#!/usr/bin/perl

# Copyright 2009-2012, Equinox Software, Inc.
#
# This program 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.
#
# This program 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 this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301, USA.

use strict;
use warnings;
use open ':utf8';

use Getopt::Long;
use MARC::Batch;
use Unicode::Normalize;
use MARC::File::XML ( BinaryEncoding => 'utf-8' );
use Equinox::Migration::SubfieldMapper;
use Equinox::Migration::Utils qw/normalize_oclc_number/;

my $conf  = {}; # configuration hashref
my $count = 0; my $scount = 0;
my $start = time;
$| = 1;

initialize($conf);

open OF, '>', $conf->{output} or die "$0: cannot open output file $conf->{output}: $!\n";
open XF, '>', $conf->{exception} or die "$0: cannot open exception file $conf->{exception}: $!\n";

for my $file (@ARGV) {
    print XF "Processing $file\n";

    my $batch = MARC::Batch->new($conf->{marctype}, $file);
    $batch->strict_off();
    $batch->warnings_off();

    my $record;
    while ( 1 ) {
        eval { $record = $batch->next; };
        if ($@) {
            import MARC::File::XML;
            print "skipping bad record: $@\n";
            next;
        }
        last unless $record;
        $count++; progress_ticker();
        my $marc = undef;
        unless ( defined $record )
          { dump_exception($marc); next; }

        my $id = $record->field($conf->{tag});
        unless ($id) {
            print XF "ERROR: Record $count in $file is missing a ",
              $conf->{tag}, " field.\n", $record->as_formatted(), "\n=====\n";
            next;
        }

        # populate and normalize marc
        $marc = populate_marc($record, $id);
        # check for manual exclusion
        next if this_record_is_excluded($record, $marc);
        normalize_marc($marc);
        unless (marc_isvalid($marc))
          { dump_exception($marc); next; }

        # if everything looks good, score it and dump fingerprints
        score_marc($marc, $record);
        dump_fingerprints($marc);
        $scount++; progress_ticker();
    }
}

print "\nSuccessfully processed:\t$count\n" unless $conf->{quiet};

=head2 populate_marc

Constructs a hash containing the relevant MARC data for a record and
returns a reference to it.

=cut

sub populate_marc {
    my ($record, $id) = @_;
    my %marc = (); $marc{isbns} = [];

    # id, stringified
    $marc{id} = $id->as_string($conf->{subfield});

    # record_type, bib_lvl
    $marc{record_type} = substr($record->leader, 6, 1);
    $marc{bib_lvl}     = substr($record->leader, 7, 1);

    # date1, date2
    my $my_008 = $record->field('008');
    $marc{tag008} = $my_008->as_string() if ($my_008);
    if (defined $marc{tag008}) {
        unless (length $marc{tag008} == 40) {
            $marc{tag008} = $marc{tag008} . ('|' x (40 - length($marc{tag008})));
            print XF ">> Short 008 padded to ",length($marc{tag008})," at rec $count\n";
        }
        $marc{date1} = substr($marc{tag008},7,4) if ($marc{tag008});
        $marc{date2} = substr($marc{tag008},11,4) if ($marc{tag008}); # UNUSED
    }
    unless ($marc{date1} and $marc{date1} =~ /\d{4}/) {
        my $my_260 = $record->field('260');
        if ($my_260 and $my_260->subfield('c')) {
            my $date1 = $my_260->subfield('c');
            $date1 =~ s/\D//g;
            if (defined $date1 and $date1 =~ /\d{4}/) {
                $marc{date1} = $date1;
                $marc{fudgedate} = 1;
                print XF ">> using 260c as date1 at rec $count\n";
            }
        }
    }

    # item_form
    if ( $marc{record_type} =~ /[gkroef]/ ) { # MAP, VIS
        $marc{item_form} = substr($marc{tag008},29,1) if ($marc{tag008});
    } else {
        $marc{item_form} = substr($marc{tag008},23,1) if ($marc{tag008});
    }

    # isbns
    my @isbns = $record->field('020') if $record->field('020');
    push @isbns, $record->field('024') if $record->field('024');
    for my $f ( @isbns ) {
        push @{ $marc{isbns} }, $1 if ( defined $f->subfield('a') and
                                        $f->subfield('a')=~/(\S+)/ );
    }

    # author
    for my $rec_field (100, 110, 111) {
        if ($record->field($rec_field)) {
            $marc{author} = $record->field($rec_field)->subfield('a');
            last;
        }
    }

