Update demos/musicgen_app.py

demos/musicgen_app.py

@@ -29,6 +29,9 @@ from audiocraft.data.audio import audio_write
 from audiocraft.models.encodec import InterleaveStereoCompressionModel
 from audiocraft.models import MusicGen, MultiBandDiffusion
 
+from pydub import AudioSegment
+import io
+
 SECRET_TOKEN = os.getenv('SECRET_TOKEN', 'default_secret')
 
 MODEL = None  # Last used model
@@ -214,18 +217,23 @@ def predict_full(secret_token, model, model_path, decoder, text, melody, duratio
     wav_path = wavs[0]
     if USE_DIFFUSION:
         wav_path = wavs[1]
-
-
     wav_base64 = ""
     
-    # Read the content of the video file and encode it to base64
-    with open(wav_path, "rb") as wav_file:
-        wav_base64 = base64.b64encode(wav_file.read()).decode('utf-8')
 
-    # Prepend the appropriate data URI header with MIME type
-    wav_base64_data_uri = 'data:audio/wav;base64,' + wav_base64
+    # Convert WAV to MP3
+    mp3_path = wav_path.replace(".wav", ".mp3")
+    sound = AudioSegment.from_wav(wav_path)
+    sound.export(mp3_path, format="mp3")
+
+    # Encode the MP3 file to base64
+    mp3_base64 = ""
+    with open(mp3_path, "rb") as mp3_file:
+        mp3_base64 = base64.b64encode(mp3_file.read()).decode('utf-8')
+
+    # Prepend the appropriate data URI header
+    mp3_base64_data_uri = 'data:audio/mp3;base64,' + mp3_base64
     
-    return wav_base64_data_uri
+    return mp3_base64_data_uri
 
 def toggle_audio_src(choice):
     if choice == "mic":