prisma-pglite

Enabling easy Prisma usage with PGlite. This includes a runtime adapter helper and a CLI Prisma wrapper:

• CLI

  • Enables prisma migrate dev with a PGlite database.
  • Enables prisma migrate reset with a PGlite database.
  • Passes all other commands to the standard prisma CLI.

• adapter helper

  • Includes and wraps pglite-prisma-adapter.
  • Automatically propagates your schema to the PGlite database (similar to prisma db push but lacking the ability to apply new migrations to an existing PGlite database).

See the full reference docs at https://electrovir.github.io/prisma-pglite

Install

npm i prisma-pglite
Copy

Usage

tl;dr

• Set the driverAdapters preview feature in your schema.prisma:

  generator client {
      provider        = "prisma-client-js"
      previewFeatures = ["driverAdapters"]
  }
  Copy

• Use createPgliteAdapter:

  import {PrismaClient} from '@prisma/client';
  import {createPgliteAdapter} from 'prisma-pglite';
  
  const prismaClient = new PrismaClient({
      adapter: await createPgliteAdapter(),
  });
  Copy

• Create new migrations with npx prisma-pglite migrate dev

CLI

Instead of running Prisma commands like this:

npx prisma migrate dev
Copy

Run them like this:

npx prisma-pglite migrate dev
Copy

The prisma-pglite CLI will:

• Intercept migrate dev commands so that they work without running a full Postgres instance.
  • Use --schema <schema-path> to customize the schema location.
  • Use --name <migration-name> to provide the migration name inline (otherwise the CLI will prompt you for one).
  • Use --migrations <migrations-dir-path> to customize the location of your migrations folder.
  • Note that this command will generate a source.snapshot (customizable with --snapshot <snapshot-file-name>) file inside each migration. You must keep and commit this file otherwise this command will not work in the future (this file is used to keep track of migration progress instead of a postgres instance, as Prisma normally uses).
• Intercept migrate reset commands so that they work with a PGlite database.
  • Use --schema <schema-path> to customize the schema location.
  • Use --database <pglite-db-parent-dir-path> to customize the location of your PGlite database that needs to be reset.
• Automatically append --no-hints to the prisma generate command.
• Pass all other commands directly to the Prisma CLI without modification.

All CLI commands are also accessible via the exported API:

• migrate dev:

  import {createPgliteMigration} from 'prisma-pglite';
  
  await createPgliteMigration({migrationName: 'my migration'});
  Copy

• migrate reset:

  import {resetPgliteDatabase} from 'prisma-pglite';
  
  await resetPgliteDatabase();
  Copy

• the whole CLI:

  import {runPrisma} from 'prisma-pglite';
  
  await runPrisma(['migrate dev']);
  await runPrisma(['generate']);
  Copy

Adapter Helper

Use this when instantiating your PrismaClient instance to connect to or create a PGlite database:

import {PrismaClient} from '@prisma/client';
import {join} from 'node:path';
import {createPgliteAdapter} from 'prisma-pglite';

const mySchemaPath = join('packages', 'backend', 'prisma', 'schema.prisma');

const prismaClient = new PrismaClient({
    adapter: await createPgliteAdapter({
        schemaFilePath: mySchemaPath,
    }),
});
Copy

NOTE: createPgliteAdapter can only push your schema to new PGlite databases. Thus, you'll need to reset your dev database whenever you create a new migration (I suggest investing in a seed script to workaround this and make your dev experience more consistent in general).

See the type PgliteAdapterParams for more details on customizing createPgliteAdapter.

Make sure that you have the driverAdapters preview feature enabled in your schema.prisma. If you don't enable this, your PrismaClient constructor won't have an adapter parameter available.

generator client {
  provider        = "prisma-client-js"
  previewFeatures = ["driverAdapters"]
}
Copy

PGlite paths

By default, the pgliteDirPath parameter of createPgliteAdapter expects multiple PGlite database to exist within itself. The default database will be in ${pgliteDirPath}/dev and all test database (triggered by passing in the testContext parameter) will be in ${pgliteDirPath}/${testName}. This structure can be ignored entirely by instead setting the directDatabaseDirPath parameter, but then you lose automatic test database paths.

Examples

• Create a new migration with some flags:

  npx prisma-pglite migrate dev --schema packages/backend/prisma/schema.prisma --name 'add user table'
  Copy

• Reset your PGlite database

  npx prisma-pglite migrate reset --database .config/pglite-db
  Copy

• Customize the PGlite adapter:

  import {PrismaClient} from '@prisma/client';
  import {join} from 'node:path';
  import {createPgliteAdapter} from 'prisma-pglite';
  
  const mySchemaPath = join('packages', 'backend', 'prisma', 'schema.prisma');
  
  const prismaClient = new PrismaClient({
      adapter: await createPgliteAdapter({
          schemaFilePath: mySchemaPath,
          pgliteDirPath: join('.dev', 'pglite'),
      }),
  });
  Copy

• Create a fresh and super fast PGlite database instance for each test:

  import {describe, it} from '@augment-vir/test';
  import {PrismaClient} from '@prisma/client';
  import {join} from 'node:path';
  import {createPgliteAdapter} from 'prisma-pglite';
  
  const mySchemaPath = join('packages', 'backend', 'prisma', 'schema.prisma');
  
  describe('my test', () => {
      it('connects to the database', async (testContext) => {
          const prismaClient = new PrismaClient({
              adapter: await createPgliteAdapter({
                  schemaFilePath: mySchemaPath,
                  pgliteDirPath: join('.dev', 'pglite'),
                  test: testContext,
              }),
          });
      });
  });
  Copy