Skip to content
Plystra Plystra Docs

Ent Database Management

Ent Database Management

Section titled “Ent Database Management”

Plystra now uses Ent as the canonical Go schema model for Core database objects.

The SQL files in migrations/ remain the ordered, checksum-verified migration history. Atlas executes those migrations and verifies the migrations/atlas.sum integrity file. Ent is used for:

  • type-safe schema definitions in ent/schema
  • generated entity builders and query APIs in ent
  • guarded, non-destructive schema apply checks through plystractl ent apply
  • schema presence checks through plystractl ent status

This split keeps production upgrades auditable while keeping business data access on Ent. Raw SQL is limited to migration metadata and Atlas migration execution.

Commands

Section titled “Commands”

Regenerate Ent code after editing ent/schema:

Terminal window
go run entgo.io/ent/cmd/ent generate ./ent/schema

Check that Ent-managed tables are present:

Terminal window
go run .\cmd\plystractl ent status

Apply the Ent schema without dropping existing columns or indexes:

Terminal window
go run .\cmd\plystractl ent apply

Apply production migrations through Atlas:

Terminal window
go run .\cmd\plystractl migrate up
go run .\cmd\plystractl migrate verify

Modeling Notes

Section titled “Modeling Notes”
  • All first-class tables use Plystra’s existing text IDs.
  • role_permissions is modeled as a first-class Ent schema with its own id, preserving the v1.0 requirement that important relationship tables stay explicit.
  • JSON/JSONB columns are modeled as Ent JSON fields.
  • Core business repositories use Ent. Keep raw SQL out of business handlers unless it is migration metadata or an explicitly documented control-plane exception.