Remove YouTube functionality completely - file upload only
Browse files- Dockerfile +4 -13
- app.py +25 -52
- requirements-flexible.txt +1 -1
- requirements.txt +1 -1
- static/js/script.js +3 -25
- templates/index.html +2 -11
- utils/youtube_downloader.py +0 -422
Dockerfile
CHANGED
|
@@ -15,12 +15,6 @@ RUN apt-get update && apt-get install -y \
|
|
| 15 |
libavcodec-dev \
|
| 16 |
libavformat-dev \
|
| 17 |
libswscale-dev \
|
| 18 |
-
curl \
|
| 19 |
-
wget \
|
| 20 |
-
ca-certificates \
|
| 21 |
-
dnsutils \
|
| 22 |
-
iputils-ping \
|
| 23 |
-
net-tools \
|
| 24 |
&& rm -rf /var/lib/apt/lists/*
|
| 25 |
|
| 26 |
# Copy requirements first to leverage Docker cache
|
|
@@ -58,13 +52,10 @@ ENV HF_HOME=/tmp/cache/huggingface
|
|
| 58 |
# DNS and network settings
|
| 59 |
ENV PYTHONDONTWRITEBYTECODE=1
|
| 60 |
ENV PYTHONPATH=/app
|
| 61 |
-
#
|
| 62 |
-
ENV
|
| 63 |
-
ENV
|
| 64 |
-
ENV
|
| 65 |
-
# Network and DNS environment variables
|
| 66 |
-
ENV PYTHONHTTPSVERIFY=0
|
| 67 |
-
ENV CURL_CA_BUNDLE=/etc/ssl/certs/ca-certificates.crt
|
| 68 |
|
| 69 |
# Expose port
|
| 70 |
EXPOSE 5000
|
|
|
|
| 15 |
libavcodec-dev \
|
| 16 |
libavformat-dev \
|
| 17 |
libswscale-dev \
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
&& rm -rf /var/lib/apt/lists/*
|
| 19 |
|
| 20 |
# Copy requirements first to leverage Docker cache
|
|
|
|
| 52 |
# DNS and network settings
|
| 53 |
ENV PYTHONDONTWRITEBYTECODE=1
|
| 54 |
ENV PYTHONPATH=/app
|
| 55 |
+
# Basic environment variables
|
| 56 |
+
ENV PYTHONUNBUFFERED=1
|
| 57 |
+
ENV FLASK_APP=app.py
|
| 58 |
+
ENV FLASK_ENV=production
|
|
|
|
|
|
|
|
|
|
| 59 |
|
| 60 |
# Expose port
|
| 61 |
EXPOSE 5000
|
app.py
CHANGED
|
@@ -25,24 +25,8 @@ os.environ['XDG_CACHE_HOME'] = '/tmp/cache'
|
|
| 25 |
os.environ['PYTHONDONTWRITEBYTECODE'] = '1'
|
| 26 |
os.environ['PYTHONPATH'] = '/app'
|
| 27 |
|
| 28 |
-
#
|
| 29 |
-
|
| 30 |
-
from utils.youtube_downloader import test_network_connectivity, check_dns_resolution, setup_network_environment
|
| 31 |
-
|
| 32 |
-
# Setup network environment first
|
| 33 |
-
setup_network_environment()
|
| 34 |
-
|
| 35 |
-
if test_network_connectivity():
|
| 36 |
-
logger.info("✅ Network connectivity test passed")
|
| 37 |
-
else:
|
| 38 |
-
logger.warning("⚠️ Network connectivity test failed")
|
| 39 |
-
|
| 40 |
-
if check_dns_resolution():
|
| 41 |
-
logger.info("✅ DNS resolution test passed")
|
| 42 |
-
else:
|
| 43 |
-
logger.warning("⚠️ DNS resolution test failed")
|
| 44 |
-
except Exception as e:
|
| 45 |
-
logger.warning(f"Could not run network tests: {e}")
|
| 46 |
|
| 47 |
# Create cache directories if they don't exist
|
| 48 |
os.makedirs('/tmp/cache/torch', exist_ok=True)
|
|
@@ -55,7 +39,8 @@ import uuid
|
|
| 55 |
from werkzeug.utils import secure_filename
|
| 56 |
|
| 57 |
from models.video_processor import VideoProcessor
|
| 58 |
-
|
|
|
|
| 59 |
|
| 60 |
# Configure clean logging
|
| 61 |
# logging.basicConfig(
|
|
@@ -106,10 +91,11 @@ def allowed_file(filename):
|
|
| 106 |
return '.' in filename and \
|
| 107 |
filename.rsplit('.', 1)[1].lower() in ALLOWED_EXTENSIONS
|
| 108 |
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
|
|
|
|
| 113 |
|
| 114 |
@app.route('/')
|
| 115 |
def index():
|
|
@@ -118,7 +104,7 @@ def index():
|
|
| 118 |
|
| 119 |
@app.route('/upload', methods=['POST'])
|
| 120 |
def upload_video():
|
| 121 |
-
"""Handle video upload
|
| 122 |
global video_processor
|
| 123 |
|
| 124 |
# Check if video processor is available, try to initialize if not
|
|
@@ -137,36 +123,23 @@ def upload_video():
|
|
| 137 |
try:
|
| 138 |
video_path = None
|
| 139 |
|
| 140 |
-
#
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
try:
|
| 148 |
-
video_path = download_youtube_video(youtube_url, app.config['UPLOAD_FOLDER'])
|
| 149 |
-
except Exception as e:
|
| 150 |
-
logger.error(f"YouTube download error: {str(e)}")
|
| 151 |
-
