How to contribute¶
This document aims to give an overview of how to contribute to PlantCV. We encourage third-party contributions in a variety of forms. There are a few guidelines that we need contributors to follow so that we can keep things working for all users.
There are many ways to contribute:
- Report bugs
- Request features
- Join the discussion of open issues, features, and bugs
- Contribute algorithms to the library
- Add unit tests
- Revise existing code
- Add or revise documentation
If you need any help, please contact us.
- Make sure you have a GitHub account.
- Search GitHub and Google to see if your issue has already been reported
- Create an issue in GitHub, assuming one does not already exist.
- Clearly describe the issue including steps to reproduce when it is a bug.
- Make sure you fill in the earliest version that you know has the issue.
- Where applicable, provide the full workflow prior to the error and the image getting analyzed.
Contributing Text or Code¶
When you add a significant new feature, please create an issue first, to allow others to comment and give feedback.
When you have created a new feature or non-trivial change to existing code, create a 'pull request'.
Branching and Pull Requests: Although core developers have write permissions to the PlantCV repository, please use the feature branch workflow (below) in order to allow pull requests, automated testing, and code review.
Using Git at the Command Line¶
Introduce your self to git, make sure you use an email associated with your GitHub account. This can also be done with the GitHub desktop application.
git config --global user.name "John Doe" git config --global user.email [email protected]
Fork the PlantCV repository to add it to your GitHub account.
Clone your fork of this repository
git clone https://github.com/<your username>/plantcv.git
Setup repository to be able to fetch from the master
git remote add upstream https://github.com/danforthcenter/plantcv.git
Adding Features and Submitting Changes¶
Always work in a branch rather than directly on the master branch. Branches should focus on fixing or adding a single feature or set of closely related features because this will make it easier to review and merge your contributions. Since multiple people work on the same repository, make sure to keep your master branch in sync with the master of the danforthcenter/plantcv repository. The master branch is the central repository for stable releases and the latest code.
Here is a simplified workflow on how add a new feature:
Get the latest version¶
Update your master branch with the main PlantCV repository (both locally and on GitHub)
git fetch upstream git checkout master git merge upstream/master git push
Create a branch to do your work¶
A good practice is to call the branch in the form of
followed by the title of the issue. This makes it easier to find out the
issue you are trying to solve and helps us to understand what is done in
the branch. Calling a branch my-work is confusing. Names of branch can
not have a space, and should be replaced with a hyphen.
git checkout -b plantcv-issuenumber-title-of-issue
Work and commit¶
Do you work, and commit as you see fit. Make your commit messages helpful.
# Stage files to commit (. for all, or specfic list of files) git add . # Commit staged files with descriptive message git commit -m "Helpful message about updates."
Push your changes up to GitHub¶
Pushing will update your fork of the PlantCV repository. If this is the
first time pushing to GitHub you will need to use the extended command
below, other wise you can simply do a
git push -u origin plantcv-issuenumber-title-of-issue
Once your new feature is ready, create a pull request from your fork/branch to the main PlantCV repository. Generating a pull request will notify the maintainers of PlantCV. GitHub will report whether the merge can be done automatically without conflict, or whether manual fixes are required. Travis-CI will also generate an automatic build report on whether or not the updates break the automated unit tests.
Guidelines for adding new features¶
In general, new contributions to PlantCV should benefit multiple users and extend the image processing or trait analysis power of PlantCV.
What should/should not be added to PlantCV:
- New validated image processing functions are highly encouraged for contribution.
- New validated trait extraction algorithms are highly encouraged for contribution.
- Image processing workflow scripts that are specific for your project should not be added to PlantCV since these are (usually) not generalized or updated to maintain compatibility with new versions.
If you are adding new image processing functions or trait extraction algorithms, make sure you read the section on testing and documenting your code below. We are very appreciative of all community contributions to PlantCV! We do hope that if you are contributing new methods to PlantCV, you are also contributing documentation and unit tests for your functions (described below), because you understand your code/functionality best. If you have questions or need help don't hesitate to ask here.
Testing and documenting your code¶
In addition to adding a new feature, test your code thoroughly:
PlantCV utilizes Travis CI, Coveralls, and Read the Docs to provide continuous integration of unit testing and documentation. Integration with Travis CI allows us to test whether new pull requests break or change the output of existing functions. Coveralls tells us the percentage of PlantCV repository code that is covered by the unit tests (better the coverage the more likely we are to catch problematic pull requests). To include new functions in Travis CI tests (to make sure they aren't broken by future pull requests) we need a 'unit test' or set of 'unit tests' (if you need more than one test to cover function options).
Existing unit tests can be found in
tests/tests.py as examples.
The data to support unit tests can be found in
If you are updating existing code, make sure that the
script passes on the function you modified. Testing locally can be done with pytest:
cd plantcv py.test -v tests/tests.py
Add documentation for your new feature (see Adding/editing Documentation for more details). A new
Markdown file should be added to the docs folder, and a reference to your new doc file should
be added to the
mkdocs.yml file. You can test that your new documentation
can be built correctly locally using mkdocs from the root of your local
mkdocs build --theme readthedocs --site-dir _site
New function style guide¶
Include commenting whenever possible.
Start code with import of modules, for example:
# Import external packages import numpy as np import cv2 import os # Import functions from within plantcv from plantcv.plantcv import params from plantcv.plantcv import plot_image from plantcv.plantcv import print_image
Here is a sample of a new function. Arguments should be defined.
Generally, functions should utilize the global
debug parameter from the params class.
Users can set
to None (default), "print" (to file), or "plot" (to screen if using Jupyter notebooks or X11).
Functions should also increment the
device number, which is a counter for image processing steps that is autoincremented by functions that use
Note that the inputs and outputs are documented with Python docstrings.
def new_function(img): """New function. Inputs: img = description of img Returns: returnval1 = return value 1 :param img: type :return returnval1: type """ # Increment the device number by 1 params.device += 1 returnval1 = some_code_on_img(img) if params.debug == "print": print_image(img, os.path.join(params.debug_outdir, str(params.device) + '_new_function.jpg')) elif params.debug == 'plot': plot_image(img) return returnval1
If you need to call other PlantCV functions from within your contributed function, be sure to disable debugging until you are ready to show your own debug images.
To do this, follow these four steps: 1. Store the value of params.debug 2. Temporarily set params.debug to None 3. Call other PlantCV functions, and take any other steps 4. Reset the debug mode to the stored value before printing/plotting your own debug image(s).
Here is a sample of a PlantCV function that calls on other PlantCV functions:
import plantcv.plantcv.params import plantcv.plantcv.example_plantcv_function import plantcv.plantcv.another_plantcv_function import plantcv.plantcv.print_image import plantcv.plantcv.plot_image def new_function_calling_plantcv(img): """New function calling another plantcv function. Inputs: img = description of img Returns: final_img = description of final img :param img: type :return final_img: type """ # Increment the device number by 1 params.device += 1 # Store debug mode debug = params.debug # Temporarily disable debugging params.debug = None # No debug images will be shown for these functions modified_img = example_plantcv_function(img) final_img = another_plantcv_function(modified_img) # Reset debug mode params.debug = debug if params.debug == 'print': print_image(final_img, os.path.join(params.debug_outdir, str(params.device) + '_new_function.jpg')) elif params.debug == 'plot': plot_image(final_img) return final_img
Parts of this contribution guide are based on the TERRA-REF Contribution Guide.