MatMoore

# Run in publishing_api_development
# These queries need reviewing to filter out links to non-live content

\copy (select l.target_content_id,coalesce(d.content_id, d2.content_id) as source_content_id, l.link_type, l.position, case when l.edition_id is null then 'document' else 'edition' end as link_scope from links l left outer join link_sets ls on l.link_set_id=ls.id left outer join documents d on ls.content_id=d.content_id left outer join editions e on e.id=l.edition_id left outer join documents d2 on d2.id=e.document_id) to 'content-links.csv' with csv header;

\copy (select d.content_id,d.locale,e.title,e.description,e.document_type,e.analytics_identifier,e.state from documents d join editions e on e.document_id=d.id and e.content_store='live') to 'content.csv' with csv header;

\copy (select d.content_id, d2.content_id as owning_content_id from documents d join documents d2 on d.owning_document_id=d2.id) to owning_documents.csv with csv header

MatMoore

Go to JENKINS_URL/script

Run this groovy:

item = Jenkins.instance.getItemByFullName("content-audit-tool/master")
item.updateNextBuildNumber(371)

We had to do this because we forked a project in github, and our release tags were based on the jenkins build number. So we wanted to start the numbering of the new CI builds where we left off.

MatMoore

1. Run a rendering app

1. Checkout https://github.com/alphagov/government-frontend, install ruby, bundle etc
2. Run ./startup.sh --live to point to the live content store
3. Run PLEK_SERVICE_STATIC_URI=static.dev.gov.uk (or wherever static is running) to point to a local instance of static
4. This will take a minute to render a page(?)

Shared templates are loaded from static via slimmer, which magically merges templates together using nokogiri.

Slimmer may cache templates (?) although I don't think it's supposed to in development mode

MatMoore

#!/usr/bin/env bash

set -euo pipefail

for i in $( ls /data/vhost|grep -v 'lost+found'); do
	echo $i
	dir="/data/vhost/$i/shared/bundle/ruby"
	versions=$(ls -1 $dir|grep -v '.backup');
	in_use=$(cat /data/vhost/$i/releases/*/.ruby-version | sed 's/[0-9]\+$/0/' | sort | uniq);
	echo -e "in use\n----\n$in_use\n"

MatMoore

{
  "annotations": {
    "list": []
  },
  "editable": true,
  "gnetId": null,
  "graphTooltip": 0,
  "hideControls": false,
  "id": 93,
  "links": [],

MatMoore

# Full code here:
# https://github.com/googleads/google-api-ads-ruby/blob/master/adwords_api/examples/solutions/conflicting_keywords/conflicting_keywords.rb

def compare_keywords(negatives, positive)
  negatives.each do |negative|
    match_type = negative.match_type.downcase
    negative_text = negative.text.downcase
    positive_text = positive.text.downcase
    match = false
    # If the negative keyword is more strict than the positive one, it cannot

MatMoore

#!/usr/bin/env ruby
require 'faraday'
require 'json'

response = Faraday.get 'https://www.gov.uk/api/content'
homepage = JSON.parse(response.body)

level_one = homepage['links']['level_one_taxons'].map do |item|
  item["api_path"]
end

MatMoore

MatMoore

People that can not distinguish colours can understand all information

avoid instruction that rely on colour don't communicate through colour alone People that have low vision are able to see content

default text size is not too small default colour contrast is good the appearance of online content can be overridden (with browser and operating system settings) - and guidance is provided on how to do this is easy to find This also helps people experiencing a situational impairment such as bright sunlight

MatMoore

#!/usr/bin/env ruby

current_statement_lines = []
primary_keys = []
other_constraints = []

current_statement = :unknown
lines = File.readlines(ARGV.first)
lines.each do |line|
	current_statement_lines << line.gsub("\r", '')