Use AWS CodeBuild with Codecov

Codecov is a tool that measures the test coverage of your code. Codecov identifies which methods and statements in your code are not tested. Use the results to determine where to write tests to improve the quality of your code. Codecov is available for three of the source repositories supported by CodeBuild: GitHub, GitHub Enterprise Server, and Bitbucket. If your build project uses GitHub Enterprise Server, you must use Codecov Enterprise.

When you run a build of a CodeBuild project that is integrated with Codecov, Codecov reports that analyzes code in your repository are uploaded to Codecov. The build logs include a link to the reports. This sample shows you how to integrate a Python and a Java build project with Codecov. For a list of languages supported by Codecov, see Codecov supported languages on the Codecov website.

Integrate Codecov into a build project

Use the following procedure to integration Codecov into a build project.

To integrate Codecov with your build project

1. Go to https://codecov.io/signup and sign up for a GitHub or Bitbucket source repository. If you use GitHub Enterprise, see Codecov Enterprise on the Codecov website.

2. In Codecov, add the repository for which you want coverage.

3. When token information is displayed, choose Copy.

   

4. Add the copied token as an environment variable named CODECOV_TOKEN to your build project. For more information, see Change a build project's settings (console).

5. Create a text file named my_script.sh in your repository. Enter the following into the file:

   #/bin/bash
   bash <(curl -s https://codecov.io/bash) -t $CODECOV_TOKEN

6. Choose the Python or Java tab, as appropriate for your build project uses, and follow these steps.

   Java

   1. Add the following JaCoCo plugin to pom.xml in your repository.

      <build>
        <plugins>
          <plugin>
            <groupId>org.jacoco</groupId>
            <artifactId>jacoco-maven-plugin</artifactId>
            <version>0.8.2</version>
            <executions>
                <execution>
                    <goals>
                        <goal>prepare-agent</goal>
                    </goals>
                </execution>
                <execution>
                    <id>report</id>
                    <phase>test</phase>
                    <goals>
                        <goal>report</goal>
                    </goals>
                </execution>
            </executions>
          </plugin>
        </plugins>
      </build>

   2. Enter the following commands in your buildspec file. For more information, see Buildspec syntax.

      build:
        - mvn test -f pom.xml -fn
      postbuild:
        - echo 'Connect to CodeCov'
        - bash my_script.sh

   Python

   Enter the following commands in your buildspec file. For more information, see Buildspec syntax.

   build:
     - pip install coverage
     - coverage run -m unittest discover
   postbuild:
     - echo 'Connect to CodeCov'
     - bash my_script.sh

7. Run a build of your build project. A link to Codecov reports generated for your project appears in your build logs. Use the link to view the Codecov reports. For more information, see Run AWS CodeBuild builds manually and Log AWS CodeBuild API calls with AWS CloudTrail. Codecov information in the build logs looks like the following:

   [Container] 2020/03/09 16:31:04 Running command bash my_script.sh
   
     _____          _
    / ____|        | |
   | |     ___   __| | ___  ___ _____   __
   | |    / _ \ / _` |/ _ \/ __/ _ \ \ / /
   | |___| (_) | (_| |  __/ (_| (_) \ V /
    \_____\___/ \__,_|\___|\___\___/ \_/
                                 Bash-20200303-bc4d7e6
   
   ·[0;90m==>·[0m AWS CodeBuild detected.
   ... The full list of Codecov log entries has been omitted for brevity ...
       ·
       ·[0;32m->·[0m View reports at ·[0;36mhttps://codecov.io/github/user/test_py/commit/commit-id·[0m
   
   [Container] 2020/03/09 16:31:07 Phase complete: POST_BUILD State: SUCCEEDED

   The reports look like the following: