Skip to main content

Namespaces

A namespace is the key you give a source in tako.config.ts:

sources: {
  pg: { use: prismaParser, options: { schema: './prisma/pg.prisma' } },
  crm: { use: prismaParser, options: { schema: './prisma/crm.prisma' } },
}

Here pg and crm are namespaces. A namespace must match ^[a-z][a-zA-Z0-9]*$ — it becomes a directory name and an import-path segment.

What a namespace does​

  • It isolates entities. pg.User and crm.User are different entities in the IR and are never merged.
  • It drives output location. Generated code for a namespace lands in its own subtree (generated/kurotako/pg/…) or, in mode B, its own npm package (@scope/pg).

What a namespace does not do​

  • It never changes generated identifiers. The entity User produces UserSchema, UserDto, UserForm whether the config has one namespace or ten. Names are deterministic and are never prefixed with the namespace.

  • Disambiguation happens through the import path, not the identifier:

    import { UserDto } from './generated/kurotako/pg';
    import { UserDto as CrmUserDto } from './generated/kurotako/crm';

One parser package, several namespaces​

The same parser package can back multiple namespaces — two Prisma schemas, a Prisma schema and a (future) Mongoose schema, and so on. Each sources entry is an independent instance with its own options.