GrimoireLab as Code for GitHub Community Analytics

A portable GitHub community analytics setup, based on GrimoireLab by CHAOSS. The project is distributed as Docker Compose, and can be reused as needed.

NOTE: The original project was developed for the WireMock Community Analytics, and now it is being reworked to a portable project template that can be reused for your custom projects. See wiremock/wiremock-community-analytics for the original project.

Configured analytics

• git
• github, github:repo, github:issues, github:pulls
• githubql
• github2:issues, github2:pulls

Running locally

1. Check out this repository
2. Create the file settings/github-api-token.txt containing only your GitHub token. It will be propagated as a Docker secret to the mordred instance
3. OPTIONAL: Run ./scripts/generate-projects-json to generate or refresh the projects configuration file. To generate test data to speedup development, use the USE_TEST_DATA variable. E.g. USE_TEST_DATA=true ./scripts/generate-projects-json
4. Run docker compose up
5. Wait... The production configuration includes many repositories, and it may take a while until GrimoireLab loads all the data. Normally it takes about 10 minutes

The service will be exposed on the port 8000 by default.

Editing organization mappings

The setup pre-configures SortingHat by the list of known organizations, originally sourced from here. This list is dated and for sure not complete.

You can add organizations from your community to settings/orgs_file.json.

Wishlist

• Add Helm charts
• Support for the GitHub application tokens. Now the menu is hardcoded.
• Non-GitHub analytics OOTB
• Actually consume the menu.yaml file
• Include the NLP sentiment server to the setup to do comments/impressions

License

GNU GPL 3 is used as the repository adopts some code and configurations from GrimoireLab and its repositories

Learn More

• Oleg Nenashev, GrimoireLab as Code. Ready-to-fly GitHub analytics, CHAOSScon 2024 (Slides)