial voice model for testing",
files=[audio_file_pth],
)
# Generate audio from text
audio = client.generate(text=prompt, voice=voice)
save(audio, f'{output_dir}/output.wav')
save_path = f'{output_dir}/output.wav'
data = get_voices(api_key)
# Find all voice IDs with the name provided by the user
trial_voice_ids = [voice.get("voice_id") for voice in data['voices'] if voice.get("name") == voice_name]
# # Delete each voice with the name provided by the user
# for voice_id in trial_voice_ids:
# status_code, response_text = delete_voice(api_key, voice_id)
# print(f"Deleted voice ID {voice_id}: Status Code {status_code}, Response {response_text}")
# if not trial_voice_ids:
# print("No voices with the name provided by the user found.")
return text_hint, save_path, audio_file_pth
# Gradio interface setup
with gr.Blocks(gr.themes.Glass()) as demo:
with gr.Row():
with gr.Column():
input_text_gr = gr.Textbox(
label="Create this",
info="One or two sentences at a time is better. Up to 200 text characters.",
value="He hoped there would be stew for dinner, turnips and carrots and bruised potatoes and fat mutton pieces to be ladled out in thick, peppered, flour-fattened sauce.",
)
style_gr = gr.Dropdown(
label="Style",
choices=['default', 'whispering', 'cheerful', 'terrified', 'angry', 'sad', 'friendly'],
info="Please upload a reference audio file that is at least 1 minute long. For best results, ensure the audio is clear. You can use Adobe Podcast Enhance(https://podcast.adobe.com/enhance) to improve the audio quality before uploading.",
max_choices=1,
value="default",
)
ref_gr = gr.Audio(
label="Original audio",
type="filepath",
sources=["upload"], # Allow only upload
)
voice_name_gr = gr.Textbox(
label="Your name and Product you bought",
value="Sam"
)
tts_button = gr.Button("Send", elem_id="send-btn", visible=True)
with gr.Column():
out_text_gr = gr.Text(label="Info")
audio_gr = gr.Audio(label="Replicated Sound", autoplay=True)
ref_audio_gr = gr.Audio(label="Original Audio Used")
tts_button.click(predict, [input_text_gr, style_gr, ref_gr, voice_name_gr], outputs=[out_text_gr, audio_gr, ref_audio_gr])
demo.queue()
demo.launch(debug=True, show_api=False, share=args.share)
# Hide Gradio footer and record button
css = """
footer {visibility: hidden}
audio .btn-container {display: none}
"""
demo.add_css(css)