f(model) - Functional Domain Modeling with TypeScript

• f(model) - Functional Domain Modeling with TypeScript
  • IOR<Library, Inspiration>
  • (command: C, state: S) => readonly E[]
  • (state: S, event: E) => S
  • Event-sourced or State-stored systems
  • Decider
  • Event-sourcing aggregate
  • State-stored aggregate
  • View
  • Materialized View
  • Saga
  • Saga Manager
  • Event Modeling
  • Structuring the data
  • C / Command / Intent to change the state of the system
  • E / Event / Fact
  • S / State / Current state of the system/aggregate/entity
  • Modeling the Behaviour
    • Decider - data type that represents the main decision-making algorithm.
    • Event-sourcing aggregate
  • Install as a dependency of your project
  • Examples
  • FModel in other languages
  • Resources
  • Credits

v2.0.0 of the library is introducing breaking changes. Check the PR! Besides keeping the focus on separating data from behavior, we want to split the responsibilities between the domain and application/adapter layers better. For example, metadata types exist only on the application layer, not leaking into the domain, as these don't benefit core logic. Example: traceId, correlationId, ... To keep it simple, v2.*.* will use the main branch going forward. v1.. will continue to be supported (bugs only, no new features)

When you’re developing an information system to automate the activities of the business, you are modeling the business. The abstractions that you design, the behaviors that you implement, and the UI interactions that you build all reflect the business — together, they constitute the model of the domain.

IOR<Library, Inspiration>

This project can be used as a library, or as an inspiration, or both. It provides just enough tactical Domain-Driven Design patterns, optimised for Event Sourcing and CQRS.

The library is fully isolated from the application and infrastructure layers. It represents a pure declaration of the program logic. It is written in TypeScript programming language.

(command: C, state: S) => readonly E[]

On a higher level of abstraction, any information system is responsible for handling the intent (Command) and based on the current State, produce new facts (Events):

• given the current State/S on the input,
• when Command/C is handled on the input,
• expect list of new Events/E to be published/emitted on the output

(state: S, event: E) => S

The new state is always evolved out of the current state S and the current event E:

• given the current State/S on the input,
• when Event/E is handled on the input,
• expect new State/S to be published on the output

Event-sourced or State-stored systems

• State-stored systems are traditional systems that are only storing the current State by overwriting the previous State in the storage.
• Event-sourced systems are storing the events in immutable storage by only appending.

Both types of systems can be designed by using only these two functions and three generic parameters

Decider is the most important datatype, but it is not the only one. Let's discuss all of them, and visualize how they fit the Onion architecture. The arrows in the image are showing the direction of the dependency. Notice that all dependencies point inwards, and that Domain is not depending on anything.

Decider

Decider is a datatype that represents the main decision-making algorithm. It belongs to the Domain layer. It has three generic parameters C, S, E , representing the type of the values that Decider may contain or use. Decider can be specialized for any type C or S or E because these types do not affect its behavior. Decider behaves the same for C=Int or C=YourCustomType, for example.

Decider is a pure domain component.

• C - Command
• S - State
• E - Event

Notice that Decider implements an interface IDecider to communicate the contract.

export class Decider<C, S, E> implements IDecider<C, S, E> {
  constructor(
    readonly decide: (c: C, s: S) => readonly E[],
    readonly evolve: (s: S, e: E) => S,
    readonly initialState: S
  ) {
  }
}

Additionally, initialState of the Decider is introduced to gain more control over the initial state of the Decider.

We can now construct event-sourcing or/and state-storing aggregate by using the same decider.

Event-sourcing aggregate

Event sourcing aggregate is using/delegating a Decider to handle commands and produce events. It belongs to the Application layer. In order to handle the command, aggregate needs to fetch the current state (represented as a list of events) via EventRepository.fetchEvents function, and then delegate the command to the decider which can produce new events as a result. Produced events are then stored via EventRepository.save function.

State-stored aggregate

State stored aggregate is using/delegating a Decider to handle commands and produce new state. It belongs to the Application layer. In order to handle the command, aggregate needs to fetch the current state via StateRepository.fetchState function first, and then delegate the command to the decider which can produce new state as a result. New state is then stored via StateRepository.save function.

View

View is a datatype that represents the event handling algorithm, responsible for translating the events into denormalized state, which is more adequate for querying. It belongs to the Domain layer. It is usually used to create the view/query side of the CQRS pattern. Obviously, the command side of the CQRS is usually event-sourced aggregate.

