#398

backend/app/services/preview/html_metadata_extractor.rb

@@ -0,0 +1,21 @@
+module Preview
+  class HtmlMetadataExtractor
+    IMAGE_SELECTORS = [
+      'meta[property="og:image"]',
+      'meta[name="twitter:image"]',
+      'meta[name="thumbnail"]'
+    ].freeze
+
+    def self.extract(response)
+      document = Nokogiri::HTML.parse(response.body)
+      image_url = IMAGE_SELECTORS.filter_map {
+        document.at_css(_1)&.[]('content')&.strip.presence
+      }.first
+
+      { title: document.at_css('title')&.text&.strip,
+        image_url: image_url ? URI.join(response.url, image_url).to_s : nil }
+    rescue URI::InvalidURIError
+      { title: document.at_css('title')&.text&.strip, image_url: nil }
+    end
+  end
+end