Encouraging more people to help develop pvlib-python is essential to our success. Therefore, we want to make it easy and rewarding for you to contribute.

There is a lot of material in this section, aimed at a variety of contributors from novice to expert. Don’t worry if you don’t (yet) understand parts of it.

Easy ways to contribute#

Here are a few ideas for how you can contribute, even if you are new to pvlib-python, git, or Python:

How to contribute new code#

The basics#

Contributors to pvlib-python use GitHub’s pull requests to add/modify its source code. The GitHub pull request process can be intimidating for new users, but you’ll find that it becomes straightforward once you use it a few times. Please let us know if you get stuck at any point in the process. Here’s an outline of the process:

  1. Create a GitHub issue and get initial feedback from users and maintainers. If the issue is a bug report, please include the code needed to reproduce the problem.

  2. Obtain the latest version of pvlib-python: Fork the pvlib-python project to your GitHub account, git clone your fork to your computer.

  3. Make some or all of your changes/additions and git commit them to your local repository.

  4. Share your changes with us via a pull request: git push your local changes to your GitHub fork, then go to GitHub make a pull request.

The Pandas project maintains an excellent contributing page that goes into detail on each of these steps. Also see GitHub’s Set Up Git and Using Pull Requests.

We strongly recommend using virtual environments for development. Virtual environments make it trivial to switch between different versions of software. This astropy guide is a good reference for virtual environments. If this is your first pull request, don’t worry about using a virtual environment.

You must include documentation and unit tests for any new or improved code. We can provide help and advice on this after you start the pull request. See the Testing section below.

Pull request scope#

This section can be summed up as “less is more”.

A pull request can quickly become unmanageable if too many lines are added or changed. “Too many” is hard to define, but as a rule of thumb, we encourage contributions that contain less than 50 lines of primary code. 50 lines of primary code will typically need at least 250 lines of documentation and testing. This is about the limit of what the maintainers can review on a regular basis.

A pull request can also quickly become unmanageable if it proposes changes to the API in order to implement another feature. Consider clearly and concisely documenting all proposed API changes before implementing any code. Modifying api.rst and/or the latest whatsnew file can help formalize this process.

Questions about related issues frequently come up in the process of addressing implementing code for a pull request. Please try to avoid expanding the scope of your pull request (this also applies to reviewers!). We’d rather see small, well-documented additions to the project’s technical debt than see a pull request languish because its scope expanded beyond what the reviewer community is capable of processing.

Of course, sometimes it is necessary to make a large pull request. We only ask that you take a few minutes to consider how to break it into smaller chunks before proceeding.

pvlib-python contains 3 “layers” of code: functions, PVSystem/Location, and ModelChain. We recommend that contributors focus their work on only one or two of those layers in a single pull request. New models are not required to be available to the higher-level API!

When should I submit a pull request?#

The short answer: anytime.

The long answer: it depends. If in doubt, go ahead and submit. You do not need to make all of your changes before creating a pull request. Your pull requests will automatically be updated when you commit new changes and push them to GitHub.

There are pros and cons to submitting incomplete pull-requests. On the plus side, it gives everybody an easy way to comment on the code and can make the process more efficient. On the minus side, it’s easy for an incomplete pull request to grow into a multi-month saga that leaves everyone unhappy. If you submit an incomplete pull request, please be very clear about what you would like feedback on and what we should ignore. Alternatives to incomplete pull requests include creating a gist or experimental branch and linking to it in the corresponding issue.

The best way to ensure that a pull request will be reviewed and merged in a timely manner is to:

  1. Start by creating an issue. The issue should be well-defined and actionable.

  2. Ask the maintainers to tag the issue with the appropriate milestone.

  3. Make a limited-scope pull request. It can be a lot of work to check all of the boxes in pull request guidelines, especially for pull requests with a lot of new primary code. See Pull request scope.

  4. Tag pvlib community members or @pvlib when the pull request is ready for review. (see Pull request reviews)

Pull request reviews#

The pvlib community and maintainers will review your pull request in a timely fashion. Please “ping” @pvlib if it seems that your pull request has been forgotten at any point in the pull request process.

