- You have a Sigrid user account.
- You have created an authentication token for using Sigrid CI.
- Python 3.7 or higher needs to be available in the CI environment if you do not use the Docker image published by SIG. The client scripts for Sigrid CI are based on Python.
On-boarding a system is done automatically when you first run Sigrid CI for that system. As long as you have a valid token, and that token is authorized to on-board systems, you will receive the message system has been on-boarded to Sigrid. Subsequent runs will then be visible in both your CI environment and sigrid-says.com.
Sigrid CI reads your Sigrid account credentials from an environment variable called SIGRID_CI_TOKEN
. To add it to your GitLab CI pipeline, follow these steps:
- Select "Settings" in your GitLab project's menu
- Select "CI/CD" in the settings menu
- Locate the section named "Variables"
- Click the "Add variable" button
- Add an environment variable
SIGRID_CI_TOKEN
and use your Sigrid authentication token as the value.
These instructions describe how to configure a single GitLab project, but you can follow the same steps to configure the entire GitLab group, which will make the environment variables available to all projects within that group.
Next, you need to edit your project's CI configuration, in order to add Sigrid CI as an extra step. There are two alternatives: use SIG's Docker image (alternative 2.a below), or download and run the Sigrid CI Python script directly on your Gitlab runners (alternative 2.b below).
Both alternatives configure two additional build steps:
- The
sigridci
job provides Sigrid feedback for pull request. This step is configured to run for every branch except the main/master branch. - The
sigridpublish
job publishes project snapshots to sigrid-says.com. This step only runs for the main/master branch.
The relevant command that starts Sigrid CI is the call to the sigridci.py
script, which starts the Sigrid CI analysis. The scripts supports a number of arguments that you can use to configure your Sigrid CI run. The scripts and its command line interface are explained in using the Sigrid CI client script.
Note that you need to perform this step for every project where you wish to use Sigrid CI. Be aware that you can set a project-specific target quality, you don't necessarily have to use the same target for every project.
The recommended approach is to run Sigrid CI using the Docker image published by SIG. This requires using either the Docker or Kubernetes executor to be available on your GitLab runners.
Open .gitlab-ci.yml
in your project's root directory and add the following:
stages:
- report
sigridci:
stage: report
image:
name: softwareimprovementgroup/sigridci
script:
- sigridci.py --customer examplecustomername --system examplesystemname --source .
allow_failure: true
artifacts:
paths:
- "sigrid-ci-output/*"
reports:
junit: "sigrid-ci-output/sigridci-junit-format-report.xml"
expire_in: 1 week
when: always
rules:
- if: $CI_MERGE_REQUEST_ID
sigridpublish:
stage: report
image:
name: softwareimprovementgroup/sigridci
script:
- sigridci.py --customer examplecustomername --system examplesystemname --source . --publish
allow_failure: true
artifacts:
paths:
- "sigrid-ci-output/*"
expire_in: 1 week
when: always
rules:
- if: $CI_COMMIT_BRANCH == $CI_DEFAULT_BRANCH
Security note: This example pulls the latest version of SIG's Sigrid CI Docker image directly from Docker Hub. That might be acceptable for many projects. However, some projects might not allow this as part of their security policy. In those cases, we recommend to either use alternative 2.b below, or clone SIG's Docker image to your own image repository.
Alternative 2b does not use a Docker image. Instead, the Sigrid CI GitLab jobs download SIG's Sigrid CI Python script directly from GitHub and runs it on your GitLab runner. This requires Github.com to be accessible from your CI environment and Python 3.7 to be available on your GitLab runners.
Sigrid CI consists of a number of Python-based client scripts, that interact with Sigrid in order to analyze your project's source code and provide feedback based on the results. These client scripts need to be available to your GitLab runners, in order to call the scripts from the CI pipeline.
Open .gitlab-ci.yml
in your project's root directory and add the following:
stages:
- report
sigridci:
stage: report
script:
- git clone https://github.com/Software-Improvement-Group/sigridci.git sigridci
- ./sigridci/sigridci/sigridci.py --customer examplecustomername --system examplesystemname --source .
allow_failure: true
artifacts:
paths:
- "sigrid-ci-output/*"
reports:
junit: "sigrid-ci-output/sigridci-junit-format-report.xml"
expire_in: 1 week
when: always
except:
- master
sigridpublish:
stage: report
script:
- git clone https://github.com/Software-Improvement-Group/sigridci.git sigridci
- ./sigridci/sigridci/sigridci.py --customer examplecustomername --system examplesystemname --source . --publish
allow_failure: true
artifacts:
paths:
- "sigrid-ci-output/*"
expire_in: 1 week
when: always
only:
- master
Security note: This example downloads the Sigrid CI client scripts directly from GitHub. That might be acceptable for some projects, and is in fact increasingly common. However, some projects might not allow this as part of their security policy. In those cases, you can simply download the sigridci
directory in this repository, and make it available to your runners (either by placing the scripts in a known location, or packaging them into a Docker container).
Sigrid will try to automatically detect the technologies you use, the component structure, and files/directories that should be excluded from the analysis. You can override the default configuration by creating a file called sigrid.yaml
and adding it to the root of your repository. You can read more about the various options for custom configuration in the configuration file documentation.
Once you have configured the integration, Sigrid CI will show up as a new step in your GitLab CI pipeline. The step will succeed if the code quality meets the specified target, and will fail otherwise.
Sigrid CI provides multiple levels of feedback. The first and fastest type of feedback is directly produced in the CI output, as shown in the following screenshot:
The output consists of the following:
- A list of refactoring candidates that were introduced in your merge request. This allows you to understand what quality issues you caused, which in turn allows you to fix them quickly. Note that quality is obviously important, but you are not expected to always fix every single issue. As long as you meet the target, it's fine.
- An overview of all ratings, compared against the system as a whole. This allows you to check if your changes improved the system, or accidentally made things worse.
- The final conclusion on whether your changes and merge request meet the quality target.
The end of the textual output provides a link to the Sigrid landing page. You can open this URL in order to use Sigrid for interpreting your analysis results.
Whether you should use the text output or the Sigrid page is largely down to personal preference: the text output is faster to acces and more concise, while Sigrid allows you to view results in a more visual and interactive way.
Finally, Sigrid CI output is also included in GitLab's CI/CD pipeline page, which contains a Tests tab. The View details button will show the same list of refactoring candidates that is shown in the aforementioned textual output and HTML report.
Feel free to contact SIG's support department for any questions or issues you may have after reading this document, or when using Sigrid or Sigrid CI. Users in Europe can also contact us by phone at +31 20 314 0953.