Add semi-support for Video/Image objects in ActivityPub (#5848)

* Add semi-support for Video/Image objects in ActivityPub

Video and Image objects will create corresponding status records
with manually crafted text contents (title + URL)

* Extract html-url-finding logic into JsonLdHelper

* Fallback to id when url missing, extract supported object types

app/helpers/jsonld_helper.rb

@@ -9,6 +9,24 @@ def first_of_value(value)
     value.is_a?(Array) ? value.first : value
   end
 
+  # The url attribute can be a string, an array of strings, or an array of objects.
+  # The objects could include a mimeType. Not-included mimeType means it's text/html.
+  def url_to_href(value, preferred_type = nil)
+    single_value = if value.is_a?(Array) && !value.first.is_a?(String)
+                     value.find { |link| preferred_type.nil? || ((link['mimeType'].presence || 'text/html') == preferred_type) }
+                   elsif value.is_a?(Array)
+                     value.first
+                   else
+                     value
+                   end
+
+    if single_value.nil? || single_value.is_a?(String)
+      single_value
+    else
+      single_value['href']
+    end
+  end
+
   def as_array(value)
     value.is_a?(Array) ? value : [value]
   end

app/lib/activitypub/activity/create.rb

@@ -1,6 +1,9 @@
 # frozen_string_literal: true
 
 class ActivityPub::Activity::Create < ActivityPub::Activity
+  SUPPORTED_TYPES = %w(Article Note).freeze
+  CONVERTED_TYPES = %w(Image Video).freeze
+
   def perform
     return if delete_arrived_first?(object_uri) || unsupported_object_type?
 
@@ -41,7 +44,7 @@ def status_params
       url: object_url || @object['id'],
       account: @account,
       text: text_from_content || '',
-      language: language_from_content,
+      language: detected_language,
       spoiler_text: @object['summary'] || '',
       created_at: @options[:override_timestamps] ? nil : @object['published'],
       reply: @object['inReplyTo'].present?,
@@ -165,40 +168,62 @@ def in_reply_to_uri
   end
 
   def text_from_content
+    return Formatter.instance.linkify([text_from_name, object_url || @object['id']].join(' ')) if converted_object_type?
+
     if @object['content'].present?
       @object['content']
-    elsif language_map?
+    elsif content_language_map?
       @object['contentMap'].values.first
     end
   end
 
-  def language_from_content
-    return LanguageDetector.instance.detect(text_from_content, @account) unless language_map?
-    @object['contentMap'].keys.first
+  def text_from_name
+    if @object['name'].present?
+      @object['name']
+    elsif name_language_map?
+      @object['nameMap'].values.first
+    end
+  end
+
+  def detected_language
+    if content_language_map?
+      @object['contentMap'].keys.first
+    elsif name_language_map?
+      @object['nameMap'].keys.first
+    elsif supported_object_type?
+      LanguageDetector.instance.detect(text_from_content, @account)
+    end
   end
 
   def object_url
     return if @object['url'].blank?
-
-    value = first_of_value(@object['url'])
-
-    return value if value.is_a?(String)
-
-    value['href']
+    url_to_href(@object['url'], 'text/html')
   end
 
-  def language_map?
+  def content_language_map?
     @object['contentMap'].is_a?(Hash) && !@object['contentMap'].empty?
   end
 
+  def name_language_map?
+    @object['nameMap'].is_a?(Hash) && !@object['nameMap'].empty?
+  end
+
   def unsupported_object_type?
-    @object.is_a?(String) || !%w(Article Note).include?(@object['type'])
+    @object.is_a?(String) || !(supported_object_type? || converted_object_type?)
   end
 
   def unsupported_media_type?(mime_type)
     mime_type.present? && !(MediaAttachment::IMAGE_MIME_TYPES + MediaAttachment::VIDEO_MIME_TYPES).include?(mime_type)
   end
 
