We welcome contributions to EarthPy. When contributing, please follow the guidelines below and adhere to the EarthPy Code of Conduct.
At this stage of development, we are developing a set of usable wrapper functions that help make working with earth systems data easier. We are open to new functionality but are currently trying to ensure our package is stable, operational and well documented.
When submitting a change to the repository, please first create an issue that covers the item that you'd like to change, update or enhance. Once a discussion has yielded a vote of support for that addition to the package, you are ready to submit a pull request.
If you are proposing a feature:
- Explain in detail how it would work.
- Keep the scope as narrow as possible, to make it easier to implement.
Ready to contribute? Here's how to set up EarthPy for local development.
To create your own copy of the repository on GitHub, navigate to the earthlab/earthpy repository and click the Fork button in the top-right corner of the page.
Use git clone
to get a local copy of your EarthPy repository on your
local filesystem:
$ git clone [email protected]:your_name_here/earthpy.git $ cd earthpy/
Using conda, there are two options.
1. The easiest option is to create an environment from the
environment.yml
file.
Note that this will only allow you to test against one version of python
locally, but this is the recommended option on Windows and MacOS:
$ conda env create -f environment.yml $ conda activate earthpy-dev
2. If you are comfortable debugging local system library installations, and want to be able to test against multiple versions of python locally, you can create an empty conda environment:
$ conda create -n earthpy-dev python=3.10 $ conda activate earthpy-dev
Or, if you prefer to use virtualenv rather than conda:
$ virtualenv earthpy-dev $ source earthpy-dev/bin/activate
Once your earthpy-dev environment is activated, install EarthPy in editable mode, along with the development requirements and pre-commit hooks:
$ pip install -e . $ pip install -r dev-requirements.txt $ pre-commit install
Use the git checkout
command to create your own branch, and pick a name
that describes the changes that you are making:
$ git checkout -b name-of-your-bugfix-or-feature
Now you can make your changes locally.
Ensure that the tests pass, and the documentation builds successfully:
$ pytest $ make docs
Note to Windows users
To use make
you will need to install and configure GNU Make for Windows,
e.g., using chocolatey: https://chocolatey.org/packages/make
To test against multiple versions of python, you can use tox.
This step is optional, as tox will test against multiple versions as part of
EarthPy's continuous integration pipeline.
To use tox, you must be able to install EarthPy's dependencies without
using the environment.yml
file, i.e., you must be able to install EarthPy
and its dependencies with pip.
Running tox is as simple as:
$ tox
If you are using conda and you get get
InterpreterNotFound
errors when running tox, you may need to
pip install tox-conda
.
Once you are sure that all tests are passing, you can commit your changes and push to GitHub:
$ git add . $ git commit -m "Your detailed description of your changes." $ git push origin name-of-your-bugfix-or-feature
When submitting a pull request:
- All existing tests should pass. Please make sure that the test suite passes, both locally and on Travis CI Status on Travis will be visible on a pull request. If you want to enable Travis CI on your own fork, please read the getting started docs.
- New functionality should include tests. Please write reasonable tests for your code and make sure that they pass on your pull request.
- Classes, methods, functions, etc. should have docstrings. The first line of a docstring should be a standalone summary. Parameters and return values should be documented explicitly.
- The API documentation is automatically generated from docstrings, which should conform to NumpPy styling. For examples, see the Napoleon docs.
- Please note that tests are also run via Travis-CI on our documentation.
So be sure that any
.rst
file submissions are properly formatted and tests are passing.
Improving the documentation and testing for code already in EarthPy is a great way to get started if you'd like to make a contribution. Please note that our documentation files are in ReStructuredText (.rst) format and format your pull request accordingly.
To create a vignette for an EarthPy workflow, create a .py
file that shows the
necessary steps to complete the workflow. Make sure the file name begins with
plot
in order to ensure that the vignette is correctly built. Store the
vignette in the examples
folder within the earthpy
directory. EarthPy
uses Sphinx Gallery to build vignettes. Help for formatting and building
vignettes can be found on their website.
To build the documentation, use the command:
$ make docs
By default make docs
will only rebuild the documentation if source
files (e.g., .py or .rst files) have changed. To force a rebuild, use
make -B docs
.
You can preview the generated documentation by opening
docs/_build/html/index.html
in a web browser.
Earthpy uses doctest to test
code in the documentation, which includes docstrings in EarthPy's modules, and
code chunks in the reStructuredText source files.
This enables the actual output of code examples to be checked against expected
output.
When the output of an example is not always identical (e.g., the
memory address of an object), use an ellipsis
(...
) to match any substring of the actual output, e.g.:
>>> print(list(range(20)))
[0, 1, ..., 18, 19]
Earthpy also uses the Matplotlib plot directive in the documentation to
generate figures.
To include a figure in an example, prefix the example with .. plot::
,
e.g.,:
.. plot:: >>> import matplotlib.pyplot as plt >>> plt.plot([1, 2, 3], [4, 5, 6])
EarthPy currently only supports Python 3 (3.6+). Please test code locally in Python 3 when possible (all supported versions will be automatically tested on Travis CI).
EarthPy uses a pre-commit hook that runs the black code autoformatter. Be sure to execute pre-commit install as described above, which will cause black to autoformat code prior to commits. If this step is skipped, black may cause build failures on Travis CI due to formatting issues.
Follow PEP 8 when possible. Some standards that we follow include:
- The first word of a comment should be capitalized with a space following
the
#
sign like this:# This is a comment here
- Variable and function names should be all lowercase with words separated
by
_
. - Class definitions should use camel case - example:
ClassNameHere
.
- The first word of a comment should be capitalized with a space following
the
Imports should be grouped with standard library imports first, 3rd-party libraries next, and EarthPy imports third following PEP 8 standards. Within each grouping, imports should be alphabetized. Always use absolute imports when possible, and explicit relative imports for local imports when necessary in tests.
A reminder for the maintainers on how to deploy. Make sure all your changes are committed, then run:
$ bumpversion patch # possible: major / minor / patch
This will increment the version according to a major release (e.g., 0.1.0 to 1.0.0), a minor release (e.g., 0.1.0 to 0.2.0), or a patch (e.g., 0.1.0 to 0.1.1), following the guidelines for semantic versioning: https://semver.org/.
Bumpversion updates the version number throughout the package, and generates a git commit along with an associated git tag for the new version. For more on bumpversion, see: https://github.com/peritus/bumpversion
To deploy EarthPy, push the commit and the version tags:
$ git push $ git push --tags
Travis will then deploy to PyPI if the build succeeds. Travis will only deploy to PyPI on tagged commits, so remember to push the tags. Once that is done, create a release on GitHub for the new version.
Once you have created a tagged version of EarthPy, you are ready to create a final release. To do this, go to the EarthPy repo on github. And click on the releases tab. Create a new release using the tagged version that you just pushed to GitHub using bumpversion. Once that release is created, a new DOI will be issued on zenodo.
If a test requires a data object such as a GeoDataFrame or numpy array, and copies of that data object are required by multiple tests, we can use pytest fixtures to cleanly create and tear down those objects independently for each test. See earthpy/tests/conftest.py for fixture definitions, and earthpy/tests/test_crop_image.py for example usage of fixtures in tests.