sqlalchemy-pydantic-codegen 0.1.0

A library to generate Pydantic models from SQLAlchemy models.

SQLAlchemy-Pydantic Codegen

A Python library for generating Pydantic models from SQLAlchemy models, providing a seamless integration between SQLAlchemy and Pydantic for data validation and serialization.

Features

• Automatic Pydantic model generation from SQLAlchemy models.
• Relationship support: Nested models for SQLAlchemy relationships.
• Enum reflection: SQLAlchemy enums become Pydantic enums.
• Field metadata: Comments become descriptions, string lengths become max_length.
• Jinja2 templating for customizable output.
• CLI for easy usage, including output directory and config options.
• Custom JSON/JSONB field mapping to your own Pydantic models.
• Post-generation cleaning to remove or replace fields as needed.
• Auto-generated __init__.py for schema packages.
• Ready for formatting tools (e.g., Black, Ruff).

Installation

uv add sqlalchemy-pydantic-codegen

or

pip install sqlalchemy-pydantic-codegen

Preparing Your SQLAlchemy Models

Before using sqlalchemy-pydantic-codegen, you may want to generate your SQLAlchemy models from an existing database using sqlacodegen. This tool introspects your database and creates SQLAlchemy model classes automatically. To generate models, run:

sqlacodegen postgresql://user:password@localhost/dbname --outfile src/db/models.py

Once your models are generated, you can proceed to use sqlalchemy-pydantic-codegen as described below.

Usage

After installation, use the CLI to generate Pydantic models from your SQLAlchemy models.

Basic Usage

sqlalchemy-pydantic-codegen --models-path my_app.db.models --output-dir src/schemas

--models-path: Dotted path to your SQLAlchemy models module (required). --output-dir: Directory for generated Pydantic schemas (default: src/schemas).

Custom Configuration

To map JSON/JSONB fields to custom Pydantic models, use the --config option.

Create a config file (e.g., codegen_config.py):

# codegen_config.py

# Maps table names to a dictionary of field names and the Pydantic model to use.
CUSTOM_JSONB_MODELS = {
    "my_table": {
        "my_jsonb_field": "MyCustomPydanticModelForJsonbField",
    },
}

# Maps the Pydantic model name to its full import statement.
CUSTOM_IMPORTS = {
    "MyCustomPydanticModelForJsonbField": "from my_app.schemas import MyCustomPydanticModelForJsonbField",
}

Then, run the command with the --config flag:

sqlalchemy-pydantic-codegen --models-path my_app.db.models --output-dir src/schemas --config codegen_config.py

Output

• One Pydantic schema file per SQLAlchemy model.
• init.py with all exports and forward references.
• Cleaned and ready-to-use Pydantic models.

Advanced

• Supports nested relationships and cyclic reference handling.
• Enum columns become Pydantic enums.
• Field comments and constraints are preserved as Pydantic metadata.
• Easily extend templates for your own conventions.