Installation¶

SFFT Installation¶

To install the latest stable version of sfft from PyPI: [recommended]

Latest

pip install sfft

One can also install sfft from GitHub:

git clone https://github.com/thomasvrussell/sfft.git

pip install .

Both will automatically install compatible versions of python dependencies.

Tip

PyPI/GitHub installation does not include the setup of GPU backend.

Enable GPU backend¶

SFFT supports both CPU (Numpy) backend and GPU (Cupy) backend. To enable the GPU backend, users need to further install CuPy according to their CUDA version. Note this backend requires GPU device(s) with double-precision support.

CUDA 12: e.g., enable the CuPy backend for CUDA 12.0 via
```
pip install cupy-cuda12x
```
CUDA 11: e.g., enable the CuPy backend for CUDA 11.5 via
```
pip install cupy-cuda115
```
CUDA 10: e.g., enable the CuPy backend for CUDA 10.1 via
```
pip install cupy-cuda101
```

Remarks

There was a PyCUDA backend in sfft but now deprecated since v1.4.0. For sfft < v1.4.0, PyCUDA backend was preserved as it consumes less GPU memory. However, the CuPy backend is now better implemented for GPU memory allocation, making the PyCUDA backend no longer useful.

AstrOmatic Dependencies¶

You need further to install additional AstrOmatic software for sfft.

SExtractor: SExtractor from AstrOmatic is required for SFFT preprocessing, as it allows SFFT to determine an appropriate pixel mask for the input image pair before performing image subtraction. This step is critical for achieving more accurate parameter solutions.
```
conda install -c conda-forge astromatic-source-extractor
```
Tip

We have wrapped SExtractor into a Python module, sfft.utils.pyAstroMatic.PYSEx, enabling users to trigger SExtractor directly within Python. Use help(sfft.utils.pyAstroMatic.PYSEx) for more details.
SWarp (optional): SWarp from AstrOmatic is not required for sfft subtraction itself. However, it is normally useful to align the input image-pair before image subtraction by SWarp.
```
conda install -c conda-forge astromatic-swarp
```
Tip

We have additionally wrapped SWarp into a Python module sfft.utils.pyAstroMatic.PYSWarp so that you can align images in a more Pythonic way. Use help(sfft.utils.pyAstroMatic.PYSWarp) for more details.