Ncdc NPM | npm.io

NCDC

NCDC (or node cdc) is a tool that takes a consumer contract (written in yaml) and tests the specified endpoints against a producer. You can also run a single command to mock those endpoints.

The point of using this tool is to ensure that each endpoint you use in development is called in the same way and responds in the same way as the endpoints in your live environment. There's not much point in testing against mock APIs that don't strongly resemble your live APIs.

This tool can easily integrate with typescript to prevent you from having to write JSON schemas for your already defined Typescript interfaces, types and enums. With this, you're guaranteed that your code, mock API endpoints and real endpoints are all consistent with eachother.

Getting started

Installation

npm install ncdc --save-dev or yarn add ncdc -D

Example config.yml

The below config.yml file contains 2 configurations

- name: Book not found
  serveOnly: true
  request:
    method: GET
    endpoints: /api/books/a-bad-id
  response:
    code: 404

- name: Book
  request:
    method: GET
    endpoints:
      - /api/books/123
      - /api/books/456
    serveEndpoint: /api/books/*
  response:
    code: 200
    headers:
      content-type: application/json
    type: Book
    serveBody:
      {
        ISBN: "9780141187761",
        ISBN_13: 978-0141187761,
        author: George Orwell,
        title: 1984 Nineteen Eighty-Four,
        inventoryId: item-87623
      }

config.yml references the type Book which is exported from mySourceFile.ts:

export interface Book {
  ISBN: string
  ISBN_13: string
  author: string
  title: string
  // You can set extra json schema properties via JSDoc comments
  /** @pattern ^item- */
  inventoryId: string
}

====> You can find more information about writing your config file HERE, as well how all of the options behave <====

Serving mocks using the config file

npx ncdc serve ./config.yml 3000

This will serve each configuration on http://localhost:3000.

When GET requests are sent to /api/books/a-bad-id, ncdc will respond with a 404 status code.

When GET requests are sent to /api/books/123 or /api/books/456, ncdc will respond with the following:

Status code as 200
content-type header as application/json
a JSON body with the value of serveBody

You can find more information about serve mode here

Testing real endpoints against the config

npx ncdc test ./config.yml https://example.com

This will run tests for each configuration except those with serveOnly set to true. For this reason, the Book not found configuration will not be tested against the real API.

For the Book configuration, ncdc will make GET requests to https://example.com/api/books/123 and https://example.com/api/books/456 and verify that both responses meet these requirements:

It responds with a 200 status code
It responds with the content-type header set to application/json
The response body is JSON containing at least:
- property ISBN with a string value
- property ISBN_13 with a string value
- property author with a string value
- property title with a string value
- property inventoryId with a string value starting with item-

You can find more information about test mode here

Generating JSON schemas

npx ncdc generate ./config.yml --output ./schemas

This will write JSON schemas to disk for each of the types you reference in your config.yml, such as Book in the above example. This can be useful for caching purposes.

You can find more information about generating schemas and why you might want to do that here

Typescript Compatability

Typescript 3 - Everything below ncdc version 1.0.0
Typescript 4 - Everything above ncdc version 1.0.0

CLI usage

$ npx ncdc --help
ncdc <command>

Commands:
  ncdc generate <configPath>        Generates a json schema for each type
                                    specified in the config file
  ncdc serve <configPath> [port]    Serves configured endpoints
  ncdc test <configPath> <baseURL>  Tests configured endpoints

Options:
  --version  Show version number                                       [boolean]
  --help     Show help                                                 [boolean]

Examples:
  ncdc generate ./config.yml                Generates json schemas for any type
                                            specified in config.yml.
  ncdc serve ./config.yml 4000              Serves the mock API endpoints
                                            defined in config.yml on port 4000.
  ncdc test ./config.yml                    Tests that the responses for the API
  https://example.com                       endpoints defined in config.yml
                                            match the configured parameters.

You can use the -f or --force flag to skip typechecking your source files. This speeds up the generate command, however, you should still use another tool e.g tsc to typecheck your project. If you try to use broken types with ncdc, ncdc will probably behave unexpectedly.

Serve

Serves mock API endpoints defined in config.yml

CLI Usage

Run npx ncdc serve --help to get contextual usage information and examples for this command.

How it works

When ncdc receives a request, it will go through each configuration in your config.yml (from top to bottom) and serve the first one that meets all of your request constraints. If the request does not match any of your configurations, ncdc will response with a 404.

Local development recipes

How to run your fake APIs locally:

package.json

  "scripts": {
    // dev script before adding ncdc
    "dev": "nodemon --watch 'src/**/*.ts' --exec 'ts-node' src/index.ts",
    // scripts after adding ncdc
    "devServer": "nodemon --watch 'src/**/*.ts' --exec 'ts-node' src/index.ts",
    "bookService": "ncdc serve ./book-service.yml 4000",
    "filmService": "ncdc serve ./film-service.yml 4001",
    // concurrently is optional, you could just run each npm script in separate shells
    "dev": "concurrently --kill-others-on-fail 'npm:bookService' 'npm:filmService' 'npm:devServer'",
  }

Test

Tests that responses from a real API match the configured parameters in config.yml

CLI Usage

Run npx ncdc test --help to get contextual usage information and examples for this command.

Generate

Generates json schemas for any type specified in config.yml

CLI Usage

Run npx ncdc generate --help to get contextual usage information and examples for this command.

How it works

This will create JSON schemas for your types and save them to disk. They can later be used with the serve and test commands, rather than re-generating them on the fly.

The reasons why you might want to do this are described in the Docker use case

Response content negotiation

alt text

Docker use case

A potential timing issue

Behind the scenes, ncdc makes use of the typescript compiler to create json schemas for each of the types referenced in your config.yml. Those schemas are then used for validation during serve mode and test mode.

Due to the use of the typescript compiler, the serve and test commands can often take a few seconds to start. The exact amount of time it takes depends on the size and configuration of your typescript project.

So, when can this become an issue? Here's a real use case. Let's say I have 2 docker-compose services:

services:
  # serves some API endpoints using ncdc
  fake-book-service:
    build: .
    command: npx ncdc serve path-to-config.yml 4000
  
  # runs some integration tests that need a book service to be available
  integration-tests:
    build: .
    command: npx jest
    depends_on:
      - fake-book-service
    environment:
      - BOOK_SERVICE_URL=http://fake-book-service:4000

The docker-compose depends_on property is only responsible for starting a docker service. It doesn't do any health checks that guarantee fake-book-service is up and serving requests. That means it's possbile for the integration tests to run before the API is ready, which would result in some (if not all) of the integration tests failing. That's where caching and the generate command become useful.

How to fix it

Solution 1

Adding the below line to the Dockerfile will generate the json scheams during the docker build.

RUN npx ncdc generate path-to-config.yml --output ./json-schema

Then the command for the fake-book-service can be changed to the following, which will skip schema generation and instead use the schemas that were produced during the docker build.

fake-book-service:
  build: .
  command: npx ncdc serve path-to-config.yml --schemaPath ./json-schema

This should speed up the start of the fake-book-service enough that the fake endpoints will be ready to use in the integration tests.

Solution 2

You can also use a wait-for script as an extra measure, which can be used in addition to Solution 1. This would be used like so:

integration-tests:
    build: .
    command: ./wait-for.sh fake-book-service:4000 -- npx jest
    depends_on:
      - fake-book-service
    environment:
      - BOOK_SERVICE_URL=http://fake-book-service:4000
    volumes:
      # the mapping of this volume depends on your own file structure
      - ./wait-for.sh:/usr/src/wait-for.sh