Keep in mind that the PV modeling community is diverse and each pvlib community member brings a different perspective when reviewing code. Some reviewers bring years of expertise in the sub-field that your code contributes to and will focus on the details of the algorithm. Other reviewers will be more focused on integrating your code with the rest of pvlib, ensuring that it is feasible to maintain, that it meets the code style guidelines, and that it is comprehensively tested. Limiting the scope of the pull request makes it much more likely that all of these reviews can be conducted and any issues can be resolved in a timely fashion.

Sometimes it’s hard for reviewers to be immediately available, so the right amount of patience is to be expected. That said, interested reviewers should do their best to not wait until the last minute to put in their two cents.

Code style#

pvlib python generally follows the PEP 8 – Style Guide for Python Code. Maximum line length for code is 79 characters.

pvlib python uses a mix of full and abbreviated variable names. See Variables and Symbols. We could be better about consistency. Prefer full names for new contributions. This is especially important for the API. Abbreviations can be used within a function to improve the readability of formulae.

Set your editor to strip extra whitespace from line endings. This prevents the git commit history from becoming cluttered with whitespace changes.

Please see Documentation for information specific to documentation style.

Remove any logging calls and print statements that you added during development. warning is ok.

We typically use GitHub’s “squash and merge” feature to merge your pull request into pvlib. GitHub will condense the commit history of your branch into a single commit when merging into pvlib-python/main (the commit history on your branch remains unchanged). Therefore, you are free to make commits that are as big or small as you’d like while developing your pull request.


Documentation must be written in numpydoc format format which is rendered using the Sphinx Napoleon extension.

The numpydoc format includes a specification for the allowable input types. Python’s duck typing allows for multiple input types to work for many parameters. pvlib uses the following generic descriptors as short-hand to indicate which specific types may be used:

  • dict-like : dict, OrderedDict, pd.Series

  • numeric : scalar, np.array, pd.Series. Typically int or float dtype.

  • array-like : np.array, pd.Series. Typically int or float dtype.

Parameters that specify a specific type require that specific input type.

Read the Docs will automatically build the documentation for each pull request. Please confirm the documentation renders correctly by following the docs/ link within the checks status box at the bottom of the pull request.

Building the documentation#

Building the documentation locally is useful for testing out changes to the documentation’s source code without having to repeatedly update a PR and have Read the Docs build it for you. Building the docs locally requires installing pvlib python as an editable library (see Installation for instructions). First, install the doc dependencies specified in the EXTRAS_REQUIRE section of An easy way to do this is with:

pip install pvlib[doc]    # on Mac:  pip install "pvlib[doc]"

Note: Anaconda users may have trouble using the above command to update an older version of docutils. If that happens, you can update it with conda (e.g. conda install docutils=0.15.2) and run the above command again.

Once the doc dependencies are installed, navigate to /docs/sphinx and execute:

make html

Be sure to skim through the output of this command because Sphinx might emit helpful warnings about problems with the documentation source code. If the build succeeds, it will make a new directory docs/sphinx/build with the documentation’s homepage located at build/html/index.html. This file can be opened with a web browser to view the local version like any other website. Other output formats are available; run make help for more information.

Note that Windows users need not have the make utility installed as pvlib includes a make.bat batch file that emulates its interface.


Developers must include comprehensive tests for any additions or modifications to pvlib. New unit test code should be placed in the corresponding test module in the pvlib/tests directory.

A pull request will automatically run the tests for you on a variety of platforms (Linux, Mac, Windows) and python versions. However, it is typically more efficient to run and debug the tests in your own local environment.

To run the tests locally, install the test dependencies specified in the file. See Installation instructions for more information.

pvlib’s unit tests can easily be run by executing pytest on the pvlib directory:

pytest pvlib

or, for a single module:

pytest pvlib/tests/

or, for a single test:

pytest pvlib/tests/

We suggest using pytest’s --pdb flag to debug test failures rather than using print or logging calls. For example:

pytest pvlib --pdb

will drop you into the pdb debugger at the location of a test failure. As described in Code style, pvlib code does not use print or logging calls, and this also applies to the test suite (with rare exceptions).

To include all network-dependent tests, include the --remote-data flag to your pytest call:

pytest pvlib --remote-data

And consider adding @pytest.mark.remote_data to any network dependent test you submit for a PR.

pvlib-python contains 3 “layers” of code: functions, PVSystem/Location, and ModelChain. Contributors will need to add tests that correspond to the layers that they modify.


