Spaces:

awacke1
/

RescuerOfStolenBikes

Sleeping

App Files Files Community

awacke1 commited on Dec 1, 2024

Commit

1ccc585

verified ·

1 Parent(s): 8edf8d2

Update app.py

Browse files

Files changed (1) hide show

app.py +159 -39

app.py CHANGED Viewed

@@ -1,4 +1,5 @@
 import streamlit as st
 import anthropic
 import openai
 import base64
@@ -13,7 +14,6 @@ import pytz
 import random
 import re
 import requests
-import streamlit.components.v1 as components
 import textract
 import time
 import zipfile
@@ -52,6 +52,142 @@ st.set_page_config(
     }
 )
 # 2. 🚲BikeAI🏆 Load environment variables and initialize clients
 load_dotenv()
@@ -872,50 +1008,34 @@ def main():
                         ["🎤 Voice Input", "💬 Chat", "📸 Media Gallery", "🔍 Search ArXiv", "📝 File Editor"],
                         horizontal=True)
-    if tab_main == "🎤 Voice Input":
-        st.subheader("Voice Recognition")
-        # 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()
-        # 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)

 import streamlit as st
+import streamlit.components.v1 as components
 import anthropic
 import openai
 import base64
 import random
 import re
 import requests
 import textract
 import time
 import zipfile
     }
 )
+def create_speech_component():
+    """Create a custom speech recognition component with bidirectional communication."""
+    speech_recognition_html = """
+    <!DOCTYPE html>
+    <html>
+    <head>
+        <title>Continuous Speech Demo</title>
+        <style>
+            /* Your existing styles here */
+        </style>
+    </head>
+    <body>
+        <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">Ready</div>
+        <div id="output"></div>
+        <script>
+            // Function to send data back to Streamlit
+            function sendToStreamlit(data) {
+                window.parent.postMessage({
+                    type: 'streamlit:setComponentValue',
+                    value: data
+                }, '*');
+            }
+            if (!('webkitSpeechRecognition' in window)) {
+                alert('Speech recognition not supported');
+            } else {
+                const recognition = new 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');
+                let fullTranscript = '';
+                // Configure recognition
+                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;
+                    // Send final transcript to Streamlit
+                    sendToStreamlit(fullTranscript);
+                };
+                clearButton.onclick = () => {
+                    fullTranscript = '';
+                    output.textContent = '';
+                    sendToStreamlit('');
+                };
+                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 + '\\n';
+                            fullTranscript += transcript + '\\n';
+                            // Send update to Streamlit
+                            sendToStreamlit(fullTranscript);
+                        } else {
+                            interimTranscript += transcript;
+                        }
+                    }
+                    output.textContent = fullTranscript + (interimTranscript ? '... ' + interimTranscript : '');
+                    output.scrollTop = output.scrollHeight;
+                };
+                recognition.onend = () => {
+                    if (!stopButton.disabled) {
+                        recognition.start();
+                    }
+                };
+                // Auto-start on load
+                window.addEventListener('load', () => {
+                    setTimeout(() => {
+                        startButton.click();
+                    }, 1000);
+                });
+            }
+        </script>
+    </body>
+    </html>
+    """
+    # Create the component with a key
+    component_value = components.html(
+        speech_recognition_html,
+        height=400,
+        key="speech_recognition"
+    )
+    return component_value
+def integrate_speech_component():
+    """Integrate the speech component into the main app."""
+    if "voice_transcript" not in st.session_state:
+        st.session_state.voice_transcript = ""
+    # Get the transcript from the component
+    transcript = create_speech_component()
+    # Update session state if there's new data
+    if transcript is not None and transcript != "":
+        st.session_state.voice_transcript = transcript
+    return st.session_state.voice_transcript
 # 2. 🚲BikeAI🏆 Load environment variables and initialize clients
 load_dotenv()
                         ["🎤 Voice Input", "💬 Chat", "📸 Media Gallery", "🔍 Search ArXiv", "📝 File Editor"],
                         horizontal=True)
+if tab_main == "🎤 Voice Input":
+    st.subheader("Voice Recognition")
+    # Get transcript from the speech component
+    current_transcript = integrate_speech_component()
+    # Display the transcript in a Streamlit text area
+    st.markdown("### Processed Voice Input:")
+    st.text_area("Voice Transcript", current_transcript, height=100)
+    # Process buttons
+    col1, col2, col3 = st.columns(3)
+    with col1:
+        if st.button("Process with GPT"):
             st.subheader("GPT-4o Response")
+            gpt_response = process_with_gpt(current_transcript)
             st.markdown(gpt_response)
+    with col2:
+        if st.button("Process with Claude"):
             st.subheader("Claude Response")
+            claude_response = process_with_claude(current_transcript)
             st.markdown(claude_response)
+    with col3:
         if st.button("Search ArXiv"):
             st.subheader("ArXiv Search Results")
+            arxiv_results = perform_ai_lookup(current_transcript)
             st.markdown(arxiv_results)