GOV.UK Navigation Prototype 2016

This app serves up prototype navigation flows for the work-in-progress GOV.UK navigation.

The master branch is deployed to https://govuk-nav-prototype.herokuapp.com/.

See also the previous prototype: https://github.com/alphagov/govuk-navigation-prototype.

Nomenclature

• Taxon: a single node within the taxonomy

Technical documentation

This is a node app built using the govuk_prototype_kit.

All the data for the prototype comes from a pre-built JSON file and scraped HTML data from GOV.UK.

Installation

We use nvm to manage node.js versions. Follow the steps below to install and use it:

• Install nvm (brew install nvm)
• follow the post-install instructions for nvm
• go into the project directory
• run nvm install
• run nvm use to verify everything is working.

You can now install the dependencies with:

npm install

Importing new content

The content in the prototype comes from inventories of education pages on GOV.UK.

Prerequisites

• Working installs of Python 2 and Ruby 2.
• A Python virtual env set up for this project. See https://virtualenvwrapper.readthedocs.io/en/latest/

Instructions

Before you import new content, delete the old content from app/content. Furthermore, check if the taxonomy has changed. If there are new taxons, of base paths have changed, fetch the new taxonomy from content-tagger and replace /lib/ruby/data/taxons.csv with it.

There are several steps to importing new content into the prototype:

0. Tag all the content to taxons using content tagger.
1. Create a list of content base paths in CSV format, with a single column heading Link. This can be done with the taxonomy:export rake task in content-tagger Save this as content.csv in the lib/python/data folder. The taxonomy:export_tagged_content rake task in content-tagger can be used to derive a list of paths from an entire taxonomy.
2. Run lib/data_import.sh <environment>, where <environment> is either staging or production. This will determine where taxonomy data is retrieved from.
3. Enter http auth credentials when requested.

The script executes the following:

• format_finder.py
  • This will fetch format information from the search API, saving the output in a file called lib/python/data/link_formats.csv.
• fetch_main_content.py
  • Scrapes the main tag html of each page from the GOV.UK mirror and store it in app/content. This uses the data from the previous step to organise the pages by format (the format directory is ignored by the prototype when rendering pages but this makes it easier to work with). See failed_pages.txt for a list of pages that encountered redirects and 404s.
• Rakefile default task
  • Fetches taxon and content page metadata (e.g. which pages link to which taxons). This will be saved in app/data/metadata_and_taxons.json.

Fetching content lists for topic pages

This information comes from rummager. We ignore any results we don't have html for, to avoid dead links.

Redirected content

When we import content we're currently skipping redirected pages.

In most cases it's fine to use the tags from the original inventory with the page the url is redirected to, but sometimes the pages are redirected to more generic information.

Checking the content

• bundle exec rake validate checks that all pages have taxons associated with them

• linkchecker crawls for dead links

  pip install LinkChecker
  linkchecker 'http://localhost:3000'
  

Dependencies

• Content store - provides taxon data
• Rummager - provides data about pages tagged to the new taxonomy

Running the application

Run npm start

Open the app at http://localhost:3000/

Licence

MIT License