Initial commit: audio-chat with fixes

- Created AGENTS.md with architecture documentation
- Fixed race conditions and async patterns
- Added conversation history to LLM prompts
- Fixed TTS audio shape handling
- Added buffer limits and graceful shutdown
- Fixed client.py with file sending support
- Removed duplicate requirements
- Added .gitignore
This commit is contained in:
2026-05-01 13:01:06 +00:00
commit 1edfd5d62f
13 changed files with 1286 additions and 0 deletions

570
static/index.html Normal file
View File

@@ -0,0 +1,570 @@
<!DOCTYPE html>
<html lang="ru">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Голосовой чат</title>
<style>
* {
margin: 0;
padding: 0;
box-sizing: border-box;
}
body {
font-family: -apple-system, BlinkMacSystemFont, 'Segoe UI', Roboto, sans-serif;
background: #0f0f23;
color: #e0e0e0;
min-height: 100vh;
display: flex;
flex-direction: column;
}
.header {
background: linear-gradient(135deg, #1a1a3e, #2d1b69);
padding: 20px;
text-align: center;
border-bottom: 1px solid #333;
}
.header h1 {
font-size: 24px;
color: #fff;
margin-bottom: 8px;
}
.status {
display: inline-flex;
align-items: center;
gap: 8px;
padding: 6px 16px;
background: rgba(255,255,255,0.1);
border-radius: 20px;
font-size: 14px;
}
.status-dot {
width: 10px;
height: 10px;
border-radius: 50%;
background: #666;
}
.status-dot.connected {
background: #4caf50;
box-shadow: 0 0 10px #4caf50;
}
.status-dot.recording {
background: #f44336;
animation: pulse 1s infinite;
}
@keyframes pulse {
0%, 100% { opacity: 1; }
50% { opacity: 0.5; }
}
.container {
flex: 1;
display: flex;
flex-direction: column;
max-width: 800px;
margin: 0 auto;
padding: 20px;
width: 100%;
}
.chat-area {
flex: 1;
overflow-y: auto;
margin-bottom: 20px;
padding: 16px;
background: rgba(255,255,255,0.05);
border-radius: 12px;
border: 1px solid rgba(255,255,255,0.1);
}
.message {
margin-bottom: 16px;
padding: 12px 16px;
border-radius: 12px;
max-width: 80%;
}
.message.user {
background: #2d1b69;
margin-left: auto;
border-bottom-right-radius: 4px;
}
.message.assistant {
background: rgba(255,255,255,0.1);
margin-right: auto;
border-bottom-left-radius: 4px;
}
.message-label {
font-size: 11px;
color: #888;
margin-bottom: 4px;
text-transform: uppercase;
letter-spacing: 1px;
}
.message-text {
font-size: 15px;
line-height: 1.5;
}
.controls {
background: rgba(255,255,255,0.05);
padding: 24px;
border-radius: 16px;
border: 1px solid rgba(255,255,255,0.1);
display: flex;
flex-direction: column;
align-items: center;
gap: 20px;
}
.visualizer {
width: 100%;
height: 80px;
background: rgba(0,0,0,0.3);
border-radius: 8px;
overflow: hidden;
}
.btn-group {
display: flex;
gap: 12px;
flex-wrap: wrap;
justify-content: center;
}
button {
padding: 14px 32px;
border: none;
border-radius: 12px;
font-size: 16px;
font-weight: 600;
cursor: pointer;
transition: all 0.3s ease;
display: inline-flex;
align-items: center;
gap: 8px;
}
.btn-primary {
background: linear-gradient(135deg, #6c3ce0, #9b59b6);
color: white;
}
.btn-primary:hover {
transform: translateY(-2px);
box-shadow: 0 4px 15px rgba(108, 60, 224, 0.4);
}
.btn-danger {
background: linear-gradient(135deg, #c0392b, #e74c3c);
color: white;
}
.btn-secondary {
background: rgba(255,255,255,0.1);
color: #ccc;
}
.btn-secondary:hover {
background: rgba(255,255,255,0.2);
}
.btn:disabled {
opacity: 0.5;
cursor: not-allowed;
transform: none !important;
}
.text-indicator {
font-size: 14px;
color: #888;
font-style: italic;
min-height: 20px;
}
.text-indicator.active {
color: #4caf50;
}
.loading {
display: none;
text-align: center;
padding: 20px;
color: #888;
}
.loading.show {
display: block;
}
.spinner {
width: 40px;
height: 40px;
border: 3px solid rgba(255,255,255,0.1);
border-top-color: #6c3ce0;
border-radius: 50%;
animation: spin 1s linear infinite;
margin: 0 auto 12px;
}
@keyframes spin {
to { transform: rotate(360deg); }
}
.config-section {
margin-bottom: 20px;
padding: 16px;
background: rgba(255,255,255,0.03);
border-radius: 8px;
}
.config-section label {
display: block;
margin-bottom: 8px;
font-size: 13px;
color: #888;
}
.config-section input {
width: 100%;
padding: 10px;
background: rgba(0,0,0,0.3);
