vidcrawler

vidcrawler

Crawls major videos sites like YouTube/Rumble/Bitchute/Brighteon for video content and outputs a json feed of all the videos that were found.

Platform Unit Tests

Scraper Tests

Note that bitchute doesn't like the github runner's IP and will fail with a 403 forbidden.

API

Command line

vidcrawler --input_crawl_json "fetch_list.json" --output_json "out_list.json"

Python

import json
from vidcrawler import crawl_video_sites
crawl_list = [
    [
        "Computing Forever",  # Can be whatever you want.
        "bitchute",  # Must be "youtube", "rumble", "bitchute" (and others).
        "hybm74uihjkf"  # The channel id on the service.
    ]
]
output = crawl_video_sites(crawl_list)
print(json.dumps(output))

"source" and "channel_id" are used to generate the video-platform-specific urls to fetch data. The "channel name" is echo'd back in the generated json feeds, but doesn't not affect the fetching process in any way.

Testing

Install vidcrawler and then the command vidcralwer_test will become available.

> pip install vidcrawler
> vidcrawler_test

youtube-pull-channel

This new command will a channel and all of it's files as mp3s. Great for transcribing and putting into an LLM.

Example input fetch_list.json

[
    [
        "Health Ranger Report",
        "brighteon",
        "hrreport"
    ],
    [
        "Sydney Watson",
        "youtube",
        "UCSFy-1JrpZf0tFlRZfo-Rvw"
    ],
    [
        "Computing Forever",
        "bitchute",
        "hybm74uihjkf"
    ],
    [
        "ThePeteSantilliShow",
        "rumble",
        "ThePeteSantilliShow"
    ],
    [
        "Macroaggressions",
        "odysee",
        "Macroaggressions"
    ]
]

Example Output:

[
  {
    "channel_name": "ThePeteSantilliShow",
    "title": "The damage this caused is now being totaled up",
    "date_published": "2022-05-17T05:00:11+00:00",
    "date_lastupdated": "2022-05-17T05:17:18.540084",
    "channel_url": "https://www.youtube.com/channel/UCXIJgqnII2ZOINSWNOGFThA",
    "source": "youtube.com",
    "url": "https://www.youtube.com/watch?v=bwqBudCzDrQ",
    "duration": 254,
    "description": "",
    "img_src": "https://i3.ytimg.com/vi/bwqBudCzDrQ/hqdefault.jpg",
    "iframe_src": "https://youtube.com/embed/bwqBudCzDrQ",
    "views": 1429
  },
  {
     "channel_name": "ThePeteSantilliShow",
     "title": "..."
  }
]

Releases

• 1.0.40: Remove shell=True to get better ctrl-c behavior in Windows.
• 1.0.39: More pinned deps problems fixed.
• 1.0.38: One of the scrapers has a pinned dependency, install it with [full]
• 1.0.37: Misc fixes.
• 1.0.36: Fixed youtube, rumble and brighteon parsers. Bitchute is still broken and now has rate limits.
• 1.0.35: Added update_yt_dlp() to allow the client to update the downloader.
• 1.0.34: Upgraded open-webdriver to version 1.5.0 to avoid yt-dlp urllib incompatibility.
• 1.0.28: youtube_pull now takes in --channel-name and --output, like the other pullers
• 1.0.27: Fixed polluting path space from multiple added static-ffmpeg
• 1.0.24: Added rumble-pull-channel
• 1.0.21: Misc fixes
• 1.0.16: Make the library downloading more robust.
• 1.0.15: Improve cleaning filepaths for brighteon_bot
• 1.0.13: New brighteon-pull-channel command
• 1.0.11: Improves youtube-pull-channel
• 1.0.10: Adds youtube-pull-channel which pulls all files down as mp3s for a channel.
• 1.0.9: Fixes crawler for rumble and minor fixes + linting fixes.
• 1.0.8: Readme correction.
• 1.0.7: Fixes Odysee scraper by including image/webp thumbnail format.
• 1.0.4: Fixes local_now() to be local timezone aware.
• 1.0.3: Bump
• 1.0.2: Updates testing
• 1.0.1: improves command line
• 1.0.0: Initial release