advanced⏱65 minutesLección 6 de 8

Repositories & Domain Services

Learn how Repositories abstract data access and Domain Services encapsulate domain operations that don't fit in entities or value objects

Repositories & Domain Services

Repositories and Domain Services are two more tactical building blocks in DDD. Repositories provide a clean abstraction for storing and retrieving aggregates. Domain Services encapsulate domain logic that does not naturally belong in an Entity or Value Object.

ℹ️Note

In DDD, a Repository is not a DAO (Data Access Object). A DAO is a technical concept for data access. A Repository is a domain concept that speaks the Ubiquitous Language — you ask it for aggregates by domain terms, not by SQL queries.

Repositories: The Collection Metaphor

A Repository acts like an in-memory collection of aggregates. You add items, remove items, and query by domain-relevant criteria. The underlying storage (database, filesystem, API) is hidden behind this interface.

💡 Tip: Use Ctrl + Mouse Wheel to zoom, or click the fullscreen icon

Repository Interface

The repository interface is defined in the domain layer. The implementation is in the infrastructure layer. This follows the Dependency Inversion Principle.

python

from abc import ABC, abstractmethod
from typing import List, Optional, Protocol


# --- Domain Layer: Repository interface ---

class OrderRepository(Protocol):
    """A repository for Order aggregates.
    Defined in the domain layer — speaks the Ubiquitous Language."""

    def find_by_id(self, order_id: str) -> Optional["Order"]:
        """Find an order by its unique identifier."""
        ...

    def find_by_customer(self, customer_id: str) -> List["Order"]:
        """Find all orders placed by a customer."""
        ...

    def find_pending_orders(self) -> List["Order"]:
        """Find all orders that have not been confirmed yet."""
        ...

    def find_orders_older_than(self, days: int) -> List["Order"]:
        """Find orders placed more than the specified days ago."""
        ...

    def save(self, order: "Order") -> None:
        """Persist an order aggregate (insert or update)."""
        ...

    def delete(self, order_id: str) -> None:
        """Remove an order from the repository."""
        ...

Repository Implementation

The implementation lives in the infrastructure layer. It translates between the domain model and the persistence mechanism.

python

# --- Infrastructure Layer: Repository Implementation ---

import sqlite3
from typing import List, Optional
from datetime import datetime, timedelta


class SQLiteOrderRepository:
    """Implements OrderRepository using SQLite.
    This is infrastructure code — it depends on the domain interface."""

    def __init__(self, db_path: str):
        self._conn = sqlite3.connect(db_path, detect_types=sqlite3.PARSE_DECLTYPES)
        self._init_tables()

    def _init_tables(self) -> None:
        self._conn.execute("""
            CREATE TABLE IF NOT EXISTS orders (
                order_id TEXT PRIMARY KEY,
                customer_id TEXT NOT NULL,
                status TEXT NOT NULL,
                placed_at TIMESTAMP NOT NULL,
                total REAL NOT NULL,
                data TEXT NOT NULL
            )
        """)
        self._conn.execute("""
            CREATE TABLE IF NOT EXISTS order_lines (
                line_id TEXT PRIMARY KEY,
                order_id TEXT NOT NULL,
                product_id TEXT NOT NULL,
                product_name TEXT NOT NULL,
                quantity INTEGER NOT NULL,
                unit_price REAL NOT NULL,
                FOREIGN KEY (order_id) REFERENCES orders(order_id)
            )
        """)
        self._conn.commit()

    def find_by_id(self, order_id: str) -> Optional["Order"]:
        cursor = self._conn.execute(
            "SELECT order_id, customer_id, status, placed_at FROM orders WHERE order_id = ?",
            (order_id,)
        )
        row = cursor.fetchone()
        if not row:
            return None
        return self._row_to_aggregate(row)

    def find_by_customer(self, customer_id: str) -> List["Order"]:
        cursor = self._conn.execute(
            "SELECT order_id, customer_id, status, placed_at FROM orders WHERE customer_id = ?",
            (customer_id,)
        )
        return [self._row_to_aggregate(row) for row in cursor.fetchall()]

    def find_pending_orders(self) -> List["Order"]:
        cursor = self._conn.execute(
            "SELECT order_id, customer_id, status, placed_at FROM orders WHERE status = ?",
            ("pending",)
        )
        return [self._row_to_aggregate(row) for row in cursor.fetchall()]

