make metadata import more flexible

datacite · Dec 27, 2018 · 7e7fe5b · 7e7fe5b
1 parent a2e3f81
commit 7e7fe5b
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/app/models/concerns/crosscitable.rb b/app/models/concerns/crosscitable.rb
@@ -92,7 +92,7 @@ def well_formed_xml(string)
     end
 
     def from_xml(string)
-      return nil unless string.start_with?('<?xml version=')
+      return nil unless string.start_with?('<?xml version=') || string.start_with?('<resource ')
 
       doc = Nokogiri::XML(string) { |config| config.strict.noblanks }
       doc.to_xml.strip

diff --git a/app/models/doi.rb b/app/models/doi.rb
@@ -345,7 +345,7 @@ def self.import_one(doi_id: nil)
       return nil
     end
 
-    string = doi.current_metadata.present? && doi.current_metadata.xml.to_s.start_with?('<?xml version=') ? doi.current_metadata.xml.force_encoding("UTF-8") : nil
+    string = doi.current_metadata.present? ? doi.from_xml(doi.current_metadata.xml).to_s.force_encoding("UTF-8").presence : nil
     unless string.present?
       logger.error "[MySQL] No metadata for DOI " + doi.doi + " found: " + doi.current_metadata.inspect
       return nil
@@ -397,7 +397,7 @@ def self.import_by_day(options={})
     Doi.where(created: from_date.midnight..from_date.end_of_day).find_each do |doi|
       begin
         # ignore broken xml
-        string = doi.current_metadata.present? && doi.current_metadata.xml.to_s.start_with?('<?xml version=') ? doi.current_metadata.xml.force_encoding("UTF-8") : nil
+        string = doi.current_metadata.present? ? doi.from_xml(doi.current_metadata.xml).to_s.force_encoding("UTF-8").presence : nil
         unless string.present?
           logger.error "[MySQL] No metadata for DOI " + doi.doi + " found."
           return nil