Skip to content

Framework which makes large scale crawling of URLs with VisibleV8 easy.

License

Notifications You must be signed in to change notification settings

wspr-ncsu/visiblev8-crawler

Repository files navigation

VisibleV8 Crawler

The VisibleV8 Crawler is a framework which makes large scale crawling of URLs with VisibleV8 much easier.

Setup

Note This tool requires Python 3.10 or above. If your OS python3 version is <3.10, you can use pyenv to setup a specific version of Python.

To setup VisibleV8 Crawler install docker and docker-compose, and run the following command

pip install -r ./scripts/requirements.txt
python ./scripts/vv8-cli.py setup

Warning Make sure that you are able to use docker and docker compose without using sudo. (instructions here)

If you plan to use visiblev8 crawler a lot, you can alias the script to the vv8cli command using:

alias vv8cli="python3 $(pwd)/scripts/vv8-cli.py" 

Note vv8 crawler cli scripts can also be used for a shared remote server by choosing the remote installation option during the setup wizard. The list of URLs (and their submission IDs) that have been run by you (and their associated submission ids) are stored locally in a sqlite3 database at ./scripts/.vv8.db

Run a single URL

python3 ./scripts/vv8-cli.py crawl -u 'https://google.com'

If you want to apply a specific vv8-postprocessor, you can use:

python3 ./scripts/vv8-cli.py crawl -u 'https://google.com' -pp 'Mfeatures'

To apply more than one postprocessor, you can instruct the postprocessor to run multiple postprocessors in a single go. For example, to use the adblock postprocessor along with the mega postprocessors, you can use:

python3 ./scripts/vv8-cli.py crawl -u 'https://google.com' -pp 'Mfeatures+adblock'

By default the postprocessed data will be written to an associated postgresql database which can be accessed using the following command if setup locally

psql --host=0.0.0.0 --port=5434 --dbname=vv8_backend --username=vv8

Note If prompted for a password, the password is by default vv8

If you want to pass more flags to the crawler (say you want to only stay on a specific page for 5s) and have the VisibleV8 binary run in the old headless mode

python3 ./scripts/vv8-cli.py crawl -u 'https://google.com' -pp 'Mfeatures' --loiter-time 5 --headless="old"

Run a list of URLs

VV8 Crawler can also be used to crawl multiple URLs in one go:

python3 ./scripts/vv8-cli.py crawl -f file.txt

Note file.txt is a file consisting of multiple urls seperated by newlines like such:

https://google.com
https://amazon.com
https://microsoft.com

Run a list of URLs in tranco format

If you have a list of URLs in the tranco CSV format, you can directly run it using

python3 ./scripts/vv8-cli.py crawl -c list.csv

Run a specific URL

python3 ./scripts/vv8-cli.py crawl -u 'https://google.com'

Monitoring the status of a crawl

The VisibleV8 crawler provides a flower Web UI to keep track of all URLs being crawled and postprocessed. The interface is accessible at http://localhost:5555.

If you are running the server locally, you can use python3 ./scripts/vv8-cli.py docker -f to get a rolling log of everything the the crawler does.

Note If you are using the crawler in a ssh session you can make use of port-forwarding to browse the web UI.

Fetch status of a crawl by URL

python3 ./scripts/vv8-cli.py fetch status 'https://google.com'

Fetch generated metadata by URL

We try to generate a har file, a screenshot and the VisibleV8 logs for every URL run and store it on mongodb, to fetch them you need to run python3 ./scripts/vv8-cli.py fetch <metadata_name> 'https://google.com'

python3 ./scripts/vv8-cli.py fetch screenshots 'https://google.com'

You can request the following things:

  • screenshots
  • raw_logs
  • hars
  • status

This command will download the files to the current directory.