use XML::Twig;
use Equinox::Migration::SubfieldMapper 1.003;
-# FIXME
-#
-# sample functionality should be extracted into a new module which
-# uses E::M::SM to drive sampling of individual datafields, and
-# reports ALL datafields which occur
-#
-# --sample should give the list of all datafields
-# --samplefile should take a SM map as teh argument and introspect the mapped datafields
-
=head1 NAME
my $m = Equinox::Migration::MapDrivenMARCXMLProc->new( mapfile => FILE,
marcfile => FILE );
-There is an optional third, argument, C<sample>, which specifies a
-arrayref of datafields to "sample" by reporting on subfields which are
-found in the data but not in the map.
-
- my $m = Equinox::Migration::MapDrivenMARCXMLProc->new( mapfile => FILE,
- marcfile => FILE,
- sample => \@TAGS
- );
-
-See L</UNMAPPED TAGS> for more info.
-
=cut
sub new {
data => { recs => undef, # X::T record objects
rptr => 0, # next record pointer
crec => undef, # parsed record storage
- stag => undef, # list of tags to sample
- umap => undef, # unmapped data samples
+ tmap => undef, # tag_id-to-tag array map
},
}, $class;
die "Can't open marc file: $!\n";
}
- # if we have a sample arg, set up the sample set and umap hash
- if (defined $args{sample}) {
- for my $s ( @{$args{sample}})
- { $self->{data}{stag}{$s} = 1 }
- $self->{data}{umap} = {};
- }
-
return $self;
}
# get the next record and wipe current parsed record
return 0 unless defined $self->{data}{recs}[ $self->{data}{rptr} ];
my $record = $self->{data}{recs}[ $self->{data}{rptr} ];
- $self->{data}{crec} = { egid => undef, bib => undef, tags => undef };
+ $self->{data}{crec} = { egid => undef, tags => undef };
+ $self->{data}{tmap} = {};
my @fields = $record->children;
for my $f (@fields)
my $map = $self->{map};
my $tag = $field->{'att'}->{'tag'};
my $crec = $self->{data}{crec};
+ my $tmap = $self->{data}{tmap};
# leader
unless (defined $tag) {
}
if ($map->has($tag)) {
push @{$crec->{tags}}, { tag => $tag, uni => undef, multi => undef };
+ push @{$tmap->{$tag}}, (@{$crec->{tags}} - 1);
my @subs = $field->children('subfield');
for my $sub (@subs)
{ $self->process_subs($tag, $sub) }
my $code = $sub->{'att'}->{'code'};
# handle unmapped tag/subs
- unless ($map->has($tag, $code)) {
- my $u = $self->{data}{umap};
- my $s = $self->{data}{stag};
- return unless (defined $s->{$tag});
-
- # set a value, total-seen count and records-seen-in count
- $u->{$tag}{$code}{value} = $sub->text unless defined $u->{$tag}{$code};
- $u->{$tag}{$code}{count}++;
- $u->{$tag}{$code}{rcnt}++ unless ( defined $u->{$tag}{$code}{last} and
- $u->{$tag}{$code}{last} == $self->{data}{rptr} );
- $u->{$tag}{$code}{last} = $self->{data}{rptr};
- return;
- }
+ return unless ($map->has($tag, $code));
# fetch our datafield struct and fieldname
my $dataf = $self->{data}{crec}{tags}[-1];
my $field = $map->field($tag, $code);
- # handle modifiers
+ # test filters
+ for my $filter ( @{$map->filters($field)} ) {
+ return if ($sub->text =~ /$filter/i);
+ }
+ # handle multi modifier
if (my $mods = $map->mods($field)) {
if ($mods->{multi}) {
my $name = $tag . $code;
}
}
- die "Multiple occurances of a non-multi field: $tag$code at rec ",($self->{data}{rptr} + 1),"\n"
- if (defined $dataf->{uni}{$code});
+ # if this were a multi field, it would be handled already. make sure its a singleton
+ die "Multiple occurances of a non-multi field: $tag$code at rec ",
+ ($self->{data}{rptr} + 1),"\n" if (defined $dataf->{uni}{$code});
+
+ # everything seems okay
$dataf->{uni}{$code} = $sub->text;
}
for my $code (@{$mods->{required}{$tag_id}}) {
my $found = 0;
- $found = 1 if ($crec->{bib}{($tag_id . $code)});
for my $tag (@{$crec->{tags}}) {
$found = 1 if ($tag->{multi}{($tag_id . $code)});
$found = 1 if ($tag->{uni}{$code});
A mapping which is not flagged as C<multi>, but which occurs more than
once per datafield will cause a fatal error.
-=head2 bib
-
-The C<bib> modifier declares that a mapping is "bib-level", and should
-be encountered once per B<record> instead of once per B<datafield> --
-which is another way of saying that it occurs in a non-repeating
-datafield or in a controlfield.
-
=head2 required
-By default, if a mapping does not occur in a datafield (or record, in
-the case of C<bib> mappings), processing continues normally. if a
-mapping has the C<required> modifier, however, it must appear, or a
-fatal error will occur.
+By default, if a mapping does not occur in a datafield, processing
+continues normally. if a mapping has the C<required> modifier,
+however, it must appear, or a fatal error will occur.
=head1 PARSED RECORDS
Then C<$rec> will look like:
{
- egid => evergreen_record_id,
- bib => {
- (tag_id . sub_code)1 => value1,
- (tag_id . sub_code)2 => value2,
- ...
- },
+ egid => evergreen_record_id,
tags => [
{
tag => tag_id,
}
That is, there is an C<egid> key which points to the Evergreen ID of
-that record, a C<bib> key which points to a hashref, and a C<tags>
-key which points to an arrayref.
-
-=head3 C<bib>
-
-A reference to a hash which holds extracted data which occurs only
-once per record (and is therefore "bib-level"; the default assumption
-is that a tag/subfield pair can occur multiple times per record). The
-keys are composed of tag id and subfield code, catenated
-(e.g. 901c). The values are the contents of that subfield of that tag.
-
-If there are no tags defined as bib-level in the mapfile, C<bib> will
-be C<undef>.
+that record, and a C<tags> key which points to an arrayref.
=head3 C<tags>
record struct. Oppose subfields which are not mapped, which will be
C<undef>.
-=head1 UNMAPPED TAGS
-
-If the C<sample> argument is passed to L</new>, there will also be a
-structure which holds data about unmapped subfields encountered in
-mapped tags which are also in the declared sample set. This
-information is collected over the life of the object and is not reset
-for every record processed (as the current record data neccessarily
-is).
-
- { tag_id => {
- sub_code => { value => VALUE,
- count => COUNT,
- rcnt => RCOUNT
- },
- ...
- },
- ...
- }
-
-For each mapped tag, for each unmapped subfield, there is a hash of
-data about that subfield containing
-
- * value - A sample of the subfield text
- * count - Total number of times the subfield was seen
- * rcnt - The number of records the subfield was seen in
=head1 AUTHOR