WikiEduDashboard

The WikiEdu Dashboard is a web application that provides data about Wikipedia educational assignments that use the course page system (the EducationProgram extension) on Wikipedia. This is a project of Wiki Education Foundation, developed in partnership with WINTR, intended for our education programs on English Wikipedia. To see it in action, visit dashboard.wikiedu.org.

What it does

The Dashboard pulls information from the EducationProgram extension's Wikipedia API to identify which users are part of courses. It then gathers information about edits those users have made and articles they have edited, and creates a dashboard for each course intended to let instructors and others quickly see key information about the work of student editors. It also creates a global dashboard to see information about many courses at once.

• The system shows information for based on list course IDs defined by a page on Wikipedia.
• The system queries the liststudents api on English Wikipedia to get basic details about each course: who the students are, when the course starts and ends, and so on.
• The system uses a set of endpoints on Wikimedia Labs (see WikiEduDashboardTools) to perform queries on a replica Wikipedia database, for information about articles and revisions related to the courses.
• The system pulls page views (from stats.grok.se) for relevant articles on a daily basis.

Contributing

Code Style

This project adheres as strictly as possible to the community Ruby Style Guide. Rubocop is used for this purpose and its associated editor integrations are highly recommended for contributors.

Tests

Tests reside in the /spec folder. Both unit and integration tests are driven by RSpec.

• Write unit tests before building new features whenever possible. This project uses RSpec in conjuction with SimpleCov for unit testing.
• Write integration tests for new interfaces. This project uses Capybara and Capybara-webkit for integration testing.
  • Integration tests require qt5. On OSX you will need to uninstall qt4, install qt5, and add a symlink. This wiki section is a useful reference.
• Install test dependencies: apt-get install pandoc

Translations

Copy translations live at /config/locales and the fallback for missing strings is en. i18n.js is used to make these translations available on the frontend. The JS files providing the translations to the front end must be regenerated whenever a change is made by running rake i18n:js:export.

Requirements

• Ruby (RVM recommended)
• Node
• NPM
• Bower

Project Setup

• Fork this repo, so that you can set it up for a new server.

• Clone the new WikiEduDashboard repo and enter that directory.

• Make sure you are in the "sudo" group.

• Install Ruby 2.1.5 (RVM is recommended)

  • From the WikiEduDashboard directory, run the curl script from rvm.io
  • rvm install ruby-2.1.5

• Install Node: apt-get install nodejs npm

• Install Gems:

  • $ gem install bundler
  • $ bundle install
  • If some gems fail to install, you may need to install some dependencies, such as: libmysqlclient-dev libpq-dev libqtwebkit-dev

• Install NPM modules:

  • $ npm install

• Install Bower and Bower modules:

  • $ sudo npm install bower -g
  • $ bower install

• Add config files:

  • Save application.example.yml and database.example.yml as application.yml and database.yml, respectively, in the config directory. Fill in your Wikipedia account login details in application.yml (for API access). The default settings in database.yml will suffice for a development environment.

• Create mysql development and test database:

  • Install mysql-server and start a mysql command line
  • CREATE DATABASE dashboard DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;
  • CREATE DATABASE dashboard_testing DEFAULT CHARACTER SET utf8 DEFAULT COLLATE utf8_general_ci;
  • Grant access to these databases to your user.
  • GRANT ALL ON dashboard.* TO <USER>@localhost identified by <PASSWORD>;
  • GRANT ALL ON dashboard_testing.* TO <USER>@localhost identified by <PASSWORD>;

Integrations

For error logging, we use Sentry. You'll need access to a Sentry server to use this functionality; add the Sentry DSN to config/application.yml.

For analytics (ie, tracking traffic), we use Piwik. You need access to a Piwik server to use this funcationality; add the url and project id to config/piwik.yml.

OAuth

In order to use and develop the authenticated features of the application (course creation, the assignment design wizard, user management, etc) you will need to create a MediaWiki OAuth consumer).

Log in to mediawiki with your username and password for mediawiki.org. Once you're logged in, click on "Preferences" in the upper right-hand corner. In the "User profile" tab under "Preferences" (selected by default), change the email address to your email. You'll need this to confirm your account and get your token and secret key in the next step.

You'll now propose an OAuth consumer. Fill out the form like so:

Clicking on the 'Propose consumer' button should return a token and secret, which you should store in your application.yml file.

Initialize

1. Start Guard $ guard

2. Migrate the database $ rake db:migrate

3. Create cohorts $ rake cohort:add_cohorts

4. Pull data from sources (optional; this could take a very long time) $ rake batch:initialize

Develop

1. Start Zeus

   $ zeus start

2. Start Guard

   $ guard

3. Start Gulp to watch JS and CSS

   $ gulp

4. The frontend is now visible at http://localhost:3000/

Maintain

The Dashboard includes several rake tasks intended to keep the database synced with Wikipedia:

• Initialize: rake batch:initialize (Only to be run manually, initializes the database)
• Constant update: rake batch:update_constantly (Runs every 15 minutes by default)
• Daily update: rake batch:update_daily (Runs once a day by default)

Deploy

For deployment, the Dashboard uses Capistrano. This requires authentication with the server via SSH.

After pushing updates to repo (on Github), run the following command(s)

• Staging: cap staging deploy (This will deploy from the "master" branch)
• Production: cap production deploy (This will deploy from "production" branch)

To run rake tasks on a server via Capistrano, use "sake":

• $ cap production sake task="batch:update_constantly"

Set up a new production server

For detailed instructions on setting up a production server — specifically on a wmflabs virtual server, but the process will be similar for other infrastructure as well — see WMFLABS_DEPLOYMENT.