    # oclc
    $marc{oclc} = [];
    if ($record->field('001') &&
        $record->field('003') &&
        $record->field('003')->as_string() =~ /OCo{0,1}LC/ &&
        defined normalize_oclc_number($record->field('001')->as_string())) {
        push @{ $marc{oclc} }, normalize_oclc_number($record->field('001')->as_string());
    }
    for ($record->field('035')) {
        my $oclc = $_->subfield('a');
        if (defined $oclc &&
            ($oclc =~ /\(OCoLC\)/ || $oclc =~ /(ocm|ocl7|ocn|on)/) &&
            defined normalize_oclc_number($oclc)) {
            push @{ $marc{oclc} }, normalize_oclc_number($oclc);
        }
    }

    if ($record->field('999')) {
        my $koha_bib_id = $record->field('999')->subfield('c');
        $marc{koha_bib_id} = $koha_bib_id if defined $koha_bib_id and $koha_bib_id =~ /^\d+$/;
    }

    # "Accompanying material" and check for "copy" (300)
    if ($record->field('300')) {
        $marc{accomp} = $record->field('300')->subfield('e');
        $marc{tag300a} = $record->field('300')->subfield('a');
    }

    # issn, lccn, title, desc, pages, pub, pubyear, edition
    $marc{lccn} = $record->field('010')->subfield('a') if $record->field('010');
    $marc{issn} = $record->field('022')->subfield('a') if $record->field('022');
    $marc{desc} = $record->field('300')->subfield('a') if $record->field('300');
    $marc{pages} = $1 if (defined $marc{desc} and $marc{desc} =~ /(\d+)/);
    $marc{title} = $record->field('245')->subfield('a')
      if $record->field('245');
    $marc{title} .= ' ' . $record->field('245')->subfield('b')
      if ($record->field('245') and 
          $record->field('245')->subfield('b') and 
          not $conf->{ignoresubtitle});
    $marc{edition} = $record->field('250')->subfield('a')
      if $record->field('250');
    if ($record->field('260')) {
        $marc{publisher} = $record->field('260')->subfield('b');
        $marc{pubyear} = $record->field('260')->subfield('c');
        $marc{pubyear} =
          (defined $marc{pubyear} and $marc{pubyear} =~ /(\d{4})/) ? $1 : '';
    }
    return \%marc;
}



=head2 normalize_marc

Gently massages your data.

=cut

sub normalize_marc {
    my ($marc) = @_;

    $marc->{record_type }= 'a' if ($marc->{record_type} eq ' ');
    if ($marc->{title}) {
        $marc->{title} = NFD($marc->{title});
        $marc->{title} =~ s/[\x{80}-\x{ffff}]//go;
        $marc->{title} = lc($marc->{title});
        $marc->{title} =~ s/\W+$//go;
    }
    if ($marc->{author}) {
        $marc->{author} = NFD($marc->{author});
        $marc->{author} =~ s/[\x{80}-\x{ffff}]//go;
        $marc->{author} = lc($marc->{author});
        $marc->{author} =~ s/\W+$//go;
        if ($marc->{author} =~ /^(\w+)/) {
            $marc->{author} = $1;
        }
    }
    if ($marc->{publisher}) {
        $marc->{publisher} = NFD($marc->{publisher});
        $marc->{publisher} =~ s/[\x{80}-\x{ffff}]//go;
        $marc->{publisher} = lc($marc->{publisher});
        $marc->{publisher} =~ s/\W+$//go;
        if ($marc->{publisher} =~ /^(\w+)/) {
            $marc->{publisher} = $1;
        }
    }
    return $marc;
}



=head2 marc_isvalid

Checks MARC record to see if neccessary fingerprinting data is
available

=cut

sub marc_isvalid {
    my ($marc) = @_;
    return 1 if ($marc->{item_form} and ($marc->{date1} =~ /\d{4}/) and
                 $marc->{record_type} and $marc->{bib_lvl} and $marc->{title});
    return 0;
}


=head2 score_marc

Assign a score to the record based on various criteria.

Score is constructed by pushing elements onto a list, via a dispatch
table.  This allows order of fingerprints in the output file to be
varied.

