Update app.py

app.py

@@ -1,5 +1,4 @@
 import streamlit as st
-import streamlit.components.v1 as components
 import anthropic
 import openai
 import base64
@@ -14,6 +13,7 @@ import pytz
 import random
 import re
 import requests
+import streamlit.components.v1 as components
 import textract
 import time
 import zipfile
@@ -31,7 +31,6 @@ from xml.etree import ElementTree as ET
 from openai import OpenAI
 import extra_streamlit_components as stx
 from streamlit.runtime.scriptrunner import get_script_run_ctx
-import extra_streamlit_components as stx
 
 
 # 1. 🚲BikeAI🏆 Configuration and Setup
@@ -53,146 +52,6 @@ st.set_page_config(
     }
 )
 
-
-
-def create_speech_component():
-    """Create speech recognition component using postMessage for communication."""
-    
-    speech_recognition_html = """
-    <div style="padding: 20px;">
-        <div class="controls">
-            <button id="start">Start Listening</button>
-            <button id="stop" disabled>Stop Listening</button>
-            <button id="clear">Clear Text</button>
-        </div>
-        <div id="status" style="margin: 10px 0; padding: 10px; background: #e8f5e9;">Ready</div>
-        <div id="output" style="white-space: pre-wrap; padding: 15px; background: #f5f5f5; min-height: 100px; max-height: 400px; overflow-y: auto;"></div>
-        <div id="debug" style="margin-top: 10px; color: #666;"></div>
-        
-        <script>
-            let currentTranscript = '';
-            const debug = document.getElementById('debug');
-            
-            function sendTranscriptUpdate() {
-                // Send transcript to parent (Streamlit)
-                window.parent.postMessage({
-                    type: 'transcript_update',
-                    data: currentTranscript
-                }, '*');
-                debug.textContent = `Last update: ${new Date().toLocaleTimeString()} - Length: ${currentTranscript.length}`;
-            }
-            
-            // Set up periodic updates
-            setInterval(sendTranscriptUpdate, 3000);  // Send update every 3 seconds
-            
-            const recognition = new (window.SpeechRecognition || window.webkitSpeechRecognition)();
-            const startButton = document.getElementById('start');
-            const stopButton = document.getElementById('stop');
-            const clearButton = document.getElementById('clear');
-            const status = document.getElementById('status');
-            const output = document.getElementById('output');
-            
-            recognition.continuous = true;
-            recognition.interimResults = true;
-            
-            startButton.onclick = () => {
-                recognition.start();
-                status.textContent = '🎤 Listening...';
-                startButton.disabled = true;
-                stopButton.disabled = false;
-            };
-            
-            stopButton.onclick = () => {
-                recognition.stop();
-                status.textContent = 'Stopped';
-                startButton.disabled = false;
-                stopButton.disabled = true;
-                sendTranscriptUpdate();  // Send final update when stopped
-            };
-            
-            clearButton.onclick = () => {
-                currentTranscript = '';
-                output.textContent = '';
-                sendTranscriptUpdate();  // Send empty transcript
-            };
-            
-            recognition.onresult = (event) => {
-                let interimTranscript = '';
-                let finalTranscript = '';
-                
-                for (let i = event.resultIndex; i < event.results.length; i++) {
-                    const transcript = event.results[i][0].transcript;
-                    if (event.results[i].isFinal) {
-                        finalTranscript += transcript + ' ';
-                        currentTranscript += transcript + ' ';
-                    } else {
-                        interimTranscript += transcript;
-                    }
-                }
-                
-                output.textContent = currentTranscript + (interimTranscript ? '... ' + interimTranscript : '');
-                output.scrollTop = output.scrollHeight;
-                
-                if (finalTranscript) {
-                    sendTranscriptUpdate();  // Send update when we have final transcript
-                }
-            };
-            
-            recognition.onend = () => {
-                if (!stopButton.disabled) {
-                    recognition.start();
-                }
-            };
-            
-            // Auto-start on load
-            window.addEventListener('load', () => {
-                setTimeout(() => startButton.click(), 1000);
-            });
-        </script>
-    </div>
-    """
-    
-    # Return both the component value
-    return components.html(
-        speech_recognition_html,
-        height=400,
-    )
-
-def integrate_speech_component():
-    """Integrate speech component with session state management."""
-    if "voice_transcript" not in st.session_state:
-        st.session_state.voice_transcript = ""
-    if "last_update" not in st.session_state:
-        st.session_state.last_update = time.time()
-    
-    # Create placeholders for display
-    transcript_container = st.empty()
-    status_container = st.empty()
-    
-    # Create component
-    component_val = create_speech_component()
-    
-    # Display current transcript
-    current_transcript = st.session_state.voice_transcript
-    transcript_container.text_area(
-        "Voice Transcript:", 
-        value=current_transcript,
-        height=100,
-        key=f"transcript_display_{int(time.time())}"
-    )
-    
-    # Show status
-    status_container.text(
-        f"Last updated: {datetime.fromtimestamp(st.session_state.last_update).strftime('%H:%M:%S')}"
-    )
-    
-    return current_transcript
-    
-
-    
-    
-
-
 # 2. 🚲BikeAI🏆 Load environment variables and initialize clients
 load_dotenv()
 
