Create rvc_service.py

rvc_service.py

@@ -0,0 +1,90 @@
+import gradio as gr
+from rvc_service import RVCService
+import asyncio
+import logging
+import numpy as np
+from scipy.io import wavfile
+import os
+
+# Initialize logging
+logging.basicConfig(level=logging.DEBUG)
+logger = logging.getLogger(__name__)
+
+# Initialize RVC Service
+rvc_service = RVCService()
+
+async def convert_tts(model_name, audio_file, slang_rate):
+    try:
+        logger.debug(f"Received request - model: {model_name}, audio: {type(audio_file)}, slang: {slang_rate}")
+        
+        if audio_file is None:
+            logger.error("No audio file provided")
+            return {"error": "No audio file uploaded."}, None
+
+        # Load and preprocess audio
+        if hasattr(audio_file, 'name'):
+            logger.debug(f"Audio file name: {audio_file.name}")
+            # Load audio file
+            sr, audio = wavfile.read(audio_file.name)
+            # Convert to mono if stereo
+            if len(audio.shape) > 1:
+                audio = np.mean(audio, axis=1)
+            
+            # Submit job to RVC service
+            job_id = await rvc_service.submit_job(
+                audio_data=audio,
+                model_name=model_name,
+                priority=1  # Default priority
+            )
+            
+            if not job_id:
+                return {"error": "Service queue is full"}, None
+                
+            # Wait for result (you might want to implement a better waiting mechanism)
+            for _ in range(30):  # Maximum 30 seconds wait
+                if job_id in rvc_service.job_queue.processing:
+                    await asyncio.sleep(1)
+                    continue
+                    
+                # Check if job completed successfully
+                output_path = f"outputs/output_{job_id}.wav"
+                if os.path.exists(output_path):
+                    return {"info": f"Processed with job ID: {job_id}"}, output_path
+                    
+            return {"error": "Processing timeout"}, None
+            
+        return {"error": "Invalid audio file"}, None
+
+    except Exception as e:
+        logger.error(f"Error in convert_tts: {str(e)}", exc_info=True)
+        return {"error": str(e)}, None
+
+# Modified interface with queue settings for better concurrency
+iface = gr.Interface(
+    fn=convert_tts,
+    inputs=[
+        gr.Dropdown(choices=get_model_names(), label="Model", interactive=True),
+        gr.Audio(label="Upload Audio", type="filepath"),
+        gr.Slider(minimum=0, maximum=1, step=0.01, label="Slang Rate"),
+    ],
+    outputs=[
+        gr.JSON(label="Info"),
+        gr.Audio(label="Converted Audio")
+    ],
+    title="Voice Conversion",
+    concurrency_limit=5,  # Limit concurrent requests
+    batch=False,  # Process requests individually
+    max_batch_size=1,
+).queue()
+
+if __name__ == "__main__":
+    # Start RVC service
+    rvc_service.start()
+    
+    # Launch Gradio interface
+    iface.launch(
+        debug=True,
+        show_error=True,
+        max_threads=10,
+        share=False
+    )