Qt for Python Getting Started

Important

This page is focused on building Qt for Python from source. If you just want to install PySide6, you need to run: pip install pyside6.

For more details, refer to our Quick Start guide. Additionally, you can check the FAQ related to the project.

General Requirements

The following prerequisites must be installed before you build Qt for Python. On Linux you might get them with your operating system package manager, on macOS you might get them with brew, and on Windows you can download the installer from each website.

Guides per platform

You can refer to the following pages for platform specific instructions:

Windows

macOS

Linux

Important

Qt for Python does not yet support WebAssembly and the mobile operating systems (Android or iOS).

Most Linux-based embedded OS provide PySide with their official package manager (for example, Raspbian and ArchlinuxARM).

A normal building command will look like this:

python setup.py install --qtpaths=/path/to/qtpaths \
                        --ignore-git \
                        --debug \
                        --build-tests \
                        --parallel=8 \
                        --verbose-build \
                        --module-subset=Core,Gui,Widgets

Which will build and install the project with debug symbols, including the tests, using ninja (instead of make), and considering only the module subset of QtCore, QtGui, and QtWidgets <PySide6.QtWidgets.

Other important options to consider are:

  • --cmake, to specify the path to the cmake binary,

  • --reuse-build, to rebuild only the modified files,

  • --openssl=/path/to/openssl/bin, to use a different path for OpenSSL,

  • --standalone, to copy over the Qt libraries into the final package to make it work on other machines,

  • --doc-build-online, to build documentation using the online template.

Testing the installation

Once the installation finishes, you will be able to execute any of our examples:

python examples/widgets/widgets/tetrix.py

Running Tests

Using the --build-tests option will enable us to run all the auto tests inside the project:

python testrunner.py test > testlog.txt

Note

On Windows, don’t forget to have qtpaths in your path (set PATH=E:\Path\to\Qt\6.0.0\msvc2019_64\bin;%PATH%)

You can also run a specific test (for example qpainter_test) by running:

ctest -R qpainter_test --verbose

Building the documentation

Starting from 5.15, there are two options to build the documentation:

1. Building rst-only documentation (no API)

The process of parsing Qt headers to generate the PySide API documentation can take several minutes, this means that modifying a specific section of the rst files we currently have, might become a hard task.

For this, you can install sphinx on a virtual environment, and execute the following command:

python setup.py build_rst_docs

which will generate a html/ directory with the following structure:

html
└── pyside6
    ├── index.html
    ├── ...
    └── shiboken6
        ├── index.html
        └── ...

so you can open the main page html/pyside6/index.html on your browser to check the generated files.

This is useful when updating the general sections of the documentation, adding tutorials, modifying the build instructions, and more.

2. Building the documentation (rst + API)

The documentation is being generated using qdoc to get the API information, and also sphinx for the local Python related notes.

The system required libxml2 and libxslt, also on the Python environment, sphinx and graphviz need to be installed before running the installation process:

pip install graphviz sphinx sphinx_tabs

After installing graphviz, the dot command needs to be in PATH, otherwise, the process will fail. Installing graphviz system-wide is also an option.

Since the process relies on a Qt installation, you need to specify where the qtbase directory is located:

export QT_SRC_DIR=/path/to/qtbase

Once the build process finishes, you can go to the generated *_build/*_release/pyside6 directory, and run:

ninja apidoc

Note

The apidoc make target builds offline documentation in QCH (Qt Creator Help) format by default. You can switch to building for the online use with the --doc-build-online configure option.

The target executes several steps:

  1. The qdoc tool is run over the Qt source code to produce documentation in WebXML format.

  2. shiboken6 is run to extract the functions for which bindings exist from WebXML and convert it into RST.

  3. sphinx is run to produce the documentation in HTML format.

Re-running the command will not execute step 1 unless the file qdoc_output/webxml/qtcore-index.webxml is removed from the build tree. Similarly, step 2 will not be executed unless the file rst/PySide6/QtCore/index.rst is removed.

Finally, you will get a html directory containing all the generated documentation. The offline help files, PySide.qch and Shiboken.qch, can be moved to any directory of your choice. You can find Shiboken.qch in the build directory, *_build\*_release\shiboken6\doc\html.

If you want to temporarily change a .rst file to examine the impact on formatting, you can re-run sphinx in the doc directory:

sphinx-build rst html

Viewing offline documentation

The offline documentation (QCH) can be viewed using the Qt Creator IDE or Qt Assistant, which is a standalone application for viewing QCH files.

To view the QCH using Qt Creator, following the instructions outlined in Using Qt Creator Help Mode. If you chose to use Qt Assistant instead, use the following command to register the QCH file before launching Qt Assistant:

assistant -register PySide.qch

Note

Qt Assistant renders the QCH content using the QTextBrowser backend, which supports a subset of the CSS styles, However, Qt Creator offers an alternative litehtml-based backend, which offers better browsing experience. At the moment, this is not the default backend, so you have to select the litehtml backend explicitly under the General tab in Qt Creator >> Tools >> Options >> Help.

Using the internal tools

A set of tools can be found under the tools/ directory inside the pyside-setup repository.

  • checklibs.py: Script to analyze dynamic library dependencies of Mach-O binaries. To use this utility, just run:

    python checklibs.py /path/to/some.app/Contents/MacOS/Some

    This script was fetched from this repository.

  • create_changelog.py: Script used to create the CHANGELOG that you can find in the dist/ directory. Usage:

    python create_changelog.py -r 6.0.1 -v v6.0.0..6.0 -t bug-fix

  • debug_windows.py: This script can be used to find out why PySide modules fail to load with various DLL errors like Missing DLL or Missing symbol in DLL.

    You can think of it as a Windows version of ldd / LD_DEBUG.

    Underneath, it uses the cdb.exe command line debugger and the gflags.exe tool, which are installed with the latest Windows Kit.

    The aim is to help developers debug issues that they may encounter using the PySide imports on Windows. The user should then provide the generated log file.

    Incidentally it can also be used for any Windows executables, not just Python. To use it just run:

    python debug_windows.py

  • missing_bindings.py: This script is used to compare the state of PySide and PyQt regarding available modules and classses. This content is displayed in our wiki page, and can be used as follows:

    python missing_bindings.py --qt-version 6.0.1 -w all

Note

The script relies on BeautifulSoup to parse the content and generate a list of the missing bindings.