@elasticpath/component-test-framework v1.1.7

Elasticpath Component Test Framework

Installation

With Yarn

yarn add --dev @elasticpath/component-test-framework

With NPM

npm install --save-dev @elasticpath/component-test-framework

Usage

The framework exposes a few classes:

Api

This is the class responsible for communicating with your API. You can use the methods directly, however, it's usually a better idea to have it as a binding.

Example

import { binding, then } from 'cucumber-tsflow'
import { Api } from 'component-test-framework'

@binding([ Api ])
export default class MyStepDefinitions {
    constructor(protected api: Api) {
    
    }
    
    @then(...)
    public async myStep(): void {
      await this.api.get('/products')
    }
    
    ...

Payloads

Because everything that comes from the Gherkin language is a string, it's impossible to specify literals such as null. The framework has a Payload class to handle such cases. Whenever you want to send a payload to the API, you can use

const payload = new Payload(table.hashes()[0])

Type conversions

The framework is clever when it comes to type conversions. Consider the following Gherkin table

| null | 123 | foo | "" | [{"foo": "bar"}] |

When sent to the framework, they are all strings

[ "null", "123", "foo", "\"\"", "[{\"foo\": \"bar\"}]"]

The new Payload method does some conversion for us. Here's a mapping of the conversion:

{
    "null": null,
    "123": "123",
    "foo": "foo",
    "\"\"": "",
    "[{"foo": "bar"}]": [{"foo": "bar"}],
}

You may have noticed the number remains a string. The framework could convert that, however there are certain cases where you need to keep it as a string (a product upc_ean can be specified like a number but is a string). Therefore, the framework introduces a special symbol to force an integer. Given the following gherkin table, the framework will parse the first value as an integer and leave the second as a string

| <i>123 | 456 |

ResourceStore

When working with the API in your component tests, it's often benificial to reference the last created resource

Example

Given An admin creates a product
When They get the created product

For this, we can use the ResourceStore. When ever you use the API module, it stores the resource that was created, or fetched in the resource store. Knowing that, we can create our step definition to easily fetch the last created resource

Example

import { ResourceStore } from 'component-test-framework'
...
const resource: Resource =  ResourceStore.LastCreated('file')

We can also get all of the resources under one type

Example

import { ResourceStore } from 'component-test-framework'
...
const resource =  ResourceStore.FetchByType('file')

And find a resource

Example

import { ResourceStore } from 'component-test-framework'
...
const resource =  ResourceStore.Find('file', 'id', 'abc-123')

Each method on the ResourceStore returns a Resource

Resource

This is a representation of a resource.

const resource = new Resource(data)

These are usually only created by the ResourceStore. However, there are a few public methods

asPayload()

Useful if you want to build an update payload for your PUT request. If no argument is supplied, then the resource is returned as a payload, which removes attributes not allowed in payloads (relationships) and nullifies values which are string "null"

If an argument is supplied, this function will merge the data from that with the current data

Example

const resource = ResourceStore.LastCreated('hierarchy')
const payload = resource.asPayload(table.hashes()[0])

Publishing

Publishing is handled via the GitLab pipeline. On an MR branch, the pipeline checks to see if the version in package.json is already a published version, and fails if so, letting you know that the version needs to be updated.

Actual publishing happens on a merge to main.

In the event that a rollback is needed, the rollback MR will fail because the version in package.json will exist on NPM. To resolve this, you will need to bump the version in package.json.