Update app.py

app.py

@@ -1,73 +1,94 @@
 import streamlit as st
 import requests
 import base64
+import json
+from io import BytesIO
+import numpy as np
+from audio_recorder_streamlit import audio_recorder
 
-API_URL = "https://startrz-proagents.hf.space/api/v1/prediction/9ad2e52d-052f-44e5-b3a0-7fd432d86506"
+def convert_audio_to_base64(audio_bytes):
+    """Convert audio bytes to base64 string"""
+    if audio_bytes is None:
+        return None
+    base64_str = base64.b64encode(audio_bytes).decode('utf-8')
+    return f"data:audio/webm;codecs=opus;base64,{base64_str}"
 
-def query(payload):
-    response = requests.post(API_URL, json=payload)
-    return response.json()
-
-def main():
-    st.title("Audio and Image Upload for Prediction")
-
-    # File uploader for audio files
-    audio_file = st.file_uploader("Upload an audio file (e.g., mp3, wav, ogg, webm)", type=["mp3", "wav", "ogg", "webm", "m4a", "aac"])
+def query_prediction_api(audio_base64):
+    """Send request to prediction API"""
+    url = "http://localhost:3000/api/v1/prediction/<chatlfowid>"
     
-    # File uploader for image files
-    image_file = st.file_uploader("Upload an image file (e.g., png, jpg, jpeg)", type=["png", "jpg", "jpeg"])
-
-    # Create payload for the query
-    payload = {}
-
-    if audio_file is not None:
-        # Convert the audio file to a Base64 string
-        audio_bytes = audio_file.read()
-        audio_base64 = base64.b64encode(audio_bytes).decode('utf-8')
-        audio_mime_type = audio_file.type  # Get the mime type
-
-        # Add audio data to the payload
-        payload["uploads"] = [
+    payload = {
+        "uploads": [
             {
-                "data": f'data:{audio_mime_type};base64,{audio_base64}',
-                "type": 'file',
-                "name": audio_file.name,
-                "mime": audio_mime_type
+                "data": audio_base64,
+                "type": "audio",
+                "name": "audio.wav",
+                "mime": "audio/webm"
             }
         ]
-
-    if image_file is not None:
-        # Convert the image file to a Base64 string
-        image_bytes = image_file.read()
-        image_base64 = base64.b64encode(image_bytes).decode('utf-8')
-        image_mime_type = image_file.type  # Get the mime type
-
-        # Add image data to the payload
-        payload.setdefault("uploads", []).append(
-            {
-                "data": f'data:{image_mime_type};base64,{image_base64}',
-                "type": 'file',
-                "name": image_file.name,
-                "mime": image_mime_type
-            }
-        )
+    }
     
-    # Add a question to the payload with a unique key
-    question = st.text_input("Enter your question", "Can you describe the image?", key="question_input")
+    try:
+        response = requests.post(
+            url,
+            headers={"Content-Type": "application/json"},
+            json=payload
+        )
+        response.raise_for_status()
+        return response.json()
+    except requests.exceptions.RequestException as e:
+        st.error(f"Error making prediction: {str(e)}")
+        return None
 
-    # Add the question to the payload if provided
-    if question:
-        payload["question"] = question
+def main():
+    st.title("Audio Prediction App")
+    st.write("Record or upload audio to get predictions")
+    
+    # Add tabs for different input methods
+    tab1, tab2 = st.tabs(["Record Audio", "Upload Audio"])
+    
+    with tab1:
+        st.write("Click the button below to record audio")
+        audio_bytes = audio_recorder()
+        
+        if audio_bytes:
+            st.audio(audio_bytes, format="audio/webm")
+            
+            if st.button("Get Prediction for Recorded Audio"):
+                with st.spinner("Getting prediction..."):
+                    audio_base64 = convert_audio_to_base64(audio_bytes)
+                    if audio_base64:
+                        prediction = query_prediction_api(audio_base64)
+                        if prediction:
+                            st.json(prediction)
+    
+    with tab2:
+        uploaded_file = st.file_uploader("Choose an audio file", type=['wav', 'mp3', 'webm'])
+        
+        if uploaded_file:
+            audio_bytes = uploaded_file.read()
+            st.audio(audio_bytes)
+            
+            if st.button("Get Prediction for Uploaded Audio"):
+                with st.spinner("Getting prediction..."):
+                    audio_base64 = convert_audio_to_base64(audio_bytes)
+                    if audio_base64:
+                        prediction = query_prediction_api(audio_base64)
+                        if prediction:
+                            st.json(prediction)
 
-    # Send the request to the API and get the output
-    if st.button("Submit"):
-        if payload:
-            with st.spinner("Sending request..."):
-                output = query(payload)
-                st.success("Request sent successfully!")
-                st.json(output)  # Display the API response
-        else:
-            st.error("Please upload at least one audio or image file.")
+    st.sidebar.markdown("""
+    ### About
+    This app allows you to:
+    - Record audio directly in the browser
+    - Upload audio files
+    - Get predictions from the API
+    
+    Supported formats:
+    - WAV
+    - MP3
+    - WebM
+    """)
 
 if __name__ == "__main__":
     main()