Database¶

Connection registry and drivers. Models, migrations, and seeds: see Models.

Register connections¶

Define the map in app/config/Database.py (data only — no Connections call at import). Pass it into Future in run.py:

from future.databases.SQLite import SQLite
from app.config.Settings import DB_DATABASE

DATABASES = {
    "default": "sqlite",
    "sqlite": SQLite(database=DB_DATABASE),
}

config = {
    "APP_DOMAIN": APP_DOMAIN,
    "APP_NAME": APP_NAME,
    "DATABASES": DATABASES,
}
app = Future(lifespan=lifespan, config=config)

CLI migrate / seed load run.py so registration matches runtime. Env loading (.env or Ansible YAML): Configuration.

Per-model connection¶

Models are not locked to one database. Set __connection__ on the model to a key from DATABASES (or "default" for DATABASES["default"]). Migrations inherit that name; see Models.

Run migrations and seeds¶

Generate files from model annotations first (future make:migration / make:seed) — details in Models. Then:

future migrate
future migrate rollback
future seed
future seed StockSeeder

Drivers¶

Driver	Schema / migrate	CRUD
SQLite	Yes	Yes
MySQL	Yes	Yes
Postgres	Yes	Yes
Elasticsearch	Yes	Yes
MongoDB	Yes	Yes
ClickHouse	Yes	Yes
Redis	N/A	Yes

Scaffold default is SQLite. Set DB_DATABASE / sqlite_database to a bare name; the driver appends .sqlite.

For local MySQL, Postgres, Redis, MongoDB, Elasticsearch, ClickHouse, and RabbitMQ, see the Docker example.