1230 lines
48 KiB
Python
1230 lines
48 KiB
Python
import argparse
|
|
import ssl
|
|
import os
|
|
import socket
|
|
import threading
|
|
import tempfile
|
|
from pathlib import Path
|
|
|
|
from flask import Flask, request, jsonify, send_file, Response
|
|
from flask_sock import Sock
|
|
from werkzeug.utils import secure_filename
|
|
import websocket as ws_client
|
|
import json
|
|
import logging
|
|
|
|
def format_time_srt(s):
|
|
hours = int(s // 3600)
|
|
minutes = int((s % 3600) // 60)
|
|
seconds = int(s % 60)
|
|
milliseconds = int((s - int(s)) * 1000)
|
|
return f"{hours:02}:{minutes:02}:{seconds:02},{milliseconds:03}"
|
|
|
|
def format_time_vtt(s):
|
|
hours = int(s // 3600)
|
|
minutes = int((s % 3600) // 60)
|
|
seconds = int(s % 60)
|
|
milliseconds = int((s - int(s)) * 1000)
|
|
return f"{hours:02}:{minutes:02}:{seconds:02}.{milliseconds:03}"
|
|
|
|
def generate_srt(segments):
|
|
output = ""
|
|
for i, segment in enumerate(segments, start=1):
|
|
start_time = format_time_srt(float(segment['start']))
|
|
end_time = format_time_srt(float(segment['end']))
|
|
text = segment['text'].strip()
|
|
output += f"{i}\n{start_time} --> {end_time}\n{text}\n\n"
|
|
return output
|
|
|
|
def generate_vtt(segments):
|
|
output = "WEBVTT\n\n"
|
|
for segment in segments:
|
|
start_time = format_time_vtt(float(segment['start']))
|
|
end_time = format_time_vtt(float(segment['end']))
|
|
text = segment['text'].strip()
|
|
output += f"{start_time} --> {end_time}\n{text}\n\n"
|
|
return output
|
|
|
|
# Configure logging
|
|
logging.basicConfig(level=logging.INFO)
|
|
logger = logging.getLogger(__name__)
|
|
|
|
def check_port_availability(port):
|
|
sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
|
|
result = sock.connect_ex(('0.0.0.0', port))
|
|
sock.close()
|
|
return result != 0
|
|
|
|
class HybridWhisperServer:
|
|
def __init__(self, websocket_port, http_port, backend="faster_whisper",
|
|
faster_whisper_custom_model_path=None, whisper_tensorrt_path=None,
|
|
trt_multilingual=False, single_model=True, ssl_context=None):
|
|
self.websocket_port = websocket_port
|
|
self.http_port = http_port
|
|
self.backend = backend
|
|
self.faster_whisper_custom_model_path = faster_whisper_custom_model_path
|
|
self.whisper_tensorrt_path = whisper_tensorrt_path
|
|
self.trt_multilingual = trt_multilingual
|
|
self.single_model = single_model
|
|
self.ssl_context = ssl_context
|
|
|
|
# Initialize Flask app
|
|
self.app = Flask(__name__)
|
|
self.app.config['MAX_CONTENT_LENGTH'] = 100 * 1024 * 1024 # 100MB max file size
|
|
self.sock = Sock(self.app)
|
|
self.setup_routes()
|
|
|
|
# Initialize WhisperLive server
|
|
from whisper_live.server import TranscriptionServer
|
|
self.whisper_server = TranscriptionServer()
|
|
|
|
# Create a shared transcriber instance for HTTP requests
|
|
self.shared_transcriber = None
|
|
if self.backend == "faster_whisper":
|
|
from whisper_live.transcriber import WhisperModel
|
|
# Use base model as default for HTTP requests
|
|
model_size = "base"
|
|
if self.faster_whisper_custom_model_path:
|
|
model_size = self.faster_whisper_custom_model_path
|
|
self.shared_transcriber = WhisperModel(model_size)
|
|
|
|
def setup_routes(self):
|
|
@self.app.route('/health', methods=['GET'])
|
|
def health_check():
|
|
return jsonify({'status': 'healthy', 'service': 'WhisperLive Hybrid Server'})
|
|
|
|
@self.app.route('/', methods=['GET'])
|
|
def serve_test_form():
|
|
"""Serve the HTML test form"""
|
|
html_content = """
|
|
<!DOCTYPE html>
|
|
<html lang="en">
|
|
|
|
<head>
|
|
<meta charset="UTF-8">
|
|
<meta name="viewport" content="width=device-width, initial-scale=1.0">
|
|
<title>WhisperLive Dashboard</title>
|
|
<link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap" rel="stylesheet">
|
|
<style>
|
|
:root {
|
|
--primary: #4f46e5;
|
|
--primary-hover: #4338ca;
|
|
--bg-color: #0f172a;
|
|
--card-bg: rgba(30, 41, 59, 0.7);
|
|
--text-main: #f8fafc;
|
|
--text-muted: #94a3b8;
|
|
--border: rgba(255, 255, 255, 0.1);
|
|
--success: #10b981;
|
|
--danger: #ef4444;
|
|
--warning: #f59e0b;
|
|
}
|
|
|
|
* {
|
|
box-sizing: border-box;
|
|
margin: 0;
|
|
padding: 0;
|
|
}
|
|
|
|
body {
|
|
font-family: 'Inter', sans-serif;
|
|
background-color: var(--bg-color);
|
|
color: var(--text-main);
|
|
min-height: 100vh;
|
|
background-image:
|
|
radial-gradient(at 0% 0%, rgba(79, 70, 229, 0.15) 0px, transparent 50%),
|
|
radial-gradient(at 100% 100%, rgba(16, 185, 129, 0.1) 0px, transparent 50%);
