# frozen_string_literal: true

# Copyright (C) 2005-2006 Laurent Sansonetti
# Copyright (C) 2007 Laurent Sansonetti and Marco Costantini
# Copyright (C) 2014, 2016 Matijs van Zuijlen
#
# Alexandria is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License as
# published by the Free Software Foundation; either version 2 of the
# License, or (at your option) any later version.
#
# Alexandria is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
# General Public License for more details.
#
# You should have received a copy of the GNU General Public
# License along with Alexandria; see the file COPYING.  If not,
# write to the Free Software Foundation, Inc., 51 Franklin Street,
# Fifth Floor, Boston, MA 02110-1301 USA.

## $Z3950_DEBUG = $DEBUG

require 'zoom'
require 'alexandria/book_providers/pseudomarc'
require 'marc'

module Alexandria
  class BookProviders
    class Z3950Provider < AbstractProvider
      include Logging
      include GetText
      GetText.bindtextdomain(Alexandria::TEXTDOMAIN, charset: 'UTF-8')

      def initialize(name = 'Z3950', fullname = 'Z39.50')
        super
        prefs.add('hostname', _('Hostname'), '')
        prefs.add('port', _('Port'), 7090)
        prefs.add('database', _('Database'), '')
        prefs.add('record_syntax', _('Record syntax'), 'USMARC', ['USMARC', 'UNIMARC', 'SUTRS'])
        prefs.add('username', _('Username'), '', nil, false)
        prefs.add('password', _('Password'), '', nil, false)
        prefs.add('charset', _('Charset encoding'), 'ISO-8859-1')

        # HACK : piggybacking support
        prefs.add('piggyback', 'Piggyback', true, [true, false])
        prefs.read
      end

      def search(criterion, type)
        prefs.read
        criterion = criterion.encode(prefs['charset'])

        # We only decode MARC at the moment.
        # SUTRS needs to be decoded separately, because each Z39.50 server has a
        # different one.
        raise NoResultsError unless marc?

        isbn = type == SEARCH_BY_ISBN ? criterion : nil
        criterion = Library.canonicalise_isbn(criterion) if type == SEARCH_BY_ISBN
        conn_count = type == SEARCH_BY_ISBN ? 1 : 10 # results to retrieve
        resultset = search_records(criterion, type, conn_count)
        log.debug { "total #{resultset.length}" }
        raise NoResultsError if resultset.length == 0

        results = books_from_marc(resultset, isbn)
        type == SEARCH_BY_ISBN ? results.first : results
      end

      def url(_book)
        nil
      end

      private

      def marc_to_book(marc_txt)
        begin
          marc = MARC::Record.new_from_marc(marc_txt, forgiving: true)
        rescue StandardError => ex
          log.error { ex.message }
          log.error { ex.backtrace.join("> \n") }
          begin
            marc = MARC::Record.new(marc_txt)
          rescue StandardError => ex2
            log.error { ex2.message }
            log.error { ex2.backtrace.join("> \n") }
            raise ex2
          end
        end

        log.debug {
          msg = 'Parsing MARC'
          msg += "\n title: #{marc.title}"
          msg += "\n authors: #{marc.authors.join(', ')}"
          msg += "\n isbn: #{marc.isbn}, #{isbn}"
          msg += "\n publisher: #{marc.publisher}"
          msg += "\n publish year: #{marc.publish_year}" if marc.respond_to?(:publish_year)
          msg += "\n edition: #{marc.edition}"
          msg
        }

        return if marc.title.nil? # or marc.authors.empty?

