Update app.py

app.py

@@ -1,25 +1,68 @@
 import streamlit as st
 import requests
+import pyaudio
+import wave
 import os
+import tempfile
 
 API_URL = "https://api-inference.huggingface.co/models/openai/whisper-large-v3-turbo"
-headers = {"Authorization": f"Bearer {st.secrets['HF_API_KEY']}"}
+headers = {"Authorization": f"Bearer {st.secrets['hf_token']}"}
 
 def query(file):
     data = file.read()
     response = requests.post(API_URL, headers=headers, data=data)
     return response.json()
 
+def record_audio(duration=5, sample_rate=44100, chunk=1024, channels=1):
+    p = pyaudio.PyAudio()
+    stream = p.open(format=pyaudio.paInt16,
+                    channels=channels,
+                    rate=sample_rate,
+                    input=True,
+                    frames_per_buffer=chunk)
+
+    st.info(f"Recording for {duration} seconds...")
+    frames = []
+    for i in range(0, int(sample_rate / chunk * duration)):
+        data = stream.read(chunk)
+        frames.append(data)
+    st.info("Recording finished.")
+
+    stream.stop_stream()
+    stream.close()
+    p.terminate()
+
+    return frames, sample_rate
+
 st.title("Speech Recognition with Whisper")
 
-uploaded_file = st.file_uploader("Choose an audio file", type=['wav', 'mp3', 'flac'])
+option = st.radio("Choose input method:", ('Upload File', 'Record from Microphone'))
 
-if uploaded_file is not None:
-    st.audio(uploaded_file, format='audio/wav')
-    
+if option == 'Upload File':
+    uploaded_file = st.file_uploader("Choose an audio file", type=['wav', 'mp3', 'flac'])
+    if uploaded_file is not None:
+        st.audio(uploaded_file, format='audio/wav')
+        file_to_transcribe = uploaded_file
+else:
+    duration = st.slider("Recording duration (seconds)", 1, 30, 5)
+    if st.button('Start Recording'):
+        frames, sample_rate = record_audio(duration=duration)
+        
+        with tempfile.NamedTemporaryFile(delete=False, suffix='.wav') as tmpfile:
+            wf = wave.open(tmpfile.name, 'wb')
+            wf.setnchannels(1)
+            wf.setsampwidth(pyaudio.PyAudio().get_sample_size(pyaudio.paInt16))
+            wf.setframerate(sample_rate)
+            wf.writeframes(b''.join(frames))
+            wf.close()
+            
+            st.audio(tmpfile.name, format='audio/wav')
+            file_to_transcribe = open(tmpfile.name, 'rb')
+
+if 'file_to_transcribe' in locals():
     if st.button('Transcribe'):
         with st.spinner('Transcribing...'):
-            result = query(uploaded_file)
+            result = query(file_to_transcribe)
             
             if 'text' in result:
                 st.success("Transcription completed!")
@@ -30,5 +73,8 @@ if uploaded_file is not None:
                 st.write("Error details:")
                 st.write(result)
 
+        if option == 'Record from Microphone':
+            os.unlink(file_to_transcribe.name)
+
 st.markdown("---")
 st.write("Note: This app uses the Whisper API from Hugging Face.")