STT / index.html
github-actions[bot]
Auto-deploy from GitHub: 1a5c5751fc1cc39af6c504a981628b4ff2f74518
b0d8269
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Audio Caption Generator</title>
<style>
* {
margin: 0;
padding: 0;
box-sizing: border-box;
}
:root {
--bg: #0a0e27;
--surface: #141b3d;
--primary: #00ff88;
--secondary: #ff00ff;
--accent: #00d4ff;
--error: #ff1744;
--text: #ffffff;
--border: 4px;
}
body {
font-family: 'Space Grotesk', 'Courier New', monospace;
background: var(--bg);
color: var(--text);
min-height: 100vh;
overflow-x: hidden;
position: relative;
}
body::before {
content: '';
position: fixed;
top: 0;
left: 0;
width: 100%;
height: 100%;
background:
radial-gradient(circle at 20% 50%, rgba(0, 255, 136, 0.1) 0%, transparent 50%),
radial-gradient(circle at 80% 80%, rgba(255, 0, 255, 0.1) 0%, transparent 50%),
radial-gradient(circle at 40% 20%, rgba(0, 212, 255, 0.1) 0%, transparent 50%);
pointer-events: none;
z-index: 0;
}
.container {
max-width: 1400px;
margin: 0 auto;
padding: 2rem;
position: relative;
z-index: 1;
}
header {
text-align: center;
margin-bottom: 3rem;
animation: slideDown 0.6s cubic-bezier(0.68, -0.55, 0.265, 1.55);
}
@keyframes slideDown {
from {
opacity: 0;
transform: translateY(-50px);
}
to {
opacity: 1;
transform: translateY(0);
}
}
h1 {
font-size: clamp(2rem, 5vw, 4rem);
font-weight: 900;
background: linear-gradient(135deg, var(--primary) 0%, var(--accent) 50%, var(--secondary) 100%);
-webkit-background-clip: text;
-webkit-text-fill-color: transparent;
background-clip: text;
text-transform: uppercase;
letter-spacing: -2px;
margin-bottom: 1rem;
position: relative;
display: inline-block;
}
h1::after {
content: '';
position: absolute;
bottom: -10px;
left: 50%;
transform: translateX(-50%);
width: 60%;
height: 6px;
background: linear-gradient(90deg, transparent, var(--primary), transparent);
animation: glow 2s ease-in-out infinite;
}
@keyframes glow {
0%,
100% {
opacity: 0.5;
}
50% {
opacity: 1;
}
}
.subtitle {
font-size: 1.2rem;
color: var(--accent);
letter-spacing: 2px;
}
.upload-section {
background: var(--surface);
border: var(--border) solid var(--primary);
box-shadow: 8px 8px 0 var(--primary);
padding: 2rem;
margin-bottom: 3rem;
position: relative;
transition: all 0.3s ease;
animation: slideUp 0.6s cubic-bezier(0.68, -0.55, 0.265, 1.55) 0.2s both;
}
@keyframes slideUp {
from {
opacity: 0;
transform: translateY(50px);
}
to {
opacity: 1;
transform: translateY(0);
}
}
.upload-section:hover {
transform: translate(-2px, -2px);
box-shadow: 12px 12px 0 var(--primary);
}
.upload-zone {
border: 3px dashed var(--accent);
padding: 3rem;
text-align: center;
cursor: pointer;
transition: all 0.3s ease;
background: rgba(0, 212, 255, 0.05);
}
.upload-zone:hover {
background: rgba(0, 212, 255, 0.1);
border-color: var(--primary);
}
.upload-zone.dragging {
background: rgba(0, 255, 136, 0.2);
border-color: var(--primary);
transform: scale(1.02);
}
input[type="file"] {
display: none;
}
.btn {
background: var(--primary);
color: var(--bg);
border: var(--border) solid var(--bg);
padding: 1rem 2rem;
font-size: 1.1rem;
font-weight: 900;
text-transform: uppercase;
cursor: pointer;
transition: all 0.2s ease;
box-shadow: 4px 4px 0 var(--bg);
letter-spacing: 1px;
position: relative;
}
.btn:hover:not(:disabled) {
transform: translate(-2px, -2px);
box-shadow: 6px 6px 0 var(--bg);
}
.btn:active:not(:disabled) {
transform: translate(2px, 2px);
box-shadow: 2px 2px 0 var(--bg);
}
.btn:disabled {
opacity: 0.6;
cursor: not-allowed;
}
.btn-secondary {
background: var(--accent);
}
.btn-small {
padding: 0.5rem 1rem;