@@ -421,6 +280,122 @@ def process_video_with_gpt(video_input, user_prompt):
     )
     return response.choices[0].message.content
 
+
+
+def process_tts(text, voice="alloy"):
+    """
+    Process text-to-speech using OpenAI's TTS API
+    Voices available: alloy, echo, fable, onyx, nova, shimmer
+    """
+    try:
+        response = openai_client.audio.speech.create(
+            model="tts-1",
+            voice=voice,
+            input=text
+        )
+        
+        # Generate a unique filename
+        filename = generate_filename("tts_output", "mp3")
+        
+        # Save the audio file
+        response.stream_to_file(filename)
+        
+        # Create audio player HTML
+        audio_html = f"""
+        <audio controls>
+            <source src="data:audio/mp3;base64,{base64.b64encode(open(filename, 'rb').read()).decode()}" type="audio/mp3">
+            Your browser does not support the audio element.
+        </audio>
+        """
+        
+        return filename, audio_html
+    except Exception as e:
+        st.error(f"TTS Error: {str(e)}")
+        return None, None
+
+def update_chat_interface():
+    """Update the chat interface to include voice selection and TTS playback"""
+    # Add voice selection to sidebar
+    st.sidebar.markdown("### 🗣️ TTS Voice Settings")
+    selected_voice = st.sidebar.selectbox(
+        "Choose TTS Voice:",
+        ["alloy", "echo", "fable", "onyx", "nova", "shimmer"],
+        help="Select the voice for text-to-speech playback"
+    )
+    
+    # Store the selected voice in session state
+    if "selected_voice" not in st.session_state:
+        st.session_state.selected_voice = selected_voice
+
+# Modify the chat processing functions to include TTS
+def process_with_gpt(text_input):
+    """Process text with GPT-4o and add TTS."""
+    if text_input:
+        st.session_state.messages.append({"role": "user", "content": text_input})
+        
+        with st.chat_message("user"):
+            st.markdown(text_input)
+        
+        with st.chat_message("assistant"):
+            completion = openai_client.chat.completions.create(
+                model=st.session_state["openai_model"],
+                messages=[
+                    {"role": m["role"], "content": m["content"]}
+                    for m in st.session_state.messages
+                ],
+                stream=False
+            )
+            return_text = completion.choices[0].message.content
+            st.write("GPT-4o: " + return_text)
+            
+            # Add TTS playback
+            filename, audio_html = process_tts(return_text, st.session_state.selected_voice)
+            if audio_html:
+                st.markdown(audio_html, unsafe_allow_html=True)
+            
+            # Original file handling
+            filename = generate_filename("GPT-4o: " + return_text, "md")
+            create_file(filename, text_input, return_text)
+            st.session_state.messages.append({"role": "assistant", "content": return_text})
+            return return_text
+
+def process_with_claude(text_input):
+    """Process text with Claude and add TTS."""
+    if text_input:
+        with st.chat_message("user"):
+            st.markdown(text_input)
+
+        with st.chat_message("assistant"):
+            response = claude_client.messages.create(
+                model="claude-3-sonnet-20240229",
+                max_tokens=1000,
+                messages=[
+                    {"role": "user", "content": text_input}
+                ]
+            )
+            response_text = response.content[0].text
+            st.write("Claude: " + response_text)
+            
+            # Add TTS playback
+            filename, audio_html = process_tts(response_text, st.session_state.selected_voice)
+            if audio_html:
+                st.markdown(audio_html, unsafe_allow_html=True)
+            
+            # Original file handling
+            filename = generate_filename("Claude: " + response_text, "md")
+            create_file(filename, text_input, response_text)
+            
+            st.session_state.chat_history.append({
+                "user": text_input,
+                "claude": response_text
+            })
+            return response_text
+
+
+
+
+
+
     
