Prerequisite

Node 18
cURL
Docker 20.10.14 or higher (to run locally)

Introduction

If you have gone through the Explore Kalix Virtaully journey, you are all set to :

However, if you would like to better understand the Kalix Developer Experience and how to build applications on Kalix, here is more details on the same-use.

Kalix Trial - eCommerce - TypeScript

Designing Kalix Services

Use case

Use case
eCommerce use case is a simple shopping cart example consisting of product stock and shopping cart. Product stock models stock (simple quantity) of products that are being sold and Shopping Cart models list of products customer intends to buy.
In this exercise the focus is on implementing Product Stock functionalities

Product Stock

Data model:

  • quantity

Note: Product stock data model does not hold the productId because the HTTP/REST API is designed around productId itself. Operations:

  • CREATE/READ/UPDATE/DELETE product stock

Kalix components

Kalix components are building blocks used to abstract functionalities.
In this use case we are going to use a Kalix component called Entity.
Entity
Entity:

  • component for modeling of data and data business logic operations
  • removes complexities around data:
    • caching
    • concurrency
    • distributed locking
  • simplifies data modeling, business logic implementation with out-of-the box scalability, resilience Kalix provides other components that are not used in this use case and more details can be found in Kalix documentation:
    Other Kalix components

Design Product Stock Kalix service

Product Stock Service

  • implements Product Stock functionalities
  • Kalix component used: Entity

Product Stock Kalix Service

ProductStock Entity

  • models one product stock instance and business logic operations over that one instance

Data model

  • quantity

API

  • HTTP/RES
  • Endpoints:
    • Create POST /product-stock/{productId}/create

      Request (JSON):

      • quantity (int)

      Response (JSON): {}

    • Read GET /product-stock/{productId}/get Request (JSON): No body

      Response (JSON):

      • quantity (int)
    • Update PUT /product-stock/{productId}/update

      Request (JSON):

      • quantity (int)

      Response (JSON): {}

    • Delete DELETE /product-stock/{productId}/delete

      Request (JSON): No body

      Response (JSON): {}

Kickstart Kalix development project

Execute in command line:

npx @kalix-io/create-kalix-entity@1.1.0 kalix-trial-shoppingcart-ts --typescript

And then:

 cd kalix-trial-shoppingcart-ts
 npm install

Generated project structure:

Node project structure

  • package.json with all pre-configured dependencies required development, testing and packaging of Kalix service code
  • node_modules all modules required for development, testing and packaging of Kalix service code
  • proto directory with Kalix protobuf examples

Note: delete Kalix protobuf example files (counter_api.proto, domain/counter_domain.proto).

Define data structure

  1. Create directory productstock in proto
  2. Create protobuf file productstock_domain.proto in src/main/protobuf/productstock
  3. Add ProductStock protobuf definition that represents ProductStock data structure that is persisted:
syntax = "proto3";
package com.example.shoppingcart.productstock;
message ProductStock{
    int32 quantity = 1;
}

Define API - Product Stock Entity API

  1. Create protobuf file productstock_api.proto in src/main/protobuf/productstock
  2. Add protobuf definitions represending Product Stock API requests:
syntax = "proto3";

import "google/protobuf/empty.proto";
import "kalix/annotations.proto";
import "google/api/annotations.proto";

import "productstock/productstock_domain.proto";

package com.example.shoppingcart.productstock;

message CreateProductStockRequest {
  string product_id = 1 [(kalix.field).entity_key = true];
  int32 quantity = 2;
}

message UpdateProductStockRequest {
  string product_id = 1 [(kalix.field).entity_key = true];
  int32 quantity = 2;
}

message DeleteProductStockRequest {
  string product_id = 1 [(kalix.field).entity_key = true];
}

message GetProductStock {
  string product_id = 1 [(kalix.field).entity_key = true];
}
  1. Add protobuf definition for a Product Stock GRPC service with corresponding rpc methods representing each endpoint
service ProductStockService {
  rpc Create(CreateProductStockRequest) returns (google.protobuf.Empty);
  rpc Get(GetProductStock) returns (ProductStock);
  rpc Update(UpdateProductStockRequest) returns (google.protobuf.Empty);
  rpc Delete(DeleteProductStockRequest) returns (google.protobuf.Empty);
  }
  1. Exposing HTTP/REST endpoints from GRPC service using GRPC Transcoding HTTP/JSON to GRPC
service ProductStockService {
  rpc Create(CreateProductStockRequest) returns (google.protobuf.Empty){
    option (google.api.http) = {
      post: "/product-stock/{product_id}/create"
      body: "*"
    };
  }
  rpc Get(GetProductStock) returns (ProductStock){
    option (google.api.http) = {
      get: "/product-stock/{product_id}/get"
    };
  }
  rpc Update(UpdateProductStockRequest) returns (google.protobuf.Empty){
    option (google.api.http) = {
      put: "/product-stock/{product_id}/update"
      body: "*"
    };
  }
  rpc Delete(DeleteProductStockRequest) returns (google.protobuf.Empty){
    option (google.api.http) = {
      delete: "/product-stock/{product_id}/delete"
    };
  }
}

Add default ACL policy

Default ACL policy needs to be added.