font-size: 0.85rem;
box-shadow: 3px 3px 0 var(--bg);
}
.btn-small:hover:not(:disabled) {
box-shadow: 4px 4px 0 var(--bg);
}
.table-section {
animation: slideUp 0.6s cubic-bezier(0.68, -0.55, 0.265, 1.55) 0.4s both;
}
.table-wrapper {
overflow-x: auto;
background: var(--surface);
border: var(--border) solid var(--secondary);
box-shadow: 8px 8px 0 var(--secondary);
}
table {
width: 100%;
border-collapse: collapse;
}
thead {
background: linear-gradient(135deg, var(--primary), var(--accent));
}
th {
padding: 1.5rem 1rem;
text-align: left;
font-weight: 900;
text-transform: uppercase;
letter-spacing: 1px;
color: var(--bg);
border-right: 3px solid var(--bg);
}
th:last-child {
border-right: none;
}
tbody tr {
border-bottom: 2px solid rgba(0, 212, 255, 0.2);
transition: all 0.3s ease;
animation: fadeIn 0.5s ease;
}
@keyframes fadeIn {
from {
opacity: 0;
}
to {
opacity: 1;
}
}
tbody tr:hover {
background: rgba(0, 255, 136, 0.1);
}
td {
padding: 1.5rem 1rem;
color: var(--text);
}
.status {
display: inline-block;
padding: 0.5rem 1rem;
border: 3px solid;
font-weight: 900;
text-transform: uppercase;
font-size: 0.85rem;
letter-spacing: 1px;
}
.status-not_started {
background: var(--bg);
border-color: var(--accent);
color: var(--accent);
}
.status-processing {
background: var(--bg);
border-color: var(--primary);
color: var(--primary);
animation: pulse 1.5s ease-in-out infinite;
}
@keyframes pulse {
0%,
100% {
opacity: 1;
}
50% {
opacity: 0.6;
}
}
.status-completed {
background: var(--primary);
border-color: var(--primary);
color: var(--bg);
}
.status-failed {
background: var(--error);
border-color: var(--error);
color: var(--text);
}
.caption-cell {
max-width: 200px;
overflow: hidden;
text-overflow: ellipsis;
white-space: nowrap;
}
.empty-state {
text-align: center;
padding: 4rem 2rem;
color: var(--accent);
font-size: 1.2rem;
}
.refresh-btn {
position: fixed;
bottom: 2rem;
right: 2rem;
width: 60px;
height: 60px;
border-radius: 50%;
background: var(--secondary);
border: var(--border) solid var(--bg);
box-shadow: 4px 4px 0 var(--bg);
cursor: pointer;
transition: all 0.3s ease;
display: flex;
align-items: center;
justify-content: center;
font-size: 1.5rem;
z-index: 1000;
}
.refresh-btn:hover {
transform: rotate(180deg) scale(1.1);
box-shadow: 6px 6px 0 var(--bg);
}
/* Loader styles */
.loader-overlay {
position: fixed;
top: 0;
left: 0;
width: 100%;
height: 100%;
background: rgba(10, 14, 39, 0.95);
display: flex;
align-items: center;
justify-content: center;
z-index: 9999;
animation: fadeIn 0.3s ease;
}
.loader {
width: 80px;
height: 80px;
border: 6px solid var(--surface);
border-top: 6px solid var(--primary);
border-right: 6px solid var(--accent);
border-bottom: 6px solid var(--secondary);
border-radius: 50%;
animation: spin 1s linear infinite;
}
@keyframes spin {
0% {
transform: rotate(0deg);
}
100% {
transform: rotate(360deg);
}
}
.loader-text {
position: absolute;
margin-top: 120px;
font-size: 1.2rem;
font-weight: 900;
color: var(--primary);
text-transform: uppercase;
letter-spacing: 2px;
}
/* Modal styles */
.modal {
display: none;
position: fixed;
top: 0;
left: 0;
width: 100%;
height: 100%;
background: rgba(10, 14, 39, 0.95);
z-index: 2000;
animation: fadeIn 0.3s ease;
overflow-y: auto;
}
.modal.active {
display: flex;
align-items: center;
justify-content: center;
padding: 2rem;
}
.modal-content {
background: var(--surface);
border: var(--border) solid var(--primary);
box-shadow: 12px 12px 0 var(--primary);
max-width: 800px;
width: 100%;
max-height: 80vh;
position: relative;
animation: modalSlideIn 0.4s cubic-bezier(0.68, -0.55, 0.265, 1.55);
display: flex;
