Skip to content

Monitor stories from news outlets for words or phrases that matter to you

License

Notifications You must be signed in to change notification settings

freedomofpress/trackthenews

Repository files navigation

Track The News

trackthenews is the script that powers @FOIAfeed, a Twitter bot that monitors news outlets for reporting that incorporates public records laws like the Freedom of Information Act (FOIA), and tweets links to and excerpts from matching articles. The underlying software can track any collection of RSS feeds for any keywords.

If you want to run your own instance of trackthenews, you can download and install the package, and run its built-in configuration process. It can be installed with pip:

Python 3.11 is what we currently test against, though it may work with other versions.

pip3 install trackthenews

or by cloning the GitHub repository and running setup.py:

python3 setup.py install

Once it is installed, you can create a configuration by running the following command in the appropriate directory:

trackthenews --config

By default, the script will place all configuration files in a new ttnconfig folder in your current working directory, but you can also designate a directory for it to use.

python3 trackthenews --config ~/foo/bar/path

That configuration process will create the necessary files and walk you through setting up a Twitter bot for matching stories. After it is configured, you'll need to use a text editor to add the matchwords and RSS feeds to their respective files.

Sample RSS feed and matchword files can be found in the project's GitHub repo. The RSS feed file is a JSON array of objects corresponding to each feed. Each object requires a url field, and should also have an outlet field.

The next two fields are optional: if you know the feed uses redirect URLs, you may set redirectLinks to true and the script will attempt to follow those redirects to store and tweet canonical URLs; if the feed uses URLs that depend on query- or hash-strings to display correctly—basically, if the content relies on text in the URL bar after a ? or #—you can set delicateURLs to true and the script will leave the URLs exactly as is.

Once you've got everything set up, you can run the program without the --config flag to check for matching articles.

trackthenews

If you designated a custom installation directory, or if you're running it from another directory (or a cron job, for example) you will need to designate the directory in which the configuration files are installed.

trackthenews ~/foo/bar/path

Settings, such as the background color for new posts, the font, and the user-agent, are all located in config.yaml, in the designated configuration directory.

How it works

Most of the script is dedicated to the Article class.

  • Articles are created based on inputs. Currently those inputs are RSS feeds, which are stored in rssfeeds.json, but in future versions other inputs will include direct URLs, news APIs, Twitter feeds, or scraped pages.
  • A series of Article methods then scrape and isolate the contents of each article (currently that cleanup is done with a Python port of Readability, but future versions may incorporate some per-site parsing), check whether it's suitable for posting, and then prepare images for tweeting.
  • Finally, the Article tweets itself.

All articles are recorded in a sqlite database.

Advanced feature: blocklist

In some cases, you may wish to suppress articles or paragraphs from being posted, even though they would otherwise match. To do so, implement a CustomBlocklist class following the abstract base class template in trackthenews/base_blocklist.py, and drop it as a file named blocklist.py in your ttnconfig directory.

You can import the bs4 library in blocklist.py for advanced parsing.

Development

Quick Start

poetry env use 3.11  # Necessary if you have a different default python version
poetry install
poetry run trackthenews sample_project
# Follow the setup script instructions
cat sample_project/matchlist-sample.txt > sample_project/matchlist.txt
cat sample_project/rssfeeds-sample.json > sample_project/rssfeeds.json
poetry run trackthenews sample_project

Detailed Instructions

To develop trackthenews, clone the repository and install the package using poetry and run the CLI tool:

# Ensure you're using Python 3.11
poetry env use 3.11
# This will create a virtual environment and install trackthenews and its dependencies
poetry install --with=dev
# This will run the setup script
poetry run trackthenews sample_project

On first run this will take you through the interactive setup process. Follow the instructions. You will need a Twitter account with application keys and/or a Mastodon account.

When the setup process is complete, the configuration for your bot will live in the sample_project directory. Fill out the matchlist.txt, matchlist_case_sensitive.txt, and rssfeeds.json files or copy configuration from the provided samples:

cat sample_project/matchlist-sample.txt > sample_project/matchlist.txt
cat sample_project/rssfeeds-sample.json > sample_project/rssfeeds.json

matchlist_case_sensitive.txt is optional and we don't provide a sample.

Now run the bot:

poetry run trackthenews sample_project

Linters

We check that all code conforms to black, flake8, and isort. To run these checks locally:

poetry run black --check .  # To automatically fix issues, exclude --check flag
poetry run flake8
poetry run isort --check-only --diff .  # To automatically fix issues, exclude --check-only and --diff flag

We also provide the following makefile shortcuts to run these commands:

make black
make flake8
make isort
make black-fix  # Automatically fix issues
make isort-fix  # Automatically fix issues
make all        # Run all three checks (check only, no fixes)

License

MIT.

About

Monitor stories from news outlets for words or phrases that matter to you

Resources

License

Code of conduct

Stars

Watchers

Forks

Packages

No packages published