=cut

sub score_marc {
    my ($marc, $record) = @_;
    my @score = ();
    my $json = '{';

    #----------------------------------
    # static criteria scoring
    #----------------------------------
    $marc->{misc_score} = 999;
    $marc->{age_score}  = 999999999999;

    # -1 if 008 has been padded, -2 if it doesn't exist
    if ($marc->{tag008})
      { $marc->{misc_score}-- if ($marc->{tag008} =~ /\|$/) }
    else
      { $marc->{misc_score} -= 2 }
    # -1 if date has been pulled from 260
    $marc->{misc_score}-- if $marc->{fudgedate};
    # -1 if this is a copy record
    $marc->{misc_score}--
      if (defined $marc->{tag300a} and $marc->{tag300a} =~ /copy/i);

    # subtract record id if we want older records to win
    #$marc->{age_score} -= $marc->{id} unless ($conf->{newwins});
    # handle arbitrary adjustments
    $marc->{age_score} = 1;
    if ($conf->{'arbitrarily-lose-above'}) {
        $marc->{age_score} = 0
          if ($marc->{id} >= $conf->{'arbitrarily-lose-above'});
    }
    if ($conf->{'arbitrarily-lose-below'}) {
        $marc->{age_score} = 0
          if ($marc->{id} <= $conf->{'arbitrarily-lose-below'});
    }

    #----------------------------------
    # dynamic calculated scoring
    #----------------------------------
    my %scores_code = (
      oclc    => sub { return $marc->{oclc}[0] ? 1 : 0 },
      dlc     => sub {
          if ($record->field('040') and $record->field('040')->subfield('a'))
            { return scalar($record->subfield( '040', 'a')) =~ /dlc/io ? 1 : 0 }
          else { return 0 }
      },
      num_650 => sub {
          if ($record->field('650')) {
              # can't say "scalar $record->field('650')"; MARC::Record
              # behaves differently in list/scalar contexts
              my @tags = $record->field('650');
              return sprintf("%04d", scalar @tags)
          } else { return '0000' }
      },
      num_tags=> sub { return sprintf( '%04d', scalar( $record->fields ) ) },
      enc_lvl => sub {
        my $enc = substr($record->leader, 17, 1) || 'u';
        my %levels = ( ' ' => 9, 1 => 8, 2 => 7,  3  => 6,  4  => 5, 5 => 4,
                       6   => 3, 7 => 2, 8 => 1, 'u' => 0, 'z' => 0 );
        return $levels{$enc} || 0;
    }
                      );

    #----------------------------------
    # assemble and store scores
    #----------------------------------
    for ( @{ $conf->{dyn_scores} } ) {
        push @score, $scores_code{$_}->($marc, $record);
        $json .= $_ . ':' . $score[-1] . ',';
    }
    $json .= 'misc:' . $marc->{misc_score} . '}';

    my $compact = join('', $marc->{age_score}, $marc->{misc_score}, @score);
    $marc->{score} = "$compact\t$json";
}

=head2 dump_fingerprints

=cut

sub dump_fingerprints {
    my ($marc) = @_;

    if ($conf->{fingerprints}{baseline}) {
        print OF join("\t", $marc->{score}, $marc->{id}, 'baseline',
                      $marc->{item_form}, $marc->{date1}, $marc->{record_type},
                      $marc->{bib_lvl}, $marc->{title}), "\n";
    }

    if ($conf->{fingerprints}{oclc} and scalar @{$marc->{oclc} }) {
        for (@{$marc->{oclc} }) {
            print OF join("\t", $marc->{score}, $marc->{id}, "oclc",
                          $marc->{item_form}, $marc->{date1},
                          $marc->{record_type}, $marc->{bib_lvl},
                          $marc->{title}, $_, "\n");
        }
    }

    if ($conf->{fingerprints}{koha_bib_id} and exists $marc->{koha_bib_id}) {
        print OF join("\t", $marc->{score}, $marc->{id}, "z_koha_bib_id",
                      $marc->{item_form}, $marc->{date1},
                      $marc->{record_type},
                      $marc->{bib_lvl}, $marc->{title},
                      $marc->{koha_bib_id}), "\n";
    }

    if ($conf->{fingerprints}{isbn}) {
        if ((scalar @{ $marc->{isbns} } > 0) and $marc->{pages}) {
            foreach my $isbn ( @{ $marc->{isbns}} ) {
                print OF join("\t", $marc->{score}, $marc->{id}, "isbn",
                              $marc->{item_form}, $marc->{date1},
                              $marc->{record_type},
                              $marc->{bib_lvl}, $marc->{title},
                              $isbn, $marc->{pages}), "\n";
            }
        }
    }