    def save(self, order: "Order") -> None:
        # Serialize order data
        self._conn.execute(
            """INSERT OR REPLACE INTO orders
               (order_id, customer_id, status, placed_at, total)
               VALUES (?, ?, ?, ?, ?)""",
            (order.id, order.customer_id, order.status.value,
             order.placed_at, order.total)
        )
        # Delete old lines, insert current ones
        self._conn.execute("DELETE FROM order_lines WHERE order_id = ?", (order.id,))
        for line in order.lines:
            self._conn.execute(
                """INSERT INTO order_lines
                   (line_id, order_id, product_id, product_name, quantity, unit_price)
                   VALUES (?, ?, ?, ?, ?, ?)""",
                (line.id, order.id, line.product_id, line.product_name,
                 line.quantity, line.unit_price)
            )
        self._conn.commit()

    def delete(self, order_id: str) -> None:
        self._conn.execute("DELETE FROM order_lines WHERE order_id = ?", (order_id,))
        self._conn.execute("DELETE FROM orders WHERE order_id = ?", (order_id,))
        self._conn.commit()

    def _row_to_aggregate(self, row) -> "Order":
        order = Order(row["customer_id"])
        # Load from row data...
        return order

    def __del__(self):
        self._conn.close()

In-Memory Repository for Testing

The testability benefit of the Repository pattern is enormous. You can write an in-memory implementation for unit tests:

python

class InMemoryOrderRepository:
    """In-memory implementation for testing.
    No database needed — tests run in milliseconds."""

    def __init__(self):
        self._orders: dict[str, "Order"] = {}

    def find_by_id(self, order_id: str) -> Optional["Order"]:
        return self._orders.get(order_id)

    def find_by_customer(self, customer_id: str) -> List["Order"]:
        return [o for o in self._orders.values()
                if o.customer_id == customer_id]

    def find_pending_orders(self) -> List["Order"]:
        return [o for o in self._orders.values()
                if o.status == "pending"]

    def save(self, order: "Order") -> None:
        self._orders[order.id] = order

    def delete(self, order_id: str) -> None:
        self._orders.pop(order_id, None)

⚠️Warning

A Repository should not be used for arbitrary queries. If you find yourself adding query methods for every possible data access pattern, you are using the Repository wrong. The Repository is for aggregate retrieval, not for reporting or data exploration.

Repository Usage in Application Services

python

class ConfirmOrderHandler:
    """Application service using the Repository."""

    def __init__(self, repo: OrderRepository, uow: "UnitOfWork"):
        self._repo = repo
        self._uow = uow

    def handle(self, order_id: str) -> None:
        self._uow.begin()
        try:
            order = self._repo.find_by_id(order_id)
            if not order:
                raise ValueError(f"Order {order_id} not found")
            order.confirm()
            self._repo.save(order)
            self._uow.commit()
        except Exception:
            self._uow.rollback()
            raise

Domain Services

A Domain Service is an operation that does not naturally belong to an Entity or Value Object. It coordinates multiple aggregates or performs calculations that involve domain logic.

When to Use a Domain Service

Situation	Example	Solution
Operation involves multiple aggregates	Transfer money between accounts	AccountTransferService
Operation involves external domain	Calculate shipping cost with carrier rules	ShippingCostService
Operation is stateless computation	Apply complex pricing rules	PricingService
Operation requires domain knowledge not owned by any entity	Detect fraud patterns	FraudDetectionService

python

from dataclasses import dataclass
from typing import List, Protocol
from decimal import Decimal


# --- Domain Service: Pricing ---

@dataclass(frozen=True)
class Money:
    amount: Decimal
    currency: str

    def __add__(self, other: "Money") -> "Money":
        if self.currency != other.currency:
            raise ValueError("Currency mismatch")
        return Money(self.amount + other.amount, self.currency)

    def __mul__(self, quantity: int) -> "Money":
        return Money(self.amount * quantity, self.currency)


class PricingService:
    """Domain Service for pricing calculations.
    Not a natural fit for Order or Product — it involves
    multiple aggregates and external rules."""

    GOLD_DISCOUNT = Decimal("0.15")
    SILVER_DISCOUNT = Decimal("0.10")
    STANDARD_VAT = Decimal("0.20")

    def calculate_order_total(
        self,
        items: List["OrderLine"],
        customer_tier: str,
        coupon_code: str | None
    ) -> Money:
        subtotal = self._calculate_subtotal(items)
        discount = self._apply_discount(subtotal, customer_tier, coupon_code)
        vat = self._calculate_vat(discount)
        return discount + vat

    def _calculate_subtotal(self, items: List["OrderLine"]) -> Money:
        total = Money(Decimal("0.00"), "USD")
        for item in items:
            total = total + item.subtotal()
        return total

    def _apply_discount(
        self, subtotal: Money, tier: str, coupon: str | None
    ) -> Money:
        discount_rate = Decimal("0")
        if tier == "gold":
            discount_rate = self.GOLD_DISCOUNT
        elif tier == "silver":
            discount_rate = self.SILVER_DISCOUNT
        if coupon == "WELCOME10":
            discount_rate += Decimal("0.10")
        discount_amount = subtotal.amount * discount_rate
        return Money(subtotal.amount - discount_amount, subtotal.currency)

    def _calculate_vat(self, amount: Money) -> Money:
        vat_amount = amount.amount * self.STANDARD_VAT
        return Money(vat_amount, amount.currency)

