# encoding: UTF-8
require 'stanford-mods/searchworks_languages'
require 'logger'
# SearchWorks specific wranglings of MODS metadata as a mixin to the Stanford::Mods::Record object
module Stanford
  module Mods

    class Record < ::Mods::Record
      
      # include langagues known to SearchWorks; try to error correct when possible (e.g. when ISO-639 disagrees with MARC standard)
      def sw_language_facet
        result = []
        @mods_ng_xml.language.each { |n| 
          # get languageTerm codes and add their translations to the result
          n.code_term.each { |ct| 
            if ct.authority.match(/^iso639/)
              begin
                vals = ct.text.split(/[,|\ ]/).reject {|x| x.strip.length == 0 } 
                vals.each do |v|
                  iso639_val = ISO_639.find(v.strip).english_name
                  if SEARCHWORKS_LANGUAGES.has_value?(iso639_val)
                    result << iso639_val
                  else
                    result << SEARCHWORKS_LANGUAGES[v.strip]
                  end
                end
              rescue => e
                # TODO:  this should be written to a logger
                p "Couldn't find english name for #{ct.text}"
                result << SEARCHWORKS_LANGUAGES[v.strip]
              end
            else
              vals = ct.text.split(/[,|\ ]/).reject {|x| x.strip.length == 0 } 
              vals.each do |v|
                result << SEARCHWORKS_LANGUAGES[v.strip]
              end
            end
          }
          # add languageTerm text values
          n.text_term.each { |tt| 
            val = tt.text.strip
            result << val if val.length > 0 && SEARCHWORKS_LANGUAGES.has_value?(val)
          }

          # add language values that aren't in languageTerm subelement
          if n.languageTerm.size == 0
            result << n.text if SEARCHWORKS_LANGUAGES.has_value?(n.text)
          end
        }
        result.uniq
      end # language_facet
      
      
      # ---- AUTHOR ----
      
      # @return [String] value for author_1xx_search field
      def sw_main_author
        main_author_w_date
      end
      
      # @return [Array<String>] values for author_7xx_search field
      def sw_addl_authors
        additional_authors_w_dates
      end
      
      # @return [Array<String>] values for author_person_facet, author_person_display
      def sw_person_authors
        personal_names_w_dates
      end
      
      # return the display_value_w_date for all <mods><name> elements that do not have type='personal'
      # @return [Array<String>] values for author_other_facet
      def sw_impersonal_authors
        @mods_ng_xml.plain_name.select {|n| n.type_at != 'personal'}.map { |n| n.display_value_w_date }
      end
      
      # @return [Array<String>] values for author_corp_display
      def sw_corporate_authors
        val=@mods_ng_xml.plain_name.select {|n| n.type_at == 'corporate'}.map { |n| n.display_value_w_date }
        val
      end
      
      # @return [Array<String>] values for author_meeting_display
      def sw_meeting_authors
        @mods_ng_xml.plain_name.select {|n| n.type_at == 'conference'}.map { |n| n.display_value_w_date }
      end
      
      # Returns a sortable version of the main_author:
      #  main_author + sorting title
      # which is the mods approximation of the value created for a marc record
      # @return [String] value for author_sort field
      def sw_sort_author
        #  substitute java Character.MAX_CODE_POINT for nil main_author so missing main authors sort last
        val = '' + (main_author_w_date ? main_author_w_date : "\u{FFFF} ") + ( sort_title ? sort_title : '')
        val.gsub(/[[:punct:]]*/, '').strip
      end
      
      # ---- TITLE ----

      # @return [String] value for title_245a_search field
      def sw_short_title
        short_titles ? short_titles.first : nil
      end
      
      # @return [String] value for title_245_search, title_display, title_full_display
      def sw_full_title
        toret = full_titles ? full_titles.find { |s| s =~ Regexp.new(Regexp.escape(sw_short_title)) } : nil
        if toret
          toret = toret.gsub(/,$/, '')
        end
        toret
      end
      
      # this includes all titles except 
      # @return [Array<String>] values for title_variant_search
      def sw_addl_titles
        full_titles.select { |s| s !~ Regexp.new(Regexp.escape(sw_short_title)) }
      end
      
