govspeak 0.8.6 → 0.8.7

data/lib/govspeak.rb

@@ -1,5 +1,6 @@
 require 'kramdown'
 require 'govspeak/header_extractor'
+require 'htmlentities'
 
 module Govspeak
 
@@ -22,6 +23,10 @@ module Govspeak
       @doc.to_html
     end
 
+    def to_text
+      HTMLEntities.new.decode(to_html.gsub(/(?:<[^>]+>|\s)+/, " ").strip)
+    end
+
     def headers
       Govspeak::HeaderExtractor.convert(@doc).first
     end

data/test/govspeak_test.rb

@@ -1,3 +1,5 @@
+# encoding: UTF-8
+
 require 'simplecov'
 require 'simplecov-rcov'
 
@@ -46,48 +48,61 @@ class GovspeakTest < Test::Unit::TestCase
     ], document.headers
   end
 
+  test "extracts text with no HTML and normalised spacing" do
+    input = "# foo\n\nbar    baz  "
+    doc = Govspeak::Document.new(input)
+    assert_equal "foo bar baz", doc.to_text
+  end
+
   test "govspark extensions" do
     markdown_regression_tests = [
 {
   input: "^ I am very informational ^",
   output: %{<div class="application-notice info-notice">
 <p>I am very informational</p>
-</div>}
+</div>},
+  text: "I am very informational"
 }, {
   input: "The following is very informational\n^ I am very informational ^",
   output: %{<p>The following is very informational</p>
 
 <div class="application-notice info-notice">
 <p>I am very informational</p>
-</div>}
+</div>},
+  text: "The following is very informational I am very informational"
 }, {
   input: "^ I am very informational",
   output: %{<div class="application-notice info-notice">
 <p>I am very informational</p>
-</div>}
+</div>},
+  text: "I am very informational"
 }, {
   input: "@ I am very important @",
   output: %{<div class="advisory"><p><strong>I am very important</strong></p>
-</div>}
+</div>},
+  text: "I am very important"
 }, {
   input: "The following is very important
 @ I am very important @",
   output: %{<p>The following is very important</p>
 
 <div class="advisory"><p><strong>I am very important</strong></p>
-</div>}
+</div>},
+  text: "The following is very important I am very important"
 }, {
   input: "% I am very helpful %",
   output:  %{<div class="application-notice help-notice">
 <p>I am very helpful</p>
-</div>}
+</div>},
+  text: "I am very helpful"
 }, {
   input: "The following is very helpful\n% I am very helpful %",
   output:  %{<p>The following is very helpful</p>
 
 <div class="application-notice help-notice">
 <p>I am very helpful</p>
-</div>}
+</div>},
+  text: "The following is very helpful I am very helpful"
 }, {
   input: "## Hello ##\n\n% I am very helpful %\r\n### Young Workers ###\n\n",
   output:  %{<h2 id="hello">Hello</h2>
@@ -96,38 +111,46 @@ class GovspeakTest < Test::Unit::TestCase
 <p>I am very helpful</p>
 </div>
 
-<h3 id="young-workers">Young Workers</h3>}
+<h3 id="young-workers">Young Workers</h3>},
+  text: "Hello I am very helpful Young Workers"
 }, {
   input: "% I am very helpful",
   output: %{<div class="application-notice help-notice">
 <p>I am very helpful</p>
-</div>}
+</div>},
+  text: "I am very helpful"
 }, {
   input: "This is a [link](http://www.google.com) isn't it?",
-  output: '<p>This is a <a href="http://www.google.com">link</a> isn&rsquo;t it?</p>'
+  output: '<p>This is a <a href="http://www.google.com">link</a> isn&rsquo;t it?</p>',
+  text: "This is a link isn’t it?"
 }, {
   input: "This is a [link with an at sign in it](http://www.google.com/@dg/@this) isn't it?",
-  output: '<p>This is a <a href="http://www.google.com/@dg/@this">link with an at sign in it</a> isn&rsquo;t it?</p>'
+  output: '<p>This is a <a href="http://www.google.com/@dg/@this">link with an at sign in it</a> isn&rsquo;t it?</p>',
+  text: "This is a link with an at sign in it isn’t it?"
 }, {
   input: "HTML
 