return jsonify({'error': f'Failed to download YouTube video: {str(e)}'}), 400
|
| 152 |
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
# Generate unique filename
|
| 161 |
-
filename = secure_filename(file.filename)
|
| 162 |
-
unique_filename = f"{uuid.uuid4()}_{filename}"
|
| 163 |
-
video_path = os.path.join(app.config['UPLOAD_FOLDER'], unique_filename)
|
| 164 |
-
file.save(video_path)
|
| 165 |
-
logger.info(f"File uploaded: {video_path}")
|
| 166 |
-
else:
|
| 167 |
-
return jsonify({'error': 'Invalid file format'}), 400
|
| 168 |
else:
|
| 169 |
-
return jsonify({'error': '
|
| 170 |
|
| 171 |
# Process video
|
| 172 |
logger.info(f"Starting video processing: {video_path}")
|
|
|
|
| 25 |
os.environ['PYTHONDONTWRITEBYTECODE'] = '1'
|
| 26 |
os.environ['PYTHONPATH'] = '/app'
|
| 27 |
|
| 28 |
+
# Network tests removed - YouTube functionality disabled
|
| 29 |
+
logger.info("📹 YouTube download functionality disabled")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
# Create cache directories if they don't exist
|
| 32 |
os.makedirs('/tmp/cache/torch', exist_ok=True)
|
|
|
|
| 39 |
from werkzeug.utils import secure_filename
|
| 40 |
|
| 41 |
from models.video_processor import VideoProcessor
|
| 42 |
+
# YouTube downloader removed
|
| 43 |
+
# from utils.youtube_downloader import download_youtube_video
|
| 44 |
|
| 45 |
# Configure clean logging
|
| 46 |
# logging.basicConfig(
|
|
|
|
| 91 |
return '.' in filename and \
|
| 92 |
filename.rsplit('.', 1)[1].lower() in ALLOWED_EXTENSIONS
|
| 93 |
|
| 94 |
+
# YouTube functionality removed
|
| 95 |
+
# def is_youtube_url(url):
|
| 96 |
+
# """Check if URL is a YouTube video"""
|
| 97 |
+
# youtube_domains = ['youtube.com', 'youtu.be', 'www.youtube.com', 'm.youtube.com']
|
| 98 |
+
# return any(domain in url.lower() for domain in youtube_domains)
|
| 99 |
|
| 100 |
@app.route('/')
|
| 101 |
def index():
|
|
|
|
| 104 |
|
| 105 |
@app.route('/upload', methods=['POST'])
|
| 106 |
def upload_video():
|
| 107 |
+
"""Handle video upload only (YouTube functionality disabled)"""
|
| 108 |
global video_processor
|
| 109 |
|
| 110 |
# Check if video processor is available, try to initialize if not
|
|
|
|
| 123 |
try:
|
| 124 |
video_path = None
|
| 125 |
|
| 126 |
+
# Handle file upload only
|
| 127 |
+
if 'video' not in request.files:
|
| 128 |
+
return jsonify({'error': 'No video file provided'}), 400
|
| 129 |
+
|
| 130 |
+
file = request.files['video']
|
| 131 |
+
if file.filename == '':
|
| 132 |
+
return jsonify({'error': 'No file selected'}), 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 133 |
|
| 134 |
+
if file and allowed_file(file.filename):
|
| 135 |
+
# Generate unique filename
|
| 136 |
+
filename = secure_filename(file.filename)
|
| 137 |
+
unique_filename = f"{uuid.uuid4()}_{filename}"
|
| 138 |
+
video_path = os.path.join(app.config['UPLOAD_FOLDER'], unique_filename)
|
| 139 |
+
file.save(video_path)
|
| 140 |
+
logger.info(f"File uploaded: {video_path}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 141 |
else:
|
| 142 |
+
return jsonify({'error': 'Invalid file format. Please upload MP4, AVI, MOV, MKV, or other supported formats.'}), 400
|
| 143 |
|
| 144 |
# Process video
|
| 145 |
logger.info(f"Starting video processing: {video_path}")
|
requirements-flexible.txt
CHANGED
|
@@ -19,7 +19,7 @@ numpy>=1.24.0,<2.0.0
|
|
| 19 |
pandas>=2.0.0
|
| 20 |
|
| 21 |
# YouTube Downloads
|
| 22 |
-
yt-dlp
|
| 23 |
|
| 24 |
# Utilities
|
| 25 |
tqdm>=4.65.0
|
|
|
|
| 19 |
pandas>=2.0.0
|
| 20 |
|
| 21 |
# YouTube Downloads
|
| 22 |
+
# yt-dlp removed - YouTube functionality disabled
|
| 23 |
|
| 24 |
# Utilities
|
| 25 |
tqdm>=4.65.0
|
requirements.txt
CHANGED
|
@@ -19,7 +19,7 @@ numpy==1.26.0
|
|
| 19 |
pandas==2.1.0
|
| 20 |
|
| 21 |
# YouTube Downloads
|
| 22 |
-
yt-dlp
|
| 23 |
|
| 24 |
# Utilities
|
| 25 |
tqdm==4.66.1
|
|
|
|
| 19 |
pandas==2.1.0
|
| 20 |
|
| 21 |
# YouTube Downloads
|
