mirror of
https://github.com/HumanAIGC-Engineering/gradio-webrtc.git
synced 2026-02-05 18:09:23 +08:00
[feat] update some feature
sync code of fastrtc, add text support through datachannel, fix safari connect problem support chat without camera or mic
This commit is contained in:
huangbinchao.hbc
486
demo/hello_computer/index.html
Normal file
486
demo/hello_computer/index.html
Normal file
@@ -0,0 +1,486 @@
|
||||
<!DOCTYPE html>
|
||||
<html lang="en">
|
||||
|
||||
<head>
|
||||
<meta charset="UTF-8">
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
||||
<title>Hello Computer 💻</title>
|
||||
<style>
|
||||
body {
|
||||
font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, sans-serif;
|
||||
background-color: #f8f9fa;
|
||||
color: #1a1a1a;
|
||||
margin: 0;
|
||||
padding: 20px;
|
||||
height: 100vh;
|
||||
box-sizing: border-box;
|
||||
}
|
||||
|
||||
.container {
|
||||
max-width: 800px;
|
||||
margin: 0 auto;
|
||||
height: calc(100% - 100px);
|
||||
}
|
||||
|
||||
.logo {
|
||||
text-align: center;
|
||||
margin-bottom: 40px;
|
||||
}
|
||||
|
||||
.chat-container {
|
||||
background: white;
|
||||
border-radius: 8px;
|
||||
box-shadow: 0 2px 4px rgba(0, 0, 0, 0.1);
|
||||
padding: 20px;
|
||||
height: 90%;
|
||||
box-sizing: border-box;
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
}
|
||||
|
||||
.chat-messages {
|
||||
flex-grow: 1;
|
||||
overflow-y: auto;
|
||||
margin-bottom: 20px;
|
||||
padding: 10px;
|
||||
}
|
||||
|
||||
.message {
|
||||
margin-bottom: 20px;
|
||||
padding: 12px;
|
||||
border-radius: 8px;
|
||||
font-size: 14px;
|
||||
line-height: 1.5;
|
||||
}
|
||||
|
||||
.message.user {
|
||||
background-color: #e9ecef;
|
||||
margin-left: 20%;
|
||||
}
|
||||
|
||||
.message.assistant {
|
||||
background-color: #f1f3f5;
|
||||
margin-right: 20%;
|
||||
}
|
||||
|
||||
.controls {
|
||||
text-align: center;
|
||||
margin-top: 20px;
|
||||
}
|
||||
|
||||
button {
|
||||
background-color: #0066cc;
|
||||
color: white;
|
||||
border: none;
|
||||
padding: 12px 24px;
|
||||
font-family: inherit;
|
||||
font-size: 14px;
|
||||
cursor: pointer;
|
||||
transition: all 0.3s;
|
||||
border-radius: 4px;
|
||||
font-weight: 500;
|
||||
}
|
||||
|
||||
button:hover {
|
||||
background-color: #0052a3;
|
||||
}
|
||||
|
||||
#audio-output {
|
||||
display: none;
|
||||
}
|
||||
|
||||
.icon-with-spinner {
|
||||
display: flex;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
gap: 12px;
|
||||
min-width: 180px;
|
||||
}
|
||||
|
||||
.spinner {
|
||||
width: 20px;
|
||||
height: 20px;
|
||||
border: 2px solid #ffffff;
|
||||
border-top-color: transparent;
|
||||
border-radius: 50%;
|
||||
animation: spin 1s linear infinite;
|
||||
flex-shrink: 0;
|
||||
}
|
||||
|
||||
@keyframes spin {
|
||||
to {
|
||||
transform: rotate(360deg);
|
||||
}
|
||||
}
|
||||
|
||||
.pulse-container {
|
||||
display: flex;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
gap: 12px;
|
||||
min-width: 180px;
|
||||
}
|
||||
|
||||
.pulse-circle {
|
||||
width: 20px;
|
||||
height: 20px;
|
||||
border-radius: 50%;
|
||||
background-color: #ffffff;
|
||||
opacity: 0.2;
|
||||
flex-shrink: 0;
|
||||
transform: translateX(-0%) scale(var(--audio-level, 1));
|
||||
transition: transform 0.1s ease;
|
||||
}
|
||||
|
||||
/* Add styles for typing indicator */
|
||||
.typing-indicator {
|
||||
padding: 8px;
|
||||
background-color: #f1f3f5;
|
||||
border-radius: 8px;
|
||||
margin-bottom: 10px;
|
||||