   *[HTML]: Hyper Text Markup Language",
-  output: %{<p><abbr title="Hyper Text Markup Language">HTML</abbr></p>}
+  output: %{<p><abbr title="Hyper Text Markup Language">HTML</abbr></p>},
+  text: "HTML"
 }, {
   input: "x[a link](http://rubyforge.org)x",
-  output: '<p><a href="http://rubyforge.org" rel="external">a link</a></p>'
+  output: '<p><a href="http://rubyforge.org" rel="external">a link</a></p>',
+  text: "a link"
 }, {
   input: "$!
 rainbow
 $!",
   output: %{<div class="summary">
 <p>rainbow</p>
-</div>}
+</div>},
+  text: "rainbow"
 }, {
   input: "$C help, send cake $C",
   output: %{<div class="contact">
 <p>help, send cake</p>
-</div>}
+</div>},
+  text: "help, send cake"
 }, {
   input: "$A
 street
@@ -135,31 +158,36 @@ road
 $A",
   output: %{<div class="address vcard"><div class="adr org fn"><p>
 street<br />road<br />
-</p></div></div>}
+</p></div></div>},
+  text: "street road"
 }, {
   input: "$P
 $I
 help
 $I
 $P",
-  output: %{<div class="place">\n<div class="information">\n<p>help</p>\n</div>\n</div>}
+  output: %{<div class="place">\n<div class="information">\n<p>help</p>\n</div>\n</div>},
+  text: "help"
 }, {
   input: "$D
 can you tell me how to get to...
 $D",
   output: %{<div class="form-download">
-<p>can you tell me how to get to&hellip;</p>
-</div>}
+<p>can you tell me how to get to…</p>
+</div>},
+  text: "can you tell me how to get to…"
 }, {
   input: "1. rod
 2. jane
 3. freddy",
-  output: "<ol>\n  <li>rod</li>\n  <li>jane</li>\n  <li>freddy</li>\n</ol>"
+  output: "<ol>\n  <li>rod</li>\n  <li>jane</li>\n  <li>freddy</li>\n</ol>",
+  text: "rod jane freddy"
 }, {
   input: "
 ((http://maps.google.co.uk/maps?q=Winkfield+Rd,+Windsor,+Berkshire+SL4+4AY&hl=en&sll=53.800651,-4.064941&sspn=17.759517,42.055664&vpsrc=0&z=14))
 ",
-  output: %{<div class="map"><iframe width="200" height="200" frameborder="0" scrolling="no" marginheight="0" marginwidth="0" src="http://maps.google.co.uk/maps?q=Winkfield+Rd,+Windsor,+Berkshire+SL4+4AY&amp;hl=en&amp;sll=53.800651,-4.064941&amp;sspn=17.759517,42.055664&amp;vpsrc=0&amp;z=14&amp;output=embed"></iframe><br /><small><a href="http://maps.google.co.uk/maps?q=Winkfield+Rd,+Windsor,+Berkshire+SL4+4AY&amp;hl=en&amp;sll=53.800651,-4.064941&amp;sspn=17.759517,42.055664&amp;vpsrc=0&amp;z=14">View Larger Map</a></small></div>}
+  output: %{<div class="map"><iframe width="200" height="200" frameborder="0" scrolling="no" marginheight="0" marginwidth="0" src="http://maps.google.co.uk/maps?q=Winkfield+Rd,+Windsor,+Berkshire+SL4+4AY&amp;hl=en&amp;sll=53.800651,-4.064941&amp;sspn=17.759517,42.055664&amp;vpsrc=0&amp;z=14&amp;output=embed"></iframe><br /><small><a href="http://maps.google.co.uk/maps?q=Winkfield+Rd,+Windsor,+Berkshire+SL4+4AY&amp;hl=en&amp;sll=53.800651,-4.064941&amp;sspn=17.759517,42.055664&amp;vpsrc=0&amp;z=14">View Larger Map</a></small></div>},
+  text: "View Larger Map"
 }, {
   input: "s1. zippy
 s2. bungle
@@ -169,14 +197,16 @@ s3. george
 <li><p>zippy</p>\n</li>
 <li><p>bungle</p>\n</li>
 <li><p>george</p>\n</li>
-</ol>}
+</ol>},
+  text: "zippy bungle george"
 }
 ]
 
