Installation Guide¶
Basic installation¶
This documentation describes PADO 1.1.0. Use Python 3.10 or later and install a PyTorch 2.10 or later build matching the intended CPU/CUDA system. Install the 1.1.0 wheel from the GitHub release:
python -m pip install "https://github.com/shwbaek/pado/releases/download/1.1.0/pado_optics-1.1.0-py3-none-any.whl"
The distribution name is pado-optics and the Python import is pado.
PyPI publication is separate; an unpinned PyPI install may still select an
earlier release. See Migrating to PADO 1.1 before comparing numerical results with PADO 1.0.1.
Optional dependencies¶
PyTorch is the only core runtime dependency. Install extras when using their corresponding APIs:
Extra |
Purpose |
Minimum |
|---|---|---|
|
NumPy arrays and NPY files |
NumPy 1.24 |
|
Plotting |
Matplotlib 3.7 |
|
MAT files |
SciPy 1.10 |
Install all optional APIs together:
python -m pip install "pado-optics[array,viz,mat] @ https://github.com/shwbaek/pado/releases/download/1.1.0/pado_optics-1.1.0-py3-none-any.whl"
Source installation¶
For development against the public repository:
git clone https://github.com/shwbaek/pado.git
cd pado
python -m pip install -e ".[array,viz,mat]"
The installed source version follows the selected public commit; it is not necessarily the latest published package. No Conda package availability is claimed here.
Next steps¶
Migrating to PADO 1.1 explains changed numerical conventions, corrections, gradients, precision costs and sampling limits.
API Reference documents the Python API.
Examples links existing examples.
License describes the license.