Skip to content

Latest commit

 

History

History
69 lines (50 loc) · 3.84 KB

CONTRIBUTING.md

File metadata and controls

69 lines (50 loc) · 3.84 KB
Raw

Contributing to pyclesperanto_prototype

This page explains how to contribute to the project. In very general, feedback, bug reports, feature requests and pull-requests are welcome.

Scope

pyclesperanto_prototype is a Python image processing library for processing 2D and 3D fluorescence microscopy imaging data using OpenCL-compatible graphics cards. Typical operations include

  • image filtering
  • spatial transforms
  • matrix operations
  • image segmentation
  • cell neighborhood graph extraction and visualization
  • quantitative feature extraction

Feature requests that are out of scope will be dealt with lower priority.

Procedure

Before making a contribution to the project, please open a github issue and let's have a discussion about your ideas. We can then guide you where in the code-base your potential addition would fit. Also check out the API reference to see if there are functions available related to your proposed addition.

How to develop

For developing the project code locally on your computer, it is recommended to fork the project and clone it using github desktop. Alternatively, you can clone your fork from the command line like this:

git clone git@github.com:your_github_account/pyclesperanto_prototype.git
cd pyclesperanto_prototype
git submodule init
git submodule update

When committing/pushing your changes to github, don't forget to create a branch first.

Code structure

The project is divided into tiers. The lower level the tier is, the closer it is to the GPU driver. Higher level tiers implement more complex functionality. For example, in tier1 the Gaussian blur is implemented. Difference-of-Gaussian is implemented in tier2 and uses the Gaussian blur from tier1.

If you program new features, make sure it complies with our coding conventions:

  • Your new function comes in a _new_function.py and is called new_function().
  • New functions contain docstrings in the numpy docstring format.
  • Add new_function to the __init__.py of the tier where it is located.
  • If the functions comes with OpenCL code, the corresponding OpenCL file is called new_function.cl and contains a function new_function. The file is located in the same folder as the corresponding python file.
  • If the OpenCL algorithm implementation differs for 2D and 3D, there should be two OpenCL-files named new_function_2d.cl and new_function_3d.cl. Only one python file should be provided and the differentiation 2D/3D should be done inside new_function().
  • Your OpenCL code is image-type agnostic. Do not call OpenCL functions such as read_imageui() but READ_IMAGE place holders instead. You find a full list of place holders and further explanation here.

How to submit pull-requests

When sending a pull-request, please make sure that you added tests for new features. For example for your _my_function.py add a test_my_function.py to the tests/ folder. Ideally, the code-coverage does not decrease when merging new functions.

Related resources

This project is part of the clesperanto project. Read more about the clesperanto mission. This project is a descendant of CLIJ. Read more development hints in the CLIJ documentation:

Support

Feel free to ask questions on https://image.sc or create a github issue.