Editing Prisma Schema

The general workflow will follow these steps:

• Create and checkout a new working branch.
• Edit schema.prisma as needed.
• Run automated formatting: npm run prisma:format
• Prepare database changes (as needed):
  • create (preview) migration: npm run prisma:create or;
  • create and apply migration: npm run prisma:apply
• Update minor version number in package.json "version": "[MAJOR].[MINOR].[PATCH]"
• Create and test new seeding function(s) if required - being cautious to avoid creating duplicate records on repeated runs.
• Raise PR for merging back into main branch.

The changes should be ready for review, merging, and publishing now.

Deploying Prisma Schema

Completing the PR (merging into main) will automatically trigger the build and release pipelines that will publish the new version of the package to the remote npm repository.

The latest version can then be installed in the application implementing the schema:

npm install @bslau/hmm_prisma_schema

NOTE that doing this on the BlueScope network will typically fail. Installing this package involves generating a fresh Prisma client file - this is the part that seems to stall/fail because it's ignorant of the system proxy settings. The workaround is to use a hotspot internet connection without any proxy settings instead.

Manual Deployment for Test Environment

Current implementations of this package don't automate migration and seeding steps after release to the Test environment. This will result in schema mismatch between the Prisma package and the Test database when updates are released.

The workaround is to manually execute these steps after release. This example describes the process for the GraphQL Apollo server, but would also work for the Event Processing service (albeit in a different directory):

• RDP into the host server: ITAVWEB91

• Open a bash terminal and navigate to the installed package:

  C:\inetpub\wwwroot\hmm\graphql_server\node_modules\@bslau\hmm_prisma_schema

• Apply any pending migrations: npm run prisma:deploy

• Run the seed function to create data: npm run prisma:seed

At this point, assuming no errors in deploying or seeding, the Test database schema should be synchronised with Prisma now.