      # Returns a sortable version of the main title
      # @return [String] value for title_sort field
      def sw_sort_title
        val = '' + ( sort_title ? sort_title : '')
        val.gsub(/[[:punct:]]*/, '').strip
      end
      
      # ---- SUBJECT ----
      
      # Values are the contents of:
      #   subject/geographic
      #   subject/hierarchicalGeographic
      #   subject/geographicCode  (only include the translated value if it isn't already present from other mods geo fields)
      # @param [String] sep - the separator string for joining hierarchicalGeographic sub elements
      # @return [Array<String>] values for geographic_search Solr field for this document or [] if none
      def sw_geographic_search(sep = ' ')
        result = term_values([:subject, :geographic]) || []
        
        # hierarchicalGeographic has sub elements
        @mods_ng_xml.subject.hierarchicalGeographic.each { |hg_node|  
          hg_vals = []
          hg_node.element_children.each { |e| 
            hg_vals << e.text unless e.text.empty?
          }
          result << hg_vals.join(sep) unless hg_vals.empty?
        }

        trans_code_vals = @mods_ng_xml.subject.geographicCode.translated_value
        if trans_code_vals
          trans_code_vals.each { |val|  
            result << val if !result.include?(val)
          }
        end

        result    
      end
      
      # Values are the contents of:
      #   subject/name/namePart
      #  "Values from namePart subelements should be concatenated in the order they appear (e.g. "Shakespeare, William, 1564-1616")"
      # @param [String] sep - the separator string for joining namePart sub elements
      # @return [Array<String>] values for names inside subject elements or [] if none
      def sw_subject_names(sep = ', ')
        result = []
        @mods_ng_xml.subject.name_el.select { |n_el| n_el.namePart }.each { |name_el_w_np|  
          parts = name_el_w_np.namePart.map { |npn| npn.text unless npn.text.empty? }.compact
          result << parts.join(sep).strip unless parts.empty?
        }
        result
      end
      
      # Values are the contents of:
      #   subject/titleInfo/(subelements)
      # @param [String] sep - the separator string for joining titleInfo sub elements
      # @return [Array<String>] values for titles inside subject elements or [] if none
      def sw_subject_titles(sep = ' ')
        result = []
        @mods_ng_xml.subject.titleInfo.each { |ti_el|
          parts = ti_el.element_children.map { |el| el.text unless el.text.empty? }.compact
          result << parts.join(sep).strip unless parts.empty?
        }
        result
      end
            
      # Values are the contents of:
      #   mods/genre
      #   mods/subject/topic
      # @return [Array<String>] values for the topic_search Solr field for this document or nil if none
      def topic_search
        @topic_search ||= begin
          vals = self.term_values(:genre) || []
          vals.concat(subject_topics) if subject_topics
          vals.empty? ? nil : vals
        end
      end
      def place
        vals = self.term_values([:origin_info,:place,:placeTerm])
        vals
      end
      def main_author_w_date_test
        result = nil
        first_wo_role = nil
        self.plain_name.each { |n|
          if n.role.size == 0
            first_wo_role ||= n
          end
          n.role.each { |r|
            if r.authority.include?('marcrelator') && 
              (r.value.include?('Creator') || r.value.include?('Author'))
              result ||= n.display_value_w_date
            end          
          }
        }
        if !result && first_wo_role
          result = first_wo_role.display_value_w_date
        end
        result
      end
      #remove trailing commas
      def sw_full_title_without_commas
        toret = self.sw_full_title
        if toret
          toret = toret.gsub(/,$/, '')
        end
        toret
      end

      def sw_logger
        @logger ||= Logger.new(STDOUT)
      end
      # Values are the contents of:
      #   subject/geographic
      #   subject/hierarchicalGeographic
      #   subject/geographicCode  (only include the translated value if it isn't already present from other mods geo fields)
      # @return [Array<String>] values for the geographic_search Solr field for this document or nil if none
      def geographic_search
        @geographic_search ||= begin
          result = self.sw_geographic_search

          # TODO:  this should go into stanford-mods ... but then we have to set that gem up with a Logger
          # print a message for any unrecognized encodings
          xvals = self.subject.geographicCode.translated_value
          codes = self.term_values([:subject, :geographicCode]) 
          if codes && codes.size > xvals.size
            self.subject.geographicCode.each { |n|
              if n.authority != 'marcgac' && n.authority != 'marccountry'
                sw_logger.info("#{druid} has subject geographicCode element with untranslated encoding (#{n.authority}): #{n.to_xml}")
              end
            }
          end

