I propose enabling the migrator to programmatically execute migrations defined as Lucid schemas nested within the migrations directory.

Currently, when provisioning tenants in my multi-tenant application, I resort to raw SQL queries via trx.raw. Ideally, I would prefer to specify a target directory containing the relevant migrations. This would allow the MigrationRunner class to automatically populate the tenant's schema.

This functionality could be implemented by:

• Allowing a directory to be specified at the configuration level.
• Adding an optional parameter to the .run() method to designate the migration directory.