diff --git a/docs/contributing.rst b/docs/contributing.rst index 9b277079b..b79560961 100644 --- a/docs/contributing.rst +++ b/docs/contributing.rst @@ -9,12 +9,11 @@ First of all: thank you! Contributions can be made in different ways, not only code! As we follow the `all-contributors`_ specification, any contribution will be recognised accordingly. -The first thing you might want to do, is to have a look at the `contributor guide `_ page as well as the `code of conduct `_. +Follow these steps to get started: -The second thing is to check you have ``git`` and ``pip`` installed in your system. - -The third thing is to install ``phys2bids`` as a developer. -This will let you run the program with the latest modification, without requiring you to re-install it every time. +1. Have a look at the `contributor guide `_ page as well as the `code of conduct `_. +2. Make sure that you have a GitHub account. You can set up a `free GitHub account `_; here are some `instructions `_. +3. If you intend to contribute code and/or use the ``physiopy`` packages in any way, check that you have ``git`` and ``pip`` installed on your system. Then install ``phys2bids`` as a developer. This will let you run the program with the latest modification, without requiring to re-install it every time. .. _`all-contributors`: https://github.com/all-contributors/all-contributors @@ -23,7 +22,6 @@ This will let you run the program with the latest modification, without requirin If it is, you might need to use ``pip`` instead of ``pip3``, although some OSs do adopt ``pip3`` anyway. If you want to check, type ``python --version`` in a terminal. - Linux and mac developer installation ------------------------------------ diff --git a/docs/contributorfile.rst b/docs/contributorfile.rst index 9f09776ef..7df84b442 100644 --- a/docs/contributorfile.rst +++ b/docs/contributorfile.rst @@ -3,9 +3,9 @@ ================================= How to contribute to ``physiopy`` ================================= -Welcome to the physiopy organisation! It’s great news you’re thinking about contributing! -Working with many people from many different places is great, but sometimes this means the code can become messy due to the many different ways a contribution can be made. For this reason, we have set up some guidelines for contributions - to help you get along ASAP! -Before you start you'll need to set up a free `GitHub `_ account and sign in. Here are some `instructions `_. +Welcome to the ``physiopy`` organisation! It’s great news that you’re thinking about contributing! + +Working with many people from many different places is great, but sometimes this means that code can become messy due to the many different ways a contribution can be made. For this reason, we have set up some guidelines for contributions - to help you get involved ASAP! If you lack knowledge in python development / github use / physiological data handling, don’t be scared! Try to jump in anyway. Most of the original contributors learned these things exactly this way - jumping in and hoping to fall in the right way without breaking too many bones. Do you want to jump in but don’t exactly know where/how? You can drop a few lines in `gitter `_, so we can help you find something that suits you! Already know what you're looking for in this guide? Jump to the following sections: @@ -18,6 +18,7 @@ Already know what you're looking for in this guide? Jump to the following sectio - `Contributing with test files <#testfile>`_ - `Contributing documentation through GitHub <#documenting>`_ - `Contributing code through GitHub <#code>`_ + - `Contributing with Pull Requests Review <#reviews>`_ - `Issues and Milestones <#issuesmilestones>`_ - `Labels <#labeltypes>`_ - `Issues & PRs labels <#issueprlabels>`_ @@ -26,20 +27,20 @@ Already know what you're looking for in this guide? Jump to the following sectio - `Good First Issues <#g1i>`_ - `Contribution workflow <#workflow>`_ - `Pull Requests <#pr>`_ +- `Reviewing PRs <#reviewing>`_ - `Style Guide <#styling>`_ - `Automatic Testing <#testing>`_ -- `Recognizing contributors <#recognising>`_ .. _aims: -Aims of physiopy ----------------- -physiopy is a **very** young project developed by a bunch of researchers from the two different sides of the Atlantic Ocean (for now). -Our main goal is to help collect, analyse and share physiological data interfacing with (MRI) neuroimaging. We’re trying to do so by: +Aims of ``physiopy`` +-------------------- +``physiopy`` is a **very** young project developed by a bunch of researchers from the two different sides of the Atlantic Ocean (for now). +Our main goal is to help collect, analyse and share physiological data, interfacing with (MRI) neuroimaging. We’re trying to do so by: -1. Write packages to make a user-friendly pipeline to deal with physiological data. -2. Organising a documentation containing tips and strategies on how to collect such data and use our packages. -3. Write packages that take into account the use of such data in combination with neuroimaging (MRI) - by getting everything ready for that analysis. +1. Writing packages to make a user-friendly pipeline to deal with physiological data. +2. Writing packages that take into account the use of this physiological data in combination with neuroimaging (MRI) analysis. +3. Providing documentation containing tips and strategies on how to collect such data and use our packages. 