How fandujar/choregate works
Project type and purpose
• Choregate is a full-stack web application for managing and automating TektonCD tasks.
• The README says it consists of a React/TypeScript front-end (“choregate-ui”), a Go back-end (in cmd/choregate and pkg/…), a Postgres database, and integrates with TektonCD.
• It targets both local development (via Docker Compose + Kind) and production (via Helm charts on Kubernetes).
File structure
• Top-level “choregate-ui” directory holds all front-end code: Vite config, pages, components, services that call the HTTP API.
• “cmd/choregate” is the Go entrypoint for the API server.
• “pkg” is organized as a layered/hexagonal architecture:
• transport: HTTP handlers (tasks, users, triggers, etc.)
• controller: orchestrates request handling
• services: business logic (tasks.go, triggers.go, …)
• repositories: persistence interfaces and implementations (in-memory & Postgres)
• providers: external integrations (auth, TektonCD)
• entities: domain models and related errors
• rbac: role-based access control logic
• “migrations” holds SQL schema files for the Postgres DB.
• “docker/” and “hack/development/” contain Dockerfiles and Docker Compose / Kind configs for local dev.
• “helm/chart” contains Kubernetes manifests/templates for production deployment.
README for additional insights
• Local setup boots a Kind cluster and runs front-end, back-end, Postgres, and TektonCD integration via Docker Compose.
• Access through HTTP port 8080.
• Emphasizes tasks creation, execution, logs—i.e., UI→API→Tekton pipeline→DB for state and logs.
• Use clear, descriptive labels for each component
• Blue rectangles for application code (UI, API)
• Rectangle with “API” for Go server
• Cloud or hexagon for TektonCD
Detail
• Migrations as a step between API server startup and DB schema.