Release: Re-thinking test coverage

Test coverage tracking support is now available for Python and Go.

Header image
By Jai on 
Share on Twitter Share on LinkedIn

Test coverage is a great metric to track which parts of the codebase is covered by tests. Measuring test coverage helps developers get clarity on how to improve the tests to make sure all functionalities of the codebase is well-tested. If executing the test suite causes 80% of the package's source statements to be run, we say that the test coverage is 80%.

Before implementing test coverage, we were on the look out for similar products solving this. We evaluated the products and looked at them from end user’s point of view, to understand how exactly is it being used by developers. We observed very similar workflows and patterns across the tools we examined. Let’s categorize them into four parts.

1. Development: How do these tools integrate with existing development workflows?

On every code change, tests run as part of CI build pipelines. A coverage file is generated and reported to tools like Coveralls, Codecov, etc.

2. Presentation: How are test coverage results shown?

All the tools provide similar types of information. First, overall test coverage of the codebase. Second, a file browser with all the files and the coverage percentile against it.

3. Interaction: How do developers interact with the results?

Developers visit the project dashboard, click on a file that has coverage below threshold, browse through the file and take a look at the lines that are marked red (lines missed).

4. Action: What are the actions post looking at the result?

Depending on the criticality of the specific part of the codebase, the developer creates an issue in their bug tracking system to work on improving the test coverage.

Introducing, the Test Coverage analyzer

At DeepSource, we strive for seamlessly integrating with native software development workflows and optimize for helping you fix issues rather than just letting you know. The all-new Test Coverage analyzer allows you to do two important things — nothing more, nothing less:

  • Keep track of the overall coverage of your project
  • Take action on lines that are not covered by tests

This is how it looks.

Test coverage analysis run

You can use our open-source CLI to report test coverage from your CI system to DeepSource — you’ll love how easy it is to set up. We’ve written guides for Travis CI and Circle CI.

Let us know what you think. Here’s what Joe, creator of Masonite framework, thinks:

“DeepSource is great! Saves a whole lot of time when it comes to code issues. It pops up alongside Masonite‚Äôs GitHub status checks so I know exactly what the issues are. I really love DeepSource!”

Ready to keep track of test coverage?
DeepSource shows you exactly what you need to know about your coverage.