Domain Service Example: Transfer Money

python

class Account:
    """Aggregate Root for a bank account."""

    def __init__(self, account_id: str, owner: str):
        self._id = account_id
        self._owner = owner
        self._balance = Money(Decimal("0.00"), "USD")
        self._is_frozen = False

    @property
    def balance(self) -> Money:
        return self._balance

    def deposit(self, amount: Money) -> None:
        if self._is_frozen:
            raise ValueError("Account is frozen")
        if amount.amount <= 0:
            raise ValueError("Deposit amount must be positive")
        self._balance = self._balance + amount

    def withdraw(self, amount: Money) -> None:
        if self._is_frozen:
            raise ValueError("Account is frozen")
        if amount.amount <= 0:
            raise ValueError("Withdrawal amount must be positive")
        if self._balance.amount < amount.amount:
            raise ValueError("Insufficient funds")
        self._balance = self._balance - amount


class TransferService:
    """Domain Service for money transfers.
    Coordinates two Account aggregates."""

    MINIMUM_TRANSFER = Money(Decimal("1.00"), "USD")
    MAXIMUM_TRANSFER = Money(Decimal("10000.00"), "USD")

    def transfer(
        self, source: Account, destination: Account,
        amount: Money, reason: str
    ) -> None:
        if amount.amount < self.MINIMUM_TRANSFER.amount:
            raise ValueError("Transfer amount below minimum")
        if amount.amount > self.MAXIMUM_TRANSFER.amount:
            raise ValueError("Transfer amount exceeds maximum")
        source.withdraw(amount)
        destination.deposit(amount)

💡Tip

| > Domain Services should be **stateless**. They hold no mutable state of their own. Their dependencies (repositories, other services) are injected through the constructor. The operations they perform are the reason they exist.

Domain Service Example: Fraud Detection

python

@dataclass
class FraudAssessment:
    risk_level: str  # "low", "medium", "high"
    score: int
    reasons: List[str]

class FraudDetectionService:
    """Domain Service that assesses fraud risk.
    Coordinates across multiple aggregates and external data."""

    def __init__(self, order_repo: "OrderRepository",
                 customer_repo: "CustomerRepository"):
        self._order_repo = order_repo
        self._customer_repo = customer_repo

    def assess_order_risk(self, order: "Order") -> FraudAssessment:
        reasons: List[str] = []
        score = 0

        # Rule 1: Large orders
        if order.total.amount > 5000:
            score += 30
            reasons.append("Order total exceeds $5,000")

        # Rule 2: Multiple orders in short time
        recent_orders = self._order_repo.find_by_customer(order.customer_id)
        recent_count = sum(
            1 for o in recent_orders
            if (datetime.now() - o.placed_at).hours < 24
        )
        if recent_count >= 3:
            score += 25
            reasons.append("More than 3 orders in 24 hours")

        # Rule 3: New customer with high value
        customer = self._customer_repo.find_by_id(order.customer_id)
        if customer and customer.account_age_days < 30 and order.total.amount > 2000:
            score += 20
            reasons.append("New customer with high-value order")

        risk = "low" if score < 30 else "medium" if score < 60 else "high"
        return FraudAssessment(risk, score, reasons)

Repository vs DAO vs Unit of Work

Pattern	Purpose	Layer
Repository	Collection-like access to aggregates	Domain interface
DAO	Low-level data access	Infrastructure
Unit of Work	Transaction management	Infrastructure
Query Object	Read-optimized queries	Application

python

# Separate read model (CQRS) from the write model repository
class OrderQueryService:
    """Optimized for reads — not a repository.
    Returns read-only DTOs, not aggregates."""

    def __init__(self, connection):
        self._conn = connection

    def get_order_summary(self, order_id: str) -> dict | None:
        cursor = self._conn.execute("""
            SELECT o.order_id, o.customer_id, o.status,
                   COUNT(ol.line_id) as item_count,
                   o.total
            FROM orders o
            LEFT JOIN order_lines ol ON o.order_id = ol.order_id
            WHERE o.order_id = ?
            GROUP BY o.order_id
        """, (order_id,))
        row = cursor.fetchone()
        if not row:
            return None
        return dict(row)

    def get_customer_order_history(self, customer_id: str) -> List[dict]:
        cursor = self._conn.execute("""
            SELECT order_id, placed_at, status, total
            FROM orders
            WHERE customer_id = ?
            ORDER BY placed_at DESC
            LIMIT 50
        """, (customer_id,))
        return [dict(row) for row in cursor.fetchall()]

