decorates 4.0.0

Decorator-driven persistence registry for Pydantic models and CLI tooling

Decorates

Decorator-driven tooling for Python:

• decorates.cli for ergonomic command-line apps
• decorates.db for Pydantic + SQLAlchemy persistence

The philosophy is simple: minimal setup, predictable behavior, and a fast path to shipping.

Install

pip install decorates

Quick Start Guide

1. Build one CLI command with a decorator.
2. Build one DB model with a decorator.
3. Use Model.objects for CRUD.

CLI in 60 seconds

import decorates.cli as cli


@cli.register(description="Greet someone")
@cli.argument("name", type=str, help="Name to greet")
@cli.option("--greet")
@cli.option("-g")
def greet(name: str) -> str:
    return f"Hello, {name}!"


if __name__ == "__main__":
    cli.run()

python app.py greet Alice
python app.py --greet Alice
python app.py -g Alice

Database + FastAPI in 5 minutes

from contextlib import asynccontextmanager
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from decorates.db import (
    RecordNotFoundError,
    UniqueConstraintError,
    database_registry,
)

DB_URL = "sqlite:///shop.db"


@database_registry(DB_URL, table_name="customers", unique_fields=["email"])
class Customer(BaseModel):
    id: int | None = None
    name: str
    email: str


@database_registry(DB_URL, table_name="products")
class Product(BaseModel):
    id: int | None = None
    name: str
    price: float


@database_registry(DB_URL, table_name="orders")
class Order(BaseModel):
    id: int | None = None
    customer_id: int
    product_id: int
    quantity: int
    total: float


class CreateCustomer(BaseModel):
    name: str
    email: str


class CreateProduct(BaseModel):
    name: str
    price: float


class CreateOrder(BaseModel):
    customer_id: int
    product_id: int
    quantity: int


@asynccontextmanager
async def lifespan(app: FastAPI):
    for model in (Customer, Product, Order):
        model.create_schema()
    yield
    for model in (Customer, Product, Order):
        model.objects.dispose()


app = FastAPI(lifespan=lifespan)


@app.post("/customers", response_model=Customer, status_code=201)
def create_customer(payload: CreateCustomer):
    try:
        return Customer.objects.create(**payload.model_dump())
    except UniqueConstraintError:
        raise HTTPException(status_code=409, detail="Email already exists")


@app.get("/customers/{customer_id}", response_model=Customer)
def get_customer(customer_id: int):
    try:
        return Customer.objects.require(customer_id)
    except RecordNotFoundError:
        raise HTTPException(status_code=404, detail="Customer not found")


@app.post("/products", response_model=Product, status_code=201)
def create_product(payload: CreateProduct):
    return Product.objects.create(**payload.model_dump())


@app.post("/orders", response_model=Order, status_code=201)
def create_order(payload: CreateOrder):
    customer = Customer.objects.get(payload.customer_id)
    if customer is None:
        raise HTTPException(status_code=404, detail="Customer not found")

    product = Product.objects.get(payload.product_id)
    if product is None:
        raise HTTPException(status_code=404, detail="Product not found")

    return Order.objects.create(
        customer_id=customer.id,
        product_id=product.id,
        quantity=payload.quantity,
        total=product.price * payload.quantity,
    )


@app.get("/orders/desc", response_model=list[Order])
def list_orders_desc(limit: int = 20, offset: int = 0):  # Filter by oldest   (1, 2, 3...n)
    return Order.objects.filter(order_by="id", limit=limit, offset=offset)


@app.get("/orders/asc", response_model=list[Order])
def list_orders_asc(limit: int = 20, offset: int = 0):  # Filter by newest  (n...3, 2, 1)
    return Order.objects.filter(order_by="-id", limit=limit, offset=offset)

Core Concepts

decorates.cli

• Register functions with module-level decorators: @register, @argument, @option.
• Run command handlers through the module registry via decorates.cli.run().
• Support positional + named argument forms (for non-bool args), with bool flags as --flag.
• Command aliases are declared with @option("-x") / @option("--long").
• Runtime wraps unexpected handler crashes as CommandExecutionError (with original exception chaining).
• Operational logs use standard Python logging namespaces under decorates.cli.*.

decorates.db

• Register BaseModel classes with @database_registry(...).
• Access all persistence through Model.objects.
• id: int | None = None gives database-managed autoincrement IDs.
• Schema helpers are available as class methods: create_schema, drop_schema, schema_exists, truncate.
• Unexpected SQLAlchemy runtime failures are normalized into SchemaError for cleaner, predictable error handling.
• Operational logs use standard Python logging namespaces under decorates.db.*.
• DB exceptions provide structured metadata (exc.context, exc.to_dict()) for production diagnostics.

decorates.db Usage Snapshot

# Filtering operators
Order.objects.filter(total__gte=100)
Customer.objects.filter(email__ilike="[e-mail]")
Order.objects.filter(quantity__in=[1, 2, 3])

# Sorting and pagination
Order.objects.filter(order_by="-id", limit=20, offset=0)

# Bulk writes
Product.objects.bulk_create([...])
Product.objects.bulk_upsert([...])

# Additive migration helpers
Customer.objects.ensure_column("phone", str | None, nullable=True)
Customer.objects.rename_table("customers_archive")

After rename_table(...) succeeds, the same Model.objects manager and schema helpers are immediately bound to the new table name.

If your model contains a field named password, password values are automatically hashed on write, and instances receive verify_password(...).

Documentation

• DB guide: src/decorates/db/USAGE.md
• CLI source API: src/decorates/cli
• DB source API: src/decorates/db

Requirements

• Python 3.10+
• pydantic>=2.0
• sqlalchemy>=2.0

Testing

• Default pytest includes SQLite plus PostgreSQL/MySQL rename-state integration tests.
• Start Docker Desktop (or another Docker engine) before running tests so docker-compose.test-db.yml services can boot.
• The decorates is backed by a rigorous, production-focused test suite (170+ tests) that covers unit, edge-case, and multi-dialect integration behavior.

License

MIT