          # FIXME:  stanford-mods should be returning [], not nil ... 
          return nil if !result || result.empty?
          result
        end
      end

      # Values are the contents of:
      #   subject/name
      #   subject/occupation  - no subelements
      #   subject/titleInfo
      # @return [Array<String>] values for the subject_other_search Solr field for this document or nil if none
      def subject_other_search
        @subject_other_search ||= begin
          vals = subject_occupations ? Array.new(subject_occupations) : []
          vals.concat(subject_names) if subject_names
          vals.concat(subject_titles) if subject_titles
          vals.empty? ? nil : vals
        end
      end

      # Values are the contents of:
      #   subject/temporal
      #   subject/genre
      # @return [Array<String>] values for the subject_other_subvy_search Solr field for this document or nil if none
      def subject_other_subvy_search
        @subject_other_subvy_search ||= begin
          vals = subject_temporal ? Array.new(subject_temporal) : []
          gvals = self.term_values([:subject, :genre])
          vals.concat(gvals) if gvals

          # print a message for any temporal encodings
          self.subject.temporal.each { |n| 
            sw_logger.info("#{druid} has subject temporal element with untranslated encoding: #{n.to_xml}") if !n.encoding.empty?
          }

          vals.empty? ? nil : vals
        end
      end
      # @return [Array<String>] values for the pub_date_group_facet
      def pub_date_groups year
        if not year
          return nil
        end
        year=year.to_i
        current_year=Time.new.year.to_i
        result = []
        if year >= current_year - 1
          result << "This year"
        else
          if year >= current_year - 3
            result << "Last 3 years"
          else
            if year >= current_year - 10
              result << "Last 10 years"
            else
              if year >= current_year - 50
                result << "Last 50 years"
              else
                result << "More than 50 years ago"
              end
            end
          end
        end
      end

      # select one or more format values from the controlled vocabulary here:
      #   http://searchworks-solr-lb.stanford.edu:8983/solr/select?facet.field=format&rows=0&facet.sort=index
      # based on the dor_content_type
      # @return [String] value in the SearchWorks controlled vocabulary
      def format
        val=[]
        formats = self.term_values(:typeOfResource)
        genres = self.term_values(:genre)
        issuance = self.term_values([:origin_info,:issuance])
        if formats
          formats.each do |form|
            case form
            when 'text'
              val << 'Thesis' if genres and genres.include? 'thesis'
              val << 'Book' if issuance and issuance.include? 'monographic'
              val << 'Journal/Periodical' if issuance and issuance.include? 'continuing'
              val << 'Journal/Periodical' if genres and genres.include? 'article'
            when 'still image'
              val << 'Image'
            when 'mixed material'
              val << 'Manuscript/Archive'
            when 'moving image'
              val << 'Video'
            when 'three dimensional object'
              val <<'Other'
            when 'cartographic'
              val << 'Map/Globe'
            when 'sound recording-musical'
              val << 'Music-Recording'
            when 'sound recording-nonmusical'
              val << 'Sound Recording'
            when 'software, multimedia'
              val << 'Computer File'      
            else
              sw_logger.warn "#{druid} has an unknown typeOfResource #{form}"
            end
          end
        end
        if val.length>0
          return val.uniq
        end
        if not self.typeOfResource or self.typeOfResource.length == 0
          sw_logger.warn "#{druid} has no valid typeOfResource"
          []
        end
      end

