ferro-orm 0.3.2

A high-performance, Rust-backed ORM for Python.

Ferro: The Pydantic x Rust ORM

Ferro is a high-performance, asynchronous ORM for Python, powered by a core engine written in Rust. Designed for simplicity and it only has one dependency: Pydantic. Delivering ergonomics familiar to the modern Pythonista and the speed and safety of Rust's SQLx and Sea-Query.

Key Features

• High-Performance Core: All SQL generation and row hydration are handled by a dedicated Rust engine, minimizing "Python Tax" on data-heavy operations.
• Async First: Built from the ground up for asynchronous applications, utilizing pyo3-async-runtimes for non-blocking I/O.
• Pydantic Integration: Leverages Pydantic V2 for schema definition and data validation, providing full IDE support and type safety.
• Zero-Copy Intent: Designed with zero-copy principles to maximize throughput during large-scale data retrieval.
• Identity Map: Ensures object consistency across your application by tracking active model instances in a thread-safe registry.

Architecture

Ferro operates through a dual-layer architecture connected via a high-performance FFI (Foreign Function Interface) bridge:

1. Python Layer: Developers define models using standard Python classes; a metaclass registers them with the backend.
2. Rust Engine: Built on SQLx and Sea-Query for GIL-free row parsing and object instantiation.

Installation

Ferro is distributed as pre-compiled wheels for macOS, Linux, and Windows.

pip install ferro-orm
# Or with migration support
pip install "ferro-orm[alembic]"

Quick Start

import asyncio
from ferro import Field, Model, connect

class User(Model):
    id: int | None = Field(default=None, primary_key=True)
    username: str
    is_active: bool = True

async def main():
    await connect("sqlite:example.db?mode=rwc", auto_migrate=True)

    # Create
    alice = await User.create(username="alice")

    # Query
    active_users = await User.where(User.is_active == True).all()
    print(f"Found {len(active_users)} active users.")

if __name__ == "__main__":
    asyncio.run(main())