js-fetch-api v1.3.5

fetch api

This package is a fetch utility for React

Install

npm i fetch-api

Why using this instead of fetch ?

• Failure errors return as JSON with status, headers, response, stackTrace, message, etc.
• Uses redux thunk to dispatch events with request/success/failure action types
• Error action dispatch:
  • 3xx: API/CALL/REDIRECT
  • over 4xx: API/CALL/REJECTED
  • 401: SERVER/UNAUTHORIZED/ERROR
  • over 500: SERVER/INTERNAL/ERROR
• Uses mocked calls in dev environment
• Simulates expected error rate, as in 50% of this calls return a 404, etc.
• Simulate answer delay
• Define expected payloads for mocked calls.
• Adds NAV_CSRF_PROTECTION cookie, ts param to overcome browsers that cache AJAX results

Can you give me an example

In your actions folder, you can have something like this:

import { ActionWithPayload, call, ThunkResult } from 'js-fetch-api'

export const getPersonInfo: ActionCreator<ThunkResult<ActionWithPayload<PersonPDL>>> = (
  id: string
): ThunkResult<ActionWithPayload<PersonPDL>> => {
  return call({
    url: '/api/person/' + id,
    expectedPayload: {
       personname: 'Demo Demo',
       username: 'demo'
    },
    expectedErrorRate: {
       200 : 0.8, 401: 0.1, 500: 0.1
    },
    maxDelayTime: 2000,
    responseType: 'json',
    cascadeFailureError: true,
    type: {
      request: 'PERSONINFO/REQUEST',
      success: 'PERSONINFO/SUCCESS',
      failure: 'PERSONINFO/FAILURE'
    }
  })
}

What does this do?

• In production environment, it will call the URL endpoint and return the result as JSON (responseType can also be 'pdf' or 'blob')
• will dispatch a request action, then a success or failure action
• If cascadeFailureError is false, in case of 500 errors, it will not dispatch failure action. IF cascade is true, it will dispatch failure action, in addition to SERVER/INTERNAL/ERROR
• expectedErrorRate: {200 : 0.8, 401: 0.1, 500: 0.1} means that, in dev environment (NOT in production), there is:
  • 80% chance for a 200 answer
  • 10% chance for a 401 answer
  • 10% chance for a 500 answer This is great to simulate backend errors / failures