search

EFForg / badger-sett

Automated training for Privacy Badger. Badger Sett automates browsers to visit websites to produce fresh Privacy Badger tracker data.
https://www.eff.org/badger-pretraining
MIT License
121 stars 15 forks source link

readme

Badger Sett

A sett or set is a badger's den which usually consists of a network of tunnels and numerous entrances. Setts incorporate larger chambers used for sleeping or rearing young.

This script is designed to raise young Privacy Badgers by teaching them about the trackers on popular sites. Every day, crawler.py visits thousands of the top sites from the Tranco List with the latest version of Privacy Badger, and saves its findings in results.json.

See the following EFF.org blog post for more information: Giving Privacy Badger a Jump Start.

Development setup

  1. Install Python 3.8+

  2. Create and activate a Python virtual environment:

    python3 -m venv venv
source ./venv/bin/activate
pip install -U pip

    For more, read this blog post.

  3. Install Python dependencies with pip install -r requirements.txt

  4. Run static analysis with prospector

  5. Run unit tests with pytest

  6. Take a look at Badger Sett commandline flags with ./crawler.py --help

  7. Git clone the Privacy Badger repository somewhere

  8. Try running a tiny scan:

    ./crawler.py firefox 5 --no-xvfb --log-stdout --pb-dir /path/to/privacybadger

Production setup with Docker

Docker takes care of all dependencies, including setting up the latest browser version.

However, Docker brings its own complexity. Problems from improper file ownership and permissions are a particular pain point.

  1. Prerequisites: have Docker installed. Make sure your user is part of the docker group so that you can build and run docker images without sudo. You can add yourself to the group with

    $ sudo usermod -aG docker $USER

  2. Clone the repository

    $ git clone https://github.com/efforg/badger-sett

  3. Run a scan

    $ BROWSER=firefox ./runscan.sh 500

    This will scan the top 500 sites on the Tranco list in Chrome with the latest version of Privacy Badger's master branch.

    To run the script with a different branch of Privacy Badger, set the PB_BRANCH variable. e.g.

    $ PB_BRANCH=my-feature-branch BROWSER=firefox ./runscan.sh 500

    You can also pass arguments to crawler.py, the Python script that does the actual crawl. Any arguments passed to runscan.sh will be forwarded to crawler.py. For example, to exclude all websites ending with .gov and .mil from your website visit list:

    $ BROWSER=edge ./runscan.sh 500 --exclude .gov,.mil

  4. Monitor the scan

    To have the scan print verbose output about which sites it's visiting, use the --log-stdout argument.

    If you don't use that argument, all output will still be logged to docker-out/log.txt, beginning after the script outputs "Running scan in Docker..."

Automatic scanning

To set up the script to run periodically and automatically update the repository with its results:

  1. Create a new ssh key with ssh-keygen. Give it a name unique to the repository.

    $ ssh-keygen
Generating public/private rsa key pair.
Enter file in which to save the key (/home/USER/.ssh/id_rsa): /home/USER/.ssh/id_rsa_badger_sett

  2. Add the new key as a deploy key with R/W access to the repo on Github. https://developer.github.com/v3/guides/managing-deploy-keys/

  3. Add a SSH host alias for Github that uses the new key pair. Create or open ~/.ssh/config and add the following:

    Host github-badger-sett
 HostName github.com
 User git
 IdentityFile /home/USER/.ssh/id_rsa_badger_sett

  4. Configure git to connect to the remote over SSH. Edit .git/config:

    [remote "origin"]
 url = ssh://git@github-badger-sett:/efforg/badger-sett

    This will have git connect to the remote using the new SSH keys by default.

  5. Create a cron job to call runscan.sh once a day. Set the environment variable RUN_BY_CRON=1 to turn off TTY forwarding to docker run (which would break the script in cron), and set GIT_PUSH=1 to have the script automatically commit and push results.json when the scan finishes. Here's an example crontab entry:

    0 0 * * *  RUN_BY_CRON=1 GIT_PUSH=1 BROWSER=chrome /home/USER/badger-sett/runscan.sh 6000 --exclude=.mil,.mil.??,.gov,.gov.??,.edu,.edu.??

  6. If everything has been set up correctly, the script should push a new version of results.json after each scan.