4. Help set a standard for these data, albeit without forcing users to use it. 5. Be an excuse for educational purposes on topics like Git/GitHub, Python3, physiology and related tools/topics. @@ -47,7 +48,7 @@ Our main goal is to help collect, analyse and share physiological data interfaci Joining the conversation ------------------------ -We’re trying to keep all the conversation related to the project development in GitHub `issues `_. +We’re trying to keep all the conversation related to the project development in GitHub `issues `_. We maintain a `gitter chat room `_ for more informal conversations and general project updates. We also have a dev call once a month - specifically the second Thursday of the month! If you want to participate, drop a line in gitter! When interacting in the common channels, please adhere to our `code of conduct `_. @@ -57,18 +58,20 @@ When interacting in the common channels, please adhere to our `code of conduct < Contributions ------------- +We welcome and recognize `all contributions `_ from documentation to testing and code development. You can see a list of current contributors in the README (kept up to date by the `all contributors bot `_). + .. _smalldocs: Contributing with small documentation changes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ If you are new to GitHub and just have a small documentation change recommendation (such as: typos detection, small improvements in the content, ...), please open an issue in the relative project, and label it with the “Documentation” label. -Chances are those types of changes are easily doable with the online editor, which means you can do them, or ask for help from the developers! +Chances are those types of changes are easily doable with GitHub's online editor, which means you can do them, or ask for help from the developers! .. _usertests: Contributing with User testing ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -Another, non-coding friendly way to contribute to physiopy is by testing the packages. +Another, non-coding friendly way to contribute to ``physiopy`` is by testing the packages. There are different kinds of tests, but to simplify things you can think mainly about automatic tests and user tests. To know more about **Automatic tests**, you can read the `testing section <#testing>`_. **User testing** are warm, human, emotional and opinionated tests that not only check that the code is doing what it needs to do, but also whether there’s a better way to do it - namely better reports, clearer screen outputs, warnings and exceptions, unexpected bugs that have to be corrected. @@ -78,10 +81,10 @@ If you want to perform one, open an issue on GitHub or drop a comment in Gitter, Contributing with test files ~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -At physiopy we always try to imagine and support every possible setting out there. However, our imagination has a limit - but if you think our packages should process a specific format/setting that you have, we’re more than glad to do so! +At ``physiopy`` we always try to imagine and support every possible setting out there. However, our imagination has a limit - but if you think our packages should process a specific format/setting that you have, we’re more than glad to do so! To make it happen, we need an example of the file we want to process, so you will have to share it with us (and the rest of the world)! The contribution can be a full file of data that you already acquired, a part of that file (pay attention to what is the minimum you need to share!), or mock data. The file contribution should come with a json file of the same name that contains the necessary information to run ``phys2bids`` on that file contribution. There is a `json blueprint in OSF `_, you can download it and adapt it. Note that the frequency list **has to be expressed in Hz** as an integer or float. -To contribute with a test file, open an Issue in GitHub and label it with *Test*. We’ll help you add the file in our +To contribute with a test file, open an Issue in GitHub and label it with *Test*. We’ll help you add the file in our `OSF `_ space. We’re extremely grateful for this type of contribution - so grateful that we asked allcontributors to add a dedicated category! @@ -89,25 +92,33 @@ We’re extremely grateful for this type of contribution - so grateful that we a Contributing documentation through GitHub ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -We use `readthedocs `_ to create our documentation. Every contribution is welcome - and it follows the steps of a code contribution. +We use `readthedocs `_ to create our documentation. Every contribution is welcome and it follows the same steps as a code contribution, explained below. .. _code: Contributing code through GitHub ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -This section covers 90% of the contributions a project like physiopy receives - code, documentation and tests. -The best way to make this kind of contributions, in a nutshell, would be: +This section covers 90% of the contributions a project like ``physiopy`` receives - code, documentation and tests. +The best way to make this kind of contributions, in a nutshell, is to: 1. Open an issue with the intended modifications. 