| 22 |
+
# yt-dlp removed - YouTube functionality disabled
|
| 23 |
|
| 24 |
# Utilities
|
| 25 |
tqdm==4.66.1
|
static/js/script.js
CHANGED
|
@@ -6,7 +6,6 @@ document.addEventListener('DOMContentLoaded', function() {
|
|
| 6 |
const resultsSection = document.getElementById('results');
|
| 7 |
const errorSection = document.getElementById('error');
|
| 8 |
const videoInput = document.getElementById('video');
|
| 9 |
-
const youtubeInput = document.getElementById('youtube_url');
|
| 10 |
|
| 11 |
// Exercise names for display
|
| 12 |
const ACTIONS = [
|
|
@@ -17,28 +16,14 @@ document.addEventListener('DOMContentLoaded', function() {
|
|
| 17 |
"shoulder press", "squat", "t bar row", "tricep Pushdown", "tricep dips"
|
| 18 |
];
|
| 19 |
|
| 20 |
-
// Clear other input when one is selected
|
| 21 |
-
videoInput.addEventListener('change', function() {
|
| 22 |
-
if (this.files.length > 0) {
|
| 23 |
-
youtubeInput.value = '';
|
| 24 |
-
}
|
| 25 |
-
});
|
| 26 |
-
|
| 27 |
-
youtubeInput.addEventListener('input', function() {
|
| 28 |
-
if (this.value.trim()) {
|
| 29 |
-
videoInput.value = '';
|
| 30 |
-
}
|
| 31 |
-
});
|
| 32 |
-
|
| 33 |
uploadForm.addEventListener('submit', async function(e) {
|
| 34 |
e.preventDefault();
|
| 35 |
|
| 36 |
// Validate inputs
|
| 37 |
const hasFile = videoInput.files.length > 0;
|
| 38 |
-
const hasYouTube = youtubeInput.value.trim();
|
| 39 |
|
| 40 |
-
if (!hasFile
|
| 41 |
-
showError('Please select a video file
|
| 42 |
return;
|
| 43 |
}
|
| 44 |
|
|
@@ -49,14 +34,7 @@ document.addEventListener('DOMContentLoaded', function() {
|
|
| 49 |
|
| 50 |
try {
|
| 51 |
const formData = new FormData();
|
| 52 |
-
|
| 53 |
-
if (hasFile) {
|
| 54 |
-
formData.append('video', videoInput.files[0]);
|
| 55 |
-
}
|
| 56 |
-
|
| 57 |
-
if (hasYouTube) {
|
| 58 |
-
formData.append('youtube_url', youtubeInput.value.trim());
|
| 59 |
-
}
|
| 60 |
|
| 61 |
const response = await fetch('/upload', {
|
| 62 |
method: 'POST',
|
|
|
|
| 6 |
const resultsSection = document.getElementById('results');
|
| 7 |
const errorSection = document.getElementById('error');
|
| 8 |
const videoInput = document.getElementById('video');
|
|
|
|
| 9 |
|
| 10 |
// Exercise names for display
|
| 11 |
const ACTIONS = [
|
|
|
|
| 16 |
"shoulder press", "squat", "t bar row", "tricep Pushdown", "tricep dips"
|
| 17 |
];
|
| 18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
uploadForm.addEventListener('submit', async function(e) {
|
| 20 |
e.preventDefault();
|
| 21 |
|
| 22 |
// Validate inputs
|
| 23 |
const hasFile = videoInput.files.length > 0;
|
|
|
|
| 24 |
|
| 25 |
+
if (!hasFile) {
|
| 26 |
+
showError('Please select a video file to upload.');
|
| 27 |
return;
|
| 28 |
}
|
| 29 |
|
|
|
|
| 34 |
|
| 35 |
try {
|
| 36 |
const formData = new FormData();
|
| 37 |
+
formData.append('video', videoInput.files[0]);
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
|
| 39 |
const response = await fetch('/upload', {
|
| 40 |
method: 'POST',
|
templates/index.html
CHANGED
|
@@ -10,26 +10,17 @@
|
|
| 10 |
<div class="container">
|
| 11 |
<header>
|
| 12 |
<h1>🏋️ Gym Exercise Recognition</h1>
|
| 13 |
-
<p>Upload a video
|
| 14 |
</header>
|
| 15 |
|
| 16 |
<div class="upload-section">
|
| 17 |
<form id="uploadForm" enctype="multipart/form-data">
|
| 18 |
<div class="input-group">
|
| 19 |
<label for="video">Upload Video File:</label>
|
| 20 |
-
<input type="file" id="video" name="video" accept="video/*">
|
| 21 |
<small>Supported formats: MP4, AVI, MOV, MKV, WebM, WMV, FLV, M4V, 3GP</small>
|
| 22 |
</div>
|
| 23 |
|
| 24 |
-
<div class="or-divider">
|
| 25 |
-
<span>OR</span>
|
| 26 |
-
</div>
|
| 27 |
-
|
| 28 |
-
<div class="input-group">
|
| 29 |
-
<label for="youtube_url">YouTube URL:</label>
|
| 30 |
-
<input type="url" id="youtube_url" name="youtube_url" placeholder="https://www.youtube.com/watch?v=...">
|
| 31 |
-
</div>
|