display: none;
|
||||
}
|
||||
|
||||
.dots {
|
||||
display: inline-flex;
|
||||
gap: 4px;
|
||||
}
|
||||
|
||||
.dot {
|
||||
width: 8px;
|
||||
height: 8px;
|
||||
background-color: #0066cc;
|
||||
border-radius: 50%;
|
||||
animation: pulse 1.5s infinite;
|
||||
opacity: 0.5;
|
||||
}
|
||||
|
||||
.dot:nth-child(2) {
|
||||
animation-delay: 0.5s;
|
||||
}
|
||||
|
||||
.dot:nth-child(3) {
|
||||
animation-delay: 1s;
|
||||
}
|
||||
|
||||
@keyframes pulse {
|
||||
|
||||
0%,
|
||||
100% {
|
||||
opacity: 0.5;
|
||||
transform: scale(1);
|
||||
}
|
||||
|
||||
50% {
|
||||
opacity: 1;
|
||||
transform: scale(1.2);
|
||||
}
|
||||
}
|
||||
|
||||
/* Add styles for toast notifications */
|
||||
.toast {
|
||||
position: fixed;
|
||||
top: 20px;
|
||||
left: 50%;
|
||||
transform: translateX(-50%);
|
||||
padding: 16px 24px;
|
||||
border-radius: 4px;
|
||||
font-size: 14px;
|
||||
z-index: 1000;
|
||||
display: none;
|
||||
box-shadow: 0 2px 5px rgba(0, 0, 0, 0.2);
|
||||
}
|
||||
|
||||
.toast.error {
|
||||
background-color: #f44336;
|
||||
color: white;
|
||||
}
|
||||
|
||||
.toast.warning {
|
||||
background-color: #ffd700;
|
||||
color: black;
|
||||
}
|
||||
</style>
|
||||
</head>
|
||||
|
||||
<body>
|
||||
<!-- Add toast element after body opening tag -->
|
||||
<div id="error-toast" class="toast"></div>
|
||||
<div class="container">
|
||||
<div class="logo">
|
||||
<h1>Hello Computer 💻</h1>
|
||||
<h2 style="font-size: 1.2em; color: #666; margin-top: 10px;">Say 'Computer' before asking your question</h2>
|
||||
</div>
|
||||
<div class="chat-container">
|
||||
<div class="chat-messages" id="chat-messages"></div>
|
||||
<div class="typing-indicator" id="typing-indicator">
|
||||
<div class="dots">
|
||||
<div class="dot"></div>
|
||||
<div class="dot"></div>
|
||||
<div class="dot"></div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
<div class="controls">
|
||||
<button id="start-button">Start Conversation</button>
|
||||
</div>
|
||||
</div>
|
||||
<audio id="audio-output"></audio>
|
||||
|
||||
<script>
|
||||
let peerConnection;
|
||||
let webrtc_id;
|
||||
const startButton = document.getElementById('start-button');
|
||||
const chatMessages = document.getElementById('chat-messages');
|
||||
|
||||
let audioLevel = 0;
|
||||
let animationFrame;
|
||||
let audioContext, analyser, audioSource;
|
||||
let messages = [];
|
||||
let eventSource;
|
||||
|
||||
function updateButtonState() {
|
||||
const button = document.getElementById('start-button');
|
||||
if (peerConnection && (peerConnection.connectionState === 'connecting' || peerConnection.connectionState === 'new')) {
|
||||
button.innerHTML = `
|
||||
<div class="icon-with-spinner">
|
||||
<div class="spinner"></div>
|
||||
<span>Connecting...</span>
|
||||
</div>
|
||||
`;
|
||||
} else if (peerConnection && peerConnection.connectionState === 'connected') {
|
||||
button.innerHTML = `
|
||||
<div class="pulse-container">
|
||||
<div class="pulse-circle"></div>
|
||||
<span>Stop Conversation</span>
|
||||
</div>
|
||||
`;
|
||||
} else {
|
||||
button.innerHTML = 'Start Conversation';
|
||||
}
|
||||
}
|
||||
|
||||
function setupAudioVisualization(stream) {
|
||||
audioContext = new (window.AudioContext || window.webkitAudioContext)();
|
||||
analyser = audioContext.createAnalyser();
|
||||
audioSource = audioContext.createMediaStreamSource(stream);
|
||||
audioSource.connect(analyser);
|
||||
analyser.fftSize = 64;
|
||||