Testing with Repositories and Domain Services

The true power of these patterns becomes evident in tests:

python

import pytest
from datetime import datetime, timedelta


class TestTransferService:
    def setup_method(self):
        self.source = Account("ACC-001", "Alice")
        self.source.deposit(Money(Decimal("1000.00"), "USD"))

        self.destination = Account("ACC-002", "Bob")
        self.destination.deposit(Money(Decimal("100.00"), "USD"))

        self.service = TransferService()

    def test_successful_transfer(self):
        self.service.transfer(
            self.source, self.destination,
            Money(Decimal("200.00"), "USD"), "Payment"
        )
        assert self.source.balance == Money(Decimal("800.00"), "USD")
        assert self.destination.balance == Money(Decimal("300.00"), "USD")

    def test_transfer_insufficient_funds(self):
        with pytest.raises(ValueError, match="Insufficient funds"):
            self.service.transfer(
                self.source, self.destination,
                Money(Decimal("2000.00"), "USD"), "Overdraft"
            )


class TestOrderLifecycleWithRepo:
    def test_confirm_order_through_repo(self):
        # Arrange
        repo = InMemoryOrderRepository()
        order = Order(customer_id="CUST-001")
        order.add_product("P1", "Widget", 2, 10.0)
        repo.save(order)

        # Act
        loaded = repo.find_by_id(order.id)
        loaded.confirm()
        repo.save(loaded)

        # Assert
        confirmed = repo.find_by_id(order.id)
        assert confirmed.status == OrderStatus.CONFIRMED

Summary: Repository and Domain Service Guidelines

Aspect	Repository	Domain Service
Purpose	Store/retrieve aggregates	Coordinate domain operations
State	Stateless	Stateless
Dependency	Depends on infrastructure	Depends on domain interfaces
Testability	Easy (in-memory impl)	Easy (unit tests)
Layer	Domain interface, Infrastructure impl	Domain
Methods	`find_by_*`, `save`, `delete`	Domain-specific verbs
Granularity	One per aggregate root	One per domain capability

💡 Tip: Use Ctrl + Mouse Wheel to zoom, or click the fullscreen icon

✅Success

Repositories keep your domain model persistence-ignorant. Domain Services keep domain logic from leaking into application services. Together, they ensure that the domain layer remains the center of your system, with infrastructure concerns pushed to the edges.

Practice Exercises

Design a Repository interface: Design a CustomerRepository for a subscription management system. Include methods for finding customers, querying by subscription status, and the standard save/delete operations. Write the interface using Protocol.
Implement an in-memory Repository: Implement the CustomerRepository from exercise 1 as an InMemoryCustomerRepository. Use it in a test that creates a customer, saves it, retrieves it by ID, and verifies the data.
Implement a SQLite Repository: Write a SQLiteCustomerRepository that implements the same interface. Include the table creation SQL and the mapping logic.
Design a Domain Service: Design a SlotAllocationService for a hospital scheduling system. The service should allocate a patient to a doctor's time slot. Consider: what happens if the slot is taken? What if the patient has a conflict? What if the doctor is overbooked?
Refactor anemic application service: The following code has domain logic leaking into the application layer. Extract the domain logic into a Domain Service:

python
class PlaceOrderController: def handle(self, request): order = Order() for item in request["items"]: if item["quantity"] > 10: raise ValueError("Cannot order more than 10") if item["quantity"] <= 0: raise ValueError("Quantity must be positive") order.add_line(item) if request["customer_tier"] == "gold": order.apply_discount(0.15) repo.save(order)
Repository vs Query Service: Your team needs to display a dashboard showing: total orders today, average order value, top 5 products, and orders by hour. Would you use a Repository or a separate Query Service? Explain why.
Domain Service with dependencies: Design a SubscriptionRenewalService that checks if a subscription should be renewed, processes the payment via a payment gateway, and sends a confirmation. List its dependencies and show the interface.
Test with in-memory repository: Write a test for this scenario: a customer can place up to 5 active orders. When they try to place a 6th, the system rejects it. Use an InMemoryOrderRepository and an OrderService that coordinates the check.

✅Success

You have completed Lesson 6. Repositories keep aggregates persistence-agnostic, and Domain Services keep operations that span multiple aggregates cleanly modeled. Together with Entities, Value Objects, and Aggregates, they form the complete tactical DDD toolbox.

Progreso75%

Aggregates & Aggregate Roots

Event-Driven Architecture

Contenido del Curso

1. Introduction to Domain-Driven Design 2. Ubiquitous Language 3. Bounded Contexts 4. Entities & Value Objects 5. Aggregates & Aggregate Roots 6. Repositories & Domain Services 7. Event-Driven Architecture 8. DDD Implementation in Python