2016 - University of Cambridge IB Group Project - Team Mike

Razvan Kusztos (razvankusz), Ben Ramchandani (Ben-Ramchandani), Dmitrij Szamozvancev (DimaSamoz), James Wood (laMudri), Laura Nechita (redls)

TWork - Server

This repository is the server for the Twork system. It's written in Java using the Play! Framework.

Installation instructions

Requirements:

• JDK 1.8 or later.
• Typesafe Activator 1.3.7.
• PostgreSQL database.

Procedure:

• Start PostgreSQL on port 5432 with username twork, password test and a database named tworkdata.
• Clone the repository.
• Unpack Typesafe Activator into the tworkserver directory.
• Move into the tworkserver directory and type ./activator run to start the server in development mode.

Code structure

All paths are given from the tworkserver directory.

Configuration and dependencies:

• build.sbt contains the dependencies for the project.
• conf/application.conf contains database configuration.

Routes:

conf/routes contains the routes for the server, that is, what URLs map to what Java code. For example it contains the line GET /job controllers.Application.job(), which Play! uses to map the request a device makes to get a job to complete. This file is well commented and acts as the API specification for communicating devices. It is worth reading if you want to become familiar with the code.

Website:

The HTML (with embedded Scala code, Java cannot be used here) is contained in app/views. The images, JavaScript and CSS used by the site is in the public directory. The supporting code for the main page of the website, including file upload, is contained in app/controllers/Web.java. The real time image result page is handled by app/controllers/Display.java with helper functions in app/sitehelper.

Job server:

The code for the job server is all contained in packages in the app/ directory, any directories not mentioned so far do not contain code written by us.

Global.java

This file contains code that runs every time the server is started.

controllers

Application.java is the entry point for requests from devices, the details of what each method should do is explained by the corresponding entry in conf/routes.

computations

This package contains both the code that is sent to devices to be run, which all implement ComputationCode and the code to generate computations, which all implement BasicComputationGenerator. Reading the relatively short PrimeComputation and PrimeComputationCode should be informative about the abstractions used in the server.

models

These classes can be stored persistently in the database. Ebean uses the field definitions and Java Persistence Annotations to generate SQL statements. UUID generation is handled automatically by Ebean when save() is first called.

twork

This package contains the bulk of the back end code that manages the computations and jobs as they move around the system.

Tests

The tworkserver/test/ directory contains a large test suite with tests covering much of the server's functionality. The tests use JUnit along with Play!'s helper functions. Along with unit tests there is a large test covering the lifetime of a PrimeComputation which runs against a test server, interacting via HTTP requests.

The tests may be run with ./activator test (from tworkserver/).

 

© 2016 Razvan Kusztos (razvankusz), Ben Ramchandani (Ben-Ramchandani), Dmitrij Szamozvancev (DimaSamoz), James Wood (laMudri), Laura Nechita (redls)