border: 1px solid rgba(255,255,255,0.1);
border-radius: 6px;
color: #fff;
font-size: 14px;
}
@media (max-width: 600px) {
.container {
padding: 12px;
}
.message {
max-width: 95%;
}
button {
padding: 12px 24px;
font-size: 14px;
}
}
</style>
</head>
<body>
<div class="header">
<h1>🎙️ Голосовой чат с ИИ</h1>
<div class="status">
<div class="status-dot" id="statusDot"></div>
<span id="statusText">Отключено</span>
</div>
</div>
<div class="container">
<div class="config-section">
<label>URL WebSocket сервера</label>
<input type="text" id="wsUrl" value="ws://localhost:8000/ws" placeholder="ws://localhost:8000/ws">
</div>
<div class="chat-area" id="chatArea">
<div class="message assistant">
<div class="message-label">Ассистент</div>
<div class="message-text">Привет! Нажми "Начать запись" и говори со мной.</div>
</div>
</div>
<div class="loading" id="loading">
<div class="spinner"></div>
<div>Генерирую ответ...</div>
</div>
<div class="controls">
<canvas class="visualizer" id="visualizer"></canvas>
<div class="text-indicator" id="textIndicator"></div>
<div class="btn-group">
<button class="btn-primary" id="btnStart" onclick="startRecording()">
🎤 Начать запись
</button>
<button class="btn-danger" id="btnStop" onclick="stopRecording()" disabled>
⏹️ Остановить
</button>
<button class="btn-secondary" onclick="resetChat()">
🔄 Сброс
</button>
</div>
</div>
</div>
<audio id="audioPlayer" style="display:none"></audio>
<script>
const audioPlayer = document.getElementById('audioPlayer');
const chatArea = document.getElementById('chatArea');
const statusDot = document.getElementById('statusDot');
const statusText = document.getElementById('statusText');
const textIndicator = document.getElementById('textIndicator');
const loading = document.getElementById('loading');
const btnStart = document.getElementById('btnStart');
const btnStop = document.getElementById('btnStop');
const visualizer = document.getElementById('visualizer');
const wsUrlInput = document.getElementById('wsUrl');
let ws = null;
let audioContext = null;
let mediaStream = null;
let analyser = null;
let isRecording = false;
let buffer = new Uint8Array();
let animationId = null;
let textTimeout = null;
let currentText = '';
// Resize visualizer
function resizeVisualizer() {
visualizer.width = visualizer.offsetWidth;
visualizer.height = visualizer.offsetHeight;
}
resizeVisualizer();
window.addEventListener('resize', resizeVisualizer);
function setStatus(status, text) {
statusDot.className = 'status-dot ' + status;
statusText.textContent = text;
}
function addMessage(role, text) {
const div = document.createElement('div');
div.className = 'message ' + role;
div.innerHTML = `
<div class="message-label">${role === 'user' ? 'Вы' : 'Ассистент'}</div>
<div class="message-text">${text}</div>
`;
chatArea.appendChild(div);
chatArea.scrollTop = chatArea.scrollHeight;
}
function drawVisualizer() {
if (!analyser) return;
const bufferLength = analyser.frequencyBinCount;
const dataArray = new Uint8Array(bufferLength);
analyser.getByteFrequencyData(dataArray);
const ctx = visualizer.getContext('2d');
ctx.clearRect(0, 0, visualizer.width, visualizer.height);
const barWidth = (visualizer.width / bufferLength) * 2.5;
let x = 0;
for (let i = 0; i < bufferLength; i++) {
const barHeight = (dataArray[i] / 255) * visualizer.height;
ctx.fillStyle = `rgba(108, 60, 224, ${dataArray[i] / 255})`;
ctx.fillRect(x, visualizer.height - barHeight, barWidth, barHeight);
x += barWidth + 1;
}
if (isRecording) {
animationId = requestAnimationFrame(drawVisualizer);
}
}
function connectWebSocket() {
const url = wsUrlInput.value || 'ws://localhost:8000/ws';
ws = new WebSocket(url);
ws.onopen = () => {
setStatus('connected', 'Подключено');
console.log('WebSocket connected');
};
ws.onclose = () => {
setStatus('', 'Отключено');
console.log('WebSocket closed');
};
ws.onerror = (error) => {
console.error('WebSocket error:', error);
setStatus('', 'Ошибка подключения');
};
ws.onmessage = (event) => {
const data = event.data;
if (typeof data === 'string') {
if (data.startsWith('TEXT:')) {
const text = data.substring(5);
currentText = text;
textIndicator.textContent = text;
textIndicator.classList.add('active');