+  def supported_object_type?
+    SUPPORTED_TYPES.include?(@object['type'])
+  end
+
+  def converted_object_type?
+    CONVERTED_TYPES.include?(@object['type'])
+  end
+
   def skip_download?
     return @skip_download if defined?(@skip_download)
     @skip_download ||= DomainBlock.find_by(domain: @account.domain)&.reject_media?

app/lib/formatter.rb

@@ -51,12 +51,7 @@ def plaintext(status)
 
   def simplified_format(account)
     return reformat(account.note).html_safe unless account.local? # rubocop:disable Rails/OutputSafety
-
-    html = encode_and_link_urls(account.note)
-    html = simple_format(html, {}, sanitize: false)
-    html = html.delete("\n")
-
-    html.html_safe # rubocop:disable Rails/OutputSafety
+    linkify(account.note)
   end
 
   def sanitize(html, config)
@@ -69,6 +64,14 @@ def format_spoiler(status)
     html.html_safe # rubocop:disable Rails/OutputSafety
   end
 
+  def linkify(text)
+    html = encode_and_link_urls(text)
+    html = simple_format(html, {}, sanitize: false)
+    html = html.delete("\n")
+
+    html.html_safe # rubocop:disable Rails/OutputSafety
+  end
+
   private
 
   def encode(html)

app/services/activitypub/fetch_remote_status_service.rb

@@ -42,7 +42,7 @@ def supported_context?
   end
 
   def expected_type?
-    %w(Note Article).include? @json['type']
+    (ActivityPub::Activity::Create::SUPPORTED_TYPES + ActivityPub::Activity::Create::CONVERTED_TYPES).include? @json['type']
   end
 
   def needs_update(actor)

app/services/activitypub/process_account_service.rb

@@ -107,12 +107,7 @@ def public_key
 
   def url
     return if @json['url'].blank?
-
-    value = first_of_value(@json['url'])
-
-    return value if value.is_a?(String)
-
-    value['href']
+    url_to_href(@json['url'], 'text/html')
   end
 
   def outbox_total_items

spec/services/activitypub/fetch_remote_status_service_spec.rb

@@ -1,6 +1,8 @@
 require 'rails_helper'
 
 RSpec.describe ActivityPub::FetchRemoteStatusService do
+  include ActionView::Helpers::TextHelper
+
   let(:sender) { Fabricate(:account) }
   let(:recipient) { Fabricate(:account) }
   let(:valid_domain) { Rails.configuration.x.local_domain }
@@ -19,6 +21,7 @@
 
   describe '#call' do
     before do
+      stub_request(:head, 'https://example.com/watch?v=12345').to_return(status: 404, body: '')
       subject.call(object[:id], prefetched_body: Oj.dump(object))
     end
 
@@ -32,5 +35,38 @@
         expect(status.text).to eq 'Lorem ipsum'
       end
     end
+
+    context 'with Video object' do
+      let(:object) do
+        {
+          '@context': 'https://www.w3.org/ns/activitystreams',
+          id: "https://#{valid_domain}/@foo/1234",
+          type: 'Video',
+          name: 'Nyan Cat 10 hours remix',
+          attributedTo: ActivityPub::TagManager.instance.uri_for(sender),
+          url: [
+            {
+              type: 'Link',
+              mimeType: 'application/x-bittorrent',
+              href: 'https://example.com/12345.torrent',
+            },
+
+            {
+              type: 'Link',
+              mimeType: 'text/html',
+              href: 'https://example.com/watch?v=12345',
+            },
+          ],
+        }
+      end
+
+      it 'creates status' do
+        status = sender.statuses.first
+
+        expect(status).to_not be_nil
+        expect(status.url).to eq 'https://example.com/watch?v=12345'
+        expect(strip_tags(status.text)).to eq 'Nyan Cat 10 hours remix https://example.com/watch?v=12345'
+      end
+    end
   end
 end