  1. Create file kalix_policy.proto in proto
  2. Add this content to the file
syntax = "proto3";
import "kalix/annotations.proto";
option (kalix.file).acl = {
    allow: {}
};

Note: more info about ACL (Access Control List)

Implementing business logic

Code generation

  1. Code generation from protobuf definitions Kalix node module uses defined protobuf definitions to build required codebase where business logic can be developed in.
    Kalix node module needs to be configured using custom kalix protobuf definition:
service ProductStockService {
  option (kalix.codegen) = {
    value_entity: {
      name: "com.example.shoppingcart.productstock.ProductStockEntity"
      entity_type: "product-stock"
      state: "com.example.shoppingcart.productstock.ProductStock"
    }
  };
  • value_entity configures key-value as a durable storage model
  • name configures the name of the Entity class that is being generated
  • entity_type configures the entity type reference name that can be used to reference Entity in other Kalix components
  • state configures the reference to ProductStock protobuf definition defined in productstock_domain.proto
  1. Generate code using Kalix node module:
npm run build

Generated code:

  • lib/generated folder containing Kalix generated help resources
  • src/index.ts file controlled by Kalix module that bootstraps
  • src/productstockentity.ts skeleton file where business logic needs to be developed in
  • test/productstockentity.test.ts skeleton mocha test file where unit tests need to be developed in

ProductStockEntity business logic

Generated file src/productstockentity.ts provides skeleton methods that need to be implemented for each endpoint.
We need to add GrpcStatus import to be able to use it in our business logic for returning NOT_FOUND error status:

import {GrpcStatus} from "@kalix-io/kalix-javascript-sdk";

create endpoint

Business logic for create is to persist product stock data if not yet exists. In other cases returns an ERROR.

Create(command, state, ctx) {
if(state == ProductStock.create({})) {
  ctx.updateState(ProductStock.create({quantiy: command.quantity}));
  return Reply.message({});
}
return Reply.failure("Already created");
}

get endpoint

Business logic for get is to get product stock data if it exists and if not return not found error.

Get(command, state, ctx) {
if(state == ProductStock.create({}))
  return Reply.failure("Not found", GrpcStatus.NotFound);
return Reply.message(state);
}

update endpoint

Business logic for update is to update product stock data if product was already created. If product is not found, return NOT FOUND error.

Update(command, state, ctx) {
if(state == ProductStock.create({})) {
  return Reply.failure("Not found", GrpcStatus.NotFound);
}
state.quantity = command.quantity;
ctx.updateState(state);
return Reply.message({});
}

delete endpoint

Business logic for delete is delete data if product stock exists and return NOT FOUND error if not. Here the soft delete is done by updating the state to ProductStock.create({}).

Delete(command, state, ctx) {
if(state == ProductStock.create({})) {
  return Reply.failure("Not found", GrpcStatus.NotFound);
}
state = ProductStock.create({});
ctx.updateState(state);
return Reply.message({});
}

Test

Kalix comes with very rich test kit for unit and integration testing of Kalix code.

Test kit provides help (custom assertions, mocks,...) with:

  • unit testing of individual Kalix components (e.g Entity) in isolation

  • integration testing in Kalix Platform simulated environment in isolation

  • Allows easy test automation with very high test coverage

Unit test

Kalix SDK uses MochaJS and for building unit tests and ChaiJS for assertion.

  1. Edit test/productstockentity.test.ts
  2. Delete all tests instead of handle command Create
  3. Add following test code:
describe("Create", () => {
    it("should...", async () => {
      const entity = new MockValueEntity(productstockentity, entityId);
      const productId = "prod1";
      let quantity = 10;
      const createResult = await entity.handleCommand("Create",
          {
              productId: productId,
              quantity: quantity
          }
      );
      expect(entity.error).to.be.undefined;
      expect(createResult).to.deep.equals({});
      // @ts-ignore
      expect(entity.state.quantity).to.deep.equal(quantity);
    });
});

MockValueEntity is used to mock the behaviour of an Entity.