2. Label it, discuss it, (self-)assign it. 3. Open a Pull Request (PR) to resolve the issue and label it. 4. Wait for a review, discuss it or comply, repeat until ready. Issues and PR chats are great to maintain track of the conversation on the contribution. They are based upon GitHub-flavoured `Markdown `_. GitHub has a helpful page on `getting started with writing and formatting Markdown on GitHub `_. +.. _reviews: + +Contributing with Pull Requests Reviews +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ +One big challenge in this kind of software development is merging code accurately but without waiting too much time. +For this reason, Reviewers for PRs are more than welcome! It is a task that requires some experience, but it's very necessary! +Read the `related section below <#reviewing>`_ to start! + .. _issuesmilestones: Issues and Milestones --------------------- -At physiopy, we use Issues and Milestones to keep track of and organise our workflow. +At ``physiopy``, we use Issues and Milestones to keep track of and organise our workflow. - **Issues** describe pieces of work that need to be completed to move the project forwards. We try to keep them as simple and clear as possible: an issue should describe a unitary, possibly small piece of work (unless it’s about refactoring). Don’t be scared of opening many issues at once, if it makes sense! Just check that what you’re proposing is not listed in a previous issue (open or closed) yet - we don’t like doubles. Issues get labelled. That helps the contributors to know what they’re about. Check the label list to know what types are there, and use them accordingly! Issues can also be **assigned**. If you want to work on an assigned issue, ask permission first! - **Milestones** set the higher level workflow. They sketch deadlines and important releases. Issues are assigned to these milestones by the maintainers. If you feel that an issue should be assigned to a specific milestone but the maintainers have not done so, discuss it in the issue chat or in Gitter! We might have just missed it, or we might not (yet) see how it aligns with the overall project structure/milestone. @@ -168,7 +179,7 @@ Other labels Good First Issues ~~~~~~~~~~~~~~~~~ -Good First Issues are issues that are either very simple, or that help knowing the programs or the language better. We use it to help contributors with less experience to learn and familiarise with Git, GitHub, Python3, and physiology. +Good First Issues are issues that are either very simple, or that help the contributor get to know the programs or the languages better. We use it to help contributors with less experience to learn and familiarise with Git, GitHub, Python3, and physiology. We invite more expert contributors to avoid those issues, leave them to beginners and possibly help them out in the resolution of the issue. However, if the issue is left unassigned or unattended for long, and it’s considered important or urgent, anyone can tackle it. .. _workflow: @@ -176,27 +187,88 @@ We invite more expert contributors to avoid those issues, leave them to beginner Contribution workflow --------------------- There are many descriptions of a good contribution workflow out there. For instance, we suggest to have a look at `tedana's workflow `_. -At physiopy, we follow a very similar workflow. The only two differences are: +At ``physiopy``, we follow a very similar workflow. The only three differences are: -- We ask you to test the code locally before merging it, and then, if possible, write some automatic tests for the code to be run in our Continuous Integration! Check the testing section below to know more. -- We suggest opening a draft PR as soon as you can - so it’s easier for us to help you! +- If you see an open issue that you would like to work on, check if it is assigned. If it is, ask the assignee if they need help before starting to work on it. +- We ask you to test the code locally before merging it, and then, if possible, write some automatic tests for the code to be run in our Continuous Integration! Check the testing section below to know more. +- We suggest opening a draft PR as soon as you can - so it’s easier for us to help you! .. _pr: Pull Requests ------------- -To improve understanding pull requests "at a glance", we use the same labels used for issues. Multiple labels can be assigned - just think which ones suit your PR the most! -When opening a pull request, assign it to at least one label. +To improve understanding pull requests "at a glance", we use the labels listed above. Multiple labels can be assigned, but it's a good idea to keep different types of contributions separated, unless they are minimal - for instance, you might want to open a PR for code, one for documentation and one for testing. +In general, if you're tempted to assign more than one label that would trigger a release, you might want to split your PR instead. +When opening a pull request, assign it at least one label. -We encourage you to open a PR as soon as possible - even before you finish working on them. This is useful especially to you - so that you can receive comments and suggestions early on, rather than having to process a lot of comments in the final review step! +We encourage you to open a PR as soon as possible - even before