related tags: fix wiki page tag extraction.

* Parse the wiki page with the actual dtext parser instead of by hand.
  This is so that wiki links inside things like [nodtext] or [code]
  blocks are handled properly.

* Only include tags that exist and are nonempty. Don't include links to
  dead pages or blank tags.

app/logical/d_text.rb

@@ -19,6 +19,18 @@ class DText
     names.uniq
   end
 
+  def self.parse_wiki_titles(text)
+    html = DTextRagel.parse(text)
+    fragment = Nokogiri::HTML.fragment(html)
+
+    titles = fragment.css("a.dtext-wiki-link").map do |node|
+      title = node["href"][%r!\A/wiki_pages/show_or_new\?title=(.*)\z!i, 1]
+      title = CGI.unescape(title)
+      title = WikiPage.normalize_title(title)
+      title
+    end
+  end
+
   def self.strip_blocks(string, tag)
     n = 0
     stripped = ""

app/logical/related_tag_query.rb

@@ -42,11 +42,7 @@ class RelatedTagQuery
   end
 
   def wiki_page_tags
-    results = wiki_page.try(:tags) || []
+    wiki_page.try(:tags) || []
-    results.reject! do |name|
-      name =~ /^(?:list_of_|tag_group|pool_group|howto:|about:|help:|template:)/
-    end
-    results
   end
 
   def other_wiki_pages

app/models/wiki_page.rb

@@ -200,13 +200,8 @@ class WikiPage < ApplicationRecord
   end
 
   def tags
-    body.scan(/\[\[(.+?)\]\]/).flatten.map do |match|
+    titles = DText.parse_wiki_titles(body)
-      if match =~ /^(.+?)\|(.+)/
+    Tag.nonempty.where(name: titles.uniq).pluck(:name)
-        $1
-      else
-        match
-      end
-    end.map {|x| x.mb_chars.downcase.tr(" ", "_").to_s}.uniq
   end
 
   def visible?