Skip to content
This repository has been archived by the owner on Sep 27, 2022. It is now read-only.

Latest commit

 

History

History
396 lines (262 loc) · 18.8 KB

README.md

File metadata and controls

396 lines (262 loc) · 18.8 KB
Raw

QEWD Example App

Example QEWD.js Back-end codebase that adheres to the RealWorld spec and API.

This repo is functionality complete — PR's and issues welcome!

Rob Tweed [email protected]
25 January 2017, M/Gateway Developments Ltd http://www.mgateway.com

Twitter: @rtweed

Google Group for discussions, support, advice etc: http://groups.google.co.uk/group/enterprise-web-developer-community

About qewd-conduit

qewd-conduit is a full implementation of the REST back-end for the RealWorld Conduit application using QEWD.js.

qewd-conduit requires QEWD.js to be installed on your server (see below).

qewd-conduit uses Redis as a Document Database and Persistent JavaScript Objects. QEWD.js itself is a Node.js-based Web Application & REST run-time platform, and is one of the simplest and most productive platforms available for Node.js application development.

Read my article that explains the rationale and objectives of QEWD.js.

Since the back-end specification of the RealWorld Conduit application is fully documented and implemented using several other technologies and/or frameworks, it provides a great way of comparing and contrasting the different development approaches used for each option.

Although QEWD.js is a Node.js-based platform, you'll see that the way in which the back-end has been able to be developed is very different from what you'd expect. Whilst it's all been written in JavaScript, there's no asynchronous logic, even for the database manipulation. That's possible due to QEWD.js's master process / queue / worker process-pool architecture.

The RealWorld Conduit initiative also allows direct comparisons to be made in terms of back-end performance. I think you'll be favourably impressed by the performance of qewd-conduit.

What may be less easy to appreciate is the speed of development when using the different RealWorld Conduit back-end technologies. I can tell you that, in the case of qewd-conduit, the entire back-end was implemented from scratch in just 2 man-days, including the time taken to read up on and understand the application's objectives, requirements and APIs. Part of the speed of development comes from not having to worry about asynchronous logic, but it's also due to its very high-level database abstraction. See Parts 17 - 27 of the QEWD.js Online Training Course.

A Quick Guided Tour of qewd-conduit

Install QEWD.js

The simplest way to do this is to use one of the pre-built installers or the QEWD.js Docker Appliance

If you prefer to install manually, use this guide.

You'll probably want to install and use Redis with qewd-conduit, but it will also work with GT.M or Caché

Install qewd-conduit

From the directory in which you installed QEWD.js (eg ~/qewd):

   npm install qewd-conduit

Master Process Startup File

We'll start with the QEWD.js Master Process startup file that you'll find in the repo at /startup/qewd.js

This file tells QEWD.js how to configure Express, the Worker Pool size and how to set up the basic URL REST routing.

The first part is a configuration object (config), the properties of which you can change before starting QEWD.js, eg:

  • managementPassword: If you also installed the qewd-monitor application as part of the QEWD.js installation (recommended so you can monitor and manage the QEWD.js run-time environment from a browser), you MUST change this password - it's what provides access to the qewd-monitor application.

  • serverName: This is the name that appears in the qewd-monitor application's banner. You can leave this alone if you wish

  • port: the port on which Express will listen. Change if you prefer a different port to 8080

  • poolSize: the maximum number of Worker Processes that QEWD.js will use. On a very busy system you may want to increase this value

  • database: currently specifies Redis. Each Worker Process will connect to Redis using its default port

The routes object should be left unchanged. It defines the basic REST routing that will be configured using Express middleware provided by QEWD.js. You'll see that it specifies that any incoming URI prefixed /api will be handled by the qewd-conduit module. Note that this handling actually takes place in the QEWD.js Worker Processes, not the QEWD.js Master Process.

The routes object also defines a custom error object response structure that conforms to the RealWorld Conduit API specification for Not Found errors.

Starting QEWD.js

Once you're happy with your QEWD.js startup file, copy it to your QEWD.js installation directory, eg to:

   ~/qewd/qewd.js

Then start it up. You can either run it manually, eg:

     cd ~/qewd   (or whatever directory you installed QEWD.js in)
     node qewd

Or use something like PM2 to run QEWD.js as a background service, eg:

     cd ~/qewd
     pm2 start qewd.js

Read this document for details of how to run QEWD.js as a service on Windows.