      # Values are the contents of:
      #  all subject subelements except subject/cartographic plus  genre top level element
      # @return [Array<String>] values for the subject_all_search Solr field for this document or nil if none
      def subject_all_search
        vals = topic_search ? Array.new(topic_search) : []
        vals.concat(geographic_search) if geographic_search
        vals.concat(subject_other_search) if subject_other_search
        vals.concat(subject_other_subvy_search) if subject_other_subvy_search
        vals.empty? ? nil : vals
      end
      def pub_date_display
        if pub_dates
          pub_dates.first
        else
          nil
        end
      end
      #get the dates from dateIssued, and dateCreated merged into 1 array.
      # @return [Array<String>] values for the issue_date_display Solr field for this document or nil if none
      def pub_dates
        vals = self.term_values([:origin_info,:dateIssued])
        if vals
          vals = vals.concat self.term_values([:origin_info,:dateCreated]) unless not self.term_values([:origin_info,:dateCreated])
        else
          vals = self.term_values([:origin_info,:dateCreated])
        end
        vals and vals.empty? ? nil : vals
      end
      def is_number?(object)
        true if Integer(object) rescue false
      end
      def is_date?(object)
        true if Date.parse(object) rescue false
      end
  
      # Get the publish year from mods
      #@return [String] 4 character year or nil if no valid date was found
      def pub_year
        #use the cached year if there is one
        if @pub_year
          if @pub_year == ''
            return nil
          end
          return @pub_year
        end
        dates=pub_dates
        if dates
          year=[]
          pruned_dates=[]
          dates.each do |f_date|
            #remove ? and [] 
            pruned_dates << f_date.gsub('?','').gsub('[','').gsub(']','')
          end
          #try to find a date starting with the most normal date formats and progressing to more wonky ones
          @pub_year=get_plain_four_digit_year pruned_dates
          return @pub_year if @pub_year
          @pub_year=get_double_digit_century pruned_dates
          return @pub_year if @pub_year
          @pub_year=get_bc_year pruned_dates
          return @pub_year if @pub_year
          @pub_year=get_three_digit_year pruned_dates
          return @pub_year if @pub_year
          @pub_year=get_single_digit_century pruned_dates
          return @pub_year if @pub_year
        end
        @pub_year=''
        sw_logger.info("#{druid} no valid pub date found in '#{dates.to_s}'")
        return nil
      end
      #creates a date suitable for sorting. Guarnteed to be 4 digits or nil
      def pub_date_sort
        pd=nil
        if pub_date
          pd=pub_date
          if pd.length == 3
            pd='0'+pd
          end
          pd=pd.gsub('--','00')
        end
        raise "pub_date_sort was about to return a non 4 digit value #{pd}!" if pd and pd.length !=4 
        pd
      end
      #The year the object was published, , filtered based on max_pub_date and min_pub_date from the config file
      #@return [String] 4 character year or nil
      def pub_date
        val=pub_year
        if val
          return val
        end
        nil
      end
      #Values for the pub date facet. This is less strict than the 4 year date requirements for pub_date
      #@return <Array[String]> with values for the pub date facet
      def pub_date_facet
        if pub_date
          if pub_date.start_with?('-')
            return (pub_date.to_i + 1000).to_s + ' B.C.'
          end
          if pub_date.include? '--'
            cent=pub_date[0,2].to_i
            cent+=1
            cent=cent.to_s+'th century'
            return cent
          else
            return pub_date
          end
        else
          nil
        end
      end
      

      # Values are the contents of:
      #   subject/topic
      #   subject/name
      #   subject/title
      #   subject/occupation
      #  with trailing comma, semicolon, and backslash (and any preceding spaces) removed
      # @return [Array<String>] values for the topic_facet Solr field for this document or nil if none 
      def topic_facet
        vals = subject_topics ? Array.new(subject_topics) : []
        vals.concat(subject_names) if subject_names
        vals.concat(subject_titles) if subject_titles
        vals.concat(subject_occupations) if subject_occupations
        vals.map! { |val| 
          v = val.sub(/[\\,;]$/, '')
          v.strip
        }
        vals.empty? ? nil : vals
      end

      # geographic_search values with trailing comma, semicolon, and backslash (and any preceding spaces) removed
      # @return [Array<String>] values for the geographic_facet Solr field for this document or nil if none 
      def geographic_facet
        geographic_search.map { |val| val.sub(/[\\,;]$/, '').strip } unless !geographic_search
      end

      # subject/temporal values with trailing comma, semicolon, and backslash (and any preceding spaces) removed
      # @return [Array<String>] values for the era_facet Solr field for this document or nil if none 
      def era_facet
        subject_temporal.map { |val| val.sub(/[\\,;]$/, '').strip } unless !subject_temporal
      end
      # @return [String] value with the numeric catkey in it, or nil if none exists
      def catkey
        catkey=self.term_values([:record_info,:recordIdentifier])
        if catkey and catkey.length>0
          return catkey.first.gsub('a','') #need to ensure catkey is numeric only
        end
        nil
      end
      def druid= new_druid
        @druid=new_druid
      end
      def druid
        @druid ? @druid : 'Unknown item'
      end

