Upload app.py with huggingface_hub

app.py

@@ -9,8 +9,7 @@ import os
 import sys
 import pickle
 import tempfile
-import numpy as np
-import scipy.io.wavfile as wavfile
+import base64
 from huggingface_hub import hf_hub_download
 
 # Add text2midi model to path
@@ -126,7 +125,7 @@ def midi_to_wav(midi_path: str, wav_path: str, sample_rate: int = 44100) -> bool
     return os.path.exists(wav_path)
 
 def generate_music(prompt: str):
-    """Generate music from text prompt. Returns audio as (sample_rate, numpy_array)."""
+    """Generate music from text prompt. Returns audio as base64-encoded dict for API compatibility."""
     if not prompt or not prompt.strip():
         return None
 
@@ -156,14 +155,19 @@ def generate_music(prompt: str):
 
         # Convert to WAV
         if SOUNDFONT_PATH and midi_to_wav(midi_path, wav_path):
-            # Read WAV file and return as numpy array
-            sample_rate, audio_data = wavfile.read(wav_path)
-            # Convert to float32 for Gradio compatibility
-            if audio_data.dtype == np.int16:
-                audio_data = audio_data.astype(np.float32) / 32768.0
-            elif audio_data.dtype == np.int32:
-                audio_data = audio_data.astype(np.float32) / 2147483648.0
-            return (sample_rate, audio_data)
+            # Read WAV file as bytes and encode as base64
+            with open(wav_path, 'rb') as f:
+                audio_bytes = f.read()
+
+            audio_base64 = base64.b64encode(audio_bytes).decode('utf-8')
+
+            # Return as dict with base64 data - this format works with Gradio's Audio component
+            # and doesn't require file download from the client
+            return {
+                "name": "generated_audio.wav",
+                "data": f"data:audio/wav;base64,{audio_base64}",
+                "is_file": False
+            }
         else:
             return None
 
@@ -193,7 +197,7 @@ demo = gr.Interface(
         placeholder="A cheerful pop song with piano and drums in C major",
         lines=2
     ),
-    outputs=gr.Audio(label="Generated Music", type="numpy"),
+    outputs=gr.Audio(label="Generated Music", type="filepath"),
     title="VR Game Music Generator",
     description="Generate music from text descriptions using AI. Enter a prompt describing the music you want.",
     examples=[