title parse improvement

lib/pleroma/web/rich_media/parsers/meta_tags_parser.ex

@@ -48,6 +48,6 @@ defp maybe_put_title(meta, html) when meta != %{} do
   defp maybe_put_title(meta, _), do: meta
 
   defp get_page_title(html) do
-    Floki.find(html, "title") |> List.first() |> Floki.text()
+    Floki.find(html, "html head title") |> List.first() |> Floki.text()
   end
 end

test/web/rich_media/parsers/twitter_card_test.exs

@@ -85,4 +85,19 @@ test "respect only first title tag on the page" do
                 image: image_path
               }}
   end
+
+  test "takes first founded title in html head if there is html markup error" do
+    html = File.read!("test/fixtures/nypd-facial-recognition-children-teenagers4.html")
+
+    assert TwitterCard.parse(html, %{}) ==
+             {:ok,
+              %{
+                site: nil,
+                title:
+                  "She Was Arrested at 14. Then Her Photo Went to a Facial Recognition Database. - The New York Times",
+                "app:id:googleplay": "com.nytimes.android",
+                "app:name:googleplay": "NYTimes",
+                "app:url:googleplay": "nytimes://reader/id/100000006583622"
+              }}
+  end
 end