Migrations with Alembic

Typically, you’ll want to use Alembic to generate and run migrations as you create and modify your tables. The Flask-Alembic extension provides integration between Flask, Flask-SQLAlchemy(-Lite), and Alembic. You can also use Alembic directly, but it will require a little more setup.

Flask-Alembic

Flask-Alembic will use the engines configured by Flask-SQLAlchemy-Lite. You need to tell it about the metadata object for your base model. Multiple databases are also supported, see the Flask-Alembic docs for details.

from flask import Flask
from flask_alembic import Alembic
from flask_sqlalchemy_lite import SQLAlchemy
from sqlalchemy.orm import DeclarativeBase, Mapped, mapped_column

class Model(DeclarativeBase):
    pass

class User(Model):
    __tablename__ = "user"
    id: Mapped[int] = mapped_column(primary_key=True)
    name: Mapped[str]

db = SQLAlchemy()
alembic = Alembic(metadatas=Model.metadata)

def create_app():
    app = Flask(__name__)
    app.config["SQLALCHEMY_ENGINES"] = {"default": "sqlite:///default.sqlite"}
    app.config.from_prefixed_env()
    db.init_app(app)
    alembic.init_app(app)
    return app

$ flask db revision 'init'
$ flask db upgrade

Plain Alembic

You’ll need to modify the migrations/env.py script that Alembic generates to tell it about your Flask application and the engine and metadata.

$ alembic init migrations

Modify parts of migrations/env.py, the ... are omitted parts of the file.

from project import create_app, Model, db

flask_app = create_app()

...

target_metadata = Model.metadata

...

def run_migrations_online() -> None:
    with flask_app.app_context():
        connectable = db.engine

    ...

...

$ alembic revision --autogenerate -m 'init'
$ alembic upgrade head