you finish working on them. This is useful especially to you - so that you can receive comments and suggestions early on, rather than having to process a lot of comments in the final review step! However, if it’s an incomplete PR, please open a **Draft PR**. That helps us process PRs by knowing which one to have a look first - and how picky to be when doing so. + +Reviewing PRs is a time consuming task, that can be stressful for both the reviewer and the author. Avoiding wasting time and the need of little fixes - such as fixing grammar mistakes and typos, styling code, or adopting conventions - is a good start for a successful (and quick) review. Before graduating a Draft PR to a PR ready for review, please check that: + +- You did all you wanted to include in your PR. If at a later stage you realise something is missing and it's not a minor thing, you will need to open a new PR. +- If your contribution contains code or tests, you run and passed all the tests locally with `pytest`. +- If you're writing documentation, you built it locally with `sphinx` and the format is what you intended. +- Your code is harmonious with the rest of the code - no repetitions of any sort! +- Your code respects the `adopted Style <#styling>`_, especially: + - Your code is lintered adequately and respects the `PEP8 `_ convention. + - Your docstrings follow the `numpydoc `_ convention. + - There is no grammar mistake or typo and the text is fluid. + - The code is sufficiently commented and the comments are clear. + - Your PR title is clear enough to be meaningful when appended to the version changelog. +- You have the correct labels. + To be merged, PRs have to: -1. Pass all the Travis CI tests. +1. Pass all the Travis CI tests, and possibly all the codecov checks. 2. Have the necessary amount of approving reviews, even if you’re a long time contributor. You can ask one (or more) contributor to do that review, if you think they align more with the content of your PR. You need **one** review for documentation, tests, and small changes, and **two** reviews for bugs, refactoring and enhancements. -3. Contain at least a unit test for your contribution, if the PR contains code (it would be better if it contains an integration or function test and all the breaking tests necessary). If you’re not confident about writing tests, it is possible to refer to an issue that asks for the test to be written, or another (Draft) PR that contains the tests required. +3. Have at least a release-related label (or a `Skip release` label). +4. Have a short title that clearly explains in one sentence the aim of the PR. +5. Contain at least a unit test for your contribution, if the PR contains code (it would be better if it contains an integration or function test and all the breaking tests necessary). If you’re not confident about writing tests, it is possible to refer to an issue that asks for the test to be written, or another (Draft) PR that contains the tests required. -As we’re trying to maintain at least a 90% code coverance, you’re strongly encouraged to write all the necessary tests not to drop below the threshold. If our coverance becomes too low, you might be asked to add more tests and/or your PR might be rejected. +As we’re trying to maintain at least a 90% code coverage, you’re strongly encouraged to write all the necessary tests not to drop below the threshold. If our coverage becomes too low, you might be asked to add more tests and/or your PR might be rejected. See the `Automatic Testing <#testing>`_ section. + +Don't merge your own pull request! That's a task for the main reviewer of your PR or the project manager. Remember that the project manager doesn't have to be a reviewer of your PR! + +.. _reviewing: + +Reviewing PRs +------------- +Reviewing PRs is an extremely important task in collaborative development. In fact, it is probably the task that requires the most time in the development, and it can be stressful for both the reviewer and the author. Remember that, as a PR Reviewer, you are granting that the codes works and integrates well with the rest of the repository, hence **you are responsible for the quality of the code and its next version release**. If it doesn't integrate well, later PR reviewers might have to ask for broader changes than expected. +Due to its importance, there are some good rules of thumbs that we need to follow while reviewing PRs: + +- Be respectful to the PR authors and be clear in what you are asking/suggesting - remember that, like you, they are contributing their spare time and doing their best job! +- If there is a Draft PR, you can comment on its development in the message board or making "Comment" reviews. Don't ask for changes, and especially, **don't approve the PR** +- If the PR graduated from Draft to full PR, check that it follows the sections `Pull requests <#pr>`_ and `Style Guide <#styling>`_ of these guidelines. If not, invite the author to do so before starting a review. +- **Don't limit your review to the parts that are changed**. Look at the entire file, see if the changes fit well in it, and see if the changes are properly addressed everywhere in the code - in the documentation, in the tests, and in other functions. Sometimes the differences reported don't show the full impact of the PR in the repository! +- Unless it's