flex-direction: column;
}
@keyframes modalSlideIn {
from {
opacity: 0;
transform: translateY(-50px) scale(0.9);
}
to {
opacity: 1;
transform: translateY(0) scale(1);
}
}
.modal-header {
display: flex;
justify-content: space-between;
align-items: center;
padding: 2rem 2rem 1rem 2rem;
border-bottom: 3px solid var(--primary);
background: var(--surface);
position: sticky;
top: 0;
z-index: 10;
}
.modal-title {
font-size: 1.5rem;
font-weight: 900;
color: var(--primary);
text-transform: uppercase;
}
.modal-close {
background: var(--error);
color: var(--text);
border: 3px solid var(--bg);
width: 40px;
height: 40px;
border-radius: 0;
cursor: pointer;
font-size: 1.5rem;
font-weight: 900;
transition: all 0.2s ease;
box-shadow: 3px 3px 0 var(--bg);
}
.modal-close:hover {
transform: translate(-2px, -2px);
box-shadow: 5px 5px 0 var(--bg);
}
.code-block {
background: var(--bg);
border: 3px solid var(--accent);
padding: 1.5rem 1.5rem 1.5rem 1.5rem;
border-radius: 0;
overflow-y: auto;
margin: 1.5rem 2rem 2rem 2rem;
position: relative;
flex: 1;
}
.code-block code {
font-family: 'Courier New', monospace;
color: var(--primary);
font-size: 0.95rem;
line-height: 1.6;
white-space: pre-wrap;
word-break: break-all;
}
.copy-btn {
position: sticky;
top: 0.5rem;
float: right;
background: var(--accent);
color: var(--bg);
border: 3px solid var(--bg);
padding: 0.5rem 1rem;
font-size: 0.8rem;
font-weight: 900;
cursor: pointer;
transition: all 0.2s ease;
box-shadow: 3px 3px 0 var(--bg);
z-index: 5;
}
.copy-btn:hover {
transform: translate(-2px, -2px);
box-shadow: 4px 4px 0 var(--bg);
}
.copy-btn.copied {
background: var(--primary);
}
@media (max-width: 768px) {
.container {
padding: 1rem;
}
.upload-section,
.table-wrapper {
box-shadow: 4px 4px 0 var(--primary);
}
th,
td {
padding: 1rem 0.5rem;
font-size: 0.9rem;
}
.caption-cell {
max-width: 100px;
}
.modal-content {
padding: 0;
}
.modal-header {
padding: 1.5rem 1.5rem 1rem 1.5rem;
}
.code-block {
margin: 1.5rem 1.5rem 1.5rem 1.5rem;
}
}
.notification {
position: fixed;
top: 2rem;
right: 2rem;
padding: 1.5rem 2rem;
background: var(--primary);
color: var(--bg);
border: var(--border) solid var(--bg);
box-shadow: 6px 6px 0 var(--bg);
font-weight: 900;
z-index: 2000;
animation: slideInRight 0.5s ease, slideOutRight 0.5s ease 3.5s;
}
@keyframes slideInRight {
from {
transform: translateX(400px);
opacity: 0;
}
to {
transform: translateX(0);
opacity: 1;
}
}
@keyframes slideOutRight {
to {
transform: translateX(400px);
opacity: 0;
}
}
</style>
</head>
<body>
<div class="container">
<header>
<h1>Audio Caption Generator</h1>
<p class="subtitle">Transcribe • Process • Analyze</p>
</header>
<div class="upload-section">
<h2 style="margin-bottom: 1.5rem; color: var(--primary);">Upload Audio File</h2>
<div class="upload-zone" id="uploadZone">
<p style="font-size: 1.2rem; margin-bottom: 1rem;">📁 Drop audio file here or click to browse</p>
<p style="color: var(--accent); font-size: 0.9rem;">Supported: WAV, MP3, FLAC, OGG, M4A, AAC</p>
<input type="file" id="audioFile" accept=".wav,.mp3,.flac,.ogg,.m4a,.aac">
</div>
<button class="btn" id="uploadBtn" style="margin-top: 1.5rem; width: 100%;">
🚀 Upload & Process
</button>
</div>
<div class="table-section">
<h2 style="margin-bottom: 1.5rem; color: var(--secondary);">Processing Queue</h2>
<div class="table-wrapper">
<table>
<thead>
<tr>
<th>Filename</th>
<th>Status</th>
<th>Est. Wait</th>
<th>Caption</th>
<th>Created</th>
<th>Processed</th>
</tr>
</thead>
<tbody id="filesTable">
<tr>
<td colspan="6" class="empty-state">No files uploaded yet. Start by uploading an audio file!