 def extract_urls(text):
     try:
@@ -776,11 +751,9 @@ speech_recognition_html = """
             const output = document.getElementById('output');
             let fullTranscript = '';
             let lastUpdateTime = Date.now();
-
             // Configure recognition
             recognition.continuous = true;
             recognition.interimResults = true;
-
             // Function to start recognition
             const startRecognition = () => {
                 try {
@@ -793,21 +766,17 @@ speech_recognition_html = """
                     status.textContent = 'Error: ' + e.message;
                 }
             };
-
             // Auto-start on load
             window.addEventListener('load', () => {
                 setTimeout(startRecognition, 1000);
             });
-
             startButton.onclick = startRecognition;
-
             stopButton.onclick = () => {
                 recognition.stop();
                 status.textContent = 'Stopped';
                 startButton.disabled = false;
                 stopButton.disabled = true;
             };
-
             clearButton.onclick = () => {
                 fullTranscript = '';
                 output.textContent = '';
@@ -815,11 +784,9 @@ speech_recognition_html = """
                     type: 'clear_transcript',
                 }, '*');
             };
-
             recognition.onresult = (event) => {
                 let interimTranscript = '';
                 let finalTranscript = '';
-
                 for (let i = event.resultIndex; i < event.results.length; i++) {
                     const transcript = event.results[i][0].transcript;
                     if (event.results[i].isFinal) {
@@ -828,7 +795,6 @@ speech_recognition_html = """
                         interimTranscript += transcript;
                     }
                 }
-
                 if (finalTranscript || (Date.now() - lastUpdateTime > 5000)) {
                     if (finalTranscript) {
                         fullTranscript += finalTranscript;
@@ -838,14 +804,10 @@ speech_recognition_html = """
                     }
                     lastUpdateTime = Date.now();
                 }            
-
                 output.textContent = fullTranscript + (interimTranscript ? '... ' + interimTranscript : '');
                 output.scrollTop = output.scrollHeight;
-
                 document.getElementById('streamlit-data').value = fullTranscript;
-
             };
-
             recognition.onend = () => {
                 if (!stopButton.disabled) {
                     try {
@@ -859,7 +821,6 @@ speech_recognition_html = """
                     }
                 }
             };
-
             recognition.onerror = (event) => {
                 console.error('Recognition error:', event.error);
                 status.textContent = 'Error: ' + event.error;
@@ -1004,48 +965,196 @@ def get_media_html(media_path, media_type="video", width="100%"):
 def set_transcript(text):
     """Set transcript in session state."""
     st.session_state.voice_transcript = text
+
 def main():
     st.sidebar.markdown("### 🚲BikeAI🏆 Claude and GPT Multi-Agent Research AI")
     