        (isbn = isbn) || marc.isbn
        isbn = Library.canonicalise_ean(isbn)

        book = Book.new(marc.title, marc.authors,
                        isbn,
                        (marc.publisher || ''),
                        marc.respond_to?(:publish_year) ? marc.publish_year.to_i : nil,
                        (marc.edition || ''))
        book
      end

      def books_from_marc(resultset, _isbn)
        results = []
        resultset[0..9].each do |record|
          marc_txt = record.render(prefs['charset'], 'UTF-8') # (prefs['record_syntax'], 'USMARC')
          log.debug { marc_txt }
          if $DEBUG
            File.open(',marc.txt', 'wb') do |f| # DEBUG
              f.write(marc_txt)
            end
          end
          book = nil
          begin
            mappings = Alexandria::PseudoMarcParser::USMARC_MAPPINGS
            mappings = Alexandria::PseudoMarcParser::BNF_FR_MAPPINGS if prefs['hostname'] == 'z3950.bnf.fr'
            # try pseudo-marc parser first (it seems to have more luck)
            book = Alexandria::PseudoMarcParser.marc_text_to_book(marc_txt,
                                                                  mappings)
            if book.nil?
              # failing that, try the genuine MARC parser
              book = marc_to_book(marc_txt)
            end
          rescue StandardError => ex
            log.warn { ex }
            log.warn { ex.backtrace }
          end

          results << [book] unless book.nil?
        end
        results
      end

      def marc?
        /MARC$/.match(prefs['record_syntax'])
      end

      def search_records(criterion, type, conn_count)
        options = {}
        unless prefs['username'].empty? || prefs['password'].empty?
          options['user'] = prefs['username']
          options['password'] = prefs['password']
        end
        hostname = prefs['hostname']
        port = prefs['port'].to_i
        log.debug { "hostname #{hostname} port #{port} options #{options}" }
        conn = ZOOM::Connection.new(options).connect(hostname, port)
        conn.database_name = prefs['database']

        # HACK: turn off piggybacking, just to see CMcG
        # #conn.piggyback = false

        conn.preferred_record_syntax = prefs['record_syntax']
        conn.element_set_name = 'F'
        conn.count = conn_count
        attr = case type
               when SEARCH_BY_ISBN     then [7]
               when SEARCH_BY_TITLE    then [4]
               when SEARCH_BY_AUTHORS  then [1, 1003]
               when SEARCH_BY_KEYWORD  then [1016]
               end
        pqf = ''
        attr.each { |att| pqf += "@attr 1=#{att} " }
        pqf += '"' + criterion.upcase + '"'
        log.debug { "pqf is #{pqf}, syntax #{prefs['record_syntax']}" }

        begin
          if prefs.variable_named('piggyback')
            unless prefs['piggyback']
              log.debug { 'setting conn.piggyback to false' }
              conn.piggyback = false
            end
          end
          conn.search(pqf)
        rescue StandardError => ex
          if /1005/ =~ ex.message
            if prefs.variable_named('piggyback') && prefs['piggyback']
              log.error { "Z39.50 search failed:: #{ex.message}" }
              log.info { 'Turning off piggybacking for this provider' }
              prefs.variable_named('piggyback').new_value = false
              search_records(criterion, type, conn_count)
              # hopefully these precautions will prevent infinite loops here
            else
              raise ex
            end
          else
            raise ex
          end
        end
      end
    end

    class LOCProvider < Z3950Provider
      # http://en.wikipedia.org/wiki/Library_of_Congress
      unabstract

      include GetText
      GetText.bindtextdomain(Alexandria::TEXTDOMAIN, charset: 'UTF-8')

      def initialize
        super('LOC', _('Library of Congress (Usa)'))
        prefs.variable_named('hostname').default_value = 'z3950.loc.gov'
        prefs.variable_named('port').default_value = 7090
        prefs.variable_named('database').default_value = 'Voyager'
        prefs.variable_named('record_syntax').default_value = 'USMARC'
        prefs.variable_named('charset').default_value = 'ISO-8859-1'
        prefs.read
      end

      def url(book)
        'http://catalog.loc.gov/cgi-bin/Pwebrecon.cgi?DB=local&CNT=25+records+per+page&CMD=isbn+' +
          Library.canonicalise_isbn(book.isbn)
      rescue StandardError => ex
        log.warn { "Cannot create url for book #{book}; #{ex.message}" }
        nil
      end
    end