-markdown_regression_tests.each do |t|
-  rendered = Govspeak::Document.new(t[:input]).to_html
-  assert_equal t[:output].strip, rendered.strip
-end
+  markdown_regression_tests.each do |t|
+    doc = Govspeak::Document.new(t[:input])
+    assert_equal t[:output].strip, doc.to_html.strip
+    assert_equal t[:text], doc.to_text
+  end
 
 end
 

metadata

@@ -1,53 +1,59 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: govspeak
-version: !ruby/object:Gem::Version 
+version: !ruby/object:Gem::Version
+  version: 0.8.7
   prerelease: 
-  version: 0.8.6
 platform: ruby
-authors: 
+authors:
 - Ben Griffiths
 - James Stewart
 autorequire: 
 bindir: bin
 cert_chain: []
-
-date: 2011-12-20 00:00:00 Z
-dependencies: 
-- !ruby/object:Gem::Dependency 
+date: 2012-01-11 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
   name: kramdown
-  prerelease: false
-  requirement: &id001 !ruby/object:Gem::Requirement 
+  requirement: &20300980 !ruby/object:Gem::Requirement
     none: false
-    requirements: 
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
+      - !ruby/object:Gem::Version
         version: 0.13.3
   type: :runtime
-  version_requirements: *id001
-- !ruby/object:Gem::Dependency 
-  name: rake
   prerelease: false
-  requirement: &id002 !ruby/object:Gem::Requirement 
+  version_requirements: *20300980
+- !ruby/object:Gem::Dependency
+  name: htmlentities
+  requirement: &20299900 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: '4'
+  type: :runtime
+  prerelease: false
+  version_requirements: *20299900
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: &20299160 !ruby/object:Gem::Requirement
     none: false
-    requirements: 
+    requirements:
     - - ~>
-      - !ruby/object:Gem::Version 
+      - !ruby/object:Gem::Version
         version: 0.8.0
   type: :development
-  version_requirements: *id002
-description: |-
-  A set of extensions to markdown layered on top of the kramdown 
-  library for use in the UK Government Single Domain project
-email: 
+  prerelease: false
+  version_requirements: *20299160
+description: ! "A set of extensions to markdown layered on top of the kramdown \nlibrary
+  for use in the UK Government Single Domain project"
+email:
 - ben@alphagov.co.uk
 - james.stewart@digital.cabinet-office.gov.uk
 executables: []
-
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
+files:
 - lib/govspeak/header_extractor.rb
 - lib/govspeak.rb
 - Gemfile
@@ -56,31 +62,28 @@ files:
 - test/test_helper.rb
 homepage: http://github.com/alphagov/govspeak
 licenses: []
-
 post_install_message: 
 rdoc_options: []
-
-require_paths: 
+require_paths:
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
+required_ruby_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      version: "0"
-required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
-  requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      version: "0"
+  requirements:
+  - - ! '>='
+    - !ruby/object:Gem::Version
+      version: '0'
 requirements: []
-
 rubyforge_project: 
-rubygems_version: 1.8.12
+rubygems_version: 1.8.11
 signing_key: 
 specification_version: 3
 summary: Markup language for single domain
-test_files: 
+test_files:
 - test/govspeak_test.rb
 - test/test_helper.rb