Stopping QEWD.js

If you're running QEWD.js in a foreground window process, simply type CTRL & C

If you're using PM2:

    pm2 stop qewd

Alternatively, you can stop QEWD.js by using the qewd-monitor application:

  • In a browser, start it up:

      http://192.168.1.100:8080/qewd-monitor/index.html

  Note: change the IP address & port to match your server and QEWD.js configuration

  • Login using the management password that you specified in the QEWD.js startup file

  • Click the red X button next to the Master Process PID

Main Routing For the RealWorld Conduit API

When REST requests for RealWorld Conduit are received (eg GET /api/tags), the request is queued by the QEWD.js master process and dispatched to an available worker process.

This worker process will load the qewd-conduit module (if it hasn't already done so), and all subsequent processing of the request is handled by qewd-conduit.

Everything starts in the file you'll find in the repo at /lib/conduit.js

When the module is first loaded, its init() function is automatically invoked. You'll see that this defines all the URI routing for RealWorld Conduit's APIs. The routing is actually managed by a sub-module named qewd-router

The processing of each specific API is handled by the module specified in the routes array, eg:

   {
    url: '/api/users/login',
    method: 'POST',    
    handler: users.authenticate
  }

This specifies that POST requests for /api/users/login will be handled by the users.authenticate module that you'll find in /lib/user/authenticate.js

At the bottom of the conduit.js file you'll see the handlers object. This is needed by QEWD.js's Worker Process to handle the 2nd-level URI options, ie:

   /api/users
   /api/user
   /api/articles
   /api/profiles
   /api/tags

So the main qewd-conduit is hopefully pretty straightforward to understand, and you can see that it is easily extensible in future to cater for new APIs.

Handlers for Specific APIs

So, as a result of the routes object in the conduit.js file (described above), each of the RealWorld Conduit APIs is handled by its own module file. eg, GET /api/tags is handled by /lib/tags/get.js

You'll find that they all conform to a pretty similar pattern:

  • if the API is a POST or PUT that requires a body payload, the body is checked to ensure all required fields are present and not empty strings, or to ensure that any optional fields are not empty strings. All the information contained in an incoming REST request is made available by QEWD.js to the handler module function via the handler function's first argument: args, eg:

    • args.req.headers: contains all the HTTP request headers

    • args.req.body: contains the body payload for POST and PUT requests (if relevant)

    • args.req.query: contains any URI QueryString name/value pairs

    • any variables specified in the route (eg /api/articles/:slug) are available as a property of args (eg args.slug)

  • if the API requires authentication, the JSON Web Token (JWT) is checked against the Secret that was used by the QEWD.js Worker when it was created or updated. The user Id and Username are extracted from the JWT payload if needed by the handler. JWT support is provided by the ewd-session module. Although this can be used by QEWD.js applications to support Sessions and associated Session Storage, qewd-conduit actually only uses QEWD.js Sessions for secure storage of the JWT Secret.

  • Any required specific field validation then takes place

  • The database containing the User, Article, Profile or Tag information is accessed. See the next section for information on how database handling is carried out in qewd-conduit. All the database activity is abstracted as object-specific API modules in the /lib/db folder., eg

    var articleId = db.articles.getIdBySlug.call(this, args.slug);

  • Finally the response is returned. This is handled by the handler function's 2nd argument: callback

The purpose of this callback function is two-fold:

  • it conveys the JSON response object that is to be returned to the QEWD.js Master Process (which forwards it to the REST Client that originally sent the request)

  • it tells QEWD.js that Worker processing has completed, which results in the Worker Process being returned to the available pool, ready to handle the next incoming request from QEWD.js's queue.

QEWD.js Database Handling

Even though QEWD.js may be using Redis, it's not using it in a conventional way. QEWD.js uses the ewd-redis-globals module to implement a style of database storage known as Global Storage. Global Storage is a very powerful database design with multi-model NoSQL capability.

QEWD.js then further abstracts these basic database APIs by using the ewd-document-store module.

The ewd-document-store module abstracts the database (eg Redis) as what are known as DocumentNode objects.

These allow the database to be simultaneously accessible as:

qewd-conduit stores the Conduit data in three Persistent Documents:

  • conduitUsers
  • conduitArticles
  • conduitComments

If you're intested in inspecting their raw storage, you can use the qewd-monitor application and click the Document Store Nav link in the banner.