| 32 |
-
|
| 33 |
<button type="submit" id="submitBtn">
|
| 34 |
<span id="btnText">Analyze Video</span>
|
| 35 |
<span id="loadingSpinner" class="spinner" style="display: none;"></span>
|
|
|
|
| 10 |
<div class="container">
|
| 11 |
<header>
|
| 12 |
<h1>🏋️ Gym Exercise Recognition</h1>
|
| 13 |
+
<p>Upload a video file to recognize gym exercises</p>
|
| 14 |
</header>
|
| 15 |
|
| 16 |
<div class="upload-section">
|
| 17 |
<form id="uploadForm" enctype="multipart/form-data">
|
| 18 |
<div class="input-group">
|
| 19 |
<label for="video">Upload Video File:</label>
|
| 20 |
+
<input type="file" id="video" name="video" accept="video/*" required>
|
| 21 |
<small>Supported formats: MP4, AVI, MOV, MKV, WebM, WMV, FLV, M4V, 3GP</small>
|
| 22 |
</div>
|
| 23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
<button type="submit" id="submitBtn">
|
| 25 |
<span id="btnText">Analyze Video</span>
|
| 26 |
<span id="loadingSpinner" class="spinner" style="display: none;"></span>
|
utils/youtube_downloader.py
DELETED
|
@@ -1,422 +0,0 @@
|
|
| 1 |
-
import os
|
| 2 |
-
import re
|
| 3 |
-
import uuid
|
| 4 |
-
import logging
|
| 5 |
-
import socket
|
| 6 |
-
from urllib.parse import urlparse, parse_qs
|
| 7 |
-
|
| 8 |
-
logger = logging.getLogger(__name__)
|
| 9 |
-
|
| 10 |
-
def check_dns_resolution():
|
| 11 |
-
"""Check if DNS resolution is working"""
|
| 12 |
-
try:
|
| 13 |
-
# Try multiple DNS servers
|
| 14 |
-
dns_servers = ['8.8.8.8', '1.1.1.1', '208.67.222.222', '9.9.9.9']
|
| 15 |
-
|
| 16 |
-
for dns_server in dns_servers:
|
| 17 |
-
try:
|
| 18 |
-
# Try to resolve YouTube with specific timeout
|
| 19 |
-
socket.setdefaulttimeout(10)
|
| 20 |
-
ip = socket.gethostbyname('www.youtube.com')
|
| 21 |
-
logger.info(f"DNS resolution successful: www.youtube.com -> {ip}")
|
| 22 |
-
return True
|
| 23 |
-
except Exception as e:
|
| 24 |
-
logger.warning(f"DNS server {dns_server} failed: {e}")
|
| 25 |
-
continue
|
| 26 |
-
|
| 27 |
-
logger.error("All DNS servers failed")
|
| 28 |
-
return False
|
| 29 |
-
|
| 30 |
-
except Exception as e:
|
| 31 |
-
logger.error(f"DNS resolution failed: {e}")
|
| 32 |
-
return False
|
| 33 |
-
|
| 34 |
-
def setup_network_environment():
|
| 35 |
-
"""Setup network environment variables"""
|
| 36 |
-
try:
|
| 37 |
-
import os
|
| 38 |
-
|
| 39 |
-
# Set DNS servers via environment
|
| 40 |
-
os.environ['DNS_SERVERS'] = '8.8.8.8 1.1.1.1 208.67.222.222'
|
| 41 |
-
|
| 42 |
-
# Disable SSL verification for testing
|
| 43 |
-
os.environ['PYTHONHTTPSVERIFY'] = '0'
|
| 44 |
-
os.environ['REQUESTS_CA_BUNDLE'] = '/etc/ssl/certs/ca-certificates.crt'
|
| 45 |
-
os.environ['SSL_CERT_FILE'] = '/etc/ssl/certs/ca-certificates.crt'
|
| 46 |
-
os.environ['CURL_CA_BUNDLE'] = '/etc/ssl/certs/ca-certificates.crt'
|
| 47 |
-
|
| 48 |
-
logger.info("Network environment variables set")
|
| 49 |
-
return True
|
| 50 |
-
except Exception as e:
|
| 51 |
-
logger.warning(f"Could not set network environment: {e}")
|
| 52 |
-
return False
|
| 53 |
-
|
| 54 |
-
def test_network_connectivity():
|
| 55 |
-
"""Test basic network connectivity"""
|
| 56 |
-
try:
|
| 57 |
-
import subprocess
|
| 58 |
-
import shutil
|
| 59 |
-
|
| 60 |
-
# Check if ping is available
|
| 61 |
-
if shutil.which('ping'):
|
| 62 |
-
# Try to ping Google DNS with shorter timeout
|
| 63 |
-
result = subprocess.run(['ping', '-c', '1', '-W', '5', '8.8.8.8'],
|
| 64 |
-
capture_output=True, text=True, timeout=8)
|
| 65 |
-
if result.returncode == 0:
|
| 66 |
-
logger.info("Network connectivity test passed")
|
| 67 |
-
return True
|
| 68 |
-
else:
|
| 69 |
-
logger.warning(f"Ping test failed: {result.stderr}")
|
| 70 |
-
return False
|
| 71 |
-
else:
|
| 72 |
-
logger.warning("Ping not available, skipping network test")
|
| 73 |
-
return True
|
| 74 |
-
|
| 75 |
-
except subprocess.TimeoutExpired:
|
| 76 |
-
logger.warning("Ping test timed out, but continuing...")
