0

My API has three endpoints: articles, websites, and users. Each article is associated with a website. A user can also share articles.

In my API, I have just created an endpoint at /website/:id/articles. This will query the database for articles associated with the given website. It then performs some manipulation on the data for each article based on who is talking to the API ("has the user shared this article?", for example).

I am now moving on to create a similar endpoint at /users/:id/shared-articles. The database query for this is slightly different, but the manipulation I want to perform on the articles data following the query is the same as before.

Here is some pseudo code for the former endpoint:

router.get('/websites/:id/articles', function (req, res) {
  articleService.find({ websiteId: req.params.id }, function (error, foundArticles) {

    async.waterfall([
      function (cb) {
        // Manipulate foundArticles…
        cb(null, manipulatedArticles)
      },
      function (articles, cb) {
        // Manipulate articles some more…
        cb(null, manipulatedArticles)
      },
    ], function (error, articles) {
      if (error) {
        return res.json(error, 400)
      }

      res.json(articles)
    })
  })
})

To create my new endpoint, /users/:id/shared-articles, I could abstract the manipulation tasks into a function that can be shared by both of my endpoints (the waterfall seen above), reducing code repetition.

router.get('/websites/:id/articles', function (req, res) {
  articleService.find({ websiteId: req.params.id }, function (error, foundArticles) {
    manipulateArticles(foundArticles, function (articles) {
      if (error) {
        return res.json(error, 400)
      }

      res.json(articles)
    })
  })
})

router.get('/users/:id/shared-articles', function (req, res) {
  shareActionService.find({ userId: req.params.id }, function (error, foundShareActions) {
    var sharedArticleIds = { _id: { $in: _.pluck(foundShareActions, 'sharedArticleId') } }

    articleService.find(sharedArticleIds, function (error, foundArticles) {
      manipulateArticles(foundArticles, function (articles) {
        if (error) {
          return res.json(error, 400)
        }

        res.json(articles)
      })
    })
  })
})

However, I figured that this sort of code re-use problem must be common when designing APIs in Node, and I would like to know if there is an obviously better solution that I am missing here.

One idea I had would be to have all article sub-resources (such as /users/:id/shared-articles or /websites/:id/links) talk to the /links API internally, which itself would deal with the manipulation I mention above. The problem then is that I would have to make /links very verbose in the query headers/parameters it needs, in order to allow for the different database queries needed (such as those by the two sub-resource endpoints demonstrated here).

Is there a better solution/abstraction here?

Improve this question

1 Answer 1

Reset to default
1

You can create a "service" layer. Abstract the link manipulation into a completely separate file and call it from each of the routes.

Create a service/links.js: 

module.exports = {
    manipulateLinks: function (response) {
        // Manipulate code
        return response
    }
}

Then in you routes, call the function:

var linkservice = require('../service/links')
var response = linkservice.manipulateLinks(response)
Improve this answer
Sign up to request clarification or add additional context in comments.

2 Comments

That is what I suggested in my original post. I would like to know if there are better ways of handling this.
No, it's pretty simple abstraction. Anything more than this would be weird.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.