Update app.py

app.py

@@ -1,36 +1,39 @@
-import gradio as gr
+from fastapi import FastAPI, HTTPException
+from pydantic import BaseModel
 from TTS.api import TTS
+import torch
+
+app = FastAPI()
 
 # Initialize the TTS model
 tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2")
+device = "cuda" if torch.cuda.is_available() else "cpu"
+tts.to(device)
 
 # Get available voices and languages
 available_voices = tts.list_speakers()
 available_languages = tts.list_languages()
 
-def generate_speech(text, voice, language):
-    # Generate speech
-    output_path = "output.wav"
-    tts.tts_to_file(text=text, speaker_wav=voice, language=language, file_path=output_path)
-    return output_path
-
-# Create Gradio interface
-with gr.Blocks() as demo:
-    gr.Markdown("# XTTS v2 Text-to-Speech Generator")
+class TextToSpeechRequest(BaseModel):
+    text: str
+    voice: str
+    language: str
 
-    with gr.Row():
-        text_input = gr.Textbox(label="Input Text", placeholder="Enter text to be synthesized")
+@app.get("/voices")
+def get_voices():
+    return {"voices": available_voices}
 
-    with gr.Row():
-        voice_dropdown = gr.Dropdown(label="Select Voice", choices=available_voices)
-        language_dropdown = gr.Dropdown(label="Select Language", choices=available_languages)
+@app.get("/languages")
+def get_languages():
+    return {"languages": available_languages}
 
-    with gr.Row():
-        generate_button = gr.Button("Generate Speech")
+@app.post("/generate")
+def generate_speech(request: TextToSpeechRequest):
+    if request.voice not in available_voices:
+        raise HTTPException(status_code=400, detail="Invalid voice selected")
+    if request.language not in available_languages:
+        raise HTTPException(status_code=400, detail="Invalid language selected")
 
-    with gr.Row():
-        audio_output = gr.Audio(label="Generated Speech")
-
-    generate_button.click(generate_speech, inputs=[text_input, voice_dropdown, language_dropdown], outputs=audio_output)
-
-demo.launch(debug=True)
+    output_path = "output.wav"
+    tts.tts_to_file(text=request.text, speaker=request.voice, language=request.language, file_path=output_path)
+    return {"message": "Speech generated successfully", "file_path": output_path}