Installing GRGL#
The simplest way to install GRGL is via pip:
pip install pygrgl
The resulting installation has two parts:
The
grgcommand, used for creating or summarizing GRGs. Seegrg --help.The Python API, accessed via
import pygrgl. The API can be used to create or access GRGs - see the Python API reference for more details.
For usage as a C++ library you’ll need to clone (git clone --recursive)
the GRGL repository. We use CMake
as our build system, so the easiest way to incorporate GRGL into your project
is via CMake’s add_subdirectory method.
Advanced: Installing from source#
You will want to install from source if you need support for BGEN file input, but otherwise the simpler pip-based installation is likely sufficient. It is generally best practice to use a virtual environment when installing Python packages.
Installing from source:
python setup.py bdist_wheel # Compiles C++, builds a wheel in the dist/ directory
pip install --force-reinstall dist/*.whl # Install from wheel
Installing from source and enabling BGEN support:
GRGL_BGEN=1 python setup.py bdist_wheel # Compiles C++, builds a wheel in the dist/ directory
pip install --force-reinstall dist/*.whl # Install from wheel
There are other environment variables that control the behavior of the GRGL build as well:
GRGL_GSL=1: Enable GNU scientific library for computing p-values with GWAS.GRGL_DEBUG=1: Build the C++ code in debug mode.