Iterate though matched patterns in document, perform a function on each match, and write result to file

Rakefile

desc "Replace email addresses in remarks with md5 hashed strings"
task :hash do |t|
  FileList.new('source/_data/remarks/*.yml').each do |path|
    File.open(path, 'r+:utf-8') do |file_name|

      require 'digest/md5'
      private :hash
      def hash(email)
        email_address = email ? email.downcase.strip : ''
        Digest::MD5.hexdigest(email_address)
      end

      contents = File.read(file_name)
      email_regex = /([_a-z0-9-]+(\.[_a-z0-9-]+)*@[a-z0-9-]+(\.[a-z0-9-]+)*(\.[a-z]{2,4}))/i
      @replace = contents.gsub(email_regex) { |m| hash(m) }

      File.open(file_name, "w") { |file| file.puts @replace }
    end
  end
end

trip-to-sf.yml

- author: John Doe
  email: [email protected]
  date: '2014-09-30 10:44:46 +0100'
  url: example.com
  content: |
    When are you going to San Francisco? I think there is a chance that I will be around there too!

- author: Mary Smith
  email: [email protected]
  date: '2014-09-30 10:57:27 +0100'
  url: example.com
  content: |
    I'll be there in August.

- author: David Smith
  email: [email protected]
  date: '2014-10-01 00:47:05 +0100'
  content: |
    I'll be around. Fancy meeting up?

With your current regex:

contents.scan(@email_regex).map{|m| m[0]}.each {|m| contents.gsub! m, hash(m)}

File.open(file_name, "w") { |file| file.puts contents}

You could lose the map by removing all the capture groups from the regex.

Actually that scan is dumb, just

contents.gsub!(@email_regex) {|m| hash m }