  1. Run the unit test:
npm run test

Integration test

Kalix test kit for integration testing runs code using test containers to simulate Kalix Platform runtime environment. Integration test uses generated grpc client.

  1. Create file test/productstock.integration-test.ts
  2. Add following test code:
import { IntegrationTestkit } from "@kalix-io/testkit";

const testkit = new IntegrationTestkit();

import {expect} from "chai"

import productstockentity from "../src/productstockentity";

testkit.addComponent(productstockentity);

function client() {
    return testkit.clients.ProductStockService;
}

describe("ProductStockService",  function () {
    this.timeout(60000);

    before(done => testkit.start(done));
    after(done => testkit.shutdown(done));

    it("Create", async () => {
        const productId = "prod1";
        let quantity = 1;
        await client().createAsync({productId: productId, quantity: quantity});

        const get = await client().getAsync({productId: productId});
        expect(get.quantity).to.deep.equal(quantity);
    });
})

Run the integration test:

npm run test

Run locally in prod-like environment

  1. Run the Kalix Proxy container:
docker-compose up
  1. Run Kalix service locally:
npm run start

This command runs the Kalix service locally and exposes it on localhost:9000.

Test

Testing using CURL:

  1. Create product:
curl -XPOST -d '{ 
  "quantity": 10
}' http://localhost:9000/product-stock/apple/create -H "Content-Type: application/json"

Result:

"{}"
  1. Get product:
curl -XGET http://localhost:9000/product-stock/apple/get

Result:

{"quantity":10}
  1. Update product:
curl -XPUT -d '{
"quantity": 20
}' http://localhost:9000/product-stock/apple/update -H "Content-Type: application/json"

Result:

"{}"
  1. Delete product:
curl -XDELETE http://localhost:9000/product-stock/apple/delete

Result:

"{}"

Deploy and run on Kalix Platform on Cloud Provider of your choice

  1. Install Kalix CLI https://docs.kalix.io/setting-up/index.html#_1_install_the_kalix_cli

  2. Kalix CLI

  3. Register (FREE)

    kalix auth signup
    

    Note: Following command will open a browser where registration information can be filled in

  4. Login

    kalix auth login
    

    Note: Following command will open a browser where authentication approval needs to be provided

  5. Create a project

    kalix projects new ecommerce --region=gcp-us-east1
    

    Note: gcp-is-east1 is currently the only available region for deploying trial projects. For non-trial projects you can select Cloud Provider and regions of your choice

  6. Authenticate local docker for pushing docker image to Kalix Container Registry (KCR)

    kalix auth container-registry configure
    

    Note: The command will output Kalix Container Registry (KCR) path that will be used to configure dockerImage in pom.xml

  7. Extract Kalix user username

    kalix auth current-login
    

Note: The command will output Kalix user details and column USERNAME will be used to configure dockerImage in pom.xml
3. Configure dockerImage path in package.json Replace my-docker-repo in dockerImage in package.json with:
Kalix Container Registry (KCR) path + / + USERNAME + /ecommerce
Example where Kalix Container Registry (KCR) path is kcr.us-east-1.kalix.io and USERNAME is myuser:

"dockerImage": "kcr.us-east-1.kalix.io/myuser/ecommerce/kalix-trial-shoppingcart-ts"
  1. Deploy service in Kalix project:
npm run deploy

This command will:

  • execute tests
  • package into a docker image
  • push the docker image to Kalix docker registry
  • trigger service deployment by invoking Kalix CLI
  1. Check deployment:
kalix service list

Result:

kalix service list                                                                         
NAME                                         AGE    REPLICAS   STATUS        IMAGE TAG                     
kalix-trial-shoppingcart-ts                  50s    0          Ready         0.0.1                

Note: When deploying service for the first time it can take up to 1 minute for internal provisioning

Proxy connection to Kalix service via Kalix CLI

  1. Proxy connection to Kalix service via Kalix CLI
kalix service proxy kalix-trial-shoppingcart-ts

Proxy Kalix CLI command will expose service proxy connection on localhost:8080.

Test

Testing using CURL:

  1. Create product:
curl -XPOST -d '{ 
  "quantity": 10
}' http://localhost:8080/product-stock/apple/create -H "Content-Type: application/json"

Result:

"{}"
  1. Get product:
curl -XGET http://localhost:8080/product-stock/apple/get

Result:

{"quantity":10}
  1. Update product:
curl -XPUT -d '{
"quantity": 20
}' http://localhost:8080/product-stock/apple/update -H "Content-Type: application/json"

Result:

"{}"
  1. Delete product:
curl -XDELETE http://localhost:8080/product-stock/apple/delete

Result:

"{}"