|
| 77 |
-
return True # Don't fail on timeout
|
| 78 |
-
except Exception as e:
|
| 79 |
-
logger.warning(f"Network connectivity test failed: {e}")
|
| 80 |
-
return True # Don't fail on network test errors
|
| 81 |
-
|
| 82 |
-
def extract_video_id(url):
|
| 83 |
-
"""Extract YouTube video ID from various URL formats"""
|
| 84 |
-
patterns = [
|
| 85 |
-
r'(?:v=|\/)([0-9A-Za-z_-]{11}).*',
|
| 86 |
-
r'(?:embed\/)([0-9A-Za-z_-]{11})',
|
| 87 |
-
r'(?:vi\/)([0-9A-Za-z_-]{11})',
|
| 88 |
-
r'(?:v\/)([0-9A-Za-z_-]{11})',
|
| 89 |
-
r'(?:\/|%3D)([0-9A-Za-z_-]{11})',
|
| 90 |
-
r'(?:watch.*v=)([0-9A-Za-z_-]{11})',
|
| 91 |
-
r'^([0-9A-Za-z_-]{11})$'
|
| 92 |
-
]
|
| 93 |
-
|
| 94 |
-
for pattern in patterns:
|
| 95 |
-
match = re.search(pattern, url)
|
| 96 |
-
if match:
|
| 97 |
-
return match.group(1)
|
| 98 |
-
|
| 99 |
-
return None
|
| 100 |
-
|
| 101 |
-
def download_youtube_video(url, output_dir):
|
| 102 |
-
"""
|
| 103 |
-
Download YouTube video using yt-dlp
|
| 104 |
-
Returns: path to downloaded video file
|
| 105 |
-
"""
|
| 106 |
-
try:
|
| 107 |
-
import yt_dlp
|
| 108 |
-
except ImportError:
|
| 109 |
-
raise ImportError("yt-dlp is required for YouTube downloads. Install with: pip install yt-dlp")
|
| 110 |
-
|
| 111 |
-
# Setup network environment
|
| 112 |
-
setup_network_environment()
|
| 113 |
-
|
| 114 |
-
# Check network connectivity first (but don't fail if it doesn't work)
|
| 115 |
-
if not test_network_connectivity():
|
| 116 |
-
logger.warning("Network connectivity test failed, but continuing with download attempt...")
|
| 117 |
-
|
| 118 |
-
# Check DNS resolution
|
| 119 |
-
if not check_dns_resolution():
|
| 120 |
-
logger.warning("DNS resolution failed, but continuing with download attempt...")
|
| 121 |
-
# Don't fail immediately, try the download anyway
|
| 122 |
-
|
| 123 |
-
# Try alternative DNS resolution methods
|
| 124 |
-
try:
|
| 125 |
-
import socket
|
| 126 |
-
import requests
|
| 127 |
-
|
| 128 |
-
# Method 1: Try with requests library
|
| 129 |
-
try:
|
| 130 |
-
response = requests.get('https://www.youtube.com', timeout=10)
|
| 131 |
-
if response.status_code == 200:
|
| 132 |
-
logger.info("YouTube connectivity confirmed via requests")
|
| 133 |
-
else:
|
| 134 |
-
logger.warning(f"YouTube response status: {response.status_code}")
|
| 135 |
-
except Exception as e:
|
| 136 |
-
logger.warning(f"Requests method failed: {e}")
|
| 137 |
-
|
| 138 |
-
# Method 2: Try with custom DNS resolution
|
| 139 |
-
try:
|
| 140 |
-
# Try to resolve with different methods
|
| 141 |
-
socket.setdefaulttimeout(10)
|
| 142 |
-
ip = socket.gethostbyname('www.youtube.com')
|
| 143 |
-
logger.info(f"Manual DNS resolution successful: {ip}")
|
| 144 |
-
except Exception as e:
|
| 145 |
-
logger.warning(f"Manual DNS resolution failed: {e}")
|
| 146 |
-
|
| 147 |
-
except Exception as e:
|
| 148 |
-
logger.warning(f"Alternative DNS methods failed: {e}")
|
| 149 |
-
|
| 150 |
-
# Extract video ID for unique filename
|
| 151 |
-
video_id = extract_video_id(url)
|
| 152 |
-
if not video_id:
|
| 153 |
-
raise ValueError("Invalid YouTube URL - cannot extract video ID")
|
| 154 |
-
|
| 155 |
-
# Create unique filename with simpler pattern
|
| 156 |
-
unique_id = str(uuid.uuid4())[:8]
|
| 157 |
-
output_filename = f"youtube_{video_id}_{unique_id}.%(ext)s"
|
| 158 |
-
output_path = os.path.join(output_dir, output_filename)
|
| 159 |
-
|
| 160 |
-
# Also create a backup simple filename
|
| 161 |
-
simple_filename = f"video_{unique_id}.%(ext)s"
|
| 162 |
-
simple_output_path = os.path.join(output_dir, simple_filename)
|
| 163 |
-
|
| 164 |
-