Tests of core pvlib functions should ensure that the function returns the desired output for a variety of function inputs. The tests should be independent of other pvlib functions (see GH394). The tests should ensure that all reasonable combinations of input types (floats, nans, arrays, series, scalars, etc) work as expected. Remember that your use case is likely not the only way that this function will be used, and your input data may not be generic enough to fully test the function. Write tests that cover the full range of validity of the algorithm. It is also important to write tests that assert the return value of the function or that the function throws an exception when input data is beyond the range of algorithm validity.


The PVSystem and Location classes provide convenience wrappers around the core pvlib functions. The tests in and should ensure that the method calls correctly wrap the function calls. Many PVSystem/Location methods pass one or more of their object’s attributes (e.g. PVSystem.module_parameters, Location.latitude) to a function. Tests should ensure that attributes are passed correctly. These tests should also ensure that the method returns some reasonable data, though the precise values of the data should be covered by function-specific tests discussed above.

We prefer to use the pytest-mock framework to write these tests. The test below shows an example of testing the PVSystem.ashraeiam method. mocker is a pytest-mock object. mocker.spy adds features to the pvsystem.ashraeiam function that keep track of how it was called. Then a PVSystem object is created and the PVSystem.ashraeiam method is called in the usual way. The PVSystem.ashraeiam method is supposed to call the pvsystem.ashraeiam function with the angles supplied to the method call and the value of b that we defined in module_parameters. The pvsystem.ashraeiam.assert_called_once_with tests that this does, in fact, happen. Finally, we check that the output of the method call is reasonable.

def test_PVSystem_ashraeiam(mocker):
    # mocker is a pytest-mock object.
    # mocker.spy adds code to a function to keep track of how it is called
    mocker.spy(pvsystem, 'ashraeiam')

    # set up inputs
    module_parameters = {'b': 0.05}
    system = pvsystem.PVSystem(module_parameters=module_parameters)
    thetas = 1

    # call the method
    iam = system.ashraeiam(thetas)

    # did the method call the function as we expected?
    # mocker.spy added assert_called_once_with to the function
    pvsystem.ashraeiam.assert_called_once_with(thetas, b=module_parameters['b'])

    # check that the output is reasonable, but no need to duplicate
    # the rigorous tests of the function
    assert iam < 1.

Avoid writing PVSystem/Location tests that depend sensitively on the return value of a statement as a substitute for using mock. These tests are sensitive to changes in the functions, which is not what we want to test here, and are difficult to maintain.


The tests in should ensure that ModelChain.__init__ correctly configures the ModelChain object to eventually run the selected models. A test should ensure that the appropriate method is actually called in the course of ModelChain.run_model. A test should ensure that the model selection does have a reasonable effect on the subsequent calculations, though the precise values of the data should be covered by the function tests discussed above. pytest-mock can also be used for testing ModelChain.

The example below shows how mock can be used to assert that the correct PVSystem method is called through ModelChain.run_model.

def test_modelchain_dc_model(mocker):
    # set up location and system for model chain
    location = location.Location(32, -111)
    system = pvsystem.PVSystem(module_parameters=some_sandia_mod_params,

    # mocker.spy adds code to the system.sapm method to keep track of how
    # it is called. use returned mock object m to make assertion later,
    # but see example above for alternative
    m = mocker.spy(system, 'sapm')

    # make and run the model chain
    mc = ModelChain(system, location,
                    aoi_model='no_loss', spectral_model='no_loss')
    times = pd.date_range('20160101 1200-0700', periods=2, freq='6H')

    # assertion fails if PVSystem.sapm is not called once

    # use `assert m.call_count == num` if function should be called
    # more than once

    # ensure that dc attribute now exists and is correct type
    assert isinstance(mc.dc, (pd.Series, pd.DataFrame))


pvlib includes a small number of performance benchmarking tests. These tests are run using the airspeed velocity tool. We do not require new performance tests for most contributions at this time. Pull request reviewers will provide further information if a performance test is necessary. See our README for instructions on running the benchmarks.

This documentation#

If this documentation is unclear, help us improve it! Consider looking at the pandas documentation for inspiration.

Code of Conduct#

All contributors are expected to adhere to the Contributor Code of Conduct.