lib/rich_text.rb

   1 # frozen_string_literal: true
   2
   3 module RichText
   4   SPAMMY_PHRASES = [
   5     "Business Description:", "Additional Keywords:"
   6   ].freeze
   7
   8   DESCRIPTION_MAX_LENGTH = 500
   9   DESCRIPTION_WORD_BREAK_THRESHOLD_LENGTH = 450
  10
  11   def self.new(format, text)
  12     case format
  13     when "html" then HTML.new(text || "")
  14     when "markdown" then Markdown.new(text || "")
  15     when "text" then Text.new(text || "")
  16     end
  17   end
  18
  19   class SimpleFormat
  20     include ActionView::Helpers::TextHelper
  21     include ActionView::Helpers::OutputSafetyHelper
  22
  23     def sanitize(text, _options = {})
  24       Sanitize.clean(text, Sanitize::Config::OSM).html_safe
  25     end
  26   end
  27
  28   class Base < String
  29     def spam_score
  30       link_count = 0
  31       link_size = 0
  32
  33       doc = Nokogiri::HTML(to_html)
  34
  35       if doc.content.empty?
  36         link_proportion = 0
  37       else
  38         doc.xpath("//a").each do |link|
  39           link_count += 1
  40           link_size += link.content.length
  41         end
  42
  43         link_proportion = link_size.to_f / doc.content.length
  44       end
  45
  46       spammy_phrases = SPAMMY_PHRASES.count do |phrase|
  47         doc.content.include?(phrase)
  48       end
  49
  50       ([link_proportion - 0.2, 0.0].max * 200) +
  51         (link_count * 40) +
  52         (spammy_phrases * 40)
  53     end
  54
  55     def image
  56       nil
  57     end
  58
  59     def image_alt
  60       nil
  61     end
  62
  63     def description
  64       nil
  65     end
  66
  67     def truncate_html(max_length = nil, img_length = 1000)
  68       html_doc = to_html
  69       return html_doc if max_length.nil?
  70
  71       doc = Nokogiri::HTML::DocumentFragment.parse(html_doc)
  72       keep_or_discards = %w[p h1 h2 h3 h4 h5 h6 pre a table ul ol dl]
  73       accumulated_length = 0
  74       exceeded_node_parent = nil
  75       truncated = false
  76
  77       doc.traverse do |node|
  78         if accumulated_length >= max_length
  79           if node == exceeded_node_parent
  80             exceeded_node_parent = node.parent
  81             node.remove if keep_or_discards.include?(node.name)
  82           else
  83             node.remove
  84           end
  85           next
  86         end
  87
  88         next unless node.children.empty?
  89
  90         if node.text?
  91           accumulated_length += node.text.length
  92         elsif node.name == "img"
  93           accumulated_length += img_length
  94         end
  95
  96         if accumulated_length >= max_length
  97           truncated = true
  98           exceeded_node_parent = node.parent
  99           node.remove
 100         end
 101       end
 102
 103       {
 104         :truncated => truncated,
 105         :html => doc.to_html.html_safe
 106       }
 107     end
 108
 109     protected
 110
 111     def simple_format(text)
 112       SimpleFormat.new.simple_format(text, :dir => "auto")
 113     end
 114
 115     def sanitize(text)
 116       Sanitize.clean(text, Sanitize::Config::OSM).html_safe
 117     end
 118
 119     def linkify(text, mode = :urls)
 120       link_attr = 'rel="nofollow noopener noreferrer" dir="auto"'
 121       Rinku.auto_link(ERB::Util.html_escape(text), mode, link_attr) do |url|
 122         url = shorten_host(url, Settings.linkify_hosts, Settings.linkify_hosts_replacement)
 123         shorten_host(url, Settings.linkify_wiki_hosts, Settings.linkify_wiki_hosts_replacement) do |path|
 124           path.sub(Regexp.new(Settings.linkify_wiki_optional_path_prefix || ""), "")
 125         end
 126       end.html_safe
 127     end
 128
 129     private
 130
 131     def shorten_host(url, hosts, hosts_replacement)
 132       %r{^(https?://([^/]*))(.*)$}.match(url) do |m|
 133         scheme_host, host, path = m.captures
 134         if hosts&.include?(host)
 135           path = yield(path) if block_given?