    class BLProvider < Z3950Provider
      # http://en.wikipedia.org/wiki/Copac
      # http://en.wikipedia.org/wiki/British_Library
      # http://www.bl.uk/catalogues/z3950fullaccess.html
      # http://www.bl.uk/catalogues/z3950copacaccess.html
      #
      # FIXME: switch from BL to Copac, which incudes the BL itself and many more
      # libraries: http://copac.ac.uk/libraries/
      #
      # Details: http://copac.ac.uk/interfaces/z39.50/
      # The SUTRS format used by Copac is different from the one used by BL
      unabstract

      include GetText
      GetText.bindtextdomain(Alexandria::TEXTDOMAIN, charset: 'UTF-8')

      def initialize
        super('BL', _('British Library'))
        prefs.variable_named('hostname').default_value = 'z3950cat.bl.uk'
        prefs.variable_named('port').default_value = 9909
        prefs.variable_named('database').default_value = 'BLAC'
        prefs.variable_named('record_syntax').default_value = 'SUTRS'
        prefs.variable_named('charset').default_value = 'ISO-8859-1'
        prefs.read
      end

      def search(criterion, type)
        return super unless prefs['record_syntax'] == 'SUTRS'

        prefs.read
        criterion = Library.canonicalise_isbn(criterion) if type == SEARCH_BY_ISBN
        conn_count = type == SEARCH_BY_ISBN ? 1 : 10 # results to retrieve
        resultset = search_records(criterion, type, conn_count)
        log.debug { "total #{resultset.length}" }
        raise NoResultsError if resultset.length == 0

        results = books_from_sutrs(resultset)
        type == SEARCH_BY_ISBN ? results.first : results
      end

      def url(book)
        'http://copac.ac.uk/openurl?isbn=' + Library.canonicalise_isbn(book.isbn)
      rescue StandardError => ex
        log.warn { "Cannot create url for book #{book}; #{ex.message}" }
        nil
      end

      private

      def books_from_sutrs(resultset)
        results = []
        resultset[0..9].each do |record|
          text = record.render(prefs['charset'], 'UTF-8')
          # File.open(',bl.marc', 'wb') {|f| f.write(text) }
          log.debug { text }

          title = isbn = publisher = publish_year = edition = nil
          authors = []

          text.split(/\n/).each do |line|
            if (md = /^Title:\s+(.*)$/.match(line))
              title = md[1].sub(/\.$/, '').squeeze(' ')
            elsif (md = /^Added Person Name:\s+(.*),[^,]+$/.match(line))
              authors << md[1]
            elsif (md = /^ME-Personal Name:\s+(.*),[^,]+$/.match(line))
              authors << md[1]
            elsif (md = /^ISBN:\s+([\dXx]+)/.match(line))
              isbn = Library.canonicalise_ean(md[1])
            elsif (md = /^Imprint:.+\:\s*(.+)\,/.match(line))
              publisher = md[1]
            end
          end

          log.debug {
            msg = 'Parsing SUTRS'
            msg += "\n title: #{title}"
            msg += "\n authors: #{authors.join(' and ')}"
            msg += "\n isbn: #{isbn}"
            msg += "\n publisher: #{publisher}"
            msg += "\n edition: #{edition}"
            msg
          }

          if title # and !authors.empty?
            book = Book.new(title, authors, isbn, (publisher || nil),
                            (publish_year || nil), (edition || nil))
            results << [book]
          end
        end
        results
      end
    end

    class SBNProvider < Z3950Provider
      # http://sbnonline.sbn.it/
      # http://it.wikipedia.org/wiki/ICCU
      unabstract

      include GetText
      GetText.bindtextdomain(Alexandria::TEXTDOMAIN, charset: 'UTF-8')

      def initialize
        super('SBN', 'Servizio Bibliotecario Nazionale (Italy)')
        prefs.variable_named('hostname').default_value = 'opac.sbn.it'
        prefs.variable_named('port').default_value = 3950
        prefs.variable_named('database').default_value = 'nopac'
        # supported 'USMARC', 'UNIMARC' , 'SUTRS'
        prefs.variable_named('record_syntax').default_value = 'USMARC'
        prefs.variable_named('charset').default_value = 'ISO-8859-1'
        prefs.read
      end

