CI/CD Pipeline

Medipal uses GitHub Actions for continuous integration and deployment across all repositories. Shared workflows are centralised in mp-github-actions.

Pipeline Overview

Every repository has its own workflow configuration. Common stages include linting, testing, building, and deploying.

Schema-Driven Generation

When mp-schema changes, CI pipelines automatically regenerate:

• TypeScript API clients — mp-frontend-api, mp-mobile-app-api, mp-mobile-app-tracker-api (via mp-typescript-api-generator)
• Python Pydantic models — mp-server-pydantic-models (via genma + yamser)
• Python SQLAlchemy ORM — mp-server-sql-alchemy-models
• FastAPI route stubs — mp-server-api
• Configuration schema — mp-server-config-schema

See Code Generation Pipeline for the full breakdown.

Docker Builds

Backend and frontend services are containerised using Docker. Images are pushed to AWS ECR and deployed via Nginx reverse proxy.

• mp-server and mp-tracker produce Docker images
• mp-frontend and mp-mobile-app are built as static assets served behind mp-frontend-nginx-proxy
• Mobile native builds use Fastlane (mp-mobile-app-builder) for iOS TestFlight and Google Play distribution

Deployment Environments

• Development — automatic deployment on push to development
• Staging — automatic deployment on push to main
• Production — manual approval required

Infrastructure is managed by mp-tf-infrastructure (Terraform on AWS).

Centralized Workflows

The mp-github-actions repository provides shared workflows and composite actions used across the platform:

Workflow	Purpose
Code generation	mp-schema push → regenerate TypeScript clients, Pydantic models, SQLAlchemy models
Build & test	ESLint, Prettier, Vitest, Playwright e2e
Publish	npm packages to GitHub Packages, Docker images to AWS ECR
Mobile	Android/iOS native builds via Fastlane, app store deployment
Deploy	EC2 pull-based deployment for frontend services

See Platform Architecture for how CI/CD fits into the overall system.