    if ($conf->{fingerprints}{edition} and $marc->{edition} and $marc->{author}) {
        print OF join("\t", $marc->{score}, $marc->{id}, "edition",
                      $marc->{item_form}, $marc->{date1},
                      $marc->{record_type}, $marc->{bib_lvl},
                      $marc->{title}, $marc->{author}, $marc->{edition}), "\n";
    }

    if ($conf->{fingerprints}{issn} and $marc->{issn}) {
        print OF join("\t", $marc->{score}, $marc->{id}, "issn",
                      $marc->{item_form}, $marc->{date1},
                      $marc->{record_type}, $marc->{bib_lvl},
                      $marc->{title}, $marc->{issn}), "\n";
    }

    if ($conf->{fingerprints}{lccn} and $marc->{lccn}) {
        print OF join("\t", $marc->{score}, $marc->{id}, "lccn",
                      $marc->{item_form}, $marc->{date1},
                      $marc->{record_type}, $marc->{bib_lvl},
                      $marc->{title}, $marc->{lccn}) ,"\n";
    }

    if ($conf->{fingerprints}{accomp} and $marc->{accomp}) {
        print OF join("\t", $marc->{score}, $marc->{id}, "accomp",
                      $marc->{item_form}, $marc->{date1},
                      $marc->{record_type}, $marc->{bib_lvl},
                      $marc->{title}, $marc->{accomp}) ,"\n";
    }

    if ($conf->{fingerprints}{authpub} and $marc->{author} and
        $marc->{publisher} and $marc->{pubyear} and $marc->{pages}) {
        print OF join("\t", $marc->{score}, $marc->{id}, "authpub",
                      $marc->{item_form}, $marc->{date1},
                      $marc->{record_type}, $marc->{bib_lvl},
                      $marc->{title}, $marc->{author},
                      $marc->{publisher}, $marc->{pubyear},
                      $marc->{pages}), "\n";
    }
}



=head2 dump_exception

Write line of exception report

=cut

sub dump_exception {
    my ($marc, $msg) = @_;
    unless (defined $marc) {
        print XF "Undefined record at line $count; likely bad XML\n";
        return;
    }

    print XF "Record ", $marc->{id}, " excluded: ";
    if (defined $msg) {
        print XF "$msg\n";
        return
    }

    print XF "missing item_form; " unless ($marc->{item_form});
    unless (defined $marc->{date1})
      { print XF "missing date1; " }
    else
      { print XF "invalid date1: '", $marc->{date1}, "'; "
          unless ($marc->{date1} =~ /\d{4}/); }
    print XF "missing record_type; " unless ($marc->{record_type});
    print XF "missing bib_lvl; " unless ($marc->{bib_lvl});
    print XF "missing title " unless ($marc->{title});
    print XF "\n";
}


=head2 this_record_is_excluded

Returns 1 if the record B<is> and 0 if the record B<is not> excluded,
according to the subfield mapping (generated via the C<--excludelist>
option).

=cut

sub this_record_is_excluded {
    my ($rec, $marc) = @_;
    return 0 unless defined $conf->{excludelist};

    for my $tag (keys %{ $conf->{excludelist}->{tags} }) {
        for my $sub (keys %{$conf->{excludelist}->{tags}{$tag}}) {
            my $f = $conf->{excludelist}->field($tag, $sub);

            # if this record doesn't have the right tag/sub, it can't be
            return 0 unless ($rec->field($tag) and $rec->field($tag)->subfield($sub));
            # but it does, so if there are no filters to check...
            unless ($conf->{excludelist}->filters($f))
              { dump_exception($marc, "exclusion $tag$sub"); return 1 }

            my $sub_contents = $rec->field($tag)->subfield($sub);
            for my $filter (@{ $conf->{excludelist}->filters($f)}) {
                if ($sub_contents =~ /$filter/i) {
                    # filter matches. no fp.
                    dump_exception($marc, "exclusion $tag$sub '$filter'");
                    return 1;
                }
                # no match, no exclude
                return 0;
            }
        }
    }
}

=head2 initialize

Performs boring script initialization. Handles argument parsing,
mostly.

