Metadata-Version: 2.4
Name: clij2-fft
Version: 0.29
Summary: A python wrapper around clij2 opencl FFT algorithms
Home-page: https://github.com/clij/clij2-fft
Author: Robert Haase, Brian Northan
Author-email: bnorthan@gmail.com
License: BSD
Description-Content-Type: text/markdown
Requires-Dist: numpy
Requires-Dist: dask
Requires-Dist: dask-image
Requires-Dist: pyopencl
Dynamic: author
Dynamic: author-email
Dynamic: description
Dynamic: description-content-type
Dynamic: home-page
Dynamic: license
Dynamic: requires-dist
Dynamic: summary

# clij2-fft (featuring non-circulant rltv deconvolution)

clij2-fft is a prototype implementation of a framework for OpenCl based FFT algorithms.  The most used algorithm from this project is the OpenCL implementation of the non-circular Richardson Lucy deconvolution algorithm with total variation regularization, which can be called as follows

```
from clij2fft.richardson_lucy import richardson_lucy_nc
decon_clij2=richardson_lucy_nc(im,psf,100,0.0002)
```

or for large image that need to be split up into chunks with dask

```
from clij2fft.richardson_lucy_dask import richardson_lucy_dask
decon=richardson_lucy_dask(img, psf, 100, 0.0001)
```

If you need support for the library please post a question on the [Image.sc Forum](https://forum.image.sc/).

Long term we hope to integrate FFT based math more closely with the [Clic project](https://github.com/clEsperanto/CLIc_prototype).  The goal is to make it easy to write algorithms such as convolution, correlation, registration and deconvolution that consist of a series of FFTs combined with other math operations. 
