DEV: Add option to keep onebox body content in post excerpt.

lib/excerpt_parser.rb

@@ -18,6 +18,7 @@ def initialize(length, options = nil)
     @keep_newlines = options[:keep_newlines] == true
     @keep_emoji_images = options[:keep_emoji_images] == true
     @keep_onebox_source = options[:keep_onebox_source] == true
+    @keep_onebox_body = options[:keep_onebox_body] == true
     @remap_emoji = options[:remap_emoji] == true
     @start_excerpt = false
     @in_details_depth = 0
@@ -34,7 +35,7 @@ def self.get_excerpt(html, length, options)
       parser.parse(html)
     end
     excerpt = me.excerpt.strip
-    excerpt = excerpt.gsub(/\s*\n+\s*/, "\n\n") if options[:keep_onebox_source]
+    excerpt = excerpt.gsub(/\s*\n+\s*/, "\n\n") if options[:keep_onebox_source] || options[:keep_onebox_body]
     excerpt = CGI.unescapeHTML(excerpt) if options[:text_entities] == true
     excerpt
   end
@@ -95,13 +96,22 @@ def start_element(name, attributes = [])
 
     when "aside"
       attributes = Hash[*attributes.flatten]
-      unless @keep_onebox_source && attributes['class'].include?('onebox')
+      unless (@keep_onebox_source || @keep_onebox_body) && attributes['class'].include?('onebox')
         @in_quote = true
       end
 
+      if @keep_onebox_body && attributes['class'].include?('quote') && attributes['data-topic'].present?
+        @in_quote = false
+      end
+
     when 'article'
-      if @keep_onebox_source && attributes.include?(['class', 'onebox-body'])
-        @in_quote = true
+      if attributes.include?(['class', 'onebox-body'])
+        @in_quote = !@keep_onebox_body
+      end
+
+    when 'header'
+      if attributes.include?(['class', 'source'])
+        @in_quote = !@keep_onebox_source
       end
 
     when "div", "span"

spec/components/excerpt_parser_spec.rb

@@ -34,4 +34,51 @@
     expect(ExcerptParser.get_excerpt(html, 3, {})).to match_html('<details class="disabled"><summary>foo</summary></details>')
     expect(ExcerptParser.get_excerpt(html, 2, {})).to match_html('<details class="disabled"><summary>fo&hellip;</summary></details>')
   end
+
+  describe "keep_onebox_body parameter" do
+    it "keeps the body content for external oneboxes" do
+      html = <<~HTML.strip
+        <aside class="onebox">
+          <header class="source">
+            <img src="https://github.githubassets.com/favicon.ico" class="site-icon" width="32" height="32">
+            <a href="https://github.com/discourse/discourse" target="_blank">GitHub</a>
+          </header>
+          <article class="onebox-body">
+            <img src="/uploads/default/original/1X/10c0f1565ee5b6ca3fe43f3183529bc0afd26003.jpeg" class="thumbnail">
+            <h3>
+              <a href="https://github.com/discourse/discourse" target="_blank">discourse/discourse</a>
+            </h3>
+            <p>A platform for community discussion. Free, open, simple. - discourse/discourse</p>
+          </article>
+        </aside>
+      HTML
+      expect(ExcerptParser.get_excerpt(html, 100, keep_onebox_body: true)).to eq(<<~HTML.strip)
+        [image]
+
+        <a href="https://github.com/discourse/discourse" target="_blank">discourse/discourse</a>
+
+        A platform for community discussion. Free, o&hellip;
+      HTML
+    end
+
+    it "keeps the content for internal oneboxes" do
+      html = <<~HTML.strip
+        <aside class="quote" data-post="1" data-topic="8">
+          <div class="title">
+            <div class="quote-controls"></div>
+            <img width="20" height="20" src="/user_avatar/localhost/system/40/2_2.png" class="avatar">
+            <a href="/t/welcome-to-discourse/8/1">Welcome to Discourse</a>
+          </div>
+          <blockquote>The first paragraph of this pinned topic will be visible as a welcome message to all new visitors on your homepage.</blockquote>
+        </aside>
+      HTML
+      expect(ExcerptParser.get_excerpt(html, 100, keep_onebox_body: true)).to eq(<<~HTML.strip)
+        [image]
+
+        <a href="/t/welcome-to-discourse/8/1">Welcome to Discourse</a>
+
+        The first paragraph of this pinned topic will be &hellip;
+      HTML
+    end
+  end
 end