=cut

sub initialize {
    my ($c) = @_;
    my @missing = ();

    # set mode on existing filehandles
    binmode(STDIN, ':utf8');

    my $rc = GetOptions( $c,
                         'exception|x=s',
                         'output|o=s',
                         'prefix|p=s',
                         'marctype|m=s',
                         'subfield|s=s',
                         'tag|t=s',
                         'fingerprints=s',
                         'scores=s',
                         'arbitrarily-lose-above=i',
                         'arbitrarily-lose-below=i',
                         'newwins',
                         'excludelist=s',
                         'ignoresubtitle|i',
                         'quiet|q',
                         'help|h',
                       );
    show_help() unless $rc;
    show_help() if ($c->{help});

    # check fingerprints list for validity
    if ($c->{fingerprints}) {
        my %fps = ();
        my %valid_fps = ( oclc => 1, isbn => 1, issn => 1, lccn => 1,
                          edition => 1, accomp => 1, authpub => 1,
                          baseline => 1, crap => 1,
                          koha_bib_id => 1,
                        );
        for (split /,/, $c->{fingerprints}) {
            die "Invalid fingerprint '$_'\n" unless $valid_fps{$_};
            $fps{$_} = 1;
        }
        $c->{fingerprints} = \%fps
    } else {
        $c->{fingerprints} = {oclc => 1, isbn => 1, edition => 1, issn => 1,
                              lccn => 1, accomp => 1, authpub => 1};
    }

    # check scores list for validity
    if ($c->{scores}) {
        my %scores = ();
        my %valid_scores = ( oclc => 1, dlc => 1, num_650 => 1,
                             num_tags => 1, enc_lvl => 1,
                           );
        for (split /,/, $c->{scores}) {
            die "Invalid score mode '$_'\n" unless $valid_scores{$_};
            $scores{$_} = 1;
        }
        $c->{dyn_scores} = [split /,/, $c->{scores}];
        $c->{scores} = \%scores;
    } else {
        $c->{scores} = {oclc => 1, dlc => 1, num_650 => 1,
                        num_tags => 1, enc_lvl => 1};
        $c->{dyn_scores} = [ qw/oclc dlc num_650 num_tags enc_lvl/ ];
    }

    # set defaults
    $c->{tag} = 903 unless defined $c->{tag};
    $c->{subfield} = 'a' unless defined $c->{subfield};
    $c->{marctype} = 'XML' unless defined $c->{marctype};
    if ($c->{prefix}) {
        $c->{output} = join('.',$c->{prefix},'fp');
        $c->{exception} = join('.',$c->{prefix},'fp','ex');
    }

    # get SFM object if excludelist was specified
    if ($c->{excludelist}) {
        $c->{excludelist} =
          Equinox::Migration::SubfieldMapper->new( file => $c->{excludelist} );
    }

    my @keys = keys %{$c};
    show_help() unless (@ARGV and @keys);
    for my $key ('tag', 'subfield', 'output', 'exception')
      { push @missing, $key unless $c->{$key} }
    if (@missing) {
        print "Required option: ", join(', ', @missing), " missing!\n";
        show_help();
    }
}


=head2 progress_ticker

=cut

sub progress_ticker {
    return if $conf->{quiet};
    printf("\r> %d recs seen; %d processed", $count, $scount);
    printf(" (%d/s)", ($count / (time - $start + 1)))
      if ($count % 500 == 0);
}

=head2 show_help

Display usage message when things go wrong

=cut

sub show_help {
print <<HELP;
Usage is: $0 [REQUIRED ARGS] [OPTIONS] <filelist>
Req'd Arguments
  --output=<FILE>      -o  Output filename
  --exceptions=<FILE>  -x  Exception report filename
       or
  --prefix=<PREFIX>>   -p  Shared prefix for output/exception files. Will
                           produce PREFIX.fp and PREFIX.fp.ex
Options
  --tag=N           -t  Which tag to use (default 903)
  --subfield=X      -s  Which subfield to use (default 'a')
  --quiet           -q  Don't write status messages to STDOUT
  --ignoresubtitle  -i  Ignore 245\$b and construct the title from 245\$a alone.

  --fingerprints=LIST  Fingerprints to generate, comma separated
                       Default: oclc,isbn,edition,issn,lccn,accomp,authpub
                       Others:  baseline,koha_bib_id
  --excludelist=FILE   Name of fingerprints exclusions file

  --scores=LIST  Scores to calculate, comma separated
                 Default: oclc,dlc,num_650,num_tags,enc_level
  --newwins      New record IDs score higher (default is old wins)
  --arbitrarily-lose-above
  --arbitrarily-lose-below
  --arbitrarily-decrease-score-by
      Modify fingerprint scoring of records whose EG id is above or below a
      given value, inclusive (so 5 is <= 5 or >= 5) such that they lose.

  --marctype=TYPE Defaults to 'XML'
HELP
exit 1;
}