As a scientific community-driven software project, ArviZ welcomes contributions from interested individuals or groups. These guidelines are provided to give potential contributors information to make their contribution compliant with the conventions of the ArviZ project, and maximize the probability of such contributions to be merged as quickly and efficiently as possible.
There are 4 main ways of contributing to the ArviZ project (in descending order of difficulty or scope):
- Adding new or improved functionality to the existing codebase
- Fixing outstanding issues (bugs) with the existing codebase. They range from low-level software bugs to higher-level design problems.
- Contributing or improving the documentation (
docs) or examples (arviz/examples) - Submitting issues related to bugs or desired enhancements
We appreciate being notified of problems with the existing ArviZ code. We prefer that issues be filed the on Github Issue Tracker, rather than on social media or by direct email to the developers.
Please verify that your issue is not being currently addressed by other issues or pull requests by using the GitHub search tool to look for key words in the project issue tracker.
While issue reporting is valuable, we strongly encourage users who are inclined to do so to submit patches for new or existing issues via pull requests. This is particularly the case for simple fixes, such as typos or tweaks to documentation, which do not require a heavy investment of time and attention.
Contributors are also encouraged to contribute new code to enhance ArviZ's functionality, also via pull requests. Please consult the ArviZ documentation to ensure that any new contribution does not strongly overlap with existing functionality.
Before starting a work on a pull request double check that no one else is working on the ticket in both issue tickets and pull requests.
If an issue exists check the ticket to ensure no one else has started work. If first to start work, comment on the ticket to make it evident to others. If the comment looks old or abandoned leave a comment asking if you may start work.
Open an issue ticket for the issue and state that you'll be solving the issue with a pull request. Optionally create a pull request and add [WIP] in the title to indicate Work in Progress.
In the event of two or more people working on the same issue, the general precedence will go to the person who first commented in the issue. If no comments it will go to the first person to submit a PR for review. Each situation will differ though, and the core contributors will make the best judgement call if needed.
The preferred workflow for contributing to ArviZ is to fork the GitHub repository, clone it to your local machine, and develop on a feature branch.
For more instructions see the Pull request checklist
For code generally follow the TensorFlow's style guide or the Google style guide Both more or less follows PEP 8.
Final formatting is done with black. For more detailed steps on a typical development workflow see the Pull request checklist
Docstrings should follow the numpy docstring guide. Extra guidance can also be found in pandas docstring guide. Please reasonably document any additions or changes to the codebase, when in doubt, add a docstring.
The different formatting and aim between numpydoc style type description and
type hints
should be noted. numpydoc style targets docstrings and aims to be human
readable whereas type hints target function definitions and .pyi files and
aim to help third party tools such as type checkers or IDEs. ArviZ does not
require functions to include type hints
however contributions including them are welcome.
If changes are made to a method documented in the
ArviZ API Guide
please consider adding inline documentation examples.
az.plot_posterior is a particularly
good example.
-
Fork the project repository by clicking on the 'Fork' button near the top right of the main repository page. This creates a copy of the code under your GitHub user account.
-
Clone your fork of the ArviZ repo from your GitHub account to your local disk, and add the base repository as a remote:
$ git clone [email protected]:<your GitHub handle>/arviz.git $ cd arviz $ git remote add upstream [email protected]:arviz-devs/arviz.git
-
Create a
featurebranch to hold your development changes:$ git checkout -b my-feature
Always use a
featurebranch. It's good practice to never routinely work on themasterbranch of any repository. -
Project requirements are in
requirements.txt, and libraries used for development are inrequirements-dev.txt. To set up a development environment, you may (probably in a virtual environment) run:$ pip install -r requirements.txt $ pip install -r requirements-dev.txt
Alternatively, there is a script to create a docker environment for development. See: Developing in Docker.
Note: Building the documentation locally requires saving Bokeh plots as
images. To do this, bokeh needs some extra dependencies
that can not be installed with pip and are therefore not in
requirements-dev.txt. To build documentation locally without having to
install these extra dependencies, Docker can be used. See: Building documentation with Docker
-
Develop the feature on your feature branch. Add changed files using
git addand thengit commitfiles:$ git add modified_files $ git commit -m "commit message here"to record your changes locally. After committing, it is a good idea to sync with the base repository in case there have been any changes:
$ git fetch upstream $ git rebase upstream/master
Then push the changes to your GitHub account with:
$ git push -u origin my-feature
-
Go to the GitHub web page of your fork of the ArviZ repo. Click the 'Pull request' button to send your changes to the project's maintainers for review. This will send an email to the committers.
We recommend that your contribution complies with the following guidelines before you submit a pull request:
-
If your pull request addresses an issue, please use the pull request title to describe the issue and mention the issue number in the pull request description. This will make sure a link back to the original issue is created.
-
All public methods must have informative docstrings with sample usage when appropriate.
-
Please prefix the title of incomplete contributions with
[WIP](to indicate a work in progress). WIPs may be useful to (1) indicate you are working on something to avoid duplicated work, (2) request broad review of functionality or API, or (3) seek collaborators. -
All other tests pass when everything is rebuilt from scratch. See Developing in Docker for information on running the test suite locally.
