Contributing to pulse2percept¶
Note
If you found a bug or want to request a feature, please open an issue in our Issue Tracker on GitHub.
We are excited that you are here and want to contribute!
Already know what you’re looking for in this guide? Jump to the following sections:
- Recommended workflow
- Contributing code
- Documenting your code
- Documenting API changes
- Testing your code
Recommended workflow¶
We appreciate all contributions, but those accepted fastest will follow a workflow similar to the following:
GitHub account: Before getting started, you will need to set up a free GitHub account.
Claim an issue on GitHub: Check the Issue Tracker to find an issue you want to work on, and add a comment announcing your intention to work on it. This allows other members of the development team to confirm that you aren’t overlapping with existing work and that everyone is on the same page with the goal of your proposed work.
These issues point to problems in the project. The bug report should already contain steps for you to reproduce the issue. Make sure to include a test case verifying that the issue was indeed solved.
These issues are asking for new features to be added to the project. Make sure to include a test case for every new function/class that you write. Each API change should be annotated in the docstring.
These issues concern the documentation / user guide of the project. Code changes should therefore be limited to the user guide and docstrings only.
These issues contain a task that a member of the team has determined we need additional help with.
These issues are good entry points if you’re new to the project. They usually require only minimal code changes or are restricted to a single file. If you run into trouble, add a comment to the issue on GitHub or reach out to @mbeyeler and/or @arokem directly.
Fork the repo: Follow the Installation Guide to fork the repo and install all developer dependencies. This is now your own unique pulse2percept copy - changes here won’t affect anyone else’s work. Make sure to keep your code up-to-date with the upstream repository.
Create a new branch: You should always work on a new branch on your fork. Once you made the code changes, “git add” and “git commit” them, and then “git push” them to your remote repository on GitHub.
Important
All code additions must be documented and tested. See Contributing code below for more detailed instructions.
Submit a pull request: You can open a pull request (PR) as soon as you have pushed a new branch on your fork. This will trigger the test suite to make sure your code does not introduce any new issues.
Choose one of the following prefixes for your PR:
- [ENH] for enhancements
- [FIX] for bug fixes
- [TST] for new or updated tests
- [DOC] for new or updated documentation
- [STY] for stylistic changes
- [REF] for refactoring existing code
Once your PR is ready, request a review from @mbeyeler and/or @arokem, who will review your changes before merging them into the main codebase.
Note
If your PR is not yet ready to be merged, click on the dropdown arrow next to the “Create pull request” button and choose “Create draft pull request” instead.
This will put your PR in draft state and block merging until you change the status of the PR to “Ready for review”.
More detailed instructions can be found below.
Contributing code¶
Perform all your work on a new branch of the repository. For example, say you want to add “feature1” to the latest version of pulse2percept:
Make sure you have the latest code:
git checkout master git pull upstream master
Note
If you get an error saying “upstream does not appear to be a git repository”, you need to run the following command first:
git remote add upstream https://github.com/pulse2percept/pulse2percept.git
Create a new branch (aptly named “feature1” or similar):
git checkout -b feature1
Add and commit your changes to this branch:
git add newfile.py git commit -m "add new feature1 file"
Then push it to your remote repository on GitHub:
git push origin feature1
Important
All code additions must be documented and tested.
Go to GitHub and submit a pull request:
Click on “compare & pull request” at the top of the page.
Choose “pulse2percept/pulse2percept” as the base repository and “master” as the base branch.
Choose “<username>/pulse2percept” as the head repository and “feature1” as the compare branch, where “<username>” is your GitHub user name.
Click on “Create pull request” (or “Create draft pull request” if your work is not ready to be merged) and describe the work you have done. Make sure to mention the issue number you are addressing (use # as prefix).
An easy way to list all the things you changed is to use a list of checkboxes (type
- [X]
; or- [ ]
for an item that has yet to be implemented).
Documenting your code¶
You are expected to document your code using NumPy docstrings. Make sure to:
- supply short and long descriptions,
- describe all input arguments to a function/method,
- describe the output of a function/method,
- provide examples of how to use your code.
For example, consider an appropriate docstring for a hypothetical function
rad2deg
:
def rad2deg(angle_rad):
"""Converts radians to degrees
This function converts an angle in radians to degrees.
Parameters
----------
angle_rad : int, float
The input angle in radians in (between 0 and 2pi)
Returns
-------
angle_deg : float
The corresponding angle in degrees (between 0 and 360 deg)
Examples
--------
Converting pi to degrees:
>>> import numpy as np
>>> rad2deg(np.pi)
180.0
.. seealso:: `deg2rad`
"""
...
You can generate the documentation yourself using Sphinx.
If you installed make
, type the following from your root directory:
make doc
Otherwise, type the following from your root directory:
cd doc
pip3 install -r requirements.txt
make html
The generated documentation can then be found in doc/_build/html
.
To see the documentation, “doc/_build/html/index.html” in your browser of
choice, e.g.:
google-chrome doc/_build/html/index.html
Documenting API changes¶
API changes that affect the user should be documented in order to help the user sort out version differences (see reST directives):
- Whenever a new API call is added, include a
.. versionadded::
statement right before listing the function parameters that mentions the pulse2percept version where the feature first appeared. - Whenever the API of a function/class is changed, include a
.. versionchanged::
statement right before listing the function parameters that explains what/how functionality changed in a particular pulse2percept version.
Testing your code¶
You are expected to test your code using pytest:
- Bug fixes should include an example that exposes the issue.
- New features should have tests that show at least a minimal example.
Running the test suite¶
pulse2percept uses pytest and numpy-testing for testing.
Every subpackage of pulse2percept (e.g., stimuli
)
has a subdirectory called “tests”.
Within the test directory, there is a “test_<subsubpackage>.py” file for every
subsubpackage of pulse2percept (e.g.,
“pulse2percept/stimuli/tests/test_pulse_trains.py” for the
pulse_trains
module).
When you contribute new code, you are expected to test your code in the corresponding test file.
You can run the test suite from your root directory with:
pip3 install -r requirements-dev.txt
pytest --doctest-modules --showlocals -v pulse2percept
Successful tasks will be marked with “PASSED”, unsuccessful ones with “FAILED”. We will usually not accept pull requests that don’t pass all tests.
Note
Whenever you submit a pull request, the test suite is automatically run in the background using GitHub Actions. This will make sure that all tests pass on all supported platforms whenever changes are made to the code.
Writing your own tests¶
If you work on code from an existing subpackage (e.g.,
pulse2percept.stimuli.pulse_trains
), open the corresponding test file
(e.g., “pulse2percept/stimuli/tests/test_pulse_trains.py”).
You can add a new test by adding a function whose name starts with “test_”, followed by the name of the class or function you want to test. For example:
def test_TimeSeries
for testing theTimeSeries
object (note that this function already exists).def test_TimeSeries_resample
for testing theresample
method of theTimeSeries
object.def test_newfunc
for a new function callednewfunc
.
Within this function, you want to make sure your code works as expected. Useful numpy-testing routines for achieving this include:
assert_equal(actual, desired)
returns anAssertionError
if two objects are not equal.assert_almost_equal(actual, desired, decimal=7)
returns anAssertionError
if two items are not equal up to desired precision (good for testing doubles).assert_raises(exception_class)
fails unless anException
of classexception_class
is thrown.
In addition, we provide
assert_warns_msg
to ensure that a
specific warning message is thrown.
See also
Thank you¶
You are awesome!
This guide is based on contributing guidelines from the `Nipype`_ project.