|
|
background-attachment: fixed;
|
|
padding: 2rem;
|
|
}
|
|
|
|
.container {
|
|
max-width: 1000px;
|
|
margin: 0 auto;
|
|
}
|
|
|
|
.header {
|
|
text-align: center;
|
|
margin-bottom: 2rem;
|
|
}
|
|
|
|
.header h1 {
|
|
font-size: 2.5rem;
|
|
font-weight: 700;
|
|
background: linear-gradient(to right, #818cf8, #34d399);
|
|
-webkit-background-clip: text;
|
|
-webkit-text-fill-color: transparent;
|
|
margin-bottom: 0.5rem;
|
|
}
|
|
|
|
.header p {
|
|
color: var(--text-muted);
|
|
}
|
|
|
|
.glass-panel {
|
|
background: var(--card-bg);
|
|
backdrop-filter: blur(12px);
|
|
-webkit-backdrop-filter: blur(12px);
|
|
border: 1px solid var(--border);
|
|
border-radius: 1rem;
|
|
padding: 1.5rem;
|
|
margin-bottom: 1.5rem;
|
|
box-shadow: 0 10px 25px -5px rgba(0, 0, 0, 0.3);
|
|
}
|
|
|
|
/* Config Section */
|
|
.config-grid {
|
|
display: grid;
|
|
grid-template-columns: 1fr 1fr;
|
|
gap: 1rem;
|
|
}
|
|
|
|
@media (max-width: 768px) {
|
|
.config-grid {
|
|
grid-template-columns: 1fr;
|
|
}
|
|
}
|
|
|
|
.form-group {
|
|
margin-bottom: 1rem;
|
|
}
|
|
|
|
.form-group label {
|
|
display: block;
|
|
font-size: 0.875rem;
|
|
font-weight: 500;
|
|
margin-bottom: 0.5rem;
|
|
color: var(--text-muted);
|
|
}
|
|
|
|
input[type="text"],
|
|
input[type="file"],
|
|
select {
|
|
width: 100%;
|
|
padding: 0.75rem 1rem;
|
|
background: rgba(15, 23, 42, 0.6);
|
|
border: 1px solid var(--border);
|
|
border-radius: 0.5rem;
|
|
color: var(--text-main);
|
|
font-size: 0.875rem;
|
|
transition: all 0.2s;
|
|
}
|
|
|
|
input[type="text"]:focus,
|
|
select:focus {
|
|
outline: none;
|
|
border-color: var(--primary);
|
|
box-shadow: 0 0 0 2px rgba(79, 70, 229, 0.2);
|
|
}
|
|
|
|
/* Tabs */
|
|
.tabs {
|
|
display: flex;
|
|
gap: 0.5rem;
|
|
margin-bottom: 1rem;
|
|
border-bottom: 1px solid var(--border);
|
|
padding-bottom: 0.5rem;
|
|
}
|
|
|
|
.tab-btn {
|
|
background: transparent;
|
|
border: none;
|
|
color: var(--text-muted);
|
|
padding: 0.75rem 1.5rem;
|
|
font-size: 1rem;
|
|
font-weight: 500;
|
|
cursor: pointer;
|
|
border-radius: 0.5rem;
|
|
transition: all 0.2s;
|
|
}
|
|
|
|
.tab-btn:hover {
|
|
color: var(--text-main);
|
|
background: rgba(255, 255, 255, 0.05);
|
|
}
|
|
|
|
.tab-btn.active {
|
|
color: var(--text-main);
|
|
background: var(--primary);
|
|
box-shadow: 0 4px 6px -1px rgba(79, 70, 229, 0.4);
|
|
}
|
|
|
|
.tab-content {
|
|
display: none;
|
|
animation: fadeIn 0.3s ease-in-out;
|
|
}
|
|
|
|
.tab-content.active {
|
|
display: block;
|
|
}
|
|
|
|
@keyframes fadeIn {
|
|
from {
|
|
opacity: 0;
|
|
transform: translateY(5px);
|
|
}
|
|
|
|
to {
|
|
opacity: 1;
|
|
transform: translateY(0);
|
|
}
|
|
}
|
|
|
|
/* Buttons */
|
|
.btn {
|
|
background: var(--primary);
|
|
color: white;
|
|
border: none;
|
|
padding: 0.75rem 1.5rem;
|
|
border-radius: 0.5rem;
|
|
font-weight: 600;
|
|
cursor: pointer;
|
|
transition: all 0.2s;
|
|
display: inline-flex;
|
|
align-items: center;
|
|
justify-content: center;
|
|
gap: 0.5rem;
|
|
width: 100%;
|
|
}
|
|
|
|
.btn:hover {
|
|
background: var(--primary-hover);
|
|
}
|
|
|
|
.btn:disabled {
|
|
opacity: 0.5;
|
|
cursor: not-allowed;
|
|
}
|
|
|
|
.btn-danger {
|
|
background: var(--danger);
|
|
}
|
|
|
|
.btn-danger:hover {
|
|
background: #dc2626;
|
|
}
|
|
|
|
.btn-success {
|
|
background: var(--success);
|
|
}
|
|
|
|
.btn-success:hover {
|
|
background: #059669;
|
|
}
|
|
|
|
/* Results / Live View */
|
|
.transcript-box {
|
|
background: rgba(15, 23, 42, 0.6);
|
|
border: 1px solid var(--border);
|
|
border-radius: 0.5rem;
|
|
padding: 1.5rem;
|
|
min-height: 200px;
|
|
max-height: 400px;
|
|
overflow-y: auto;
|
|
margin-top: 1rem;
|
|
line-height: 1.6;
|
|
}
|
|
|
|
.segment {
|
|
margin-bottom: 0.75rem;
|
|
padding-bottom: 0.75rem;
|
|
border-bottom: 1px solid rgba(255, 255, 255, 0.05);
|
|
}
|
|
|
|
.segment:last-child {
|
|
border-bottom: none;
|
|
margin-bottom: 0;
|
|
padding-bottom: 0;
|
|
}
|
|
|
|
.segment-time {
|
|
font-size: 0.75rem;
|
|
color: var(--primary);
|
|
font-weight: 600;
|
|
margin-bottom: 0.25rem;
|
|
}
|
|
|
|
.status-badge {
|
|
display: inline-flex;
|
|
align-items: center;
|
|
gap: 0.3rem;
|
|
padding: 0.25rem 0.75rem;
|
|
border-radius: 9999px;
|
|
font-size: 0.75rem;
|
|
font-weight: 600;
|
|
}
|
|
|
|
.status-offline {
|
|
background: rgba(239, 68, 68, 0.2);
|
|
color: #fca5a5;
|
|
}
|
|
|
|
.status-online {
|
|
background: rgba(16, 185, 129, 0.2);
|
|
color: #6ee7b7;
|
|
}
|
|
|
|
.status-recording {
|
|