-
When adding additional functionality, provide at least one example script or Jupyter Notebook in the
arviz/examples/folder. Have a look at other examples for reference. Examples should demonstrate why the new functionality is useful in practice and, if possible, compare it to other methods available in ArviZ. -
Added tests follow the pytest fixture pattern
-
Documentation and high-coverage tests are necessary for enhancements to be accepted.
-
Documentation follows Numpy style guide
-
Run any of the pre-existing examples in
docs/source/notebooksthat contain analyses that would be affected by your changes to ensure that nothing breaks. This is a useful opportunity to not only check your work for bugs that might not be revealed by unit test, but also to show how your contribution improves ArviZ for end users. -
If modifying a plot, render your plot to inspect for changes and copy image in the pull request message on Github
You can also check for common programming errors with the following tools:
-
Save plots as part of tests. Plots will save to a directory named test_images by default
$ pytest arviz/tests/base_tests/<name of test>.py --save
-
Optionally save plots to a user named directory. This is useful for comparing changes across branches
$ pytest arviz/tests/base_tests/<name of test>.py --save user_defined_directory
-
Code coverage cannot decrease. Coverage can be checked with pytest-cov package:
$ pip install pytest pytest-cov coverage $ pytest --cov=arviz --cov-report=html arviz/tests/
-
Your code has been formatted with black with a line length of 100 characters. Note that black only runs in Python 3.6+
$ pip install black $ black arviz/ examples/ asv_benchmarks/
-
Your code passes pylint
$ pip install pylint $ pylint arviz/
-
No code style warnings, check with:
$ ./scripts/lint.sh
We have provided a Dockerfile which helps for isolating build problems, and local development.
Install Docker for your operating system, clone this repo. Docker will generate an environment with your local copy of arviz with all the packages in Dockerfile.
Predefined docker commands can be run with a ./scripts/container.sh (on Linux and macOS)
and with ./scripts/container.ps1. The scripts enables developer easily to call predefined docker commands.
User can use one or multiple flags.
They are executed on the following order: clear-cache, build, test, docs, shell, notebook, lab
$ ./scripts/container.sh --clear-cache
$ ./scripts/container.sh --build
$ ./scripts/container.sh --test
$ ./scripts/container.sh --docs
$ ./scripts/container.sh --shell
$ ./scripts/container.sh --notebook
$ ./scripts/container.sh --lab
$ powershell.exe -File ./scripts/container.ps1 --clear-cache
$ powershell.exe -File ./scripts/container.ps1 --build
$ powershell.exe -File ./scripts/container.ps1 --test
$ powershell.exe -File ./scripts/container.ps1 --docs
$ powershell.exe -File ./scripts/container.ps1 --shell
$ powershell.exe -File ./scripts/container.ps1 --notebook
$ powershell.exe -File ./scripts/container.ps1 --lab
Testing the code using docker consists of executing the same file 3 times (you may need root privileges to run it).
First run ./scripts/container.sh --clear-cache. Then run ./scripts/container.sh --build. This starts a local docker image called arviz. Finally run the tests with ./scripts/container.sh --test. This should be quite close to how the tests run on TravisCI.
NOTE: If you run into errors due to __pycache__ files (i.e. while testing in
docker after testing locally or installing with pip after testing with
docker), try running ./scripts/container.sh --clear-cache before the errored
command.
Once the Docker image is built with ./scripts/container.sh --build, interactive containers can also be run. Therefore, code can be edited and executed using the docker container, but modifying directly the working directory of the host machine.
To start a bash shell inside Docker, run:
$ docker run --mount type=bind,source="$(pwd)",target=/opt/arviz/ -it arviz bash
and for Windows, use %CD% on cmd.exe and $pwd.Path on powershell.
$ docker run --mount type=bind,source=%CD%,target=/opt/arviz/ -it arviz bash
Alternatively, to start a jupyter notebook, there are two steps, first run:
$ docker run --mount type=bind,source="$(pwd)",target=/opt/arviz/ --name jupyter-dock -it -d -p 8888:8888 arviz
$ docker exec jupyter-dock bash -c "pip install jupyter"
$ docker exec -it jupyter-dock bash -c "jupyter notebook --ip 0.0.0.0 --no-browser --allow-root"
and the same on Windows
$ docker run --mount type=bind,source=%CD%,target=/opt/arviz/ --name jupyter-dock -it -d -p 8888:8888 arviz
$ docker exec jupyter-dock bash -c "pip install jupyter"
$ docker exec -it jupyter-dock bash -c "jupyter notebook --ip 0.0.0.0 --no-browser --allow-root"
This will output something similar to http://(<docker container id> or <ip>):8888/?token=<token id>, and can be accessed at http://localhost:8888/?token=<token id>.
The documentation can be build with Docker by running ./scripts/container.sh --docs. The docker image contains by default all dependencies needed
for building the documentation. After having build the docs in the Docker
container, they can be checked at doc/build.
To run the benchmark tests do the following:
$ pip install asv
$ cd arviz
$ asv run