CONTRIBUTING.rst

We welcome any assistance from the community in maintaining and building this project. It's impossible for the maintainers to keep up with all opt-out methods, so we're open for new Rules to be added and suggested by the community. Of course, we're also open to assistance in tracking down issues, resolving bugs, and improving simplicity.

Working together, we can make this project a standard in the ML community, and help to make sure all consent is respected throughout AI orgs around the world.

Contributing

Welcome to datadiligence contributor's guide.

This document focuses on getting any potential contributor familiarized with the development processes, but we're also open to other forms of contribution (docs, tests, issues, etc).

If you are new to using git or have never collaborated in a project previously, please have a look at contribution-guide.org. Other resources are also listed in the excellent guide created by FreeCodeCamp [1].

Please notice, all users and contributors are expected to be open, considerate, reasonable, and respectful. When in doubt, Python Software Foundation's Code of Conduct is a good reference in terms of behavior guidelines.

Issue Reports

If you experience bugs or general issues with datadiligence, please have a look on the issue tracker. If you don't see anything useful there, please feel free to fire an issue report.

Tip

Please don't forget to include the closed issues in your search. Sometimes a solution was already reported, and the problem is considered solved.

New issue reports should include information about your programming environment (e.g., operating system, Python version) and steps to reproduce the problem. Please try also to simplify the reproduction steps to a very minimal example that still illustrates the problem you are facing. By removing other factors, you help us to identify the root cause of the issue.

Documentation Improvements

You can help improve datadiligence docs by making them more readable and coherent, or by adding missing information and correcting mistakes.

datadiligence documentation uses Sphinx and reStructuredText as its main documentation compiler. This means that the docs are kept in the same repository as the project code, and that any documentation update is done in the same way was a code contribution.

Tip

Please notice that the GitHub web interface provides a quick way of propose changes in datadiligence's files. While this mechanism can be tricky for normal code contributions, it works perfectly fine for contributing to the docs, and can be quite handy.

If you are interested in trying this method out, please navigate to the docs folder in the source repository, find which file you would like to propose changes and click in the little pencil icon at the top, to open GitHub's code editor. Once you finish editing the file, please write a message in the form at the bottom of the page describing which changes have you made and what are the motivations behind them and submit your proposal.

When working on documentation changes in your local machine, you can compile them using tox:

tox -e docs

and use Python's built-in web server for a preview in your web browser (http://localhost:8000):

python3 -m http.server --directory 'docs/_build/html'

Code Contributions

In general, most contributions should be as either new Evaluators or new Rules. Other, structural changes can be made for optimization or ease-of-use, but should be discussed. Additionally, any impact or inconvenience to the user should be avoided. We want to increase adoption, and any barriers to entry would harm that goal.

Submit an issue

Before you work on any non-trivial code contribution it's best to first create a report in the issue tracker to start a discussion on the subject. This often provides additional considerations and avoids unnecessary work.

Create an environment

Before you start coding, we recommend creating an isolated virtual environment to avoid any problems with your installed Python packages. This can easily be done via either virtualenv:

virtualenv <PATH TO VENV>
source <PATH TO VENV>/bin/activate

Clone the repository

1. Create an user account on GitHub if you do not already have one.

2. Fork the project repository: click on the Fork button near the top of the page. This creates a copy of the code under your account on GitHub.

3. Clone this copy to your local disk:

   git clone git@github.com:YourLogin/datadiligence.git
   cd datadiligence
   

4. You should run:

   pip install -U pip setuptools -e .
   

   to be able to import the package under development in the Python REPL.

Install dependencies

1. Install tox with pip install tox or pipx.

2. Install the development dependencies with:

   tox -e dev
   

   This will install all the dependencies needed to run the tests and build the documentation.

Implement your changes

1. Create a branch to hold your changes:

   git checkout -b my-feature
   

   and start making changes. Never work on the main branch!

2. Start your work on this branch. Don't forget to add docstrings to new functions, modules and classes, especially if they are part of public APIs.

3. Add yourself to the list of contributors in AUTHORS.rst.

4. When you’re done editing, do:

   git add <MODIFIED FILES>
   git commit
   

   to record your changes in git.

   Important

   Don't forget to add unit tests and documentation in case your contribution adds an additional feature and is not just a bugfix.

   Moreover, writing a descriptive commit message is highly recommended. In case of doubt, you can check the commit history with:

   git log --graph --decorate --pretty=oneline --abbrev-commit --all
   

   to look for recurring communication patterns.

5. Please check that your changes don't break any unit tests with:

   tox
   

   (after having installed tox with pip install tox or pipx).

   You can also use tox to run several other pre-configured tasks in the repository. Try tox -av to see a list of the available checks.

Submit your contribution

1. If everything works fine, push your local branch to GitHub with:

   git push -u origin my-feature
   

2. Go to the web page of your fork and click "Create pull request" to send your changes for review.

Find more detailed information in creating a PR. You might also want to open the PR as a draft first and mark it as ready for review after the feedbacks from the continuous integration (CI) system or any required fixes.

Troubleshooting

The following tips can be used when facing problems to build or test the package:

1. Make sure to fetch all the tags from the upstream repository. The command git describe --abbrev=0 --tags should return the version you are expecting. If you are trying to run CI scripts in a fork repository, make sure to push all the tags. You can also try to remove all the egg files or the complete egg folder, i.e., .eggs, as well as the *.egg-info folders in the src folder or potentially in the root of your project.

2. Sometimes tox misses out when new dependencies are added, especially to setup.cfg and docs/requirements.txt. If you find any problems with missing dependencies when running a command with tox, try to recreate the tox environment using the -r flag. For example, instead of:

   tox -e docs
   

   Try running:

   tox -r -e docs
   

3. Make sure to have a reliable tox installation that uses the correct Python version (e.g., 3.7+). When in doubt you can run:

   tox --version
   # OR
   which tox
   

   If you have trouble and are seeing weird errors upon running tox, you can also try to create a dedicated virtual environment with a tox binary freshly installed. For example:

   virtualenv .venv
   source .venv/bin/activate
   .venv/bin/pip install tox
   .venv/bin/tox -e all
   

4. Pytest can drop you in an interactive session in the case an error occurs. In order to do that you need to pass a --pdb option (for example by running tox -- -k <NAME OF THE FALLING TEST> --pdb). You can also setup breakpoints manually instead of using the --pdb option.

Maintainer tasks

Releases

If you are part of the group of maintainers and have correct user permissions on PyPI, the following steps can be used to release a new version for datadiligence:

1. Make sure all unit tests are successful.
2. Tag the current commit on the main branch with a release tag, e.g., git tag 0.1.7.
3. Push the new tag to the upstream repository, e.g., git push origin 0.1.7
4. Clean up the dist and build folders with tox -e clean (or rm -rf dist build) to avoid confusion with old builds and Sphinx docs.
5. Run tox -e build and check that the files in dist have the correct version (no .dirty or git hash) according to the git tag. Also check the sizes of the distributions, if they are too big (e.g., > 500KB), unwanted clutter may have been accidentally included.
6. Run tox -e publish -- --repository pypi and check that everything was uploaded to PyPI correctly. This will prompt you for pypi authentication. We only use PyPi API Keys, so the user will be __token__ and the password is value of the API key. Optionally, you can set the TWINE_PASSWORD environment variable to the value of the API key.

[1]	Even though, these resources focus on open source projects and communities, the general ideas behind collaborating with other developers to collectively create software are general and can be applied to all sorts of environments, including private companies and proprietary code bases.