+    # Main navigation
     tab_main = st.radio("Choose Action:", 
-                      ["🎤 Voice Input", "💬 Chat", "📸 Media Gallery", "🔍 Search ArXiv", "📝 File Editor"],
-                      horizontal=True)
+                        ["🎤 Voice Input", "💬 Chat", "📸 Media Gallery", "🔍 Search ArXiv", "📝 File Editor"],
+                        horizontal=True)
     
     if tab_main == "🎤 Voice Input":
         st.subheader("Voice Recognition")
         
-        try:
-            # Initialize speech component
-            current_transcript = integrate_speech_component()
+        # Initialize session state for the transcript
+        if 'voice_transcript' not in st.session_state:
+            st.session_state.voice_transcript = ""
+        
+        # Display speech recognition component and capture returned value
+        transcript = st.components.v1.html(speech_recognition_html, height=400)
+        
+        # Update session state if there's new data
+        if transcript is not None and transcript != "":
+            st.session_state.voice_transcript = transcript
+        
+        # Display the transcript in a Streamlit text area
+        st.markdown("### Processed Voice Input:")
+        st.text_area("Voice Transcript", st.session_state.voice_transcript, height=100)
+        
+        # Add functionality to process the transcript
+        if st.button("Process Transcript"):
+            st.subheader("AI Response to Transcript")
+            gpt_response = process_with_gpt(st.session_state.voice_transcript)
+            st.markdown(gpt_response)
+        
+        # Option to clear the transcript
+        if st.button("Clear Transcript"):
+            st.session_state.voice_transcript = ""
+            st.rerun()
             
-            # Show last update time
-            st.text(f"Last updated: {datetime.fromtimestamp(st.session_state.last_update).strftime('%H:%M:%S')}")
+
+        # Buttons to process the transcript
+        if st.button("Search with GPT"):
+            st.subheader("GPT-4o Response")
+            gpt_response = process_with_gpt(st.session_state.voice_transcript)
+            st.markdown(gpt_response)
+
+        if st.button("Search with Claude"):
+            st.subheader("Claude Response")
+            claude_response = process_with_claude(st.session_state.voice_transcript)
+            st.markdown(claude_response)
+
+        if st.button("Search ArXiv"):
+            st.subheader("ArXiv Search Results")
+            arxiv_results = perform_ai_lookup(st.session_state.voice_transcript)
+            st.markdown(arxiv_results)
             
-            # Process buttons if we have a transcript
-            if current_transcript:
-                col1, col2, col3 = st.columns(3)
-                
-                with col1:
-                    if st.button("Process with GPT"):
-                        with st.spinner("Processing with GPT..."):
-                            response = process_with_gpt(current_transcript)
-                            st.markdown(response)
-                
-                with col2:
-                    if st.button("Process with Claude"):
-                        with st.spinner("Processing with Claude..."):
-                            response = process_with_claude(current_transcript)
-                            st.markdown(response)
-                
-                with col3:
-                    if st.button("Search ArXiv"):
+        
+        # Display last voice input
+        if st.session_state.last_voice_input:
+            st.text_area("Last Voice Input:", st.session_state.last_voice_input, height=100)
+
+
+        # Model Selection
+        model_choice = st.sidebar.radio(
+            "Choose AI Model:",
+            ["GPT-4o", "Claude-3", "GPT+Claude+Arxiv"]
+        )
+        
+        # Chat Interface
+        user_input = st.text_area("Message:", height=100)
+        
+        if st.button("Send 📨"):
+            if user_input:
+                if model_choice == "GPT-4o":
+                    gpt_response = process_with_gpt(user_input)
+                elif model_choice == "Claude-3":
+                    claude_response = process_with_claude(user_input)
+                else:  # Both
+                    col1, col2, col3 = st.columns(3)
+                    with col2:
+                        st.subheader("Claude-3.5 Sonnet:")
+                        try:
+                            claude_response = process_with_claude(user_input)
+                        except:
+                            st.write('Claude 3.5 Sonnet out of tokens.')
+                    with col1:
+                        st.subheader("GPT-4o Omni:")
+                        try:
+                            gpt_response = process_with_gpt(user_input)  
+                        except:
+                            st.write('GPT 4o out of tokens')
+                    with col3:
+                        st.subheader("Arxiv and Mistral Research:")
                         with st.spinner("Searching ArXiv..."):
-                            results = perform_ai_lookup(current_transcript)
+                            #results = search_arxiv(user_input)
+                            results = perform_ai_lookup(user_input)
+                            
                             st.markdown(results)
+                            
+        # Display Chat History
+        st.subheader("Chat History 📜")
+        tab1, tab2 = st.tabs(["Claude History", "GPT-4o History"])
         