background: rgba(239, 68, 68, 0.2);
|
|
color: #fca5a5;
|
|
animation: pulse 2s infinite;
|
|
}
|
|
|
|
@keyframes pulse {
|
|
0% {
|
|
box-shadow: 0 0 0 0 rgba(239, 68, 68, 0.4);
|
|
}
|
|
|
|
70% {
|
|
box-shadow: 0 0 0 10px rgba(239, 68, 68, 0);
|
|
}
|
|
|
|
100% {
|
|
box-shadow: 0 0 0 0 rgba(239, 68, 68, 0);
|
|
}
|
|
}
|
|
|
|
/* Code snippets */
|
|
pre {
|
|
background: #1e293b;
|
|
padding: 1rem;
|
|
border-radius: 0.5rem;
|
|
overflow-x: auto;
|
|
font-size: 0.875rem;
|
|
color: #e2e8f0;
|
|
border: 1px solid var(--border);
|
|
margin-bottom: 1rem;
|
|
}
|
|
|
|
code {
|
|
font-family: 'Courier New', Courier, monospace;
|
|
}
|
|
|
|
.loading-spinner {
|
|
display: none;
|
|
width: 24px;
|
|
height: 24px;
|
|
border: 3px solid rgba(255, 255, 255, 0.3);
|
|
border-radius: 50%;
|
|
border-top-color: white;
|
|
animation: spin 1s ease-in-out infinite;
|
|
}
|
|
|
|
@keyframes spin {
|
|
to {
|
|
transform: rotate(360deg);
|
|
}
|
|
}
|
|
</style>
|
|
</head>
|
|
|
|
<body>
|
|
<div class="container">
|
|
<div class="header">
|
|
<h1>WhisperLive</h1>
|
|
<p>High-Performance Real-Time Audio Transcription</p>
|
|
</div>
|
|
|
|
<!-- Configuration Panel -->
|
|
<div class="glass-panel">
|
|
<h3 style="margin-bottom: 1rem; font-size: 1.1rem;">Connection Settings</h3>
|
|
<div class="config-grid">
|
|
<div class="form-group">
|
|
<label>HTTP API URL (For File Upload & API)</label>
|
|
<input type="text" id="httpUrl" value="https://whisperlive.classroomcopilot.ai">
|
|
</div>
|
|
<div class="form-group">
|
|
<label>WebSocket URL (For Live Audio)</label>
|
|
<input type="text" id="wsUrl" value="wss://whisperlive.classroomcopilot.ai/ws">
|
|
</div>
|
|
</div>
|
|
<div style="margin-top: 0.5rem; font-size: 0.8rem; color: var(--text-muted);">
|
|
HTTP Status: <span id="httpStatus" class="status-badge status-offline">Checking...</span>
|
|
</div>
|
|
</div>
|
|
|
|
<!-- Main Workspace -->
|
|
<div class="glass-panel">
|
|
<div class="tabs">
|
|
<button class="tab-btn active" onclick="switchTab('file-tab')">File Upload</button>
|
|
<button class="tab-btn" onclick="switchTab('live-tab')">Live Microphone</button>
|
|
<button class="tab-btn" onclick="switchTab('api-tab')">API Usage</button>
|
|
</div>
|
|
|
|
<!-- Tab 1: File Upload -->
|
|
<div id="file-tab" class="tab-content active">
|
|
<form id="fileForm">
|
|
<div class="form-group">
|
|
<label>Audio File</label>
|
|
<input type="file" id="audioFile" accept=".wav,.mp3,.flac,.m4a,.ogg,.webm,.opus" required>
|
|
</div>
|
|
|
|
<div class="config-grid">
|
|
<div class="form-group">
|
|
<label>Language</label>
|
|
<select id="fileLanguage">
|
|
<option value="">Auto-detect</option>
|
|
<option value="en">English</option>
|
|
<option value="es">Spanish</option>
|
|
<option value="fr">French</option>
|
|
</select>
|
|
</div>
|
|
<div class="form-group">
|
|
<label>Task</label>
|
|
<select id="fileTask">
|
|
<option value="transcribe">Transcribe</option>
|
|
<option value="translate">Translate to English</option>
|
|
</select>
|
|
</div>
|
|
</div>
|
|
|
|
<button type="submit" class="btn" id="fileSubmitBtn">
|
|
<span>Transcribe File</span>
|
|
<div class="loading-spinner" id="fileSpinner"></div>
|
|
</button>
|
|
</form>
|
|
|
|
<div id="fileResult" style="display: none;">
|
|
<div class="transcript-box" id="fileTranscript"></div>
|
|
</div>
|
|
</div>
|
|
|
|
<!-- Tab 2: Live Recording -->
|
|
<div id="live-tab" class="tab-content">
|
|
<div class="config-grid" style="margin-bottom: 1.5rem;">
|
|
<div class="form-group">
|
|
<label>Language</label>
|
|
<select id="liveLanguage">
|
|
<option value="en">English</option>
|
|
<option value="es">Spanish</option>
|
|
<option value="fr">French</option>
|
|
</select>
|
|
</div>
|
|
<div class="form-group">
|
|
<label>Task</label>
|
|
<select id="liveTask">
|
|
<option value="transcribe">Transcribe</option>
|
|
<option value="translate">Translate to English</option>
|
|
</select>
|
|
</div>
|
|
</div>
|
|
|
|
<div style="display: flex; gap: 1rem; align-items: center;">
|
|
<button id="recordBtn" class="btn btn-success" style="width: auto;">
|
|
<span id="recordIcon">🎤</span> <span id="recordText">Start Recording</span>
|
|
</button>
|
|
<span id="liveStatus" class="status-badge status-offline" style="display: none;">Not
|
|
connected</span>
|
|
</div>
|
|
|
|
<div class="transcript-box" id="liveTranscript">
|
|
<div style="color: var(--text-muted); text-align: center; margin-top: 3rem;">
|
|
Click Start Recording to begin live transcription...