# yt-dlp options with multiple output templates and enhanced network settings
|
| 165 |
-
ydl_opts = {
|
| 166 |
-
'outtmpl': [
|
| 167 |
-
output_path, # Primary template
|
| 168 |
-
simple_output_path, # Backup simple template
|
| 169 |
-
os.path.join(output_dir, '%(title)s.%(ext)s'), # Title-based template
|
| 170 |
-
os.path.join(output_dir, '%(id)s.%(ext)s'), # ID-based template
|
| 171 |
-
],
|
| 172 |
-
'format': 'best[height<=720]/best', # Prefer 720p or lower for faster processing
|
| 173 |
-
'noplaylist': True,
|
| 174 |
-
'quiet': False, # Enable output for debugging
|
| 175 |
-
'no_warnings': False, # Show warnings for debugging
|
| 176 |
-
'nocheckcertificate': True, # Skip SSL certificate verification
|
| 177 |
-
'ignoreerrors': True, # Continue on download errors
|
| 178 |
-
'socket_timeout': 120, # Increase socket timeout significantly
|
| 179 |
-
'retries': 10, # More retries
|
| 180 |
-
'fragment_retries': 10, # More fragment retries
|
| 181 |
-
'progress_hooks': [lambda d: logger.info(f"Download progress: {d.get('status', 'unknown')}") if d.get('status') == 'finished' else None],
|
| 182 |
-
'http_headers': {
|
| 183 |
-
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36',
|
| 184 |
-
'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
|
| 185 |
-
'Accept-Language': 'en-US,en;q=0.5',
|
| 186 |
-
'Accept-Encoding': 'gzip, deflate',
|
| 187 |
-
'Connection': 'keep-alive',
|
| 188 |
-
},
|
| 189 |
-
'extractor_args': {
|
| 190 |
-
'youtube': {
|
| 191 |
-
'skip': ['dash', 'live'],
|
| 192 |
-
'player_client': ['android', 'web'],
|
| 193 |
-
}
|
| 194 |
-
},
|
| 195 |
-
}
|
| 196 |
-
|
| 197 |
-
logger.info(f"Downloading YouTube video: {url}")
|
| 198 |
-
|
| 199 |
-
try:
|
| 200 |
-
with yt_dlp.YoutubeDL(ydl_opts) as ydl:
|
| 201 |
-
# Get video info first
|
| 202 |
-
try:
|
| 203 |
-
info = ydl.extract_info(url, download=False)
|
| 204 |
-
if info is None:
|
| 205 |
-
raise ValueError("Could not extract video information from URL")
|
| 206 |
-
|
| 207 |
-
duration = info.get('duration', 0)
|
| 208 |
-
|
| 209 |
-
# Check duration limit (10 minutes max for demo)
|
| 210 |
-
if duration > 600: # 10 minutes
|
| 211 |
-
raise ValueError(f"Video too long ({duration}s). Please use videos shorter than 10 minutes.")
|
| 212 |
-
|
| 213 |
-
logger.info(f"Video info extracted: duration={duration}s")
|
| 214 |
-
|
| 215 |
-
except Exception as info_error:
|
| 216 |
-
logger.warning(f"Could not extract video info: {info_error}")
|
| 217 |
-
# Continue with download anyway
|
| 218 |
-
info = None
|
| 219 |
-
|
| 220 |
-
# Download the video
|
| 221 |
-
logger.info("Starting video download...")
|
| 222 |
-
logger.info(f"Output template: {output_path}")
|
| 223 |
-
logger.info(f"Output directory: {output_dir}")
|
| 224 |
-
|
| 225 |
-
# List files before download
|
| 226 |
-
try:
|
| 227 |
-
files_before = os.listdir(output_dir)
|
| 228 |
-
logger.info(f"Files before download: {files_before}")
|
| 229 |
-
except Exception as e:
|
| 230 |
-
logger.warning(f"Could not list files before download: {e}")
|
| 231 |
-
|
| 232 |
-
ydl.download([url])
|
| 233 |
-
|
| 234 |
-
# List files after download
|
| 235 |
-
try:
|
| 236 |
-
files_after = os.listdir(output_dir)
|
| 237 |
-
logger.info(f"Files after download: {files_after}")
|
| 238 |
-
except Exception as e:
|
| 239 |
-
logger.warning(f"Could not list files after download: {e}")
|
| 240 |
-
|
| 241 |
-
# Try fallback download if no file found
|
| 242 |
-
if not any(f.lower().endswith(('.mp4', '.mkv', '.webm', '.avi', '.m4v', '.mov')) for f in files_after):
|
| 243 |
-
logger.warning("No video file found, trying fallback download...")
