NGTube 1.0.0

A Python library for scraping YouTube video data

NGTube

A comprehensive Python library for scraping YouTube data, including videos, comments, and channel profiles.

⚠️ Disclaimer

This library is provided for educational and research purposes only. Scraping YouTube data may violate YouTube's Terms of Service. Use at your own risk. The authors are not responsible for any misuse or legal consequences. Always respect robots.txt and implement appropriate rate limiting.

Features

• Video Extraction: Extract detailed metadata from YouTube videos (title, views, likes, duration, tags, description, etc.)
• Comment Extraction: Extract comments from videos, including loading additional comments via YouTube's internal API
• Channel Extraction: Extract complete channel profile data (subscribers, description, featured video, video list with continuation support)
• Flexible Video Loading: Load specific number of videos or all available videos from a channel
• Clean Data Output: Structured JSON-compatible data output
• Modular Design: Separate classes for different extraction tasks

Installation

Option 1: Install as Package (Recommended)

1. Clone or download the repository.
2. Navigate to the project directory.
3. Install the package using pip:

pip install .

This will install NGTube as a Python package with all dependencies automatically handled.

Option 2: Manual Installation

1. Clone or download the repository.
2. Ensure you have Python 3.6+ installed.
3. Install required dependencies:

pip install requests demjson3

4. Copy the NGTube folder to your project directory or add it to your Python path.

Using setup.py

The setup.py file is used for packaging and installation. You can also install manually:

python setup.py install

However, using pip install . is recommended as it handles modern Python packaging better.

Quick Start

Extract Video Metadata

from NGTube import Video

url = "https://www.youtube.com/watch?v=y1XrJyFF1O0"
video = Video(url)
metadata = video.extract_metadata()

print("Title:", metadata['title'])
print("Views:", metadata['view_count'])
print("Likes:", metadata['like_count'])
print("Duration:", metadata['duration_seconds'], "seconds")

Extract Comments

from NGTube import Comments

url = "https://www.youtube.com/watch?v=y1XrJyFF1O0"
comments = Comments(url)
comment_data = comments.get_comments()

print(f"Total comments: {len(comment_data['comments'])}")
for comment in comment_data['comments'][:3]:
    print(f"{comment['author']}: {comment['text'][:50]}...")

Extract Channel Profile

from NGTube import Channel

url = "https://www.youtube.com/@HandOfUncut"
channel = Channel(url)

# Load first 10 videos
profile = channel.extract_profile(max_videos=10)

print("Channel Title:", profile['title'])
print("Subscribers:", profile['subscribers'])
print("Videos loaded:", profile['loaded_videos_count'])

# Load all videos
profile_all = channel.extract_profile(max_videos='all')
print("Total videos:", profile_all['loaded_videos_count'])

Detailed Usage

Video Class

from NGTube import Video

video = Video("https://www.youtube.com/watch?v=VIDEO_ID")
metadata = video.extract_metadata()

# Available metadata keys:
# - title, view_count, like_count, duration_seconds
# - channel_name, channel_id, subscriber_count
# - description, tags, category, is_private
# - upload_date, published_time_text

Comments Class

from NGTube import Comments

comments = Comments("https://www.youtube.com/watch?v=VIDEO_ID")
data = comments.get_comments()

# Returns dictionary with:
# - 'top_comment': list of top comments
# - 'comments': list of regular comments

# Each comment contains:
# - author, text, like_count, published_time_text
# - author_thumbnail, comment_id, reply_count

Channel Class

from NGTube import Channel

channel = Channel("https://www.youtube.com/@ChannelHandle")

# Extract profile with specific number of videos
profile = channel.extract_profile(max_videos=50)

# Extract profile with all videos (may take time)
profile = channel.extract_profile(max_videos='all')

# Available profile data:
# - title, description, channel_id, channel_url
# - keywords, is_family_safe, links
# - subscriber_count_text, view_count_text, video_count_text
# - subscribers, total_views, video_count (parsed numbers)
# - featured_video (dict with videoId, title, description)
# - videos (list of video dictionaries)
# - loaded_videos_count

Examples

See the examples/ directory for complete working examples:

• basic_usage.py: Extract video metadata and comments
• batch_processing.py: Process multiple videos
• channel_usage.py: Extract channel profile data

Run any example:

python examples/basic_usage.py

API Reference

Core Classes

YouTubeCore

Base class for YouTube interactions.