for typo fixes or similar, invite the author of the PR to make changes before actually doing them yourself. Request changes via comments or in the message board or by checking out the PR locally, making changes and then submitting a PR to the author's branch. +- If you're reviewing documentation, build it locally with `sphinx`. +- If you're asking for changes, **don't approve the PR**. Approve it only after everything was sufficiently addressed. Someone else might merge the PR in taking your word for granted. +- If you are the main reviewer, and the last reviewer required to approve the PR, merge the PR! + +Before approving and/or merging PRs, be sure that: + +- All the tests in Travis CI pass without errors. +- Prefereably, codecov checks pass as well - if they don't, pin the project manager. +- The title describes the content of the PR clearly enough to be meaningful on its own - remember that it will appear in the version changelog! +- The PR has the appropriate labels to trigger the appropriate version release and update the contributors table. + +.. _mainreviewer: + +Main reviewer +~~~~~~~~~~~~~ +At ``physiopy`` we use the "Assignees" section of a PR to mark the **main reviewer** for that PR. +The main reviewer: + +- Takes care of the reviewing process of the PR, in particular: + - Invites the reviewers to finish their review in a relatively short time. + - Checks that this document was respected in all its part, especially the part about `Pull Requests <#pr>`_. + +- Is the one that is going to merge the PR. +- After the PR got merged and a new release was triggered, checks that: + - The documentation was updated correctly (if changed). + - The Pipy version of the repository coincides with the new release (if changed). + - New contributors or forms of contributions were correctly added in the README (if changed). .. _styling: @@ -205,13 +277,14 @@ Style Guide Docstrings should follow `numpydoc `_ convention. We encourage extensive documentation. The python code itself should follow `PEP8 `_ convention whenever possible: there are continuous integration tests checking that! You can use linters to help you write your code following style conventions. Linters are add-ons that you can run on the written script file. We suggest the use of **flake8** for Python 3. Many editors (Atoms, VScode, Sublimetext, ...) support addons for online lintering, which means you’ll see warnings and errors while you write the code - check out if your does! +Since we adopt `auto `_, the PR title will be automatically reported as part of the changelog when updating versions. Try to describe in one (short) sentence what your PR is about - possibly using the imperative and starting with a capital letter. For instance, a good PR title could be: ``Implement support for files`` or ``Reorder dictionary entries``, rather than `` support`` or ``reorders keys``. .. _testing: Automatic Testing ----------------- -physiopy uses Continuous Integration (CI) to make life easier. In particular, we use the `Travis CI `_ platform to run automatic testing! -**Automatic tests** are cold, robotic, emotionless, and opinionless tests that check that the program is doing what it is expected to. They are written by the developers and run (by Travis CI) every time they send a Pull Request to physiopy repositories. They complement the warm, human, emotional and opinionated **user tests**, as they tell us if a piece of code is failing. +``physiopy`` uses Continuous Integration (CI) to make life easier. In particular, we use `Travis CI `_ and `Azure `_ to run automatic testing! +**Automatic tests** are cold, robotic, emotionless, and opinionless tests that check that the program is doing what it is expected to. They are written by the developers and run (by Travis CI) every time they send a Pull Request to ``physiopy`` repositories. They complement the warm, human, emotional and opinionated **user tests**, as they tell us if a piece of code is failing. Travis CI uses `pytest `_ to run the tests. The great thing about it is that you can run it in advance on your local version of the code! We can measure the amount of code that is tested with [codecov]8https://docs.pytest.org/en/latest/), which is an indication of how reliable our packages are! We try to maintain a 90% code coverage, and for this reason, PR should contain tests! The four main type of tests we use are: @@ -225,12 +298,6 @@ The four main type of tests we use are: 4. Functional tests If integration tests and unit tests could have babies, those would be functional tests. In practice, this kind of tests check that an output is produced, and *also* that it contains what it should contain. If a function should output a new file or an object, this test passes only if the file exists *and* it is like we expect it to be. They are run on real or mock data, and call the program itself or a function. -.. _recognising: - -Recognising contributors ------------------------- -We welcome and recognize `all contributions `_ from documentation to testing to code development. You can see a list of current contributors in the README (kept up to date by the `all contributors bot `_). - **Thank you!** -*— Based on contributing guidelines from the `STEMMRoleModels `_ project.* +*— Based on contributing guidelines from the `STEMMRoleModels `_ project.*