Update app.py

app.py

@@ -1,22 +1,27 @@
-import gradio as gr
 import os
 import requests
 import torch
 import zipfile
 from TTS.api import TTS
 from pydub import AudioSegment
+import gradio as gr
 
+# Environment variable
 os.environ["COQUI_TOS_AGREED"] = "1"
 
+# Constants
 MODEL_PATH = "tts_models/multilingual/multi-dataset/xtts_v2"
 LANGUAGES = ["en", "es", "fr", "de", "it", "pt", "pl", "tr", "ru", "nl", "cs", "ar", "zh-cn", "ja", "hu", "ko", "hi"]
 AUDIO_FORMATS = [".wav", ".mp3", ".flac", ".mp4"]
 
+# Device setup
 device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 print(f"Using device: {device}")
 
+# Initialize TTS model
 tts = TTS(MODEL_PATH).to(device)
 
+# Function to download audio file
 def download_audio_file(url):
  try:
  response = requests.get(url)
@@ -29,6 +34,7 @@ def download_audio_file(url):
  print(f"Error downloading audio file: {e}")
  return None
 
+# Function to extract zip file
 def extract_zip_file(zip_file):
  try:
  with zipfile.ZipFile(zip_file, 'r') as zip_ref:
@@ -38,6 +44,7 @@ def extract_zip_file(zip_file):
  print(f"Error extracting zip file: {e}")
  return False
 
+# Function to convert audio file to WAV
 def convert_to_wav(input_audio_file):
  file_extension = os.path.splitext(input_audio_file)[-1].lower()
  if file_extension!= ".wav":
@@ -47,11 +54,13 @@ def convert_to_wav(input_audio_file):
  return "temp.wav"
  return input_audio_file
 
+# Function to synthesize text
 def synthesize_text(text, input_audio_file, language):
  input_audio_file = convert_to_wav(input_audio_file)
  tts.tts_to_file(text=text, speaker_wav=input_audio_file, language=language, file_path="./output.wav")
  return "./output.wav"
 
+# Function to clone audio
 def clone(text, input_file, language, url=None, use_url=False):
  if use_url:
  if url is None:
@@ -72,9 +81,15 @@ def clone(text, input_file, language, url=None, use_url=False):
  else:
  input_audio_file = input_file.name
 
- output_file_path = synthesize_text(text, input_audio_file, language)
+ # Cache audio file to avoid re-downloading or re-extracting
+ audio_cache = {}
+ if input_audio_file not in audio_cache:
+ audio_cache[input_audio_file] = convert_to_wav(input_audio_file)
+
+ output_file_path = synthesize_text(text, audio_cache[input_audio_file], language)
  return output_file_path
 
+# Gradio interface
 iface = gr.Interface(
  fn=clone,
  inputs=["text", gr.File(label="Input File", file_types=[".zip", *AUDIO_FORMATS]), gr.Dropdown(choices=LANGUAGES, label="Language"), gr.Text(label="URL"), gr.Checkbox(label="Use URL", value=False)],