It has two generic parameters S, E, representing the type of the values that View may contain or use. View can be specialized for any type of S, E because these types do not affect its behavior. View behaves the same for E=Int or E=YourCustomType, for example.

View is a pure domain component.

• S - State
• E - Event

Notice that View implements an interface IView to communicate the contract.

export class View<S, E> implements IView<S, E> {
  constructor(readonly evolve: (s: S, e: E) => S, readonly initialState: S) {
  }
}

Materialized View

A Materialized view is using/delegating a View to handle events of type E and to maintain a state of denormalized projection(s) as a result. Essentially, it represents the query/view side of the CQRS pattern. It belongs to the Application layer.

In order to handle the event, materialized view needs to fetch the current state via ViewStateRepository.fetchState function first, and then delegate the event to the view, which can produce new state as a result. New state is then stored via ViewStateRepository.save function.

Saga

Saga is a datatype that represents the central point of control, deciding what to execute next (A). It is responsible for mapping different events from many aggregates into action results AR that the Saga then can use to calculate the next actions A to be mapped to commands of other aggregates.

Saga is stateless, it does not maintain the state.

It has two generic parameters AR, A, representing the type of the values that Saga may contain or use. Saga can be specialized for any type of AR, A because these types do not affect its behavior. Saga behaves the same for AR=Int or AR=YourCustomType, for example.

Saga is a pure domain component.

• AR - Action Result
• A - Action

Notice that Saga implements an interface ISaga to communicate the contract.

export class Saga<AR, A> implements ISaga<AR, A>{
    constructor(readonly react: (ar: AR) => readonly A[]) {}
}

Saga Manager

Saga manager is a stateless process orchestrator. It belongs to the Application layer. It is reacting on Action Results of type AR and produces new actions A based on them.

Saga manager is using/delegating a Saga to react on Action Results of type AR and produce new actions A which are going to be published via ActionPublisher.publish function.

Event Modeling

Event Modeling is:

• a method of describing systems using an example of how information has changed within them over time.
• a scenario-based and UX-driven approach to defining requirements.

Structuring the data

TypeScript adopts a structural type system which determines type compatibility and equivalence based on the type structure or definition rather than the declarative relationship between types and interfaces, which contrasts with nominal type system.

In TypeScript, we can use Algebraic Data Types (ADTs) to model our application's domain entities and relationships in a functional way, clearly defining the set of possible values and states. TypeScript has two main types of ADTs: union types ("|" operator), intersection types ("&" operator), tuples and records

• union types is used to define a type that can take on one of several possible variants - modeling a sum/OR type.
• intersection types, tuples and records are used to combine several types into one - modeling a product/AND type.

ADTs will help with

• representing the business domain in the code accurately
• enforcing correctness
• reducing the likelihood of bugs.

In FModel, we extensively use ADTs to model the data.

C / Command / Intent to change the state of the system

// Be precise and explicit about the types
export type SchemaVersion = number;
export type RestaurantId = string;
export type OrderId = string;
export type MenuItemId = string;
export type RestaurantName = string;
export type RestaurantMenuId = string;
export type MenuItemName = string;
export type MenuItemPrice = string;

export type Command = RestaurantCommand | OrderCommand;

export type RestaurantCommand = CreateRestaurantCommand | ChangeRestaurantMenuCommand | PlaceOrderCommand;

export type CreateRestaurantCommand = {
  readonly decider: "Restaurant";
  readonly kind: "CreateRestaurantCommand";
  readonly id: RestaurantId;
  readonly name: RestaurantName;
  readonly menu: RestaurantMenu;
};

export type ChangeRestaurantMenuCommand = {
  readonly decider: "Restaurant";
  readonly kind: "ChangeRestaurantMenuCommand";
  readonly id: RestaurantId;
  readonly menu: RestaurantMenu;
};

export type PlaceOrderCommand = {
  readonly decider: "Restaurant";
  readonly kind: "PlaceOrderCommand";
  readonly id: RestaurantId;
  readonly orderId: OrderId;
  readonly menuItems: MenuItem[];
};

export type OrderCommand = CreateOrderCommand | MarkOrderAsPreparedCommand;

export type CreateOrderCommand = {
  decider: "Order";
  kind: "CreateOrderCommand";
  id: OrderId;
  restaurantId: RestaurantId;
  menuItems: MenuItem[];
};

export type MarkOrderAsPreparedCommand = {
  decider: "Order";
  kind: "MarkOrderAsPreparedCommand";
  id: OrderId;
};