      # protected ----------------------------------------------------------

      # convenience method for subject/name/namePart values (to avoid parsing the mods for the same thing multiple times)
      def subject_names
        @subject_names ||= self.sw_subject_names
      end

      # convenience method for subject/occupation values (to avoid parsing the mods for the same thing multiple times)
      def subject_occupations
        @subject_occupations ||= self.term_values([:subject, :occupation])
      end

      # convenience method for subject/temporal values (to avoid parsing the mods for the same thing multiple times)
      def subject_temporal
        @subject_temporal ||= self.term_values([:subject, :temporal])
      end

      # convenience method for subject/titleInfo values (to avoid parsing the mods for the same thing multiple times)
      def subject_titles
        @subject_titles ||= self.sw_subject_titles
      end

      # convenience method for subject/topic values (to avoid parsing the mods for the same thing multiple times)
      def subject_topics
        @subject_topics ||= self.term_values([:subject, :topic])
      end
  
      #get a 4 digit year like 1865 from the date array
      def get_plain_four_digit_year dates
        dates.each do |f_date|
          matches=f_date.scan(/\d{4}/)
          if matches.length == 1
            @pub_year=matches.first 
            return matches.first
          else
            #if there are multiples, check for ones with CE after them
            matches.each do |match|
              #look for things like '1865-6 CE'
              pos = f_date.index(Regexp.new(match+'...CE'))
              pos = pos ? pos.to_i : 0
              if f_date.include?(match+' CE') or pos > 0
                @pub_year=match
                return match
              end 
            end
          end
        end
        return nil
      end
  
      #get a double digit century like '12th century' from the date array
      def get_double_digit_century dates
        dates.each do |f_date|
          matches=f_date.scan(/\d{2}th/)
          if matches.length == 1
            @pub_year=((matches.first[0,2].to_i)-1).to_s+'--'
            return @pub_year
          end
          #if there are multiples, check for ones with CE after them
          if matches.length > 0
            matches.each do |match|
              pos = f_date.index(Regexp.new(match+'...CE'))
              pos = pos ? pos.to_i : f_date.index(Regexp.new(match+' century CE'))
              pos = pos ? pos.to_i : 0
              if f_date.include?(match+' CE') or pos > 0
                @pub_year=((match[0,2].to_i) - 1).to_s+'--'
                return @pub_year
              end 
            end
          end
        end
        return nil
      end
  
      #get a 3 digit year like 965 from the date array
      def get_three_digit_year dates
        dates.each do |f_date|
          matches=f_date.scan(/\d{3}/)
          if matches.length > 0
            return matches.first
          end
        end
        return nil
      end
      #get the 3 digit BC year, return it as a negative, so -700 for 300 BC. Other methods will translate it to proper display, this is good for sorting.
      def get_bc_year dates
        dates.each do |f_date|
          matches=f_date.scan(/\d{3} B.C./)
          if matches.length > 0   
            bc_year=matches.first[0..2]
            return (bc_year.to_i-1000).to_s
          end
        end
        return nil
      end
  
      #get a single digit century like '9th century' from the date array
      def get_single_digit_century dates
        dates.each do |f_date|
          matches=f_date.scan(/\d{1}th/)
          if matches.length == 1
            @pub_year=((matches.first[0,2].to_i)-1).to_s+'--'
            return @pub_year
          end
          #if there are multiples, check for ones with CE after them
          if matches.length > 0
            matches.each do |match|
              pos = f_date.index(Regexp.new(match+'...CE'))
              pos = pos ? pos.to_i : f_date.index(Regexp.new(match+' century CE'))
              pos = pos ? pos.to_i : 0
              if f_date.include?(match+' CE') or pos > 0
                @pub_year=((match[0,1].to_i) - 1).to_s+'--'
                return @pub_year
              end 
            end
          end
        end 
        return nil
      end      
    end # class Record
  end # Module Mods
end # Module Stanford