QtPy is part of the Spyder IDE Github org, and is developed with standard Github flow.
If you're not comfortable with at least the basics of git
and GitHub, we recommend reading beginner tutorials such as GitHub's Git Guide, its introduction to basic Git commands and its guide to the fork workflow, or (if you prefer) their video equivalents.
However, this contributing guide should fill you in on most of the basics you need to know.
Let us know if you have any further questions, and we look forward to your contributions!
Discover a bug? Want a new feature? Open an issue! Make sure to describe the bug or feature in detail, with reproducible examples and references if possible, what you are looking to have fixed/added. While we can't promise we'll fix everything you might find, we'll certainly take it into consideration, and typically welcome pull requests to resolve accepted issues.
Note: You may need to substitute python3
for python
in the commands below on some Linux distros where python
isn't mapped to python3
(yet).
First, navigate to the project repository in your web browser and press the Fork
button to make a personal copy of the repository on your own Github account.
Then, click the Clone or Download
button on your repository, copy the link and run the following on the command line to clone the repo:
git clone <LINK-TO-YOUR-REPO>
Finally, set the upstream remote to the official QtPy repo with:
git remote add upstream https://github.com/spyder-ide/qtpy.git
Particularly for development installs, we highly recommend you create and activate a virtual environment to avoid any conflicts with other packages on your system or causing any other issues. Of course, you're free to use any environment management tool of your choice (conda, virtualenvwrapper, pyenv, etc).
To do so with Conda (recommended), simply execute the following:
conda create -c conda-forge -n qtpy-env python=3.9
And activate it with
conda activate qtpy-env
With pip/venv, you can create a virtual environment with
python -m venv qtpy-env
And activate it with the following on Linux and macOS,
source qtpy-env/bin/activate
or on Windows (cmd),
.\qtpy-env\Scripts\activate.bat
Regardless of the tool you use, make sure to remember to always activate your environment before using it.
Before installing QtPy itself, make sure you have the Qt binding(s) you wish to develop against. For example, for PyQt5 on Conda, you'd run:
conda install -c conda-forge pyqt=5
Or for the same using pip, you'd execute:
python -m pip install pyqt5==5.* PyQtWebEngine==5.*
While having separate environments for each binding is recommended, you can install multiple in one environment and select between them using the QT_API
environment variable, as described in the Readme (for example, setting it to pyqt5
to select PyQt5, if it is installed).
Finally, to install the QtPy package itself in editable ("development") mode, where updates to the source files will be reflected in the installed package, and include any additional dependencies used for development, run
python -m pip install -e .[test]
You can then import and use QtPy as normal. When you make changes in your local copy of the git repository, they will be reflected in your installed copy as soon as you re-run Python.
We use pre-commit to run some checks before each commit. To install it in local environment, run:
pip install pre-commit
or globally with pipx:
pipx install pre-commit
or from conda:
conda install -c conda-forge pre-commit
Then, install the pre-commit hooks with:
pre-commit install
If you do not want to run the hooks locally the pre-commit.ci
workflow will run them for you on GitHub.
When you start to work on a new pull request (PR), you need to be sure that your work is done on top of the correct branch, and that you base your PR on Github against it.
To guide you, issues on Github are marked with a milestone that indicates the correct branch to use. If not, follow these guidelines:
- Use the latest release branch (e.g.
1.x
) to fix security issues and critical bugs only (if in any doubt, ask first) - Use
master
branch for anything else, particularly introducing new features or breaking compatibility with previous versions
Of course, if a bug is only present in master
, please base bugfixes on that branch.
To start working on a new PR, you need to execute these commands, filling in the branch names where appropriate (<BASE-BRANCH>
is the branch you're basing your work against, e.g. master
, while <FEATURE-BRANCH>
is the branch you'll be creating to store your changes, e.g. fix-startup-bug
or add-widget-support
:
git checkout <BASE-BRANCH>
git pull upstream <BASE-BRANCH>
git checkout -b <FEATURE-BRANCH>
Once you've made and tested your changes, commit them with a descriptive, unique message of 74 characters or less written in the imperative tense, with a capitalized first letter and no period at the end. Try to make your commit message understandable on its own, giving the reader a high-level idea of what your changes accomplished without having to dig into the diffs. For example:
git commit -am "Fix bug reading env variable when importing package on Windows"
If your changes are complex (more than a few dozen lines) and can be broken into discrete steps/parts, its often a good idea to make multiple commits as you work.
On the other hand, if your changes are fairly small (less than a dozen lines), its usually better to make them as a single commit, and then use the git -a --amend
(followed by git push -f
, if you've already pushed your work) if you spot a bug or a reviewer requests a change.
These aren't hard and fast rules, so just use your best judgment, and if there does happen to be a significant issue we'll be happy to help.
Once you've made your changes (or ideally, before), you'll want to run the full test suite and write new tests of your own, if you haven't already done so.
This package uses the Pytest framework for its unit and integration tests, which are located inside the package alongside the tested code, in the tests/
subdirectory.
We strongly suggest you run the full test suite before every commit (it should only take a few seconds to run on most machines).
In general, any new major functionality should come with tests, and we welcome contributing to expand our coverage, increase reliability, and ensure we don't experience any regressions. If you need help writing tests, please let us know, and we'll be happy to guide you.
To run the tests, install the development dependencies as above, and then simply execute
pytest
The pytest.ini
config file configures a variety of settings and command line options for you, so you shouldn't need to pass any further options to pytest unless you have a specific use case.
For a more rigorous run mirroring what is executed on our CIs, execute the following:
cd qtpy
python -bb -X dev -W error -m pytest
cd ..
Now that your changes are ready to go, you'll need to push them to the appropriate remote. All contributors, including core developers, should push to their personal fork and submit a PR from there, to avoid cluttering the upstream repo with feature branches. To do so, run:
git push -u origin <FEATURE-BRANCH>
Where <FEATURE-BRANCH>
is the name of your feature branch, e.g. fix-startup-bug
.
Finally, create a pull request to the spyder-ide/qtpy repository on Github.
Make sure to set the target branch to the one you based your PR off of (master
or X.x
).
We'll then review your changes, and after they're ready to go, your work will become an official part of QtPy.
Thanks for taking the time to read and follow this guide, and we look forward to your contributions!