Update app.py

app.py

@@ -1,6 +1,9 @@
 import streamlit as st
 import outetts
 from scipy.io.wavfile import write
+import tempfile
+import os
+from pydub import AudioSegment
 
 # Initialize model configuration
 model_config = outetts.HFModelConfig_v1(
@@ -17,15 +20,26 @@ st.write("Enter text below to generate speech.")
 
 # Sidebar for reference voice
 st.sidebar.title("Voice Cloning")
-reference_audio = st.sidebar.file_uploader("Upload a reference audio (wav)", type=["wav"])
+reference_audio = st.sidebar.file_uploader("Upload a reference audio (any format)", type=["wav", "mp3", "ogg", "flac", "m4a"])
+
+# Function to convert audio to WAV format
+def convert_to_wav(audio_file):
+    temp_audio = tempfile.NamedTemporaryFile(delete=False, suffix=".wav")
+    audio = AudioSegment.from_file(audio_file)
+    audio.export(temp_audio.name, format="wav")
+    return temp_audio.name
 
 if reference_audio:
-    ref_audio_path = "reference.wav"
-    with open(ref_audio_path, "wb") as f:
-        f.write(reference_audio.read())
+    ref_audio_path = convert_to_wav(reference_audio)
 else:
     ref_audio_path = None
 
+# Recording functionality
+if ref_audio_path is None:
+    st.sidebar.write("Or record your voice below:")
+    if st.sidebar.button("Record Voice"):
+        st.sidebar.warning("Recording functionality not implemented yet. Please upload a file.")
+
 text_input = st.text_area("Text to convert to speech:", "Hello, this is an AI-generated voice.")
 
 if st.button("Generate Speech"):
@@ -46,3 +60,7 @@ if st.button("Generate Speech"):
         # Play the audio in the Streamlit app
         st.audio(output_path, format="audio/wav")
         st.success("Speech generated successfully!")
+
+# Clean up temporary files
+if ref_audio_path:
+    os.remove(ref_audio_path)