Thank you for contributing to Concise!
The easiest way to get help or ask a question regarding Concise is in the EOSDIS Slack #podaac channel, or by asking a question on the PO.DAAC Forum.
Bugs can be reported by creating a new Github issue in this repository.
Please use the bug label on any bug issue created. Provide detailed
instructions on how to reproduce the bug if possible. Include the version
of Concise that produced this bug and specific error messages.
Feature/enhancement requests are tracked with Github issues. Please use an appropriate label:
documentation- Docstring or sphinx documentation changes
automation- Anything related to CI/CD or metadata
enhancement- Improvement of an existing feature
new-feature- An entirely new feature
Feature and enhancement requests should contain a detailed description of the desired behavior and rationale.
Please ensure all pull requests follow the PR template. This is to ensure
- A reviewer understands what is being changed
- A reviewer understands why it is being changed
- A reviewer understands how the changes have been verified
- A reviewer is confident the changes work as expected and won't break existing functionality.
The name of the pull request should be meaningful, and if the pull request is addressing a Github issue, should match the name of the Github issue.
If any performance improvements are being made, include graphs and charts.
develop(protected)- Code actively being developed
main(protected)- Matches the version currently in prod
release/#.#.#- Release branch being considered for delivery to prod.
- Matches the version currently in test
feature/issue-#- Work for enhancements and new features should be done in a branch with this naming convention
- The issue number should match the associated Github issue number
bugfix/issue-#- Work for bug fixes should be done in a branch with this naming convention
- The issue number should match the associated Github issue number
hotfix/issue-#orhotfix/short-fix-description- Rare/special case to address a special anomaly.
- The issue number should match the associated Github issue number,
unless no such issue exists. If not, use a short description of the
issue e.g.
hotfix/fix-request-url
Concise maintains a changelog. See Keep a Changelog for more information.
The top of the changelog will contain an [Unreleased] section. Add
any changes made in the current PR to the appropriate section of the
changelog.
Please keep changelog messages relatively short (1-2 sentences).
In almost all cases unit tests should be either created or updated in every PR.
Every Python module contains an associated test file named
test_nameofmodule.py. If resources are needed for unit tests, they
should be placed in tests/data. Test coverage reports are generated
as part of the CI/CD pipeline.
If a change impacts the harmony adapter code, the code should be tested with Harmony. There are instructions here on deploying a local instance of Harmony.
Another valuable way to contribute to the project is to review pull requests. Pull requests reviews are appreciated and valued by anybody at any skill level.
This list is intended for people conducting code reviews of pull requests. When reviewing a pull request the reviewer should use the following checklist to verify that the code will not break clients or cause problems and that the code/tests are correct, complete, follow conventions/guidelines, and can be merged into the master branch without issue. Some of the items on the list are very specific, e.g., no TODOs, while others are more open ended (well structured tests). The list is intended to make the job of reviewing pull requests easier and more repeatable by identifying areas in pull requests that commonly need addressing.
When reviewing a pull request, start at the top of the list and consider the most important things that could lead to problems. These are outlined as a set of questions in the Most Important section. Check off the boxes when you are satisfied that each of the criteria is met. Then proceed to the more specific items below in the General, Testing, and Documentation sections. Check off each box that is satisfied by the pull request and add comments to the pull request for those that are not. If all the boxes are checked then approve the request. Otherwise, mark it as needing work, or, if there are critical errors, decline it.
- Could this change break clients?
- Do the changes handle data that may have been previously saved or indexed by an older version of the code?
- What is the operational impact of this change - are there any potential issues such as special deployment procedures, performance issues, etc.?
- Are there tests for all cases (including edge cases)?
- What could go wrong?
- Does the code do what it's supposed to do?
- Have they implemented all of the acceptance criteria?
- There are no overly-long or complicated functions that should to be broken up for readability
- Are web API parameters validated?
- Are symbols used rather than “magic number” constants or string constants? (OK in tests, particularly for error messages or response codes)
- There is no repeated or copy-and-paste code / tests
- There are no TODOs
- There are no stray comment blocks, commented out code, capture/reveals, proto-saves, printlns, or unnecessary logging
- All namespaces and non-trivial defs/defns/defmacros have docstrings
- There are no dangling requires, i.e., requires that were added and not used or requires that are no longer necessary due to code removal
- Code conventions are followed
- If the issue is a bug fix - a test was added that reproduces the conditions that triggered the bug
- The tests are well structured and follow current practices
- Documentation (api_docs.md, README.md, etc) was added for any new features or old documentation updated for any changed features
- Code/curl examples or sample data have been updated as necessary
- Has the CHANGELOG been updated?
Concise follows PEP8 as much as possible. Reference the pylint and flake8 configuration files for specific expectations. The CI/CD pipeline will fail unless there are 0 flake8 warnings and anything other than a 10/10 pylint score.
Please keep any Harmony or CMR-specific code out of the actual Concise code. The Concise service should be able to run locally without any reliance on outside services.
Concise uses Numpy docstrings. All functions should contain a docstring, though short or trivial function may contain a 1-line docstring.
If adding a new module, ensure it has been added to index.rst for inclusion in auto-generated Sphinx docs.