FEATURE: Add a onebox_locale site setting. (#30655)

Following on from f369db5ae9, this change adds the ability to choose a custom locale to send to onebox providers.

If this setting is left blank, it will fall back to using default_locale.

app/models/onebox_locale_site_setting.rb

@@ -0,0 +1,11 @@
+# frozen_string_literal: true
+
+# onebox_locale is just like any other locale setting, except it allows for an empty value,
+# which is used to indicate that the default_locale should be used for onebox_locale.
+class OneboxLocaleSiteSetting < LocaleSiteSetting
+  def self.valid_value?(val)
+    supported_locales.include?(val) || val == ""
+  end
+
+  @lock = Mutex.new
+end

config/locales/server.en.yml

@@ -1803,6 +1803,7 @@ en:
     block_onebox_on_redirect: "Prevent oneboxing for URLs that lead to a redirecting page. This configuration stops the creation of a visual card (onebox) for any URL that redirects to a different destination, ensuring that direct, non-redirecting URLs are prioritized for oneboxing."
     allowed_inline_onebox_domains: "A list of domains that will be oneboxed in miniature form if linked without a title"
     enable_inline_onebox_on_all_domains: "Ignore allowed_inline_onebox_domains site setting and allow inline onebox on all domains."
+    onebox_locale: "The locale sent to onebox providers. If left blank, the default locale will be used."
     force_custom_user_agent_hosts: "Hosts for which to use the custom onebox user agent on all requests. (Especially useful for hosts that limit access by user agent)."
     max_oneboxes_per_post: "Set the maximum number of oneboxes that can be included in a single post. Oneboxes provide a preview of linked content within the post."
     facebook_app_access_token: "A token generated from your Facebook app ID and secret. Used to generate Instagram oneboxes."
@@ -3021,7 +3022,6 @@ en:
       delete_user_max_post_age: ""
       delete_user_self_max_post_count: ""
       desktop_category_page_style: ""
-      mobile_category_page_style: ""
       detailed_404: ""
       detect_custom_avatars: ""
       digest_logo: ""
@@ -3214,11 +3214,11 @@ en:
       here_mention_allowed_groups: ""
       hidden_post_visible_groups: ""
       hide_email_address_taken: ""
+      hide_new_user_profiles: ""
       hide_post_sensitivity: ""
       hide_suspension_reasons: ""
       hide_user_activity_tab: ""
       hide_user_profiles_from_public: ""
-      hide_new_user_profiles: ""
       high_trust_flaggers_auto_hide_posts: ""
       highlighted_languages: ""
       history_hours_high: ""
@@ -3356,6 +3356,7 @@ en:
       min_trust_to_send_messages: ""
       min_username_length: ""
       minimum_topics_similar: ""
+      mobile_category_page_style: ""
       mobile_logo: ""
       mobile_logo_dark: ""
       moderators_change_post_ownership: ""
@@ -3387,6 +3388,7 @@ en:
       num_tl3_users_to_silence_new_user: ""
       num_users_to_silence_new_user: ""
       old_post_notice_days: ""
+      onebox_locale: ""
       opengraph_image: ""
       page_loading_indicator: ""
       password_unique_characters: ""

config/site_settings.yml

@@ -2284,6 +2284,10 @@ onebox:
   onebox_user_agent:
     default: ""
     hidden: true
+  onebox_locale:
+    default: ""
+    type: enum
+    enum: "OneboxLocaleSiteSetting"
   force_custom_user_agent_hosts:
     default: "http://codepen.io"
     type: list

lib/onebox/helpers.rb

@@ -41,10 +41,11 @@ module Onebox
         canonical_uri = Addressable::URI.parse(canonical_link)
         if canonical_link && canonical_uri &&
              "#{canonical_uri.host}#{canonical_uri.path}" != "#{uri.host}#{uri.path}"
-          canonical_options = Oneboxer.get_final_destination_options(canonical_link)
+          uri =
-          canonical_options["extra_headers"] = { "Accept-Language" => accept_language }
+            FinalDestination.new(
-
+              canonical_link,
-          uri = FinalDestination.new(canonical_link, canonical_options).resolve
+              Oneboxer.get_final_destination_options(canonical_link),
+            ).resolve
           if uri.present?
             response =
               (
@@ -104,7 +105,7 @@ module Onebox
         headers ||= {}
 