// Add user message after delay
clearTimeout(textTimeout);
textTimeout = setTimeout(() => {
addMessage('user', text);
textIndicator.textContent = '';
textIndicator.classList.remove('active');
currentText = '';
}, 1500);
} else if (data === 'RESET') {
console.log('Session reset');
}
} else if (data instanceof ArrayBuffer) {
const bytes = new Uint8Array(data);
if (bytes[0] === 0x4F) { // 'O'
const audioData = bytes.slice(1);
playAudio(audioData);
}
}
};
}
function playAudio(audioData) {
// Convert 24kHz WAV to 48kHz for browser playback
const sampleRate = 24000;
const targetSampleRate = 48000;
const wavHeaderSize = 44;
const audioContent = audioData.slice(wavHeaderSize);
const numSamples = audioContent.length / 2;
// Create AudioBuffer
const audioCtx = new (window.AudioContext || window.webkitAudioContext)();
const outputBuffer = audioCtx.createBuffer(1, numSamples * (targetSampleRate / sampleRate), targetSampleRate);
const outputData = outputBuffer.getChannelData(0);
// Resample
for (let i = 0; i < numSamples; i++) {
const inputSample = parseInt(audioContent.slice(i * 2, i * 2 + 2).reverse());
const outputSample = inputSample / 32768;
const outputIndex = i * (targetSampleRate / sampleRate);
for (let j = 0; j < (targetSampleRate / sampleRate); j++) {
const idx = Math.floor(outputIndex + j);
if (idx < outputBuffer.length) {
outputData[idx] = outputSample;
}
}
}
// Create source and play
const source = audioCtx.createBufferSource();
source.buffer = outputBuffer;
source.connect(audioCtx.destination);
source.start();
loading.classList.remove('show');
}
async function startRecording() {
try {
// Connect WebSocket
connectWebSocket();
// Request microphone access
mediaStream = await navigator.mediaDevices.getUserMedia({
audio: {
channelCount: 1,
sampleRate: 48000,
echoCancellation: true,
noiseSuppression: true
}
});
// Setup audio context
audioContext = new (window.AudioContext || window.webkitAudioContext)();
const source = audioContext.createMediaStreamSource(mediaStream);
// Setup analyser for visualization
analyser = audioContext.createAnalyser();
analyser.fftSize = 256;
source.connect(analyser);
isRecording = true;
btnStart.disabled = true;
btnStop.disabled = false;
setStatus('recording', 'Запись...');
textIndicator.textContent = 'Слушаю...';
textIndicator.classList.add('active');
drawVisualizer();
// Resample from 48kHz to 16kHz and send
const scriptProcessor = audioContext.createScriptProcessor(4096, 1, 1);
scriptProcessor.onaudioprocess = (event) => {
if (!isRecording) return;
const input = event.inputBuffer.getChannelData(0);
// Resample from 48kHz to 16kHz (3:1 ratio)
const resampled = [];
for (let i = 0; i < input.length; i += 3) {
resampled.push(input[i]);
}
// Convert to PCM 16-bit
const pcm = new Uint8Array(resampled.length * 2);
for (let i = 0; i < resampled.length; i++) {
const sample = Math.max(-1, Math.min(1, resampled[i]));
const int16 = sample < 0 ? sample * 32768 : sample * 32767;
pcm[i * 2] = int16 & 0xFF;
pcm[i * 2 + 1] = (int16 >> 8) & 0xFF;
}
// Send via WebSocket
if (ws && ws.readyState === WebSocket.OPEN) {
ws.send(new Uint8Array([0x41]).concat(pcm).buffer);
}
};
source.connect(scriptProcessor);
scriptProcessor.connect(audioContext.destination);
} catch (err) {
console.error('Error starting recording:', err);
alert('Не удалось получить доступ к микрофону: ' + err.message);
isRecording = false;
btnStart.disabled = false;
btnStop.disabled = true;
}
}
function stopRecording() {
isRecording = false;
if (animationId) {
cancelAnimationFrame(animationId);
}
if (mediaStream) {
mediaStream.getTracks().forEach(t => t.stop());
}
if (audioContext) {
audioContext.close();
}
btnStart.disabled = false;
btnStop.disabled = true;
setStatus('connected', 'Подключено');
textIndicator.textContent = '';
textIndicator.classList.remove('active');
}
async function resetChat() {
if (ws && ws.readyState === WebSocket.OPEN) {
ws.send(new Uint8Array([0x52]).buffer); // 'R'
}
chatArea.innerHTML = '';
addMessage('assistant', 'Чат сброшен. Говорите со мной!');
}
// Auto-connect on page load
window.addEventListener('load', () => {
connectWebSocket();
});
</script>
</body>
</html>