1 package Equinox::Migration::MapDrivenMARCXMLProc;
7 use Equinox::Migration::SubfieldMapper 1.004;
12 Equinox::Migration::MapDrivenMARCXMLProc
20 our $VERSION = '1.005';
24 my @modlist = qw( multi ignoremulti bib required first concatenate );
35 use Equinox::Migration::MapDrivenMARCXMLProc;
43 Takes two required arguments: C<mapfile> (which will be passed along
44 to L<Equinox::Migration::SubfieldMapper> as the basis for its map),
45 and C<marcfile> (the MARC data to be processed).
47 my $m = Equinox::Migration::MapDrivenMARCXMLProc->new( mapfile => FILE,
53 my ($class, %args) = @_;
55 $verbose = 1 if $args{verbose};
57 my $self = bless { multis => \$multis,
60 # initialize map and taglist
61 die "Argument 'mapfile' must be specified\n" unless ($args{mapfile});
62 $sfmap = Equinox::Migration::SubfieldMapper->new( file => $args{mapfile},
65 # initialize datastore
67 $reccount = 0; # next record ptr
68 $dstore->{tags} = $sfmap->tags; # list of all tags
69 $self->{data} = $dstore;
72 die "Argument 'marcfile' must be specified\n" unless ($args{marcfile});
73 if (-r $args{marcfile}) {
74 my $xmltwig = XML::Twig->new( twig_handlers => { record => \&parse_record } );
75 $xmltwig->parsefile( $args{marcfile} );
77 die "Can't open marc file: $!\n";
85 Extracts data from the next record, per the mapping file.
90 my ($twig, $record) = @_;
91 my $crec = {}; # current record
93 my @fields = $record->children;
95 { process_field($f, $crec) }
97 # fill in blank values if needed
98 for my $mappedtag ( @{ $sfmap->tags }) {
99 unless (exists $crec->{tmap}{$mappedtag}) {
100 push @{ $crec->{tags} }, {};
101 for my $mappedsub ( @{ $sfmap->subfields($mappedtag) } ) {
102 my $fieldname = $sfmap->field($mappedtag, $mappedsub);
103 my $mods = $sfmap->mods($fieldname);
104 next if $mods->{multi};
105 $crec->{tags}[-1]{uni}{$mappedsub} = '';
106 $crec->{tags}[-1]{multi} = undef;
107 $crec->{tags}[-1]{tag} = $mappedtag;
109 push @{ $crec->{tmap}{$mappedtag} }, $#{ $crec->{tags} };
113 # cleanup memory and increment pointer
117 # check for required fields
118 check_required($crec);
119 push @{ $dstore->{recs} }, $crec;
121 print STDERR "$reccount\n"
122 if ($verbose and !($reccount % 1000));
126 my ($field, $crec) = @_;
127 my $tag = $field->{'att'}->{'tag'};
130 unless (defined $tag) {
137 my $sub = $field->first_child('subfield');
138 $crec->{egid} = $sub->text;
141 if ($sfmap->has($tag)) {
142 push @{$crec->{tags}}, { tag => $tag, uni => undef, multi => undef };
143 push @{$crec->{tmap}{$tag}}, (@{$crec->{tags}} - 1);
144 my @subs = $field->children('subfield');
146 { process_subs($tag, $sub, $crec) }
148 # check map to ensure all declared tags and subs have a value
149 for my $mappedsub ( @{ $sfmap->subfields($tag) } ) {
150 my $fieldname = $sfmap->field($tag, $mappedsub);
151 my $mods = $sfmap->mods($fieldname);
152 next if $mods->{multi};
153 $crec->{tags}[-1]{uni}{$mappedsub} = ''
154 unless defined $crec->{tags}[-1]{uni}{$mappedsub};
160 my ($tag, $sub, $crec) = @_;
161 my $code = $sub->{'att'}->{'code'};
163 # handle unmapped tag/subs
164 return unless ($sfmap->has($tag, $code));
166 # fetch our datafield struct and field and mods
167 my $dataf = $crec->{tags}[-1];
168 my $field = $sfmap->field($tag, $code);
169 my $sep = $sfmap->sep($field);
170 $allmods{$field} = $sfmap->mods($field) unless $allmods{$field};
171 my $mods = $allmods{$field};
174 for my $filter ( @{$sfmap->filters($field)} ) {
175 return if ($sub->text =~ /$filter/i);
178 # handle multi modifier
179 if ($mods->{multi}) {
180 $multis->{$tag}{$code} = 1;
181 if ($mods->{concatenate}) {
182 if (exists($dataf->{multi}{$code})) {
183 $dataf->{multi}{$code}[0] .= $sep . $sub->text;
185 push @{$dataf->{multi}{$code}}, $sub->text;
187 $multis->{$tag}{$code} = 1;
189 push @{$dataf->{multi}{$code}}, $sub->text;
195 if ($mods->{concatenate}) {
196 if (exists($dataf->{uni}{$code})) {
197 $dataf->{uni}{$code} .= $sep . $sub->text;
199 $dataf->{uni}{$code} = $sub->text;
204 # if this were a multi field, it would be handled already. make sure its a singleton
205 die "Multiple occurances of a non-multi field: $tag$code at rec ",
206 ($reccount + 1),"\n" if (defined $dataf->{uni}{$code} and !$mods->{ignoremulti});
208 # everything seems okay
209 $dataf->{uni}{$code} = $sub->text;
215 my $mods = $sfmap->mods;
217 for my $tag_id (keys %{$mods->{required}}) {
218 for my $code (@{$mods->{required}{$tag_id}}) {