const dataArray = new Uint8Array(analyser.frequencyBinCount);
|
||||
|
||||
function updateAudioLevel() {
|
||||
analyser.getByteFrequencyData(dataArray);
|
||||
const average = Array.from(dataArray).reduce((a, b) => a + b, 0) / dataArray.length;
|
||||
audioLevel = average / 255;
|
||||
|
||||
const pulseCircle = document.querySelector('.pulse-circle');
|
||||
if (pulseCircle) {
|
||||
pulseCircle.style.setProperty('--audio-level', 1 + audioLevel);
|
||||
}
|
||||
|
||||
animationFrame = requestAnimationFrame(updateAudioLevel);
|
||||
}
|
||||
updateAudioLevel();
|
||||
}
|
||||
|
||||
function showError(message) {
|
||||
const toast = document.getElementById('error-toast');
|
||||
toast.textContent = message;
|
||||
toast.className = 'toast error';
|
||||
toast.style.display = 'block';
|
||||
|
||||
// Hide toast after 5 seconds
|
||||
setTimeout(() => {
|
||||
toast.style.display = 'none';
|
||||
}, 5000);
|
||||
}
|
||||
|
||||
function handleMessage(event) {
|
||||
const eventJson = JSON.parse(event.data);
|
||||
const typingIndicator = document.getElementById('typing-indicator');
|
||||
|
||||
if (eventJson.type === "error") {
|
||||
showError(eventJson.message);
|
||||
} else if (eventJson.type === "send_input") {
|
||||
fetch('/input_hook', {
|
||||
method: 'POST',
|
||||
headers: {
|
||||
'Content-Type': 'application/json',
|
||||
},
|
||||
body: JSON.stringify({
|
||||
webrtc_id: webrtc_id,
|
||||
chatbot: messages,
|
||||
state: messages
|
||||
})
|
||||
});
|
||||
} else if (eventJson.type === "log") {
|
||||
if (eventJson.data === "pause_detected") {
|
||||
typingIndicator.style.display = 'block';
|
||||
chatMessages.scrollTop = chatMessages.scrollHeight;
|
||||
} else if (eventJson.data === "response_starting") {
|
||||
typingIndicator.style.display = 'none';
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
async function setupWebRTC() {
|
||||
const config = __RTC_CONFIGURATION__;
|
||||
peerConnection = new RTCPeerConnection(config);
|
||||
|
||||
const timeoutId = setTimeout(() => {
|
||||
const toast = document.getElementById('error-toast');
|
||||
toast.textContent = "Connection is taking longer than usual. Are you on a VPN?";
|
||||
toast.className = 'toast warning';
|
||||
toast.style.display = 'block';
|
||||
|
||||
// Hide warning after 5 seconds
|
||||
setTimeout(() => {
|
||||
toast.style.display = 'none';
|
||||
}, 5000);
|
||||
}, 5000);
|
||||
|
||||
try {
|
||||
const stream = await navigator.mediaDevices.getUserMedia({
|
||||
audio: true
|
||||
});
|
||||
|
||||
setupAudioVisualization(stream);
|
||||
|
||||
stream.getTracks().forEach(track => {
|
||||
peerConnection.addTrack(track, stream);
|
||||
});
|
||||
|
||||
const dataChannel = peerConnection.createDataChannel('text');
|
||||
dataChannel.onmessage = handleMessage;
|
||||
|
||||
const offer = await peerConnection.createOffer();
|
||||
await peerConnection.setLocalDescription(offer);
|
||||
|
||||
await new Promise((resolve) => {
|
||||
if (peerConnection.iceGatheringState === "complete") {
|
||||
resolve();
|
||||
} else {
|
||||
const checkState = () => {
|
||||
if (peerConnection.iceGatheringState === "complete") {
|
||||
peerConnection.removeEventListener("icegatheringstatechange", checkState);
|
||||
resolve();
|
||||
}
|
||||
};
|
||||
peerConnection.addEventListener("icegatheringstatechange", checkState);
|
||||
}
|
||||
});
|
||||
|
||||
peerConnection.addEventListener('connectionstatechange', () => {
|
||||
