1 files changed, 31 insertions, 0 deletions
diff --git a/lib/gitlab/github_import/markdown_text.rb b/lib/gitlab/github_import/markdown_text.rb
index 692016bd005..bf2856bc77f 100644
--- a/lib/gitlab/github_import/markdown_text.rb
+++ b/lib/gitlab/github_import/markdown_text.rb
@@ -1,5 +1,8 @@
 # frozen_string_literal: true
 
+# This class includes overriding Kernel#format method
+# what makes impossible to use it here
+# rubocop:disable Style/FormatString
 module Gitlab
   module GithubImport
     class MarkdownText
@@ -8,6 +11,21 @@ module Gitlab
       ISSUE_REF_MATCHER = '%{github_url}/%{import_source}/issues'
       PULL_REF_MATCHER = '%{github_url}/%{import_source}/pull'
 
+      MEDIA_TYPES = %w[gif jpeg jpg mov mp4 png svg webm].freeze
+      DOC_TYPES = %w[
+        csv docx fodg fodp fods fodt gz log md odf odg odp ods
+        odt pdf pptx tgz txt xls xlsx zip
+      ].freeze
+      ALL_TYPES = (MEDIA_TYPES + DOC_TYPES).freeze
+
+      # On github.com we have base url for docs and CDN url for media.
+      # On github EE as far as we can know there is no CDN urls and media is placed on base url.
+      # To no escape the escaping symbol we use single quotes instead of double with interpolation.
+      # rubocop:disable Style/StringConcatenation
+      CDN_URL_MATCHER = '(!\[.+\]\(%{github_media_cdn}/\d+/(\w|-)+\.(' + MEDIA_TYPES.join('|') + ')\))'
+      BASE_URL_MATCHER = '(\[.+\]\(%{github_url}/.+/.+/files/\d+/.+\.(' + ALL_TYPES.join('|') + ')\))'
+      # rubocop:enable Style/StringConcatenation
+
       class << self
         def format(*args)
           new(*args).to_s
@@ -24,8 +42,20 @@ module Gitlab
               .gsub(pull_ref_matcher, url_helpers.project_merge_requests_url(project))
         end
 
+        def fetch_attachment_urls(text)
+          cdn_url_matcher = CDN_URL_MATCHER % { github_media_cdn: Regexp.escape(github_media_cdn) }
+          doc_url_matcher = BASE_URL_MATCHER % { github_url: Regexp.escape(github_url) }
+
+          text.scan(Regexp.new(cdn_url_matcher)).map(&:first) +
+            text.scan(Regexp.new(doc_url_matcher)).map(&:first)
+        end
+
         private
 
+        def github_media_cdn
+          'https://user-images.githubusercontent.com'
+        end
+
         # Returns github domain without slash in the end
         def github_url
           oauth_config = Gitlab::Auth::OAuth::Provider.config_for('github') || {}
@@ -63,3 +93,4 @@ module Gitlab
     end
   end
 end
+# rubocop:enable Style/FormatString