i18n-tasks custom scanner for enums

I would like to create a custom scanner for i18n-tasks that can detect enums declared as hashes in models.

My enum declaration pattern will always be like this:

class Conversation < ActiveRecord::Base
  enum status: { active: 0, archived: 1}, _prefix: true
  enum subject: { science: 0, literature: 1, music: 2, art: 3 }, _prefix: true
end

The enums will always be declared as hashes, and will always have a numerical hash value, and will always have the option _prefix: true at the end of the declaration. There can be any number of values in the hash.

My custom scanner currently looks like this:

require 'i18n/tasks/scanners/file_scanner'
class ScanModelEnums < I18n::Tasks::Scanners::FileScanner
  include I18n::Tasks::Scanners::OccurrenceFromPosition

  # @return [Array<[absolute key, Results::Occurrence]>]
  def scan_file(path)
    text = read_file(path)
    text.scan(/enum\s([a-zA-Z]*?):\s\{.*\W(\w+):.*\}, _prefix: true$/).map do |prefix, attribute|
      occurrence = occurrence_from_position(
          path, text, Regexp.last_match.offset(0).first)
      model = File.basename(path, ".rb") #.split('/').last
      name = prefix + "_" + attribute
      ["activerecord.attributes.%s.%s" % [model, name], occurrence]
    end
  end
end

I18n::Tasks.add_scanner 'ScanModelEnums'

However this is only returning the very last element of each hash:

• activerecord.attributes.conversation.status_archived
• activerecord.attributes.conversation.subject_art

How can I return all the elements of each hash? I am wanting to see a result like this:

• activerecord.attributes.conversation.status_active
• activerecord.attributes.conversation.status_archived
• activerecord.attributes.conversation.subject_science
• activerecord.attributes.conversation.subject_literature
• activerecord.attributes.conversation.subject_music
• activerecord.attributes.conversation.subject_art

For reference, the i18n-tasks github repo offers an example of a custom scanner.

The file scanner class that it uses can be found here.

This works:

def scan_file(path)
  result = []
  text = read_file(path)

  text.scan(/enum\s([a-zA-Z]*?):\s\{(.*)}, _prefix: true$/).each do |prefix, body|
    occurrence = occurrence_from_position(path, text, 
                                            Regexp.last_match.offset(0).first)

    body.scan(/(\w+):/).flatten.each do |attr|
      model = File.basename(path, ".rb")
      name = "#{prefix}_#{attr}" 
      result << ["activerecord.attributes.#{model}.#{name}", occurrence]
    end
  end
  result
end

It's similar to your 'answer' approach, but uses the regex to get all the contents between '{...}', and then uses another regex to grab each enum key name.

The probable reason your 'answer' version raises an error is that it is actually returning a three-dimensional array, not two:

1. The outer .map is an array of all iterations.
2. Each iteration returns retval, which is an array.
3. Each element of retail is an array of ['key', occurrence] pairs.