         headers["User-Agent"] ||= user_agent if user_agent
-        headers["Accept-Language"] ||= accept_language if accept_language
+        headers["Accept-Language"] ||= Oneboxer.accept_language
 
         request = Net::HTTP::Get.new(uri.request_uri, headers)
         start_time = Time.now
@@ -236,14 +237,6 @@ module Onebox
       user_agent
     end
 
-    def self.accept_language
-      if SiteSetting.default_locale == "en"
-        "en;q=0.9, *;q=0.5"
-      else
-        "#{SiteSetting.default_locale.gsub(/_/, "-")};q=0.9, en;q=0.8, *;q=0.5"
-      end
-    end
-
     # Percent-encodes a URI string per RFC3986 - https://tools.ietf.org/html/rfc3986
     def self.uri_encode(url)
       return "" unless url

lib/oneboxer.rb

@@ -152,7 +152,7 @@ module Oneboxer
   end
 
   def self.redis_cached_response_body_key(uri)
-    "CACHED_RESPONSE_#{SiteSetting.default_locale}_#{uri}"
+    "CACHED_RESPONSE_#{onebox_locale}_#{uri}"
   end
 
   # Parse URLs out of HTML, returning the document when finished.
@@ -677,6 +677,9 @@ module Oneboxer
       force_custom_user_agent_hosts: force_custom_user_agent_hosts,
       preserve_fragment_url_hosts: preserve_fragment_url_hosts,
       timeout: 5,
+      headers: {
+        "Accept-Language" => accept_language,
+      },
     }
 
     uri = URI(url)
@@ -705,4 +708,16 @@ module Oneboxer
 
     fd_options
   end
+
+  def self.onebox_locale
+    SiteSetting.onebox_locale.presence || SiteSetting.default_locale
+  end
+
+  def self.accept_language
+    if onebox_locale == "en"
+      "en;q=0.9, *;q=0.5"
+    else
+      "#{onebox_locale.gsub(/_/, "-")};q=0.9, en;q=0.8, *;q=0.5"
+    end
+  end
 end

spec/lib/oneboxer_spec.rb

@@ -957,11 +957,18 @@ RSpec.describe Oneboxer do
     end
 
     it "separates cache by default_locale" do
-      preview = Oneboxer.preview(url, invalidate_oneboxes: true)
+      Oneboxer.preview(url, invalidate_oneboxes: true)
       expect(Oneboxer.cached_response_body_exists?(url)).to eq(true)
       SiteSetting.default_locale = "fr"
       expect(Oneboxer.cached_response_body_exists?(url)).to eq(false)
     end
+
+    it "separates cache by onebox_locale, when set" do
+      Oneboxer.preview(url, invalidate_oneboxes: true)
+      expect(Oneboxer.cached_response_body_exists?(url)).to eq(true)
+      SiteSetting.onebox_locale = "fr"
+      expect(Oneboxer.cached_response_body_exists?(url)).to eq(false)
+    end
   end
 
   describe "register_local_handler" do

spec/models/onebox_locale_site_setting_spec.rb

@@ -0,0 +1,17 @@
+# frozen_string_literal: true
+
+RSpec.describe OneboxLocaleSiteSetting do
+  describe ".valid_value?" do
+    it "returns true for a locale that we have translations for" do
+      expect(OneboxLocaleSiteSetting.valid_value?("en")).to eq(true)
+    end
+
+    it "returns true for an empty value" do
+      expect(OneboxLocaleSiteSetting.valid_value?("")).to eq(true)
+    end
+
+    it "returns false for a locale that we do not have translations for" do
+      expect(OneboxLocaleSiteSetting.valid_value?("swedish-chef")).to eq(false)
+    end
+  end
+end