</td>
</tr>
</tbody>
</table>
</div>
</div>
</div>
<button class="refresh-btn" id="refreshBtn" title="Refresh">🔄</button>
<!-- Loader -->
<div class="loader-overlay" id="loader" style="display: none;">
<div>
<div class="loader"></div>
<div class="loader-text">Uploading...</div>
</div>
</div>
<!-- Modal -->
<div class="modal" id="captionModal">
<div class="modal-content">
<div class="modal-header">
<div class="modal-title">📄 Caption Details</div>
<button class="modal-close" onclick="closeModal()">×</button>
</div>
<div class="code-block">
<button class="copy-btn" onclick="copyCaption()">📋 Copy</button>
<code id="captionCode"></code>
</div>
</div>
</div>
<script>
const API_URL = '/api';
let selectedFile = null;
// Store captions by file ID to avoid inline JSON issues with large texts
const captionStore = new Map();
// Upload zone interactions
const uploadZone = document.getElementById('uploadZone');
const fileInput = document.getElementById('audioFile');
const loader = document.getElementById('loader');
const uploadBtn = document.getElementById('uploadBtn');
uploadZone.addEventListener('click', () => fileInput.click());
uploadZone.addEventListener('dragover', (e) => {
e.preventDefault();
uploadZone.classList.add('dragging');
});
uploadZone.addEventListener('dragleave', () => {
uploadZone.classList.remove('dragging');
});
uploadZone.addEventListener('drop', (e) => {
e.preventDefault();
uploadZone.classList.remove('dragging');
const file = e.dataTransfer.files[0];
if (file) {
fileInput.files = e.dataTransfer.files;
selectedFile = file;
showNotification(`Selected: ${file.name}`);
}
});
fileInput.addEventListener('change', (e) => {
selectedFile = e.target.files[0];
if (selectedFile) {
showNotification(`Selected: ${selectedFile.name}`);
// Auto-upload if triggered by upload button click
if (pendingUpload) {
pendingUpload = false;
uploadFile();
}
}
});
// Flag to track if upload was triggered without file selection
let pendingUpload = false;
// Upload file function
async function uploadFile() {
if (!selectedFile) {
return;
}
const formData = new FormData();
formData.append('audio', selectedFile);
// Show loader
loader.style.display = 'flex';
uploadBtn.disabled = true;
try {
const response = await fetch(`${API_URL}/upload`, {
method: 'POST',
body: formData
});
const data = await response.json();
if (response.ok) {
showNotification('File uploaded successfully! 🎉');
selectedFile = null;
fileInput.value = '';
loadFiles();
} else {
showNotification(data.error || 'Upload failed', 'error');
}
} catch (error) {
showNotification('Network error: ' + error.message, 'error');
} finally {
// Hide loader
loader.style.display = 'none';
uploadBtn.disabled = false;
}
}
// Upload button - one-click flow
uploadBtn.addEventListener('click', async () => {
if (!selectedFile) {
// No file selected - trigger file picker and auto-upload after selection
pendingUpload = true;
fileInput.click();
return;
}
// File already selected - upload directly
uploadFile();
});
// Load files
async function loadFiles() {
try {
const response = await fetch(`${API_URL}/files`);
const files = await response.json();
const tbody = document.getElementById('filesTable');
// Clear old captions and store new ones
captionStore.clear();
if (files.length === 0) {
tbody.innerHTML = '<tr><td colspan="6" class="empty-state">No files uploaded yet. Start by uploading an audio file!</td></tr>';
return;
}
tbody.innerHTML = files.map(file => {
// Store caption in Map to avoid inline JSON issues with large texts
if (file.caption) {
captionStore.set(file.id, file.caption);
}
const captionPreview = file.caption ?