      def search(criterion, type)
        prefs.read

        isbn = type == SEARCH_BY_ISBN ? criterion : nil
        criterion = canonicalise_isbn_with_dashes(criterion)
        resultset = search_records(criterion, type, 0)
        log.debug { "total #{resultset.length}" }
        raise NoResultsError if resultset.length == 0

        results = books_from_marc(resultset, isbn)
        type == SEARCH_BY_ISBN ? results.first : results
      end

      def url(book)
        'http://sbnonline.sbn.it/cgi-bin/zgw/BRIEF.pl?displayquery=' \
          '%253CB%253E%253Cfont%2520color%253D%2523000064%253E' \
          'Codice%2520ISBN%253C%2FB%253E%253C%2Ffont%253E%2520' \
          'contiene%2520%2522%2520%253CFONT%2520COLOR%253Dred%253E' +
          canonicalise_isbn_with_dashes(book.isbn) +
          '%253C%2FFONT%253E%2522&session=&zurl=opac&zquery=%281%3D7+4%3D2+2%3D3+5%3D100+6%3D1+3%3D3+%22' +
          canonicalise_isbn_with_dashes(book.isbn) +
          '%22%29&language=it&maxentries=10&target=0&position=1'
      rescue StandardError => ex
        log.warn { "Cannot create url for book #{book}; #{ex.message}" }
        nil
      end

      private

      def canonicalise_isbn_with_dashes(isbn)
        # The reference for the position of the dashes is
        # http://www.isbn-international.org/converter/ranges.htm

        isbn = Alexandria::Library.canonicalise_isbn(isbn)

        if isbn[0..1] == '88'
          # Italian speaking area
          if isbn > '8895000' && (isbn <= '8899999996')
            return isbn[0..1] + '-' + isbn[2..6] + '-' + isbn[7..8] + '-' + isbn[9..9]
          elsif isbn > '88900000'
            return isbn[0..1] + '-' + isbn[2..7] + '-' + isbn[8..8] + '-' + isbn[9..9]
          elsif isbn > '8885000'
            return isbn[0..1] + '-' + isbn[2..6] + '-' + isbn[7..8] + '-' + isbn[9..9]
          elsif isbn > '886000'
            return isbn[0..1] + '-' + isbn[2..5] + '-' + isbn[6..8] + '-' + isbn[9..9]
          elsif isbn > '88200'
            return isbn[0..1] + '-' + isbn[2..4] + '-' + isbn[5..8] + '-' + isbn[9..9]
          elsif isbn > '8800'
            return isbn[0..1] + '-' + isbn[2..3] + '-' + isbn[4..8] + '-' + isbn[9..9]
          else
            raise 'Invalid ISBN'
          end

        else
          return isbn
        end
      end
      #
      # Remarks about SBN
      #
      # This provider requires that value of conn.count is 0.
      # It's a Yaz option "Number of records to be retrieved".
      # This provider requires to specify the value of conn.element_set_name = 'F'.
      # It's a Yaz option "Element-Set name of records".
      # See http://www.indexdata.dk/yaz/doc/zoom.resultsets.tkl
      #
      # Dashes:
      # this database requires that Italian books are searched with dashes :(
      # However, they have also books with dashes in wrong positions, for instance 88-061-4934-2
      #
      # References:
      # http://opac.internetculturale.it/cgi-bin/main.cgi?type=field
      # http://www.internetculturale.it/
      # http://sbnonline.sbn.it/zgw/homeit.html
      # http://www.iccu.sbn.it/genera.jsp?id=124
      # with link at http://www.iccu.sbn.it/upload/documenti/cartecsbn.pdf
      # http://www.loc.gov/cgi-bin/zgstart?ACTION=INIT&FORM_HOST_PORT=/prod/www/data/z3950/iccu.html,opac.sbn.it,2100
      # http://gwz.cilea.it/cgi-bin/reportOpac.cgi
      #
    end
  end
end