E / Event / Fact

export type Event = RestaurantEvent | OrderEvent;

export type RestaurantEvent =
  | RestaurantCreatedEvent
  | RestaurantNotCreatedEvent
  | RestaurantMenuChangedEvent
  | RestaurantMenuNotChangedEvent
  | RestaurantOrderPlacedEvent
  | RestaurantOrderNotPlacedEvent;

export type RestaurantCreatedEvent = {
  readonly version: SchemaVersion;
  readonly decider: "Restaurant";
  readonly kind: "RestaurantCreatedEvent";
  readonly id: RestaurantId;
  readonly name: RestaurantName;
  readonly menu: RestaurantMenu;
  readonly final: boolean;
};

export type RestaurantNotCreatedEvent = {
  readonly version: SchemaVersion;
  readonly decider: "Restaurant";
  readonly kind: "RestaurantNotCreatedEvent";
  readonly id: RestaurantId;
  readonly name: RestaurantName;
  readonly menu: RestaurantMenu;
  readonly reason: Reason;
  readonly final: boolean;
};

export type RestaurantMenuChangedEvent = {
  readonly version: SchemaVersion;
  readonly decider: "Restaurant";
  readonly kind: "RestaurantMenuChangedEvent";
  readonly id: RestaurantId;
  readonly menu: RestaurantMenu;
  readonly final: boolean;
};

export type RestaurantMenuNotChangedEvent = {
  readonly version: SchemaVersion;
  readonly decider: "Restaurant";
  readonly kind: "RestaurantMenuNotChangedEvent";
  readonly id: RestaurantId;
  readonly menu: RestaurantMenu;
  readonly reason: Reason;
  readonly final: boolean;
};

export type RestaurantOrderPlacedEvent = {
  readonly version: SchemaVersion;
  readonly decider: "Restaurant";
  readonly kind: "RestaurantOrderPlacedEvent";
  readonly id: RestaurantId;
  readonly orderId: OrderId;
  readonly menuItems: MenuItem[];
  readonly final: boolean;
};

export type RestaurantOrderNotPlacedEvent = {
  readonly version: SchemaVersion;
  readonly decider: "Restaurant";
  readonly kind: "RestaurantOrderNotPlacedEvent";
  readonly id: RestaurantId;
  readonly orderId: OrderId;
  readonly menuItems: MenuItem[];
  readonly reason: Reason;
  readonly final: boolean;
};

export type OrderEvent =
  | OrderCreatedEvent
  | OrderNotCreatedEvent
  | OrderPreparedEvent
  | OrderNotPreparedEvent;

export type OrderCreatedEvent = {
  version: SchemaVersion;
  decider: "Order";
  kind: "OrderCreatedEvent";
  id: OrderId;
  restaurantId: RestaurantId;
  menuItems: MenuItem[];
  final: boolean;
};

export type OrderNotCreatedEvent = {
  version: SchemaVersion;
  decider: "Order";
  kind: "OrderNotCreatedEvent";
  id: OrderId;
  restaurantId: RestaurantId;
  reason: Reason;
  menuItems: MenuItem[];
  final: boolean;
};

export type OrderPreparedEvent = {
  version: SchemaVersion;
  decider: "Order";
  kind: "OrderPreparedEvent";
  id: OrderId;
  final: boolean;
};

export type OrderNotPreparedEvent = {
  version: SchemaVersion;
  decider: "Order";
  kind: "OrderNotPreparedEvent";
  id: OrderId;
  reason: Reason;
  final: boolean;
};

S / State / Current state of the system/aggregate/entity

/**
 * Restaurant state / a data type that represents the current state of the Restaurant
 */
export type Restaurant = {
  readonly restaurantId: RestaurantId;
  readonly name: RestaurantName;
  readonly menu: RestaurantMenu;
};

/**
 * Order state / a data type that represents the current state of the Order
 */
export type Order = {
  readonly orderId: OrderId;
  readonly restaurantId: RestaurantId;
  readonly menuItems: MenuItem[];
  readonly status: OrderStatus;
};

Modeling the Behaviour

• algebraic data types form the structure of our entities (commands, state, and events).
• functions/lambda offers the algebra of manipulating the entities in a compositional manner, effectively modeling the behavior.