|
|
</div>
|
|
</div>
|
|
</div>
|
|
|
|
<!-- Tab 3: API Usage -->
|
|
<div id="api-tab" class="tab-content">
|
|
<h3 style="margin-bottom: 1rem;">OpenAI Compatible API</h3>
|
|
<p style="color: var(--text-muted); margin-bottom: 1rem; font-size: 0.9rem;">
|
|
WhisperLive acts as a drop-in replacement for OpenAI's Whisper API. You can use any standard OpenAI
|
|
client by changing the base URL.
|
|
</p>
|
|
|
|
<h4 style="margin-bottom: 0.5rem; color: #cbd5e1;">Python (openai package)</h4>
|
|
<pre><code id="pythonSnippet">from openai import OpenAI
|
|
|
|
client = OpenAI(
|
|
api_key="sk-no-key-required",
|
|
base_url="https://whisperlive.classroomcopilot.ai/v1/"
|
|
)
|
|
|
|
with open("audio.wav", "rb") as file:
|
|
transcription = client.audio.transcriptions.create(
|
|
file=file,
|
|
model="base",
|
|
response_format="verbose_json"
|
|
)
|
|
|
|
print(transcription.text)</code></pre>
|
|
|
|
<h4 style="margin-bottom: 0.5rem; color: #cbd5e1;">cURL</h4>
|
|
<pre><code id="curlSnippet">curl https://whisperlive.classroomcopilot.ai/v1/audio/transcriptions \
|
|
-H "Content-Type: multipart/form-data" \
|
|
-F file="@audio.wav" \
|
|
-F model="base" \
|
|
-F response_format="verbose_json"</code></pre>
|
|
</div>
|
|
</div>
|
|
</div>
|
|
|
|
<script>
|
|
// DOM Elements
|
|
const httpUrlInput = document.getElementById('httpUrl');
|
|
const wsUrlInput = document.getElementById('wsUrl');
|
|
const httpStatus = document.getElementById('httpStatus');
|
|
|
|
// Initialization
|
|
window.onload = () => {
|
|
// Check if on same domain to set default URL intelligently, else leave defaults
|
|
if (window.location.hostname !== '' && window.location.hostname !== 'localhost') {
|
|
httpUrlInput.value = window.location.origin;
|
|
wsUrlInput.value = window.location.origin.replace(/^http/, 'ws') + '/ws';
|
|
}
|
|
checkHealth();
|
|
updateSnippets();
|
|
};
|
|
|
|
httpUrlInput.addEventListener('change', () => { checkHealth(); updateSnippets(); });
|
|
|
|
// Tab Switching
|
|
function switchTab(tabId) {
|
|
document.querySelectorAll('.tab-content').forEach(t => t.classList.remove('active'));
|
|
document.querySelectorAll('.tab-btn').forEach(b => b.classList.remove('active'));
|
|
document.getElementById(tabId).classList.add('active');
|
|
event.target.classList.add('active');
|
|
}
|
|
|
|
// Health Check
|
|
async function checkHealth() {
|
|
try {
|
|
const res = await fetch(`${httpUrlInput.value}/health`);
|
|
if (res.ok) {
|
|
httpStatus.className = 'status-badge status-online';
|
|
httpStatus.textContent = '✅ Online';
|
|
} else throw new Error();
|
|
} catch (e) {
|
|
httpStatus.className = 'status-badge status-offline';
|
|
httpStatus.textContent = '❌ Offline';
|
|
}
|
|
}
|
|
|
|
// Update Code Snippets
|
|
function updateSnippets() {
|
|
const baseUrl = httpUrlInput.value.endsWith('/') ? httpUrlInput.value.slice(0, -1) : httpUrlInput.value;
|
|
document.getElementById('pythonSnippet').textContent = `from openai import OpenAI\n\nclient = OpenAI(\n api_key="sk-no-key-required",\n base_url="${baseUrl}/v1/"\n)\n\nwith open("audio.wav", "rb") as file:\n transcription = client.audio.transcriptions.create(\n file=file,\n model="base",\n response_format="verbose_json"\n )\n \nprint(transcription.text)`;
|
|
document.getElementById('curlSnippet').textContent = `curl ${baseUrl}/v1/audio/transcriptions \\\n -H "Content-Type: multipart/form-data" \\\n -F file="@audio.wav" \\\n -F model="base" \\\n -F response_format="verbose_json"`;
|
|
}
|
|
|
|
// Utility: Format Time
|
|
function formatTime(seconds) {
|
|
if (!seconds) return "0:00";
|
|
const mins = Math.floor(seconds / 60);
|
|
const secs = (seconds % 60).toFixed(2);
|
|
return `${mins}:${secs.padStart(5, '0')}`;
|
|
}
|
|
|
|
// ==========================================
|
|
// FEATURE 1: FILE TRANSCRIPTION
|
|
// ==========================================
|
|
document.getElementById('fileForm').addEventListener('submit', async (e) => {
|
|
e.preventDefault();
|
|
const file = document.getElementById('audioFile').files[0];
|
|
if (!file) return;
|
|
|
|
const btn = document.getElementById('fileSubmitBtn');
|
|
const spinner = document.getElementById('fileSpinner');
|
|
const resultBox = document.getElementById('fileResult');
|
|
const transcriptBox = document.getElementById('fileTranscript');
|
|
|
|
btn.disabled = true;
|
|
spinner.style.display = 'block';
|
|
resultBox.style.display = 'none';
|
|
|
|
const formData = new FormData();
|
|
formData.append('file', file);
|
|
formData.append('model', 'base');
|
|
formData.append('response_format', 'verbose_json');
|
|
|
|
const lang = document.getElementById('fileLanguage').value;
|
|
if (lang) formData.append('language', lang);
|
|
|
|
const task = document.getElementById('fileTask').value;