console.log('connectionstatechange', peerConnection.connectionState);
|
||||
if (peerConnection.connectionState === 'connected') {
|
||||
clearTimeout(timeoutId);
|
||||
const toast = document.getElementById('error-toast');
|
||||
toast.style.display = 'none';
|
||||
}
|
||||
updateButtonState();
|
||||
});
|
||||
|
||||
webrtc_id = Math.random().toString(36).substring(7);
|
||||
|
||||
const response = await fetch('/webrtc/offer', {
|
||||
method: 'POST',
|
||||
headers: { 'Content-Type': 'application/json' },
|
||||
body: JSON.stringify({
|
||||
sdp: peerConnection.localDescription.sdp,
|
||||
type: peerConnection.localDescription.type,
|
||||
webrtc_id: webrtc_id
|
||||
})
|
||||
});
|
||||
|
||||
const serverResponse = await response.json();
|
||||
|
||||
if (serverResponse.status === 'failed') {
|
||||
showError(serverResponse.meta.error === 'concurrency_limit_reached'
|
||||
? `Too many connections. Maximum limit is ${serverResponse.meta.limit}`
|
||||
: serverResponse.meta.error);
|
||||
stop();
|
||||
return;
|
||||
}
|
||||
|
||||
await peerConnection.setRemoteDescription(serverResponse);
|
||||
|
||||
eventSource = new EventSource('/outputs?webrtc_id=' + webrtc_id);
|
||||
eventSource.addEventListener("output", (event) => {
|
||||
const eventJson = JSON.parse(event.data);
|
||||
console.log(eventJson);
|
||||
messages.push(eventJson.message);
|
||||
addMessage(eventJson.message.role, eventJson.audio ?? eventJson.message.content);
|
||||
});
|
||||
} catch (err) {
|
||||
clearTimeout(timeoutId);
|
||||
console.error('Error setting up WebRTC:', err);
|
||||
showError('Failed to establish connection. Please try again.');
|
||||
stop();
|
||||
}
|
||||
}
|
||||
|
||||
function addMessage(role, content) {
|
||||
const messageDiv = document.createElement('div');
|
||||
messageDiv.classList.add('message', role);
|
||||
|
||||
if (role === 'user') {
|
||||
// Create audio element for user messages
|
||||
const audio = document.createElement('audio');
|
||||
audio.controls = true;
|
||||
audio.src = content;
|
||||
messageDiv.appendChild(audio);
|
||||
} else {
|
||||
// Text content for assistant messages
|
||||
messageDiv.textContent = content;
|
||||
}
|
||||
|
||||
chatMessages.appendChild(messageDiv);
|
||||
chatMessages.scrollTop = chatMessages.scrollHeight;
|
||||
}
|
||||
|
||||
function stop() {
|
||||
if (eventSource) {
|
||||
eventSource.close();
|
||||
eventSource = null;
|
||||
}
|
||||
|
||||
if (animationFrame) {
|
||||
cancelAnimationFrame(animationFrame);
|
||||
}
|
||||
if (audioContext) {
|
||||
audioContext.close();
|
||||
audioContext = null;
|
||||
analyser = null;
|
||||
audioSource = null;
|
||||
}
|
||||
if (peerConnection) {
|
||||
if (peerConnection.getTransceivers) {
|
||||
peerConnection.getTransceivers().forEach(transceiver => {
|
||||
if (transceiver.stop) {
|
||||
transceiver.stop();
|
||||
}
|
||||
});
|
||||
}
|
||||
|
||||
if (peerConnection.getSenders) {
|
||||
peerConnection.getSenders().forEach(sender => {
|
||||
if (sender.track && sender.track.stop) sender.track.stop();
|
||||
});
|
||||
}
|
||||
peerConnection.close();
|
||||
}
|
||||
updateButtonState();
|
||||
audioLevel = 0;
|
||||
}
|
||||
|
||||
startButton.addEventListener('click', () => {
|
||||
if (!peerConnection || peerConnection.connectionState !== 'connected') {
|
||||
setupWebRTC();
|
||||
} else {
|
||||
stop();
|
||||
}
|
||||
});
|
||||
</script>
|
||||
</body>
|
||||
|
||||
</html>
|
||||
Reference in New Issue
Block a user