221 for my $tag (@{$crec->{tags}}) {
222 $found = 1 if ($tag->{multi}{($tag_id . $code)});
223 $found = 1 if ($tag->{uni}{$code});
226 die "Required mapping $tag_id$code not found in rec ",$reccount,"\n"
235 Returns current record number (starting from zero)
239 sub recno { my ($self) = @_; return $self->{data}{rcnt} }
243 Returns mapped fieldname when passed a tag, and code
245 my $name = $m->name(999,'a');
249 sub name { my ($self, $t, $c) = @_; return $sfmap->field($t, $c) }
253 Returns whether mapped fieldname is to be applied only to first
259 my ($self, $t, $c) = @_;
260 my $field = $sfmap->field($t, $c);
261 my $mods = $sfmap->mods($field);
262 return exists($mods->{first});
267 Returns hashref of C<{tag}{code}> for all mapped multi fields
278 MapDrivenMARCXMLProc implements the following modifiers, and passes
279 them to L<Equinox::Migration::SubfieldMapper>, meaning that specifying
280 any other modifiers in a MDMP map file will cause a fatal error when
285 If a mapping is declared to be C<multi>, then MDMP expects to see more
286 than one instance of that subfield per datafield, and the data is
287 handled accordingly (see L</PARSED RECORDS> below).
289 Occurring zero or one time is legal for a C<multi> mapping.
291 A mapping which is not flagged as C<multi>, but which occurs more than
292 once per datafield will cause a fatal error.
296 By default, if a mapping does not occur in a datafield, processing
297 continues normally. if a mapping has the C<required> modifier,
298 however, it must appear, or a fatal error will occur.
300 =head1 PARSED RECORDS
304 my $m = Equinox::Migration::MapDrivenMARCXMLProc->new(ARGUMENTS);
305 $rec = $m->parse_record;
307 Then C<$rec> will look like:
310 egid => evergreen_record_id,
314 multi => { code => [ val1, val2, ... ] },
315 uni => { code => value, code2 => value2, ... },
319 tmap => { tag_id => [ INDEX_LIST ], tag_id2 => [ INDEX_LIST ], ... }
322 That is, there is an C<egid> key which points to the Evergreen ID of
323 that record, a C<tags> key which points to an arrayref, and a C<tmap>
324 key which points to a hashref.
328 A reference to a list of anonymous hashes, one for each instance of
329 each tag which occurs in the map.
331 Each tag hash holds its own id (e.g. C<998>), and two references to
332 two more hashrefs, C<multi> and C<uni>.
334 The C<multi> hash holds the extracted data for tag/sub mappings which
335 have the C<multiple> modifier on them. The keys in C<multi> subfield
336 codes. The values are arrayrefs containing the content of all
337 instances of that subfield in that instance of that tag. If no tags
338 are defined as C<multi>, it will be C<undef>.
340 The C<uni> hash holds data for tag/sub mappings which occur only once
341 per instance of a tag (but may occur multiple times in a record due to
342 there being multiple instances of that tag in a record). Keys are
343 subfield codes and values are subfield content.
345 All C<uni> subfields occuring in the map are guaranteed to be
346 defined. Sufields which are mapped but do not occur in a particular
347 datafield will be given a value of '' (the null string) in the current
348 record struct. Oppose subfields which are not mapped, which will be
353 A hashref, where each key (a tag id like "650") points to a listref
354 containing the index (or indices) of C<tags> where that tag has
357 The intended use of this is to simplify the processing of data from
358 tags which can appear more than once in a MARC record, like
359 holdings. If your holdings data is in 852, C<tmap->{852}> will be a
360 listref with the indices of C<tags> which hold the data from the 852
363 Complimentarily, C<tmap> prevents data from singular datafields from
364 having to be copied for every instance of a multiple datafield, as it
365 lets you get the data from that record's one instance of whichever
366 field you're looking for.
370 Shawn Boyette, C<< <sboyette at esilibrary.com> >>
374 Please report any bugs or feature requests to the above email address.
378 You can find documentation for this module with the perldoc command.
380 perldoc Equinox::Migration::MapDrivenMARCXMLProc
383 =head1 COPYRIGHT & LICENSE
385 Copyright 2009 Equinox, all rights reserved.
387 This program is free software; you can redistribute it and/or modify it
388 under the same terms as Perl itself.
393 1; # End of Equinox::Migration::MapDrivenMARCXMLProc