Update app.py

app.py

@@ -91,10 +91,8 @@ class VoicesResponse(BaseModel):
 async def generate_speech_async(text: str, voice: str, pitch: str, rate: str, output_file: str) -> bool:
     """Generate speech asynchronously"""
     try:
-        # Create SSML with pitch and rate adjustments
-        # ssml_text = f'<speak><prosody pitch="{pitch}" rate="{rate}">{text}</prosody></speak>'
-        
-        communicate = edge_tts.Communicate(text = text, voice = voice, rate = rate, pitch = pitch)
+        # Use edge_tts.Communicate with direct parameters (no SSML needed)
+        communicate = edge_tts.Communicate(text=text, voice=voice, rate=rate, pitch=pitch)
         await communicate.save(output_file)
         return True
     except Exception as e:
@@ -170,12 +168,12 @@ async def synthesize_speech(request: TTSRequest):
         if not os.path.exists(output_file):
             raise HTTPException(status_code=500, detail="Audio file was not generated")
         
-        # Return the audio file
+        # Return the audio file directly
         return FileResponse(
             output_file,
             media_type="audio/mpeg",
             filename=f"speech_{file_id}.mp3",
-            background=cleanup_file(output_file)  # Cleanup after response
+            background=None  # Don't cleanup immediately, let the response complete first
         )
         
     except HTTPException:
@@ -253,10 +251,15 @@ if __name__ == "__main__":
     print("API Documentation will be available at: http://localhost:7860/")
     print("Health check: http://localhost:7860/health")
     print("Available voices: http://localhost:7860/voices")
-    print("\nExample usage:")
+    print("\nExample usage (saves audio file locally):")
     print("curl -X POST 'http://localhost:7860/synthesize' \\")
     print("  -H 'Content-Type: application/json' \\")
     print("  -d '{\"text\":\"Hello from Hugging Face!\",\"voice\":\"en-GB-SoniaNeural\",\"pitch\":\"-10Hz\",\"rate\":\"+15%\"}' \\")
     print("  --output speech.mp3")
+    print("\nFor your deployed space:")
+    print("curl -X POST 'https://nitinbot001-tts-api.hf.space/synthesize' \\")
+    print("  -H 'Content-Type: application/json' \\")
+    print("  -d '{\"text\":\"hello my name is nitin\",\"voice\":\"en-US-AriaNeural\",\"pitch\":\"+0Hz\",\"rate\":\"+0%\"}' \\")
+    print("  --output speech.mp3")
     
     uvicorn.run(app, host="0.0.0.0", port=7860)