Contributions to pyqtgraph are welcome! Be kind and respectful! See our Code of Conduct for details.
Please use the following guidelines when preparing changes:
First thing to do is fork the repository, and clone your own fork to your local computer.
git clone https://github.com/<username>/pyqtgraph.git
cd pyqtgraph
While there is nothing preventing users from using conda
environments, as a general
principle, we recommend using the venv
module for creating an otherwise empty virtual
environment.
python3.12 -m venv .venv
source .venv/bin/activate
# on windows this would be .venv/Scripts/activate
python -m pip install numpy scipy pyqt6 -e .
PyQtGraph supports PySide2, PyQt5, PySide6 and PyQt6 bindings, but if a contributor is
to test against only one set of bindings, we suggest that it be PyQt6. PyQt6 is the
only bindings that enforces the use of fully scoped enums. For example, apart from PyQt6, QGraphicsItem.ItemIgnoresTransformations
would be recognized, but that would error in PyQt6. The cross-binding compatible reference to that enum would be
QGraphicsItem.GraphicsItemFlags.ItemIgnoresTransformations
.
Before making changes to the code-base, create a different branch with a name that should be unique (this makes it easier for maintainers to examine the proposed changes locally).
git switch -c my-new-feature
The target of the pull request should be the master
branch in the pyqtgraph repo.
Pull requests should include only a focused and related set of changes. Mixed features and unrelated changes may be rejected.
For major changes, it is recommended to discuss your plans on the mailing list or in a github issue/discussion before putting in too much effort.
PyQtGraph has adopted NEP-29 which governs the timeline for phasing out support for numpy and python versions.
- Writing proper documentation is highly encouraged.
- Documentation is generated with sphinx, and usage of numpy-docstyle is encouraged (many places in the library do not use this docstring style at present, it's a gradual process to migrate).
- The docs built for this PR can be previewed by clicking on the "Details" link for the read-the-docs entry in the checks section of the PR conversation page.
To build the documentation locally, users will need to install
graphviz in such a way that the dot
executable is in the
users's PATH
.
pip install -r doc/requirements.txt
cd doc
make html
To view the result, from the ./doc
directory, open ./build/html/index.html
in a web
browser.
To validate the documentation against numpydoc, a fork of numpydoc needs to be installed
pip install git+https://github.com/j9ac9k/numpydoc.git@see-more-accepts-sphinx-crosslinks
To verify your documentation changes conform to numpydoc, run:
numpydoc lint path/to/file.py
Every reasonable effort should be made to address the issues the linter outputs.
- PyQtGraph prefers PEP8 for most style issues, but this is not enforced rigorously as long as the code is clean and readable.
- Variable and Function/Methods that are intended to be part of the public API should be camelCase.
- "Private" methods/variables should have a leading underscore (
_
) before the name.
PyQtGraph developers are highly encouraged to (but not required) to use
pre-commit
. pre-commit
does a number of checks when
attempting to commit the code to being committed, such as ensuring no large files are
accidentally added, address mixed-line-endings types and so on. Check the
pre-commit documentation on how to setup.
- pytest
- Optional: pytest-xdist (used to run tests in parallel)
- Optional: pytest-xvfb (used on linux with headless displays)
To run the test suite, after installing the above dependencies run
pytest tests
In addition, the examples can be tested as well.
pytest pyqtgraph/examples
The examples will each run for 1 second, and automatically close. If no unhandled exception is created, it's considered a success.
As PyQtGraph supports a wide array of Qt-bindings, and python versions, we make use of
tox
to test against as many supported configurations as feasible. With tox installed,
simply run tox
and it will run through all the configurations. This should be done if
there is uncertainty regarding changes working on specific combinations of PyQt bindings
and/or python versions.
For our Continuous Integration, pyqtgraph tests across all OSs for a variety of combinations of supported Qt bindings, python and numpy versions. For submitted changes to be merged, it is expected that the CI passes.
( Still under development ) To ensure this library is performant, we use Air Speed Velocity (asv) to run benchmarks. For developing on core functions and classes, be aware of any impact your changes have on their speed. To configure and run asv:
pip install asv
python setup.py asv_config
asv run
( TODO publish results )