|
|
const baseUrl = httpUrlInput.value.endsWith('/') ? httpUrlInput.value.slice(0, -1) : httpUrlInput.value;
|
|
const endpoint = task === 'translate' ? `${baseUrl}/v1/audio/translations` : `${baseUrl}/v1/audio/transcriptions`;
|
|
|
|
try {
|
|
const response = await fetch(endpoint, { method: 'POST', body: formData });
|
|
const data = await response.json();
|
|
|
|
resultBox.style.display = 'block';
|
|
if (response.ok) {
|
|
let html = '';
|
|
if (data.segments && data.segments.length > 0) {
|
|
data.segments.forEach(seg => {
|
|
html += `<div class="segment"><div class="segment-time">${formatTime(seg.start)} - ${formatTime(seg.end)}</div><div class="segment-text">${seg.text}</div></div>`;
|
|
});
|
|
} else if (data.text) {
|
|
html += `<div class="segment"><div class="segment-text">${data.text}</div></div>`;
|
|
}
|
|
transcriptBox.innerHTML = html;
|
|
} else {
|
|
transcriptBox.innerHTML = `<div style="color: var(--danger)">Error: ${data.error?.message || JSON.stringify(data.error)}</div>`;
|
|
}
|
|
} catch (error) {
|
|
resultBox.style.display = 'block';
|
|
transcriptBox.innerHTML = `<div style="color: var(--danger)">Network Error: ${error.message}</div>`;
|
|
} finally {
|
|
btn.disabled = false;
|
|
spinner.style.display = 'none';
|
|
}
|
|
});
|
|
|
|
// ==========================================
|
|
// FEATURE 2: LIVE WEBSOCKET TRANSCRIPTION
|
|
// ==========================================
|
|
let ws = null;
|
|
let audioContext = null;
|
|
let mediaStream = null;
|
|
let processor = null;
|
|
let isRecording = false;
|
|
|
|
const recordBtn = document.getElementById('recordBtn');
|
|
const liveStatus = document.getElementById('liveStatus');
|
|
const liveTranscript = document.getElementById('liveTranscript');
|
|
|
|
recordBtn.addEventListener('click', async () => {
|
|
if (isRecording) {
|
|
stopRecording();
|
|
} else {
|
|
startRecording();
|
|
}
|
|
});
|
|
|
|
async function startRecording() {
|
|
liveTranscript.innerHTML = '';
|
|
liveStatus.style.display = 'inline-flex';
|
|
liveStatus.className = 'status-badge status-offline';
|
|
liveStatus.textContent = 'Connecting...';
|
|
|
|
try {
|
|
// 1. Connect WebSocket
|
|
ws = new WebSocket(wsUrlInput.value);
|
|
|
|
ws.onopen = () => {
|
|
// Send options to server
|
|
const options = {
|
|
uid: "web-" + Math.random().toString(36).substring(7),
|
|
language: document.getElementById('liveLanguage').value,
|
|
task: document.getElementById('liveTask').value,
|
|
model: "base",
|
|
use_vad: true
|
|
};
|
|
ws.send(JSON.stringify(options));
|
|
};
|
|
|
|
ws.onmessage = async (event) => {
|
|
const data = JSON.parse(event.data);
|
|
|
|
if (data.message === "SERVER_READY") {
|
|
liveStatus.className = 'status-badge status-recording';
|
|
liveStatus.innerHTML = '🔴 Recording';
|
|
await startAudioCapture();
|
|
} else if (data.segments) {
|
|
renderLiveSegments(data.segments);
|
|
} else if (data.status === "WAIT") {
|
|
liveStatus.textContent = `Waiting in queue (Est: ${data.message} min)`;
|
|
} else if (data.message === "DISCONNECT") {
|
|
stopRecording();
|
|
liveStatus.className = 'status-badge status-offline';
|
|
liveStatus.textContent = 'Disconnected by server';
|
|
}
|
|
};
|
|
|
|
ws.onerror = (err) => {
|
|
console.error('WebSocket Error', err);
|
|
stopRecording();
|
|
liveStatus.className = 'status-badge status-offline';
|
|
liveStatus.textContent = 'Connection Error';
|
|
};
|
|
|
|
ws.onclose = () => {
|
|
stopRecording();
|
|
};
|
|
|
|
// Update UI
|
|
isRecording = true;
|
|
recordBtn.className = 'btn btn-danger';
|
|
document.getElementById('recordIcon').textContent = '⏹';
|
|
document.getElementById('recordText').textContent = 'Stop Recording';
|
|
|
|
} catch (err) {
|
|
console.error(err);
|
|
liveStatus.className = 'status-badge status-offline';
|
|
liveStatus.textContent = 'Microphone Error';
|
|
stopRecording();
|
|
}
|
|
}
|
|
|
|
async function startAudioCapture() {
|
|
mediaStream = await navigator.mediaDevices.getUserMedia({ audio: true });
|
|
audioContext = new (window.AudioContext || window.webkitAudioContext)({ sampleRate: 16000 });
|
|
const source = audioContext.createMediaStreamSource(mediaStream);
|
|
|
|
// Create a ScriptProcessorNode with bufferSize of 4096 and a single input/output channel
|
|
processor = audioContext.createScriptProcessor(4096, 1, 1);
|
|
|
|
processor.onaudioprocess = function (e) {
|
|
if (!isRecording || ws.readyState !== WebSocket.OPEN) return;
|
|
const float32Array = e.inputBuffer.getChannelData(0);
|
|
ws.send(float32Array.buffer);
|
|
};
|
|
|
|
source.connect(processor);
|
|
processor.connect(audioContext.destination);
|
|
}
|
|
|
|
function stopRecording() {
|
|
isRecording = false;
|
|
|
|
if (processor) {
|
|
processor.disconnect();