(file.caption.length > 50 ? file.caption.substring(0, 50) + '...' : file.caption) :
'—';
// Format estimated wait time
let estWait = '—';
if (file.status === 'not_started' && file.estimated_start_seconds !== null) {
const seconds = file.estimated_start_seconds;
if (seconds < 60) {
estWait = `${seconds}s`;
} else if (seconds < 3600) {
const mins = Math.floor(seconds / 60);
const secs = seconds % 60;
estWait = secs > 0 ? `${mins}m ${secs}s` : `${mins}m`;
} else {
const hours = Math.floor(seconds / 3600);
const mins = Math.floor((seconds % 3600) / 60);
estWait = mins > 0 ? `${hours}h ${mins}m` : `${hours}h`;
}
// Add queue position
if (file.queue_position) {
estWait = `#${file.queue_position} (${estWait})`;
}
} else if (file.status === 'processing') {
estWait = '⏳ Processing...';
}
return `
<tr>
<td><strong>${file.filename}</strong></td>
<td><span class="status status-${file.status}">${file.status.replace('_', ' ')}</span></td>
<td>${estWait}</td>
<td class="caption-cell">
${file.caption ?
`<button class="btn btn-small btn-secondary" onclick="showCaption('${file.id}')" style="margin-left: 0.5rem;">Show</button>`
: '—'}
</td>
<td>${new Date(file.created_at).toLocaleString()}</td>
<td>${file.processed_at ? new Date(file.processed_at).toLocaleString() : '—'}</td>
</tr>
`;
}).join('');
} catch (error) {
console.error('Error loading files:', error);
}
}
// Modal functions
function showCaption(fileId) {
const caption = captionStore.get(fileId);
if (!caption) {
console.error('Caption not found for file:', fileId);
return;
}
const modal = document.getElementById('captionModal');
const codeBlock = document.getElementById('captionCode');
// Parse JSON if it's a string
let formattedCaption = caption;
try {
const parsed = JSON.parse(caption);
formattedCaption = JSON.stringify(parsed, null, 2);
} catch (e) {
// If not JSON, use as is
formattedCaption = caption;
}
codeBlock.textContent = formattedCaption;
modal.classList.add('active');
}
function closeModal() {
const modal = document.getElementById('captionModal');
modal.classList.remove('active');
}
function copyCaption() {
const codeBlock = document.getElementById('captionCode');
const copyBtn = event.target;
navigator.clipboard.writeText(codeBlock.textContent).then(() => {
const originalText = copyBtn.textContent;
copyBtn.textContent = '✓ Copied!';
copyBtn.classList.add('copied');
setTimeout(() => {
copyBtn.textContent = originalText;
copyBtn.classList.remove('copied');
}, 2000);
});
}
// Close modal on background click
document.getElementById('captionModal').addEventListener('click', (e) => {
if (e.target.id === 'captionModal') {
closeModal();
}
});
// Close modal on ESC key
document.addEventListener('keydown', (e) => {
if (e.key === 'Escape') {
closeModal();
}
});
// Refresh button
document.getElementById('refreshBtn').addEventListener('click', loadFiles);
// Auto-refresh every 10 minutes
setInterval(loadFiles, 1000 * 60 * 10);
// Show notification
function showNotification(message, type = 'success') {
const notification = document.createElement('div');
notification.className = 'notification';
notification.textContent = message;
if (type === 'error') {
notification.style.background = 'var(--error)';
}
document.body.appendChild(notification);
setTimeout(() => {
notification.remove();
}, 4000);
}
// Initial load
loadFiles();
</script>
</body>
</html>