Update app.py

app.py

@@ -1,6 +1,5 @@
-import gradio as gr
 import os
-import requests
+import gradio as gr
 import torch
 import zipfile
 from TTS.api import TTS
@@ -15,75 +14,39 @@ device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
 print(f"Using device: {device}")
 
 # TTS model setup
+os.environ["COQUI_TOS_AGREED"] = "1"
 MODEL_PATH = "tts_models/multilingual/multi-dataset/xtts_v2"
 tts = TTS(MODEL_PATH).to(device)
 
-class AudioProcessor:
-    def __init__(self):
-        pass
-
-    def convert_to_wav(self, input_audio_file: str) -> str:
-        file_extension = os.path.splitext(input_audio_file)[-1].lower()
-        if file_extension!= ".wav":
-            audio = AudioSegment.from_file(input_audio_file)
-            audio.export("temp.wav", format="wav")
-            os.remove(input_audio_file)
-            return "temp.wav"
-        return input_audio_file
+def generate_audio(text, language, speed, pitch, volume):
+    # Prepare input
+    input_text = {"text": text, "language": language}
+    tts.prepare_input(input_text)
 
-    def synthesize_text(self, text: str, input_audio_file: str, language: str) -> str:
-        input_audio_file = self.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"
+    # Generate audio
+    audio = tts.generate_audio(input_text, speed=speed, pitch=pitch, volume=volume)
 
-def download_audio_file(url: str) -> str:
-    try:
-        response = requests.get(url)
-        file_extension = os.path.splitext(url)[-1].lower()
-        if file_extension not in AUDIO_FORMATS:
-            raise ValueError(f"Unsupported file extension: {file_extension}")
-        file_name = f"temp{file_extension}"
-        with open(file_name, "wb") as f:
-            f.write(response.content)
-        return file_name
-    except requests.exceptions.RequestException as e:
-        print(f"Error downloading audio file: {e}")
-        return None
+    # Save audio
+    audio_path = "output.wav"
+    tts.save_audio(audio_path, audio)
 
-def extract_zip_file(zip_file: str) -> bool:
-    try:
-        with zipfile.ZipFile(zip_file, 'r') as zip_ref:
-            zip_ref.extractall()
-        return True
-    except zipfile.BadZipfile as e:
-        print(f"Error extracting zip file: {e}")
-        return False
+    # Convert to mp3
+    audio_segment = AudioSegment.from_wav(audio_path)
+    audio_segment.export(audio_path[:-4] + ".mp3", format="mp3")
 
-def synthesize_audio(text: str, input_file: gr.File, language: str) -> str:
-    audio_processor = AudioProcessor()
-    if input_file is None:
-        return None
-    if input_file.name.endswith(".zip"):
-        if extract_zip_file(input_file):
-            input_audio_file = [f for f in os.listdir('.') if os.path.isfile(f) and f.endswith(tuple(AUDIO_FORMATS))]
-            if len(input_audio_file) == 1:
-                input_audio_file = input_audio_file[0]
-            else:
-                return "Error: Please select a single audio file from the extracted files."
-    else:
-        input_audio_file = input_file.name
-    output_file_path = audio_processor.synthesize_text(text, input_audio_file, language)
-    return output_file_path
+    # Return audio path
+    return audio_path[:-4] + ".mp3"
 
 iface = gr.Interface(
-    fn=synthesize_audio,
-    inputs=["text", gr.File(label="Input File", file_types=[".zip", *AUDIO_FORMATS]), gr.Dropdown(choices=LANGUAGES, label="Language")],
-    outputs=gr.Audio(type='filepath'),
-    title='Voice Clone',
-    description=""" by [Angetyde](https://youtube.com/@Angetyde?si=7nusP31nTumIkPTF) and [Tony Assi](https://www.tonyassi.com/ ) use this colab with caution <3. Clone any voice with a model and generate a speech waveform.""",
-    examples=[["Hello! My name is Voice Clone. What is your name?", None, "en"]],
-    height=600,
-    width=1200,
+    generate_audio,
+    inputs=["text", "language", "speed", "pitch", "volume"],
+    outputs="audio",
+    audio=True,
+    audio_format="mp3",
+    title="Text-to-Speech",
+    description="Convert text to speech in multiple languages.",
+    allow_flagging=False,
+    cache_examples=False,
 )
 
 iface.launch()