|
|
processor = null;
|
|
}
|
|
if (mediaStream) {
|
|
mediaStream.getTracks().forEach(track => track.stop());
|
|
mediaStream = null;
|
|
}
|
|
if (audioContext) {
|
|
audioContext.close();
|
|
audioContext = null;
|
|
}
|
|
if (ws) {
|
|
if (ws.readyState === WebSocket.OPEN) {
|
|
ws.send("END_OF_AUDIO");
|
|
setTimeout(() => ws.close(), 1000);
|
|
}
|
|
ws = null;
|
|
}
|
|
|
|
recordBtn.className = 'btn btn-success';
|
|
document.getElementById('recordIcon').textContent = '🎤';
|
|
document.getElementById('recordText').textContent = 'Start Recording';
|
|
|
|
if (liveStatus.textContent === '🔴 Recording') {
|
|
liveStatus.className = 'status-badge status-offline';
|
|
liveStatus.textContent = 'Stopped';
|
|
}
|
|
}
|
|
|
|
let liveSegments = [];
|
|
function renderLiveSegments(segments) {
|
|
let html = '';
|
|
segments.forEach(seg => {
|
|
const timeHtml = (seg.start !== undefined && seg.end !== undefined)
|
|
? `<div class="segment-time">${formatTime(seg.start)} - ${formatTime(seg.end)}</div>`
|
|
: '';
|
|
html += `<div class="segment">${timeHtml}<div class="segment-text">${seg.text}</div></div>`;
|
|
});
|
|
liveTranscript.innerHTML = html;
|
|
liveTranscript.scrollTop = liveTranscript.scrollHeight;
|
|
}
|
|
|
|
</script>
|
|
</body>
|
|
|
|
</html>
|
|
"""
|
|
return html_content, 200, {'Content-Type': 'text/html'}
|
|
|
|
@self.app.route('/transcribe', methods=['POST'])
|
|
def transcribe_file():
|
|
try:
|
|
if 'file' not in request.files:
|
|
return jsonify({'error': 'No file provided'}), 400
|
|
|
|
file = request.files['file']
|
|
if file.filename == '':
|
|
return jsonify({'error': 'No file selected'}), 400
|
|
|
|
# Get optional parameters
|
|
language = request.form.get('language', None)
|
|
task = request.form.get('task', 'transcribe') # 'transcribe' or 'translate'
|
|
model_size = request.form.get('model', 'base')
|
|
|
|
# For now, we'll use the shared transcriber regardless of the requested model size
|
|
# In the future, we could create different transcriber instances for different models
|
|
|
|
# Validate file type
|
|
allowed_extensions = {'wav', 'mp3', 'flac', 'm4a', 'ogg', 'webm', 'opus', 'oga'}
|
|
if not file.filename.lower().endswith(tuple('.' + ext for ext in allowed_extensions)):
|
|
return jsonify({'error': f'Unsupported file type. Allowed: {", ".join(allowed_extensions)}'}), 400
|
|
|
|
# Save file temporarily
|
|
with tempfile.NamedTemporaryFile(delete=False, suffix=Path(file.filename).suffix) as temp_file:
|
|
file.save(temp_file.name)
|
|
temp_path = temp_file.name
|
|
|
|
try:
|
|
# Transcribe the file using WhisperLive
|
|
if self.backend == "faster_whisper":
|
|
# Use the shared transcriber instance
|
|
if self.shared_transcriber is None:
|
|
return jsonify({'error': 'Transcriber not initialized'}), 500
|
|
|
|
segments, info = self.shared_transcriber.transcribe(
|
|
temp_path,
|
|
language=language,
|
|
task=task
|
|
)
|
|
else:
|
|
# For other backends, use the server's transcriber
|
|
# This would need to be adapted based on your specific backend setup
|
|
return jsonify({'error': 'Backend not yet supported for file transcription'}), 501
|
|
|
|
# Convert segments to serializable format
|
|
transcript_segments = []
|
|
for segment in segments:
|
|
transcript_segments.append({
|
|
'start': segment.start,
|
|
'end': segment.end,
|
|
'text': segment.text,
|
|
'no_speech_prob': segment.no_speech_prob
|
|
})
|
|
|
|
# Get transcription info
|
|
transcription_info = {
|
|
'language': info.language,
|
|
'language_probability': info.language_probability,
|
|
'duration': info.duration,
|
|
'duration_after_vad': info.duration_after_vad,
|
|
'transcription_options': info.transcription_options
|
|
}
|
|
|
|
return jsonify({
|
|
'success': True,
|
|
'segments': transcript_segments,
|
|
'info': transcription_info,
|
|
'filename': file.filename
|
|
})
|
|
|
|
finally:
|
|
# Clean up temporary file
|
|
if os.path.exists(temp_path):
|
|
os.unlink(temp_path)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error transcribing file: {str(e)}")
|
|
return jsonify({'error': f'Transcription failed: {str(e)}'}), 500
|
|
|
|
@self.app.route('/transcribe/url', methods=['POST'])
|
|
def transcribe_url():
|
|
try:
|
|
data = request.get_json()
|
|
if not data or 'url' not in data:
|
|
return jsonify({'error': 'No URL provided'}), 400
|
|
|
|
url = data['url']
|
|
language = data.get('language', None)
|
|
task = data.get('task', 'transcribe')
|
|
model_size = data.get('model', 'base')
|
|
|
|
# Validate URL
|
|
if not url.startswith(('http://', 'https://', 'rtsp://', 'hls://')):
|
|
return jsonify({'error': 'Invalid URL format'}), 400
|
|
|
|
# For now, we'll return a message that this endpoint is available
|
|
# but the actual implementation would depend on your specific needs
|