All database access in the qewd-conduit application is implemented via a set of APIs that you'll find in the /lib/db/ folder. So, for example, all the APIs for accessing the articles Document are in /lib/db/articles.js.

Here's some examples of their use in the qewd-conduit API handler modules:

  • In /lib/db/articles.js, you'll find the getTags() function, which spins through the conduitArticles document byTag index, using the DocumentNode's forEachChild method:

    function getTags() {
  var tags = [];
  var tagsDoc = new this.documentStore.DocumentNode('conduitArticles', ['byTag']);
  tagsDoc.forEachChild(function(tag) {
    tags.push(tag);
  });
  return tags;
}

Read this for more details on traversing DocumentNode objects.

  • Also in /lib/db/articles.js, look at the create() function where it creates the various indices for a new article:

      // save to database

  var articlesDoc = new this.documentStore.DocumentNode('conduitArticles');

  articlesDoc.$(['byId', articleId]).setDocument(article);

  // create indices

  articlesDoc.$(['bySlug', slug]).value = articleId;
  articlesDoc.$(['byAuthor', authorId, articleId]).value = articleId;
  articlesDoc.$(['byTimestamp', ts]).value = articleId;

The setDocument() method is an example of the Document Database capability in action, whilst setting the value shows how another part of the same DocumentNode object can be handled as a Persistent Javascript Object.

  • In /lib/db/comments.js, is the del() function which implements how a comment is deleted from the database:

      function del(commentId, unlinkArticle) {
    if (typeof unlinkArticle === 'undefined') unlinkArticle = true;
    var commentDoc = new this.documentStore.DocumentNode('conduitComments', ['byId', commentId]);
    var articleId = commentDoc.$('articleId').value;
    commentDoc.delete();
    if (unlinkArticle) unlinkComment.call(this, articleId, commentId);
  }

You'll notice that, even though QEWD.js is a Node.js-based platform, all the database handling is done synchronously. This is possible due to the Master Process/Queue/Worker Pool architecture of QEWD.js.

You'll hopefully agree that the approach to database handling adopted by QEWD.js is straightforward, intuitive, yet very powerful. You'll also find that it's very fast.

That's basically the qewd-conduit application in a nutshell. To find out more, I'd encourage you to dig into the code and consult the training slide decks.

Demo System

I've made a live instance of qewd-conduit available at the endpoint:

  http://178.62.26.29:8080/api

Feel free to try it out against the RealWorld Conduit PostMan Collection

I've also installed the React/Redux version of the Conduit UI on this same server, so you can see the qewd-conduit back-end working with it.

Point a browser at http://178.62.26.29:8080 to try it out.

Is QEWD.js an Express or Koa.js Framework?

One of the unique features of QEWD.js is that it isn't tied to any other particular technologies (apart from Node.js!). So, for example, by default, QEWD.js's outward-facing HTTP(S) interface is provided by Express. However, simply by making a simple change to your QEWD.js startup file, you can switch it to use Koa.js.
See this article which explains what you need to do to use Koa.js with QEWD.js.

You can try out a QEWD.js/Koa.js Conduit demo system. It uses exactly the same qewd-conduit back-end code as the Express-based version described in the previous section.
One interesting difference is that the Koa.js version's HTTP response headers show the QEWD.js response time (excluding any network latency, but including database access times). Inspect these response headers in your browser's JavaScript Console and you'll see just how fast QEWD.js is: you should typically see response times between 3ms - 9ms, and this is on a very modest Amazon EC2 t2.micro instance.

Is QEWD.js a REST or WebSockets Platform?

REST support is just one of QEWD.js's capabilities. "Out of the Box" it also supports real-time, secure WebSocket messaging to/from clients (eg browsers and mobile devices). Indeed, a single instance of QEWD.js can simultaneously and concurrently support incoming REST and WebSocket requests, from the same or different clients.

WebSocket support in QEWD.js is provided by Socket.io, which is automatically installed and configured for you when you install QEWD.js.

To demonstrate these capabilities, I've included a modified version of the Conduit UI that communicates with the qewd-conduit back-end via WebSockets and that invokes the very same back-end database logic as the corresponding REST requests. See the /www folder in this repository for more details of the modified UI, and see the /lib/ws folder for the back-end WebSocket message handler functions.

License

Copyright (c) 2017 M/Gateway Developments Ltd,
Redhill, Surrey UK.
All rights reserved.

http://www.mgateway.com
Email: [email protected]

Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at

  http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and
limitations under the License.