Demo to prevent software regressions
This project contains a basic API to perform the following tasks:
- Create a new task
- Get all tasks
- Get a specific task by ID
- Delete all tasks
Pipeline automation
This state diagram shows the change process from the view of the developer. The developer starts by opening a branch to implement a small branch to implement the changes. After he pushed all changes and tests to git, he creates a pull request. The pipeline runs automatically. It is not possible to merge the pull request without a successful pipeline. When the tests in the pipeline are good, regressions are prevented. After a successful pipeline run, the developer can merge the small change into the main branch. In the case of a project with more contributors, a pull request approval should be added as a requirement to merge as well.
Available tests/pipeline steps
The automated pipeline runs in every pull request and is configured as a mandatory requirement to merge a pull request. The coverage of the code with unit and API-tests is configured to be >90%.
Unit tests
Unit tests provide fast developer feedback and test single functions or classes. In the case of this API, the unit tests validate that the functionality of the task service is working as expected.
API tests
API tests are testing the end-to-end flow of the functionality of the API. In this case the CRUD functionality of the task api
Static code scans
With static code scan plugins like prospector which is used in this pipeline, common errors can be prevented and the quality of the code increases because of specified rules which are checked automatically. In addition, static security scans like bandit can spot potential security issues in the code and those could be fixed directly.