This leads to modularity in design and a clear separation of the entity’s structure and functions/behaviour of the entity.

Fmodel library offers generic and abstract components to specialize in for your specific case/expected behavior:

Decider - data type that represents the main decision-making algorithm.

/**
 * Restaurant - `pure` command handler / a decision-making component
 * ___
 * A pure command handling algorithm, responsible for evolving the state of the restaurant.
 * It does not produce any side effects, such as I/O, logging, etc.
 * It utilizes type narrowing to make sure that the command is handled exhaustively.
 * https://www.typescriptlang.org/docs/handbook/2/narrowing.html#exhaustiveness-checking
 * ___
 * @param c - command type that is being handled - `RestaurantCommand`
 * @param s - state type that is being evolved - `Restaurant | null`
 * @param e - event type that is being produced / a fact / an outcome of the decision - `RestaurantEvent`
 */
export const restaurantDecider: Decider<
  RestaurantCommand,
  Restaurant | null,
  RestaurantEvent
> = new Decider<RestaurantCommand, Restaurant | null, RestaurantEvent>(
  (command, currentState) => {
    switch (command.kind) {
      case "CreateRestaurantCommand":
        return currentState == null
          ? [
            {
              version: 1,
              decider: "Restaurant",
              kind: "RestaurantCreatedEvent",
              id: command.id,
              name: command.name,
              menu: command.menu,
              final: false,
            },
          ]
          : [
            {
              version: 1,
              decider: "Restaurant",
              kind: "RestaurantNotCreatedEvent",
              id: command.id,
              name: command.name,
              menu: command.menu,
              reason: "Restaurant already exist!",
              final: false,
            },
          ];
      case "ChangeRestaurantMenuCommand":
        return currentState !== null
          ? [
            {
              version: 1,
              decider: "Restaurant",
              kind: "RestaurantMenuChangedEvent",
              id: currentState.restaurantId,
              menu: command.menu,
              final: false,
            },
          ]
          : [
            {
              version: 1,
              decider: "Restaurant",
              kind: "RestaurantMenuNotChangedEvent",
              id: command.id,
              menu: command.menu,
              reason: "Restaurant does not exist!",
              final: false,
            },
          ];
      case "PlaceOrderCommand":
        return currentState !== null
          ? [
            {
              version: 1,
              decider: "Restaurant",
              kind: "RestaurantOrderPlacedEvent",
              id: command.id,
              orderId: command.orderId,
              menuItems: command.menuItems,
              final: false,
            },
          ]
          : [
            {
              version: 1,
              decider: "Restaurant",
              kind: "RestaurantOrderNotPlacedEvent",
              id: command.id,
              orderId: command.orderId,
              menuItems: command.menuItems,
              reason: "Restaurant does not exist!",
              final: false,
            },
          ];
      default:
        // Exhaustive matching of the command type
        const _: never = command;
        return [];
    }
  },
  (currentState, event) => {
    switch (event.kind) {
      case "RestaurantCreatedEvent":
        return { restaurantId: event.id, name: event.name, menu: event.menu };
      case "RestaurantNotCreatedEvent":
        return currentState;
      case "RestaurantMenuChangedEvent":
        return currentState !== null
          ? {
            restaurantId: currentState.restaurantId,
            name: currentState.name,
            menu: event.menu,
          }
          : currentState;
      case "RestaurantMenuNotChangedEvent":
        return currentState;
      case "RestaurantOrderPlacedEvent":
        return currentState;
      case "RestaurantOrderNotPlacedEvent":
        return currentState;
      default:
        // Exhaustive matching of the event type
        const _: never = event;
        return currentState;
    }
  },
  null,
);

