Skip to content

Latest commit

 

History

History
150 lines (136 loc) · 3.64 KB

README.md

File metadata and controls

150 lines (136 loc) · 3.64 KB
Raw

“The server”

This is a little mock REST API that works with about a 1000 hard-coded names and responds to network requests with responses that bear a payload which is semantically homomorphic to the one specified in the gist. I however cleared away all the clutter, so the server responds with no more data than what is absolutely necessary for the client application to function.

Disclaimer: I did not concentrate a lot on this part of the project, hence the not too neatly organized code, no separation of concerns, lack of tests etc.

API

There are no endpoints. There are no routes either.
The server catches every GET request coming its way and responds with:

  • 200 and a list of employees, together with some metadata and pageing info.
  • 400 or 418 and an error message.

Parameters

In order to query “the database” and get one page of results, you must supply the following GET arguments:

  • q – the name you are looking for, use john+doe format when providing multiple word queries
  • page – the page number for pagination of suggestions
  • per_page – the number of items per page for pagination of suggestions

So a request URL could for example look like:

http://localhost:3001/?q=jo+a&page=1&per_page=5

And it would let’s say generate the following output:

{
  "data": [
    {
      "id": "73",
      "attributes": {
        "name": "Anahi Joseph",
        "avatar": null
      }
    },
    {
      "id": "102",
      "attributes": {
        "name": "Ariel Johnston",
        "avatar": null
      }
    },
    {
      "id": "362",
      "attributes": {
        "name": "Francisco Joseph",
        "avatar": null
      }
    },
    {
      "id": "420",
      "attributes": {
        "name": "Irene Johnson",
        "avatar": null
      }
    },
    {
      "id": "506",
      "attributes": {
        "name": "Joel Castro",
        "avatar": null
      }
    }
  ],
  "included": [
    {
      "id": "73",
      "attributes": {
        "email": "anahi.joseph@example.com"
      }
    },
    {
      "id": "102",
      "attributes": {
        "email": "ariel.johnston@example.com"
      }
    },
    {
      "id": "362",
      "attributes": {
        "email": "francisco.joseph@example.com"
      }
    },
    {
      "id": "420",
      "attributes": {
        "email": "irene.johnson@example.com"
      }
    },
    {
      "id": "506",
      "attributes": {
        "email": "joel.castro@example.com"
      }
    }
  ],
  "meta": {
    "page": {
      "total": 30
    }
  },
  "links": {
    "next": "http://localhost:3001/?per_page=5&page=2&q=jo"
  }
}

In case there are no results, the response payload would look like this:

{
  "data": [],
  "included": [],
  "meta": {
    "page": {
      "total": 0
    }
  },
  "links": {
    "next": "http://localhost:3001/?per_page=5&page=2&q=joasd"
  }
}

You will get a 400 if you fail to provide the required GET parameters:

{
  "error": "You must provide the following GET parameters: page, page_number, q"
}

You can also squeeze out a nasty 418 error of the server, if you include err in the GET parameter q:

{
  "error": "I’m almost, but not quite, entirely unlike a teapot"
}

Run the server

Start the server:

node .

This will make the server listen closely on port :3001.

Improvement ideas

  • Maybe this would deserve an actual npm project, now it’s just 3 JavaScript files lying around.
  • Some tests wouldn’t hurt.
  • An optional fuzzy search functionality would be killer, so that e,g, Barack Obama would come up to barobm.