This repo contains a small module and 3 CLI scripts.
To run update-reviewers
you will need to run both update-reviews
and update-contributors
first to create the required input .pickle
files.
The scripts run as follows:
This script parses data from all-contributors bot .json
files in the following repos:
- software-submission repo where peer review happens
- python-package-guide repo
- peer review guide repo
- pyopensci.github.io (website) repo
- update-package-meta repo (this repo)
Running this this script:
- Parses through all of the all-contributors bot
.json
files across our pyOpenSci repos to gather contributors to our organization. This allows us to acknowledge contributors on our website who are not always making explicit code contributions (thus might not have commits). These contributors are reviewing guidebooks, participating in peer review, and performing other important tasks that are critical to our pyOpenSci mission. We acknowledge all contributions at pyOpenSci regardless of volume or size. - Updates the existing contributors.yml file found in our website repo with new contributors and the contributor role (package guides, code workflows, peer review, etc).
a. If you run the script using
--update update_all
, this script will also use the GitHub api to update the users metadata from their GitHub profile.
This script parses through our (accepted) review issues to find packages that have been accepted. It then grabs each reviews editor, reviewers and package authors / maintainers. This information allows us to
- Update a contributor's peer review metadata in the contributors.yml file in the third script.
- Update the pyOpenSci website's package listing with the package's DOI, documentation URL.
- Update the package's stats including stars, contributors, etc. using the GitHub API
This final script is a bridge between the first it uses pickle files outputted from the first two scripts to update each contributors peer review contributions including
- packages submitted or reviewed
- packages in which the contributor served as editors
- contributor types associated with peer review including:
- peer-review
- package-maintainer
- package-reviewer
- package-editor
These general contributor types are used to drive our website's contributor search and filter functionality that you can see here.
To begin:
- Create a local environment and activate it.
- Install the required dependencies via the
requirement.txt
file by running the following command;pip install -r requirements.txt
- Install the pyosmeta package in development/editable mode:
pip install -e .
To run this you need to create a TOKEN that can be used to access the GitHub API.
After obtaining a token;
- Duplicate the
.env-default
file and rename the copy to.env
- Assign your token to the
GITHUB_TOKEN
variable in the.env
file.
Each script is available through the command line through entry points specified in our pyproject.toml
file.
To run:
update-contributors
update-contributors --update update_all
The update-contributors
script does the following:
- It grabs the
all-contribs.json
files from each repository and turns that json data into a dictionary of all unique contributors across repos. Repos include:- peer review guide
- packaging guide
- website
- software-review
- update-package-meta (this repo)
- It then:
If you use the --update update_all
flag, it will:
- Update contrib profile information including name using whatever information is available their public github account for website, location, organization, twitter, etc).
- Check to see that the website in their profile works, if not removes it so it doesn't begin to fail our website ci tests.
Without the update
flag, running update-contributors
will only add any new users that
are not already in the website contributors.yml
file to a output .pickle
file.
all-contributors.pickle
file that will be used in the final update-reviewers script to update all reviewer contribution data.
To run:
update-reviews
or
update-reviews --update update_all
- This script parses through all pyOpenSci issues in the software-submissions repo issues where the issue has a label of 6/
pyOS-approved πππ
. - Grabs crucial metadata including the reviewers and editors for each.
- Adds people who have participated in peer review who are NOT listed currently in the website
contributors.yml
file
It then collects the GitHub id and user information for:
- reviewers,
- submitting authors,
- editors and
- maintainers.
Finally, it updates GitHub statistics for each packages including stars, last commit date and more repo metadata.
This returns a packages.pickle
file that will be used in the final script which bridges data between the first two scripts.
This script is a bridge between update-contributors
and update-reviews
. It parses each review in the output
update-reviews.pickle
file and
- updates contributor name in the review data (often the github username is there but the first and last name is missing). This allows us to publish the maintainer names (rather than github usernames) on our website package listing.
- Updates each review issue contributor's name in the contributor metadata. This allows us to ensure we have updated contributor types, package submission information etc, in the contributor.yml file on our website.
To run:
update_reviewers
This final script uses the two pickle files to update information. It then returns two output files:
_data/contributors.yml
_data/packages.yml
Each are stored in the /_data/file.yml
directory to mimic the directory structure of our website.
The scripts above are called in the GitHub actions located here. These actions can be run manually via workflow dispatch but also have a cron job to update our metadata periodically.
TODO: right now this isn't an issue but it will be in the future I suspect.... Rate limiting - how to handle this...
Create environment:
mamba env create -f environment.yml
Thanks goes to these wonderful people (emoji key):
This project follows the all-contributors specification. Contributions of any kind welcome!