/**
 * Order  - `pure` command handler / a decision-making component
 * ___
 * A pure command handling algorithm, responsible for evolving the state of the order.
 * It does not produce any side effects, such as I/O, logging, etc.
 * It utilizes type narrowing to make sure that the command is handled exhaustively.
 * https://www.typescriptlang.org/docs/handbook/2/narrowing.html#exhaustiveness-checking
 * ___
 * @param c - command type that is being handled - `OrderCommand`
 * @param s - state type that is being evolved - `Order | null`
 * @param e - event type that is being produced / a fact / an outcome of the decision - `Order`Event`
 */
export const orderDecider: Decider<OrderCommand, Order | null, OrderEvent> =
  new Decider<OrderCommand, Order | null, OrderEvent>(
    (command, currentState) => {
      switch (command.kind) {
        case "CreateOrderCommand":
          return currentState == null
            ? [
              {
                version: 1,
                decider: "Order",
                kind: "OrderCreatedEvent",
                id: command.id,
                restaurantId: command.restaurantId,
                menuItems: command.menuItems,
                final: false,
              },
            ]
            : [
              {
                version: 1,
                decider: "Order",
                kind: "OrderNotCreatedEvent",
                id: command.id,
                restaurantId: command.restaurantId,
                menuItems: command.menuItems,
                final: false,
                reason: "Order already exist!",
              },
            ];
        case "MarkOrderAsPreparedCommand":
          return currentState !== null
            ? [
              {
                version: 1,
                decider: "Order",
                kind: "OrderPreparedEvent",
                id: currentState.orderId,
                final: false,
              },
            ]
            : [
              {
                version: 1,
                decider: "Order",
                kind: "OrderNotPreparedEvent",
                id: command.id,
                reason: "Order does not exist!",
                final: false,
              },
            ];
        default:
          // Exhaustive matching of the command type
          const _: never = command;
          return [];
      }
    },
    (currentState, event) => {
      switch (event.kind) {
        case "OrderCreatedEvent":
          return {
            orderId: event.id,
            restaurantId: event.restaurantId,
            menuItems: event.menuItems,
            status: "CREATED",
          };
        case "OrderNotCreatedEvent":
          return currentState;
        case "OrderPreparedEvent":
          return currentState !== null
            ? {
              orderId: currentState.orderId,
              restaurantId: currentState.restaurantId,
              menuItems: currentState.menuItems,
              status: "PREPARED",
            }
            : currentState;
        case "OrderNotPreparedEvent":
          return currentState;
        default:
          // Exhaustive matching of the event type
          const _: never = event;
          return currentState;
      }
    },
    null,
  );

The logic execution will be orchestrated by the outside components that use the domain components (decider, view) to do the computations. These components will be responsible for fetching and saving the data (repositories).

The arrows in the image (adapters->application->domain) show the direction of the dependency. Notice that all dependencies point inward and that Domain does not depend on anybody or anything.

Pushing these decisions from the core domain model is very valuable. Being able to postpone them is a sign of good architecture.

Event-sourcing aggregate

// We are adding new types on this layer: Metadatada and Version. Observe how these types are not leaking into the Domain layer (decider), and not influencing the core logic.
export type StreamVersion = { event_id: string };
export type CommandMetadata = { tenant: string };
export type EventMetadata = {
  tenant: string;
  command_id: string;
  event_id: string;
};

/**
 * An aggregate that handles the command and produces new events / A convenient type alias for the Fmodel's `EventSourcingAggregate`
 * 
 * This aggregate can handle all the commands of the system. Observe how two deciders (restaurant and order) are combined into one.
 */
export type ApplicationAggregate = EventSourcingAggregate<
  Command,
  Restaurant & Order,
  Event,
  StreamVersion,
  CommandMetadata,
  EventMetadata
>;

// Parse the command from the request
const command: Command = JSON.parse(await req.json());
console.log("Handling command: ", command);
// We can combine deciders to create a new decider that can handle both restaurant and order commands
const decider = restaurantDecider.combine(orderDecider);
// Create a repository for the events of all types
const eventRepository = new EventRepository(supabaseClient);
// Create an aggregate to handle the commands of all types!
const aggregate: ApplicationAggregate = new EventSourcingAggregate(
  decider,
  eventRepository,
);

// Handle the command
const result = await aggregate.handle(command);

Install as a dependency of your project

npm i @fraktalio/fmodel-ts

Available on https://www.npmjs.com/package/@fraktalio/fmodel-ts

Examples

• Why don't you start by browsing tests?
• Event sourcing and event streaming with Axon - Gift Card domain
• Event sourcing on Deno runtime, with Deno KV as an event store - Restaurant Management domain

FModel in other languages

• FModel in Kotlin
• FModel in Rust

Resources

• The Blog - Domain modeling
• Event Modeling - What is it?

Credits

Special credits to Jérémie Chassaing for sharing his research and Adam Dymitruk for hosting the meetup.

Created with :heart: by Fraktalio