Update app.py

app.py

@@ -3,6 +3,21 @@ from responser import responsr
 from gtts import gTTS
 from io import BytesIO
 
+import streamlit as st
+from responser import responsr
+from gtts import gTTS
+from io import BytesIO
+import whisper
+import pyaudio
+import numpy as np
+import time
+
+# Load Whisper model
+whisper_model = whisper.load_model("base")
+
+# Initialize PyAudio
+p = pyaudio.PyAudio()
+
 # Function to convert text to speech and return audio file
 def text_to_speech(text):
     tts = gTTS(text)
@@ -11,55 +26,87 @@ def text_to_speech(text):
     audio_file.seek(0)
     return audio_file
 
+# Function to record audio from the microphone
+def record_audio(duration=5, fs=16000):
+    stream = p.open(format=pyaudio.paInt16, channels=1, rate=fs, input=True, frames_per_buffer=1024)
+    frames = []
+    for _ in range(int(fs / 1024 * duration)):
+        data = stream.read(1024)
+        frames.append(data)
+    stream.stop_stream()
+    stream.close()
+    audio_data = b''.join(frames)
+    return np.frombuffer(audio_data, dtype=np.int16)
+
+# Function to recognize speech using Whisper
+def recognize_speech(audio_data):
+    # Assuming audio_data is a numpy array of int16
+    audio_data = audio_data.astype(np.float32) / 32768.0
+    result = whisper_model.transcribe(audio_data)
+    return result['text']
+
 def main():
     # Layout with three columns
     col1, col2, col3 = st.columns([3, 1, 1])
     
     with col1:
         # Title with custom CSS styling for top margin
-        st.markdown('<div style="margin-top: -5px;margin-left: -200px;" class="title-wrapper"><h1 style="text-align: center;">ChatBot</h1></div>', unsafe_allow_html=True)
+        st.markdown('<div style="margin-top: -5px;" class="title-wrapper"><h1 style="text-align: center;">ChatBot</h1></div>', unsafe_allow_html=True)
     
     # Initialize chat history if not already initialized
     if "chat_messages" not in st.session_state:
         st.session_state.chat_messages = []
 
-    # Display chat history with audio only
+    # Display chat history with audio and text
     for message in st.session_state.chat_messages:
-        if message["role"] == "assistant":
-            st.audio(message["audio"], format="audio/mp3")
-        else:
-            st.audio(message["audio"], format="audio/mp3")
+        with col1:
+            if message["role"] == "assistant":
+                st.markdown(f"**Assistant:** {message['content']}")
+                st.audio(message["audio"], format="audio/mp3")
+            else:
+                st.markdown(f"**User:** {message['content']}")
+                st.audio(message["audio"], format="audio/mp3")
 
-    # User input
-    if prompt := st.chat_input("Welcome - How can I help you?"):
-        # Convert user input to speech (optional: if you want to hear user input)
-        user_audio = text_to_speech(prompt)
-        
-        # Add user message (as audio) to chat history
-        st.session_state.chat_messages.append({
-            "role": "user",
-            "content": prompt,
-            "audio": user_audio.getvalue()
-        })
-        
-        # Get AI response using responsr function
-        response = responsr(prompt)
+    # Button to record audio input
+    if st.button('Record Audio'):
+        st.write("Recording...")
+        audio_data = record_audio(duration=5)  # Adjust duration as needed
+        st.write("Processing...")
+        user_input = recognize_speech(audio_data)
         
-        # Convert AI response to speech
-        response_audio = text_to_speech(response)
-        
-        # Add assistant's response (as audio) to chat history
-        st.session_state.chat_messages.append({
-            "role": "assistant",
-            "content": response,
-            "audio": response_audio.getvalue()
-        })
-        
-        # Display the audio files for both user input and AI response
-        st.audio(user_audio, format="audio/mp3")
-        st.audio(response_audio, format="audio/mp3")
+        if user_input:
+            # Convert user input to speech
+            user_audio = text_to_speech(user_input)
+            
+            # Add user message (as audio) to chat history
+            st.session_state.chat_messages.append({
+                "role": "user",
+                "content": user_input,
+                "audio": user_audio.getvalue()
+            })
+            
+            # Get AI response using responsr function
+            response = responsr(user_input)
+            
+            # Convert AI response to speech
+            response_audio = text_to_speech(response)
+            
+            # Add assistant's response (as audio) to chat history
+            st.session_state.chat_messages.append({
+                "role": "assistant",
+                "content": response,
+                "audio": response_audio.getvalue()
+            })
+            
+            # Display the audio files for both user input and AI response
+            with col1:
+                st.markdown(f"**User:** {user_input}")
+                st.audio(user_audio, format="audio/mp3")
+                st.markdown(f"**Assistant:** {response}")
+                st.audio(response_audio, format="audio/mp3")
 
 if __name__ == "__main__":
     main()
 
 
+