|
| 244 |
-
|
| 245 |
-
# Try with simpler options
|
| 246 |
-
fallback_opts = {
|
| 247 |
-
'outtmpl': os.path.join(output_dir, '%(id)s.%(ext)s'),
|
| 248 |
-
'format': 'best',
|
| 249 |
-
'noplaylist': True,
|
| 250 |
-
'quiet': False,
|
| 251 |
-
'no_warnings': False,
|
| 252 |
-
'nocheckcertificate': True,
|
| 253 |
-
'ignoreerrors': True,
|
| 254 |
-
}
|
| 255 |
-
|
| 256 |
-
try:
|
| 257 |
-
with yt_dlp.YoutubeDL(fallback_opts) as ydl_fallback:
|
| 258 |
-
ydl_fallback.download([url])
|
| 259 |
-
|
| 260 |
-
# List files after fallback download
|
| 261 |
-
try:
|
| 262 |
-
files_after_fallback = os.listdir(output_dir)
|
| 263 |
-
logger.info(f"Files after fallback download: {files_after_fallback}")
|
| 264 |
-
except Exception as e:
|
| 265 |
-
logger.warning(f"Could not list files after fallback download: {e}")
|
| 266 |
-
|
| 267 |
-
except Exception as fallback_error:
|
| 268 |
-
logger.error(f"Fallback download also failed: {fallback_error}")
|
| 269 |
-
|
| 270 |
-
# Try curl as last resort
|
| 271 |
-
logger.warning("Trying curl as last resort...")
|
| 272 |
-
try:
|
| 273 |
-
import subprocess
|
| 274 |
-
curl_cmd = [
|
| 275 |
-
'curl', '-L', '-o',
|
| 276 |
-
os.path.join(output_dir, f'{video_id}.mp4'),
|
| 277 |
-
'--connect-timeout', '30',
|
| 278 |
-
'--max-time', '300',
|
| 279 |
-
'--retry', '3',
|
| 280 |
-
'--retry-delay', '2',
|
| 281 |
-
url
|
| 282 |
-
]
|
| 283 |
-
result = subprocess.run(curl_cmd, capture_output=True, text=True, timeout=300)
|
| 284 |
-
if result.returncode == 0:
|
| 285 |
-
logger.info("Curl download successful")
|
| 286 |
-
else:
|
| 287 |
-
logger.error(f"Curl download failed: {result.stderr}")
|
| 288 |
-
except Exception as curl_error:
|
| 289 |
-
logger.error(f"Curl download also failed: {curl_error}")
|
| 290 |
-
|
| 291 |
-
# Find the actual downloaded file
|
| 292 |
-
logger.info("Searching for downloaded file...")
|
| 293 |
-
|
| 294 |
-
downloaded_file = None
|
| 295 |
-
|
| 296 |
-
# Method 1: Try all possible output templates
|
| 297 |
-
possible_templates = [
|
| 298 |
-
output_path.replace('.%(ext)s', ''),
|
| 299 |
-
simple_output_path.replace('.%(ext)s', ''),
|
| 300 |
-
os.path.join(output_dir, video_id),
|
| 301 |
-
os.path.join(output_dir, f"youtube_{video_id}"),
|
| 302 |
-
os.path.join(output_dir, f"video_{unique_id}"),
|
| 303 |
-
]
|
| 304 |
-
|
| 305 |
-
possible_extensions = ['.mp4', '.mkv', '.webm', '.avi', '.m4v', '.mov']
|
| 306 |
-
|
| 307 |
-
for template in possible_templates:
|
| 308 |
-
for ext in possible_extensions:
|
| 309 |
-
candidate = template + ext
|
| 310 |
-
if os.path.exists(candidate) and os.path.isfile(candidate):
|
| 311 |
-
downloaded_file = candidate
|
| 312 |
-
logger.info(f"Found file with template {template} + {ext}: {candidate}")
|
| 313 |
-
break
|
| 314 |
-
if downloaded_file:
|
| 315 |
-
break
|
| 316 |
-
|
| 317 |
-
# Method 2: Search by pattern in output directory
|
| 318 |
-
if not downloaded_file:
|
| 319 |
-
patterns = [
|
| 320 |
-
f"youtube_{video_id}",
|
| 321 |
-
f"video_{unique_id}",
|
| 322 |
-
video_id,
|
| 323 |
-
"youtube_",
|
| 324 |
-
"video_"
|
| 325 |
-
]
|
| 326 |
-
|
| 327 |
-
try:
|
| 328 |
-
files_in_dir = os.listdir(output_dir)
|
| 329 |
-
logger.info(f"Files in directory: {files_in_dir}")
|
| 330 |
-
|
| 331 |
-
for pattern in patterns:
|
| 332 |
-
for file in files_in_dir:
|
| 333 |
-
if (file.startswith(pattern) and
|
| 334 |
-
not file.endswith('.part') and
|
| 335 |
-
any(file.lower().endswith(ext) for ext in possible_extensions)):
|
| 336 |
-
candidate = os.path.join(output_dir, file)
|
| 337 |
-
if os.path.isfile(candidate):
|
| 338 |
-
downloaded_file = candidate
|
| 339 |
-
logger.info(f"Found file by pattern '{pattern}': {candidate}")
|
| 340 |
-
break
|
| 341 |
-
if downloaded_file:
|
| 342 |
-
break
|
| 343 |
-
except Exception as e:
|
| 344 |
-
logger.warning(f"Error listing directory: {e}")
|
| 345 |
-
|
| 346 |
-
# Method 3: Find any video file in directory (last resort)
|
| 347 |
-
if not downloaded_file:
|
| 348 |
-
logger.info("Searching for any video file in directory...")