 136           if hosts_replacement
 137             "#{hosts_replacement}#{path}"
 138           else
 139             "#{scheme_host}#{path}"
 140           end
 141         end || url
 142       end || url
 143     end
 144   end
 145
 146   class HTML < Base
 147     def to_html
 148       linkify(simple_format(self))
 149     end
 150
 151     def to_text
 152       to_s
 153     end
 154   end
 155
 156   class Markdown < Base
 157     def to_html
 158       linkify(sanitize(document.to_html), :all)
 159     end
 160
 161     def to_text
 162       to_s
 163     end
 164
 165     def image
 166       @image_element = first_image_element(document.root) unless defined? @image_element
 167       @image_element.attr["src"] if @image_element
 168     end
 169
 170     def image_alt
 171       @image_element = first_image_element(document.root) unless defined? @image_element
 172       @image_element.attr["alt"] if @image_element
 173     end
 174
 175     def description
 176       return @description if defined? @description
 177
 178       @description = first_truncated_text_content(document.root)
 179     end
 180
 181     private
 182
 183     def document
 184       return @document if @document
 185
 186       @document = Kramdown::Document.new(self)
 187
 188       should_get_dir_auto = lambda do |el|
 189         dir_auto_types = [:p, :header, :codespan, :codeblock, :pre, :ul, :ol, :table, :dl, :math]
 190         return true if dir_auto_types.include?(el.type)
 191         return true if el.type == :a && el.children.length == 1 && el.children[0].type == :text && el.children[0].value == el.attr["href"]
 192
 193         false
 194       end
 195
 196       add_dir = lambda do |element|
 197         element.attr["dir"] ||= "auto" if should_get_dir_auto.call(element)
 198         element.children.each(&add_dir)
 199       end
 200       add_dir.call(@document.root)
 201
 202       @document
 203     end
 204
 205     def first_image_element(element)
 206       return element if image?(element) && element.attr["src"].present?
 207
 208       element.children.find do |child|
 209         nested_image = first_image_element(child)
 210         break nested_image if nested_image
 211       end
 212     end
 213
 214     def first_truncated_text_content(element)
 215       if paragraph?(element)
 216         truncated_text_content(element)
 217       else
 218         element.children.find do |child|
 219           text = first_truncated_text_content(child)
 220           break text unless text.nil?
 221         end
 222       end
 223     end
 224
 225     def truncated_text_content(element)
 226       text = +""
 227
 228       append_text = lambda do |child|
 229         if child.type == :text
 230           text << child.value
 231         else
 232           child.children.each do |c|
 233             append_text.call(c)
 234             break if text.length > DESCRIPTION_MAX_LENGTH
 235           end
 236         end
 237       end
 238       append_text.call(element)
 239
 240       return nil if text.blank?
 241
 242       text_truncated_to_word_break = text.truncate(DESCRIPTION_MAX_LENGTH, :separator => /(?<!\s)\s+/)
 243
 244       if text_truncated_to_word_break.length >= DESCRIPTION_WORD_BREAK_THRESHOLD_LENGTH
 245         text_truncated_to_word_break
 246       else
 247         text.truncate(DESCRIPTION_MAX_LENGTH)
 248       end
 249     end
 250
 251     def image?(element)
 252       element.type == :img || (element.type == :html_element && element.value == "img")
 253     end
 254
 255     def paragraph?(element)
 256       element.type == :p || (element.type == :html_element && element.value == "p")
 257     end
 258   end
 259
 260   class Text < Base
 261     def to_html
 262       linkify(simple_format(ERB::Util.html_escape(self)))
 263     end
 264
 265     def to_text
 266       to_s
 267     end
 268   end
 269 end