maacck / sqlagen

sqlagen

Model Generator from projects using SQLAlchemy as their orm.

This package is helpful while building web apps with FastAPI and SQLAlchemy.

It provides capability which can:

• generate SQLAlchemy from database schema
• generates Pydantic Model from SQLAlchemy Model.

Project was originally forked from https://github.com/agronholm/sqlacodegen

However, we design to modify some of the features and add new features to the project to fit our own project needs, so we decided to create a new project.

Installation

To install,

pip install sqlagen

Quickstart

Please be aware that we are using pydantic >= 2.0.0 and sqlalchemy >= 2.0 in this project.

Generate SQLAlchemy Models

You may use generate_db_models function directly to generate Pydantic Model from SQLAlchemy Model. It takes Generator's init arguments and init a Generator class.

Examples:

from sqlagen import generate_db_models

generate_db_models(db_url="pymysql+mysql://localhost:6306/test")

from sqlagen import DeclarativeGenerator
from sqlalchemy.engine import create_engine
from sqlalchemy.schema import MetaData

metadata = MetaData()
engine = create_engine("pymysql+mysql://localhost:6306/test")
generator = DeclarativeGenerator(metadata=metadata, engine=engine, options={})
print(generator.generate())

Generate Pydantic Schema Models

You may use generate_db_models function directly to generate Pydantic Model from SQLAlchemy Model. It takes Generator's init arguments and init a Generator class.

Examples:

from sqlagen import generate_schema_models
    
generate_schema_models(models=[MyModel], base_model=CustomBaseModel)

from sqlagen import SchemaModelGenerator

generator = SchemaModelGenerator(base_model=CustomBaseModel)
generator.generate_from_module(models=my_models_module, output_path="schemas.py")

Generator Class takes following init arguments

• split_models: Whether to split models into Base, Create, Update and Read models. Default is Fakse.
• base_model: Base model to inherit from. Default is BaseModel from pydantic.
• restrict_fields: Which takes a set of fields to restrict. Default is None. This is useful when you want to restrict some fields to be readonly such as id, created_at, updated_at.
• indentation: Indentation to use in generated code.

More Examples

Since most of our projects use sqlalchemy and pydantic, so we often need to generate models from database schema and generate pydantic models from sqlalchemy models.

from sqlagen import generate_db_models, SchemaModelGenerator

generate_db_models(db_url="pymysql+mysql://localhost:6306/test",outfile_path="./models")
generate_schema_models(models=[MyModel], base_model=CustomBaseModel, outfile_path="./schemas")

RoadMap

• Strict typing, such as using conint for limiting Integer size and constr for String length.
• Probably, generate relationships as well.