-        except Exception as e:
-            st.error(f"Error in voice input: {str(e)}")
-            
+        with tab1:
+            for chat in st.session_state.chat_history:
+                st.text_area("You:", chat["user"], height=100)
+                st.text_area("Claude:", chat["claude"], height=200)
+                st.markdown(chat["claude"])
+        
+        with tab2:
+            for message in st.session_state.messages:
+                with st.chat_message(message["role"]):
+                    st.markdown(message["content"])
+
+
+# -------------------------------------------------------  *************************  --->
+
+    
+    
+    if tab_main == "💬 Chat":
+        # Model Selection
+        model_choice = st.sidebar.radio(
+            "Choose AI Model:",
+            ["GPT-4o", "Claude-3", "GPT+Claude+Arxiv"]
+        )
+        
+        # Chat Interface
+        user_input = st.text_area("Message:", height=100)
+        
+        if st.button("Send 📨"):
+            if user_input:
+                if model_choice == "GPT-4o":
+                    gpt_response = process_with_gpt(user_input)
+                elif model_choice == "Claude-3":
+                    claude_response = process_with_claude(user_input)
+                else:  # Both
+                    col1, col2, col3 = st.columns(3)
+                    with col2:
+                        st.subheader("Claude-3.5 Sonnet:")
+                        try:
+                            claude_response = process_with_claude(user_input)
+                        except:
+                            st.write('Claude 3.5 Sonnet out of tokens.')
+                    with col1:
+                        st.subheader("GPT-4o Omni:")
+                        try:
+                            gpt_response = process_with_gpt(user_input)  
+                        except:
+                            st.write('GPT 4o out of tokens')
+                    with col3:
+                        st.subheader("Arxiv and Mistral Research:")
+                        with st.spinner("Searching ArXiv..."):
+                            #results = search_arxiv(user_input)
+                            results = perform_ai_lookup(user_input)
+                            
+                            st.markdown(results)
+                            
+        # Display Chat History
+        st.subheader("Chat History 📜")
+        tab1, tab2 = st.tabs(["Claude History", "GPT-4o History"])
+        
+        with tab1:
+            for chat in st.session_state.chat_history:
+                st.text_area("You:", chat["user"], height=100)
+                st.text_area("Claude:", chat["claude"], height=200)
+                st.markdown(chat["claude"])
+        
+        with tab2:
+            for message in st.session_state.messages:
+                with st.chat_message(message["role"]):
+                    st.markdown(message["content"])
+    
+    elif tab_main == "📸 Media Gallery":
+        create_media_gallery()
+    
+    elif tab_main == "🔍 Search ArXiv":
+        query = st.text_input("Enter your research query:")
+        if query:
+            with st.spinner("Searching ArXiv..."):
+                results = search_arxiv(query)
+                st.markdown(results)
+    
+    elif tab_main == "📝 File Editor":
+        if hasattr(st.session_state, 'current_file'):
+            st.subheader(f"Editing: {st.session_state.current_file}")
+            new_content = st.text_area("Content:", st.session_state.file_content, height=300)
+            if st.button("Save Changes"):
+                with open(st.session_state.current_file, 'w', encoding='utf-8') as file:
+                    file.write(new_content)
+                st.success("File updated successfully!")
+    
 
     # Always show file manager in sidebar
     display_file_manager()