|
return jsonify({
|
|
'message': 'URL transcription endpoint available',
|
|
'url': url,
|
|
'note': 'This endpoint is ready for implementation based on your specific requirements'
|
|
})
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error processing URL transcription request: {str(e)}")
|
|
return jsonify({'error': f'URL transcription failed: {str(e)}'}), 500
|
|
|
|
def handle_openai_audio_request(task_type):
|
|
try:
|
|
if 'file' not in request.files:
|
|
return jsonify({'error': {'message': 'No file provided', 'type': 'invalid_request_error', 'code': 'invalid_parameters'}}), 400
|
|
|
|
file = request.files['file']
|
|
if file.filename == '':
|
|
return jsonify({'error': {'message': 'No file selected', 'type': 'invalid_request_error', 'code': 'invalid_parameters'}}), 400
|
|
|
|
# Get OpenAI specific parameters
|
|
language = request.form.get('language', None)
|
|
model_size = request.form.get('model', 'base')
|
|
prompt = request.form.get('prompt', None)
|
|
response_format = request.form.get('response_format', 'json')
|
|
temperature = request.form.get('temperature', 0)
|
|
|
|
try:
|
|
temperature = float(temperature)
|
|
except ValueError:
|
|
temperature = 0.0
|
|
|
|
allowed_extensions = {'wav', 'mp3', 'flac', 'm4a', 'ogg', 'webm', 'mp4', 'mpeg', 'mpga', 'opus', 'oga'}
|
|
if not file.filename.lower().endswith(tuple('.' + ext for ext in allowed_extensions)):
|
|
return jsonify({'error': {'message': 'Unsupported file type.', 'type': 'invalid_request_error', 'code': 'invalid_file_format'}}), 400
|
|
|
|
with tempfile.NamedTemporaryFile(delete=False, suffix=Path(file.filename).suffix) as temp_file:
|
|
file.save(temp_file.name)
|
|
temp_path = temp_file.name
|
|
|
|
try:
|
|
if self.backend == "faster_whisper":
|
|
if self.shared_transcriber is None:
|
|
return jsonify({'error': {'message': 'Transcriber not initialized', 'type': 'internal_server_error'}}), 500
|
|
|
|
kwargs = {
|
|
"language": language,
|
|
"task": task_type,
|
|
"temperature": temperature
|
|
}
|
|
if prompt:
|
|
kwargs["initial_prompt"] = prompt
|
|
|
|
segments, info = self.shared_transcriber.transcribe(temp_path, **kwargs)
|
|
else:
|
|
return jsonify({'error': {'message': 'Backend not yet supported for file transcription', 'type': 'internal_server_error'}}), 501
|
|
|
|
transcript_segments = []
|
|
full_text = ""
|
|
for segment in segments:
|
|
text = segment.text
|
|
full_text += text
|
|
transcript_segments.append({
|
|
'id': segment.id,
|
|
'seek': segment.seek,
|
|
'start': segment.start,
|
|
'end': segment.end,
|
|
'text': text,
|
|
'tokens': segment.tokens,
|
|
'temperature': segment.temperature,
|
|
'avg_logprob': segment.avg_logprob,
|
|
'compression_ratio': segment.compression_ratio,
|
|
'no_speech_prob': segment.no_speech_prob
|
|
})
|
|
|
|
full_text = full_text.strip()
|
|
|
|
if response_format == 'json':
|
|
return jsonify({'text': full_text})
|
|
elif response_format == 'text':
|
|
return Response(full_text, mimetype='text/plain')
|
|
elif response_format == 'srt':
|
|
return Response(generate_srt(transcript_segments), mimetype='text/plain')
|
|
elif response_format == 'vtt':
|
|
return Response(generate_vtt(transcript_segments), mimetype='text/plain')
|
|
elif response_format == 'verbose_json':
|
|
return jsonify({
|
|
'task': task_type,
|
|
'language': info.language,
|
|
'duration': info.duration,
|
|
'text': full_text,
|
|
'segments': transcript_segments
|
|
})
|
|
else:
|
|
return jsonify({'text': full_text})
|
|
|
|
finally:
|
|
if os.path.exists(temp_path):
|
|
os.unlink(temp_path)
|
|
|
|
except Exception as e:
|
|
logger.error(f"Error processing OpenAI audio request: {str(e)}")
|
|
return jsonify({'error': {'message': f'Transcription failed: {str(e)}', 'type': 'internal_server_error'}}), 500
|
|
|
|
@self.app.route('/v1/audio/transcriptions', methods=['POST'])
|
|
def openai_transcriptions():
|
|
return handle_openai_audio_request('transcribe')
|
|
|
|
@self.app.route('/v1/audio/translations', methods=['POST'])
|
|
def openai_translations():
|
|
return handle_openai_audio_request('translate')
|
|
|
|
@self.app.route('/v1/models', methods=['GET'])
|
|
def list_models():
|
|
# Standard Whisper models supported by faster-whisper
|
|
model_names = [
|
|
"whisper-1", "tiny", "tiny.en", "base", "base.en",
|
|
"small", "small.en", "medium", "medium.en",
|
|
"large", "large-v1", "large-v2", "large-v3"
|
|
]
|
|
|
|
models = []
|
|
for name in model_names:
|
|
models.append({
|
|
"id": name,
|
|
"object": "model",
|
|
"created": 1677532384,
|
|
"owned_by": "openai" if name == "whisper-1" else "local",
|
|
"permission": [],
|
|
"root": name,
|
|
"parent": None
|
|
})
|
|
|
|
return jsonify({
|
|
"object": "list",
|
|
"data": models
|
|
})
