In BE Dockerfile is set up in a way that it has to be built form to monorepo root. Why? In the monorepo there is a types package that introduces shared types between BE and FE. To include this package in the built Docker image, we have to provide the full monorepo context to the build command. In the Dockerfile then we copy the backend and the types packages only. This prevents using the --immutable flag on yarn install as the lock.file contains the dependencies for the FE also. However until now I see no issue with this. If you have other thoughts, let me know.
docker build -f packages/backend/Dockerfile . -t your-backend-service-tag
To run the backend in Docker:
docker-compose up --build
To run frontend, in a standalone terminal run the command below from root:
yarn start:frontend
Add Shadcn UI components to frontend workspace
yarn workspace frontend exec -- npx shadcn-ui@latest add <component-name>
- Implement Postgres DB with Prisma and add Articles model
- Add mutation to feature to Articles trough GraphGL
- Add user registration with Oauth 2 and Google
- Restrict the Articles mutation to logged in users
- Add for each user a custom domain where they can add their blog Articles
- Add better env variables management - NestJs has a built in support, check how it works
- Add HTTPS for complete application
- Create a monorepo with NX.dev and connect it with the FE part - optional
- Rewrite FE to Remix