FIX: Wikipedia onebox images and sections (#31384)

Both article images and section extraction were
not working for wikipedia oneboxes, this commit
fixes both and updates our spec fixture responses
to use the new HTML

lib/onebox/engine/wikipedia_onebox.rb

@@ -30,14 +30,18 @@ module Onebox
         if m_url_hash.nil? # no hash found in url
           paras = raw.search("p") # default get all the paras
         else
-          section_header_title = raw.xpath("//span[@id='#{CGI.unescape(m_url_hash_name)}']")
+          section_header_title =
+            raw.xpath(
+              "//*[@id=\"#{CGI.unescape(m_url_hash_name)}\"][self::h1 or self::h2 or self::h3 or self::h4 or self::h5 or self::h6]",
+            )
 
           if section_header_title.empty?
             paras = raw.search("p") # default get all the paras
           else
             section_title_text = section_header_title.inner_text
-            section_header = section_header_title[0].parent # parent element of the section span element should be an <h3> node
-            cur_element = section_header
+
+            # Get .mw-heading which wraps the h* element
+            cur_element = section_header_title[0].parent
 
             # p|text|div covers the general case. We assume presence of at least 1 P node. if section has no P node we may end up with a P node from the next section.
             # div tag is commonly used as an assets wraper in an article section. often as the first element holding an image.
@@ -95,7 +99,7 @@ module Onebox
           description: text,
         }
 
-        img = raw.css(".image img")
+        img = raw.css(".infobox-image img")
 
         if img && img.size > 0
           img.each do |i|

spec/lib/onebox/engine/wikipedia_onebox_spec.rb

@@ -21,7 +21,7 @@ RSpec.describe Onebox::Engine::WikipediaOnebox do
     end
 
     it "includes summary" do
-      expect(html).to include("Billy Jack is a 1971 action/drama")
+      expect(html).to include("Billy Jack is a 1971 American action drama independent film")
     end
   end
 
@@ -34,9 +34,7 @@ RSpec.describe Onebox::Engine::WikipediaOnebox do
   end
 
   describe "url with url-encoded section hash" do
-    let(:wp_link) do
-      "https://fr.wikipedia.org/wiki/Th%C3%A9ologie#La_th%C3%A9ologie_selon_Aristote"
-    end
+    let(:wp_link) { "https://fr.wikipedia.org/wiki/Th%C3%A9ologie#L'ontoth%C3%A9ologie" }
 
     before do
       stub_request(:get, "https://fr.wikipedia.org/wiki/Th%C3%A9ologie").to_return(
@@ -46,7 +44,7 @@ RSpec.describe Onebox::Engine::WikipediaOnebox do
     end
 
     it "includes summary" do
-      expect(html).to include("Le terme est repris par")
+      expect(html).to include("investigation rationnelle sur les substances divines")
     end
   end