|
|
|
|
# ===== WebSocket Bridge =====
|
|
# Bridges browser WebSocket connections on the HTTP port (8080)
|
|
# to the internal WhisperLive WebSocket server (port 5000).
|
|
# This allows live transcription through a single HTTPS port via NPM.
|
|
@self.sock.route('/ws')
|
|
def ws_bridge(ws):
|
|
"""Bridge WebSocket from HTTP port to internal WhisperLive WS server"""
|
|
internal_url = f"ws://127.0.0.1:{self.websocket_port}"
|
|
logger.info(f"WebSocket bridge: new connection, proxying to {internal_url}")
|
|
|
|
internal = None
|
|
try:
|
|
internal = ws_client.create_connection(internal_url)
|
|
|
|
# Thread: internal server → browser
|
|
def server_to_browser():
|
|
try:
|
|
while True:
|
|
opcode, data = internal.recv_data()
|
|
if opcode == ws_client.ABNF.OPCODE_TEXT:
|
|
ws.send(data.decode('utf-8'))
|
|
elif opcode == ws_client.ABNF.OPCODE_BINARY:
|
|
ws.send(data)
|
|
elif opcode in (ws_client.ABNF.OPCODE_CLOSE, ):
|
|
break
|
|
except Exception:
|
|
pass
|
|
|
|
relay_thread = threading.Thread(target=server_to_browser, daemon=True)
|
|
relay_thread.start()
|
|
|
|
# Main thread: browser → internal server
|
|
while True:
|
|
data = ws.receive()
|
|
if data is None:
|
|
break
|
|
if isinstance(data, bytes):
|
|
internal.send_binary(data)
|
|
else:
|
|
if data == "END_OF_AUDIO":
|
|
internal.send_binary(b"END_OF_AUDIO")
|
|
else:
|
|
internal.send(data)
|
|
|
|
except Exception as e:
|
|
logger.error(f"WebSocket bridge error: {e}")
|
|
finally:
|
|
if internal:
|
|
try:
|
|
internal.close()
|
|
except Exception:
|
|
pass
|
|
logger.info("WebSocket bridge: connection closed")
|
|
|
|
def run_websocket_server(self):
|
|
"""Run the WebSocket server in a separate thread"""
|
|
logger.info(f"Starting WebSocket server on port {self.websocket_port}")
|
|
self.whisper_server.run(
|
|
"0.0.0.0",
|
|
port=self.websocket_port,
|
|
backend=self.backend,
|
|
faster_whisper_custom_model_path=self.faster_whisper_custom_model_path,
|
|
whisper_tensorrt_path=self.whisper_tensorrt_path,
|
|
trt_multilingual=self.trt_multilingual,
|
|
single_model=self.single_model,
|
|
ssl_context=self.ssl_context
|
|
)
|
|
|
|
def run_http_server(self):
|
|
"""Run the HTTP server"""
|
|
logger.info(f"Starting HTTP server on port {self.http_port}")
|
|
self.app.run(host='0.0.0.0', port=self.http_port, debug=False, threaded=True)
|
|
|
|
def start(self):
|
|
"""Start both servers"""
|
|
# Start WebSocket server in a separate thread
|
|
websocket_thread = threading.Thread(target=self.run_websocket_server, daemon=True)
|
|
websocket_thread.start()
|
|
|
|
# Start HTTP server in main thread
|
|
self.run_http_server()
|
|
|
|
if __name__ == "__main__":
|
|
parser = argparse.ArgumentParser(description='WhisperLive Hybrid Server (WebSocket + HTTP)')
|
|
parser.add_argument('--websocket-port', '-wp',
|
|
type=int,
|
|
default=int(os.getenv('PORT_WHISPERLIVE', 9090)),
|
|
help="WebSocket port to run the server on.")
|
|
parser.add_argument('--http-port', '-hp',
|
|
type=int,
|
|
default=int(os.getenv('HTTP_PORT', 8080)),
|
|
help="HTTP port to run the server on.")
|
|
parser.add_argument('--backend', '-b',
|
|
type=str,
|
|
default='faster_whisper',
|
|
help='Backends from ["tensorrt", "faster_whisper"]')
|
|
parser.add_argument('--faster_whisper_custom_model_path', '-fw',
|
|
type=str, default=None,
|
|
help="Custom Faster Whisper Model")
|
|
parser.add_argument('--trt_model_path', '-trt',
|
|
type=str,
|
|
default=None,
|
|
help='Whisper TensorRT model path')
|
|
parser.add_argument('--trt_multilingual', '-m',
|
|
action="store_true",
|
|
help='Boolean only for TensorRT model. True if multilingual.')
|
|
parser.add_argument('--ssl_cert_path', '-ssl',
|
|
type=str,
|
|
default=None,
|
|
help='Path to cert.pem and key.pem if ssl should be used.')
|
|
parser.add_argument('--omp_num_threads', '-omp',
|
|
type=int,
|
|
default=1,
|
|
help="Number of threads to use for OpenMP")
|
|
parser.add_argument('--no_single_model', '-nsm',
|
|
action='store_true',
|
|
help='Set this if every connection should instantiate its own model. Only relevant for custom model, passed using -trt or -fw.')
|
|
|
|
args = parser.parse_args()
|
|
|
|
if args.backend == "tensorrt":
|
|
if args.trt_model_path is None:
|
|
raise ValueError("Please Provide a valid tensorrt model path")
|
|
|
|
websocket_port = args.websocket_port
|
|
http_port = args.http_port
|
|
|
|
if not check_port_availability(websocket_port):
|
|
print(f"Warning: WebSocket port {websocket_port} might already be in use!")
|
|
if not check_port_availability(http_port):
|
|
print(f"Warning: HTTP port {http_port} might already be in use!")
|
|
|
|
ssl_context = None
|
|
if args.ssl_cert_path is not None:
|
|
try:
|
|
ssl_context = ssl.SSLContext(ssl.PROTOCOL_TLS_SERVER)
|
|
ssl_context.load_cert_chain(
|
|
certfile=f"{args.ssl_cert_path}/cert.pem",
|
|
keyfile=f"{args.ssl_cert_path}/privkey.pem"
|
|
)
|
|
print("SSL context created successfully")
|
|
except Exception as e:
|
|
print(f"Failed to load SSL certificates: {str(e)}")
|
|
raise
|
|
|
|
if "OMP_NUM_THREADS" not in os.environ:
|
|
print(f"Setting OMP_NUM_THREADS to {args.omp_num_threads}")
|
|
os.environ["OMP_NUM_THREADS"] = str(args.omp_num_threads)
|
|
|
|
print(f"Running hybrid server with args: {args}")
|
|
server = HybridWhisperServer(
|
|
websocket_port=websocket_port,
|
|
http_port=http_port,
|
|
backend=args.backend,
|
|
faster_whisper_custom_model_path=args.faster_whisper_custom_model_path,
|
|
whisper_tensorrt_path=args.trt_model_path,
|
|
trt_multilingual=args.trt_multilingual,
|
|
single_model=not args.no_single_model,
|
|
ssl_context=ssl_context
|
|
)
|
|
|
|
print(f"Starting hybrid server with WebSocket on port {websocket_port} and HTTP on port {http_port}")
|
|
print(f"Backend: {args.backend}, SSL: {args.ssl_cert_path is not None}")
|
|
|
|
server.start()
|