This was my solution to [R.O.]'s challenge. Coffeescript and mocha were completely new to me.
To see just my edits, look at this commit.
— J. Eiten
This exam should cover the following skills:
- Familiarity with cloning a git repository
- Familiarity with algorithmic thinking
- Familiarity with unit testing concepts
- Familiarity with web application structure
- Familiarity with basic JavaScript libraries
-
Set up git and clone the repository INSERT REPO LOCATION HERE. We'll leave the details for you to work out.
-
Install the latest Node.js and ensure that you have the Node Package Manager (npm) installed as well.
Hint: You should be able to run the following commands:
$ node -v #(prints the version of node.js installed) $ npm -v #(prints the version of npm installed) -
Install the Node.js CoffeeScript module globally using npm:
$ npm install -g coffee-script #(depending on your system, this may require the use of sudo) -
Install the Node.js Mocha module globally using npm:
$ npm install -g mocha #(you will likely need to sudo again) -
For this application we depend on the ExpressJS web application framework, the UnderscoreJS functional programming library, and the hbs Handlebars wrapper for ExpressJS applications. In order to locally install all of the project dependencies, run the following command in the root directory of the cloned project:
$ npm install #(performs local installation of Node project dependencies)Note: You can view the project dependencies and other information in the
package.jsonfile. Locally installed modules and their dependencies can be found in the node_modules folder. -
Verify that everything works and is ready to go by running the following commands:
#(prints the version of coffeescript installed) $ coffee -v #(executes the unit test procedures and result in 15 test failures) $ mocha --compilers coffee:coffee-script #Note that due to a bug (or feature) in the latest version of mocha # you may need to run the following command instead $ mocha --recursive --compilers coffee:coffee-script/register #(will start the web application and print `Server listening on port 3000`) $ coffee web.coffeeHint: Use ctrl-c to terminate the web app process.
You're working for a company that has an online movie database. Users can submit integer valued ratings of movies that they have watched. The lead analyst wants to try out a new way of aggregating all of these ratings by using a trimmed average. For a trimmed average, you must remove all minimal and maximal values from a mean before calculating the average. So, for example,
TrimmedAverage(1,1,2,2,3,4,4,5,5,5) = Average(2,2,3,4,4) = 3
Since this analyst is a really good product owner, she was nice enough to write you a set of unit tests! You can find
the tests in the test directory of the app: test/movie-rater-test.coffee. She's even written you a function stub to
get started (see app/movie-rater.coffee). You'll just have to implement the trimmed average inside of the function
--taking care to throw Errors when not enough ratings are given or when the input is invalid--and you're done once the
movie rater unit tests pass (run $ mocha --compilers coffee:coffee-script and you should not see any tests fail
with movieRater() in the name).
If you're really an over-achiever (bonus points here), you'll figure out how to implement this function with a constant amount of memory with respect to the elements in the ratings parameter AND with a single pass over the ratings.
...that the lead engineer wants you to go back and unit test the movie database's data access object
app/movie-ratings.coffee. He has already written unit test stubs in test/movie-ratings-test.coffee. You will know you
are done when you can run $ mocha --compilers coffee:coffee-script multiple times and all of the tests pass.
Hint: The MovieRatingsResource does not guarantee the order that ratings are returned is the same as the order they are inserted, and your unit tests should handle these cases.
NOTE: You should not modify any code outside of app/movie-rater.coffee and test/movie-ratings-test.coffee for this
section!
Now that you have implemented the movie rater and tested the movie ratings data access object, your company is ready to
unveil their product! It is a web app that displays the stored movie ratings and the trimmed average of those ratings.
The web app runs on Node.js using the Express.js framework. The engineering team has built a
basic RESTful API for accessing and manipulating stored movie ratings data that uses the movie ratings access object
that you helped test. There is also an http endpoint for calculating the trimmed average of a collection of ratings that
uses the movie rater that you created. web.coffee contains all of your application's routes, and you can start the
server through the command $ coffee web.coffee. You can view the client by navigating to localhost:3000 in
your web browser.
Boy, that sure is a beautiful UI, isn't it? Now, it is your chance to show how creative you are by improving it! The
UI is a client side JavaScript application (compiled from CoffeeScript). The client app uses RequireJS for modular code
and comes equipped with jQuery, Underscore, and
Handlebars. The only limit is that you cannot modify app/views/layout.hbs.
Some useful hints:
- To compile the CoffeeScript files into JavaScript files use the
$ coffee --compile --output ./assets/javascripts/ ./app/assets/coffeescripts/command. You can also use the--watchflag to have the compiler run in the background and compile whenever you save changes to your CoffeeScript. - You can use the Sass to create simplified stylesheets that are compiled into CSS.
- It would be great to upgrade the movie-ratings-service-client to take advantage of the full RESTful API provided by the application, and then provide the client UI to allow the user to add/modify/delete ratings interactively.