Skip to content

README.md

Latest commit

 

History

History
150 lines (111 loc) · 3.55 KB

README.md

File metadata and controls

150 lines (111 loc) · 3.55 KB

Building and Installing pyamtrack on Linux

⚠️ This guide is intended for developers building and testing pyamtrack, not for end-users. For the main user-oriented documentation, see the main README.

Prerequisites

Ensure you have the following installed:

  • Python 3 (with development headers, at least 3.9)
  • CMake (3.16 or later)
  • Ninja (recommended for faster builds)
  • GSL (GNU Scientific Library)
  • Git (for fetching dependencies)

Installing GSL

On Debian/Ubuntu:

sudo apt-get install libgsl-dev

On Fedora/RHEL:

sudo dnf install gsl-devel

Installation Instructions

Follow these steps to build and install the pyamtrack Python module as a wheel package.

Building and Installing the Project Locally

To build and install the project locally, follow these steps:

  1. Create a Virtual Environment: Create and activate a Python virtual environment:

    python3 -m venv .venv
    source .venv/bin/activate

  2. Install Development Dependencies: Install the required dependencies for building the project:

    pip install -r requirements-dev.txt

  3. Build the Project: Run the following command to build the project:

    python -m build --wheel --no-isolation --config-setting=build-dir=./build

    This command speeds up the build process for local development by avoiding recompilation of the libamtrack source code in every iteration.

  4. Install the Built Wheel: After building, install the wheel:

    pip install dist/*.whl

  5. Test the Installation: Verify the installation by running:

    python -c "import pyamtrack; print(pyamtrack.converters.beta_from_energy(150))"

Expected output:

Python bindings for libamtrack

Using the scripts/build_linux.sh Script

Alternatively, you can use the scripts/build_linux.sh script to perform the same steps as above. This script automates the process of creating a virtual environment, installing dependencies, building the project, and testing the installation:

./scripts/build_linux.sh

Code Formatting and Pre-commit Hooks

This project uses pre-commit. More information about that can be found here.

Uninstallation

Remove the Installed Package

pip uninstall pyamtrack

Rebuilding the Package

Clean Previous Builds

rm -rf build dist

Rebuild the Wheel

python -m build --wheel

Now you can use pyamtrack in your Python scripts:

import pyamtrack
print(pyamtrack.converters.beta_from_energy(150))

Building with cibuildwheel

For a clean and isolated build environment, you can use cibuildwheel. This ensures that the compilation happens in a very clean environment, which is ideal for producing distributable wheels:

pip install cibuildwheel

Then, run the following command to build the wheel for a specific Python version and platform:

cibuildwheel --only "cp312-manylinux_x86_64"

To test the installed package, type in the virtual environment:

pip install wheelhouse/*whl

Then:

>>> import pyamtrack
>>> pyamtrack.stopping.electron_range(120)
0.0006700734300203331
>>> pyamtrack.__version__
'0.14.1'
>>>

CI of this project consists of quick, automatic jobs. There are also long manual jobs, which are described here.