|
| 349 |
-
try:
|
| 350 |
-
for file in os.listdir(output_dir):
|
| 351 |
-
if any(file.lower().endswith(ext) for ext in possible_extensions):
|
| 352 |
-
candidate = os.path.join(output_dir, file)
|
| 353 |
-
if os.path.isfile(candidate) and not file.endswith('.part'):
|
| 354 |
-
downloaded_file = candidate
|
| 355 |
-
logger.info(f"Found video file: {candidate}")
|
| 356 |
-
break
|
| 357 |
-
except Exception as e:
|
| 358 |
-
logger.warning(f"Error searching for video files: {e}")
|
| 359 |
-
|
| 360 |
-
if not downloaded_file or not os.path.exists(downloaded_file):
|
| 361 |
-
logger.error(f"Downloaded file not found. Searched in: {output_dir}")
|
| 362 |
-
logger.error(f"Video ID: {video_id}")
|
| 363 |
-
logger.error(f"Unique ID: {unique_id}")
|
| 364 |
-
logger.error(f"Possible templates: {possible_templates}")
|
| 365 |
-
raise FileNotFoundError("Downloaded file not found after exhaustive search")
|
| 366 |
-
|
| 367 |
-
logger.info(f"YouTube video downloaded: {downloaded_file}")
|
| 368 |
-
return downloaded_file
|
| 369 |
-
|
| 370 |
-
except Exception as e:
|
| 371 |
-
logger.error(f"YouTube download failed: {str(e)}")
|
| 372 |
-
|
| 373 |
-
# Provide more specific error messages
|
| 374 |
-
if "Video unavailable" in str(e):
|
| 375 |
-
raise ValueError("This video is unavailable or private. Please check the URL.")
|
| 376 |
-
elif "Sign in" in str(e):
|
| 377 |
-
raise ValueError("This video requires sign-in. Please use a public video.")
|
| 378 |
-
elif "copyright" in str(e).lower():
|
| 379 |
-
raise ValueError("This video has copyright restrictions and cannot be downloaded.")
|
| 380 |
-
elif "DNS" in str(e) or "resolve" in str(e):
|
| 381 |
-
raise ConnectionError("Network connection failed. Please check your internet connection.")
|
| 382 |
-
else:
|
| 383 |
-
# Clean up any partial downloads
|
| 384 |
-
try:
|
| 385 |
-
base_path = output_path.replace('.%(ext)s', '')
|
| 386 |
-
for ext in ['.mp4', '.mkv', '.webm', '.avi', '.part']:
|
| 387 |
-
candidate = base_path + ext
|
| 388 |
-
if os.path.exists(candidate):
|
| 389 |
-
os.remove(candidate)
|
| 390 |
-
logger.info(f"Cleaned up partial download: {candidate}")
|
| 391 |
-
except Exception as cleanup_error:
|
| 392 |
-
logger.warning(f"Could not clean up partial downloads: {cleanup_error}")
|
| 393 |
-
|
| 394 |
-
raise
|
| 395 |
-
|
| 396 |
-
def get_video_info(url):
|
| 397 |
-
"""Get basic info about YouTube video without downloading"""
|
| 398 |
-
try:
|
| 399 |
-
import yt_dlp
|
| 400 |
-
|
| 401 |
-
ydl_opts = {
|
| 402 |
-
'quiet': True,
|
| 403 |
-
'no_warnings': True,
|
| 404 |
-
'nocheckcertificate': True,
|
| 405 |
-
'ignoreerrors': True,
|
| 406 |
-
}
|
| 407 |
-
|
| 408 |
-
with yt_dlp.YoutubeDL(ydl_opts) as ydl:
|
| 409 |
-
info = ydl.extract_info(url, download=False)
|
| 410 |
-
if info is None:
|
| 411 |
-
logger.error("Could not extract video information")
|
| 412 |
-
return None
|
| 413 |
-
|
| 414 |
-
return {
|
| 415 |
-
'title': info.get('title', 'Unknown'),
|
| 416 |
-
'duration': info.get('duration', 0),
|
| 417 |
-
'uploader': info.get('uploader', 'Unknown'),
|
| 418 |
-
'view_count': info.get('view_count', 0)
|
| 419 |
-
}
|
| 420 |
-
except Exception as e:
|
| 421 |
-
logger.error(f"Failed to get video info: {str(e)}")
|
| 422 |
-
return None
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|