• __init__(url: str): Initialize with YouTube URL
• fetch_html() -> str: Fetch HTML content
• extract_ytinitialdata(html: str) -> dict: Extract ytInitialData
• make_api_request(endpoint: str, payload: dict) -> dict: Make API requests

Video

Extract video metadata.

• __init__(url: str): Initialize with video URL
• extract_metadata() -> dict: Extract and return video metadata

Comments

Extract video comments.

• __init__(url: str): Initialize with video URL
• get_comments() -> dict: Extract and return comments data

Channel

Extract channel profile and videos.

• __init__(url: str): Initialize with channel URL
• extract_profile(max_videos: int | str = 200) -> dict: Extract profile data
  • max_videos: Number of videos to load, or 'all' for all videos

Utils Module

• extract_number(text: str) -> int: Extract numbers from text (handles German formatting)
• extract_links(text: str) -> list: Extract URLs from text

Data Structures

Video Metadata

{
  "title": "Video Title",
  "view_count": 299955,
  "duration_in_seconds": 6994,
  "description": "Video description...",
  "tags": ["tag1", "tag2"],
  "video_id": "VIDEO_ID",
  "channel_id": "UC...",
  "is_owner_viewing": false,
  "is_crawlable": true,
  "thumbnail": {...},
  "allow_ratings": true,
  "author": "Channel Name",
  "is_private": false,
  "is_unplugged_corpus": false,
  "is_live_content": false,
  "like_count": 8547,
  "channel_name": "Channel Name",
  "category": "Gaming",
  "publish_date": "2023-12-01",
  "upload_date": "2023-12-01",
  "family_safe": true,
  "channel_url": "https://...",
  "subscriber_count": 1400000
}

Comment Data

{
  "top_comment": [...],
  "comments": [
    {
      "author": "Username",
      "text": "Comment text",
      "likeCount": 196,
      "publishedTimeText": "vor 1 Tag",
      "authorThumbnail": "https://...",
      "commentId": "...",
      "replyCount": 1
    }
  ]
}

Channel Profile

{
  "title": "Channel Title",
  "description": "Channel description...",
  "channelId": "UC...",
  "channelUrl": "https://...",
  "keywords": "keyword1 keyword2",
  "isFamilySafe": true,
  "links": ["https://..."],
  "subscriberCountText": "159.000 Abonnenten",
  "viewCountText": "84.770 Aufrufe",
  "videoCountText": "2583 Videos",
  "subscribers": 159000,
  "total_views": 84770,
  "video_count": 2583,
  "featured_video": {
    "videoId": "...",
    "title": "Featured Video Title",
    "description": "Featured video description..."
  },
  "videos": [
    {
      "videoId": "...",
      "title": "Video Title",
      "publishedTimeText": "vor 1 Tag",
      "viewCountText": "40.773 Aufrufe",
      "lengthText": "1:02:58",
      "thumbnails": [...]
    }
  ],
  "loaded_videos_count": 1
}

Limitations

• Rate Limiting: YouTube may rate-limit requests. Add delays between requests for bulk operations.
• Comment Limits: Without authentication, typically 40-50 comments can be loaded per video.
• Video Limits: Channel video extraction may be limited by YouTube's pagination.
• Terms of Service: This library is for educational purposes. Respect YouTube's Terms of Service and robots.txt.

Troubleshooting

• Import Errors: Ensure NGTube folder is in your Python path
• API Errors: YouTube changes their internal APIs frequently. The library uses current endpoints as of December 2025.
• Missing Data: Some videos/channels may have restricted data access

Contributing

This library is maintained for educational purposes. Feel free to submit issues or improvements.

License

This project can be used by anyone with attribution.