improvements to the mbox import script

* ignores dot-files and empty emails
* new setting to prefer HTML over plaintext emails during import
* restore original site settings at the end of import
* elided content of HTML mails was not put inside details block

script/import_scripts/mbox/support/indexer.rb

@@ -102,10 +102,12 @@ module ImportScripts::Mbox
 
         if @split_regex.present?
           each_mail(filename) do |raw_message, first_line_number, last_line_number|
-            yield read_mail_from_string(raw_message), filename, first_line_number, last_line_number
+            receiver = read_mail_from_string(raw_message)
+            yield receiver, filename, first_line_number, last_line_number if receiver.present?
           end
         else
-          yield read_mail_from_file(filename), filename
+          receiver = read_mail_from_file(filename)
+          yield receiver, filename if receiver.present?
         end
 
         mark_as_fully_indexed(category_name, filename)
@@ -161,7 +163,7 @@ module ImportScripts::Mbox
     end
 
     def read_mail_from_string(raw_message)
-      Email::Receiver.new(raw_message)
+      Email::Receiver.new(raw_message) unless raw_message.blank?
     end
 
     def extract_reply_message_ids(mail)
@@ -208,7 +210,12 @@ module ImportScripts::Mbox
     end
 
     def ignored_file?(filename, checksums)
-      File.directory?(filename) || metadata_file?(filename) || fully_indexed?(filename, checksums)
+      File.directory?(filename) || hidden_file?(filename) ||
+        metadata_file?(filename) || fully_indexed?(filename, checksums)
+    end
+
+    def hidden_file?(filename)
+      File.basename(filename).start_with?('.')
     end
 
     def metadata_file?(filename)

script/import_scripts/mbox/support/settings.rb

@@ -11,12 +11,14 @@ module ImportScripts::Mbox
     attr_reader :split_regex
     attr_reader :batch_size
     attr_reader :trust_level
+    attr_reader :prefer_html
 
     def initialize(yaml)
       @data_dir = yaml['data_dir']
       @split_regex = Regexp.new(yaml['split_regex']) unless yaml['split_regex'].empty?
       @batch_size = 1000 # no need to make this actually configurable at the moment
       @trust_level = yaml['default_trust_level']
+      @prefer_html = yaml['prefer_html']
     end
   end
 end