use strict;
use XML::Twig;
-use Equinox::Migration::SubfieldMapper;
+use Equinox::Migration::SubfieldMapper 1.002;
+
+# 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
}, $class;
# initialize map and taglist
+ die "Argument 'mapfile' must be specified\n" unless (defined $args{mapfile});
my @mods = keys %{$self->{mods}};
$self->{map} = Equinox::Migration::SubfieldMapper->new( file => $args{mapfile},
mods => \@mods );
- $self->{tags} = $self->{map}->tags;
+ $self->{data}{tags} = $self->{map}->tags;
# initialize twig
die "Argument 'marcfile' must be specified\n" unless (defined $args{marcfile});
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 $self;
}
=head2 parse_record
-Extracts data from the next record, per the mapping file. Returns 1 on
-success, 0 otherwise.
+Extracts data from the next record, per the mapping file. Returns a
+normalized datastructure (see L</format_record> for details) on
+success; returns 0 otherwise.
- while ($m->parse_record) {
+ while (my $rec = $m->parse_record) {
# handle extracted record data
}
# 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} = {};
+ $self->{data}{crec} = { egid => undef, bib => undef, tags => undef };
my @fields = $record->children;
for my $f (@fields)
# cleanup memory and increment pointer
$record->purge;
$self->{data}{rptr}++;
-}
-=head2 process_field
-
-=cut
+ return $self->{data}{crec};
+}
sub process_field {
my ($self, $field) = @_;
my $map = $self->{map};
my $tag = $field->{'att'}->{'tag'};
- my $parsed = $self->{data}{crec};
+ my $crec = $self->{data}{crec};
+
+ # leader
+ unless (defined $tag) {
+ #FIXME
+ return;
+ }
+ # datafields
if ($tag == 903) {
my $sub = $field->first_child('subfield');
- $parsed->{egid} = $sub->text;;
- } elsif ($map->has($tag)) {
- push @{$parsed->{tags}}, { tag => $tag };
+ $crec->{egid} = $sub->text;
+ return;
+ }
+ if ($map->has($tag)) {
+ push @{$crec->{tags}}, { tag => $tag, uni => undef, multi => undef };
my @subs = $field->children('subfield');
for my $sub (@subs)
{ $self->process_subs($tag, $sub) }
+ # check map to ensure all declared subs have a value
+ my $mods = $map->mods($field);
+ for my $mappedsub ( @{ $map->subfields($tag) } ) {
+ next if $mods->{multi};
+ $crec->{tags}[-1]{uni}{$mappedsub} = ''
+ unless defined $crec->{tags}[-1]{uni}{$mappedsub};
+ }
}
}
-=head2 process_subs
-
-=cut
-
sub process_subs {
my ($self, $tag, $sub) = @_;
my $map = $self->{map};
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;
}
- my $data = $self->{data}{crec}{tags}[-1];
+ # fetch our datafield struct and fieldname
+ my $dataf = $self->{data}{crec}{tags}[-1];
my $field = $map->field($tag, $code);
- if ($map->mod($field) eq 'multi') {
- my $name = $tag . $code;
- push @{$data->{multi}{$name}}, $sub->text;
+
+ # handle modifiers, or slug data in normally
+ if (my $mods = $map->mods($field)) {
+ if ($mods->{multi}) {
+ my $name = $tag . $code;
+ push @{$dataf->{multi}{$name}}, $sub->text;
+ }
} else {
- $data->{uni}{$code} = $sub->text;
+ die "Multiple occurances of a non-multi field: \n"
+ if (defined $dataf->{uni}{$code});
+ $dataf->{uni}{$code} = $sub->text;
}
}
+=head1 MODIFIERS
+
+MapDrivenMARCXMLProc implements the following modifiers, and passes
+them to L<Equinox::Migration::SubfieldMapper>, meaning that specifying
+any other modifiers in a MDMP map file will cause a fatal error when
+it is processed.
+
+=head2 multi
+
+If a mapping is declared to be C<multi>, then MDMP expects to see more
+than one instance of that subfield per datafield, and the data is
+handled accordingly (see L</PARSED RECORDS> below).
+
+Occurring zero or one time is legal for a C<multi> mapping.
+
+A mapping which is not flagged as C<multi>, but which occurs more than
+once per datafield will cause a fatal error.
+
+=head2 bib
+
+=head2 required
+
=head1 PARSED RECORDS
+Given:
+
+ my $m = Equinox::Migration::MapDrivenMARCXMLProc->new(ARGUMENTS);
+ $rec = $m->parse_record;
+
+Then C<$rec> will look like:
+
{
egid => evergreen_record_id,
bib => {
=head3 C<bib>
-This hashref holds extracted data which should occur once per record
-(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.
+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>.
=head3 C<tags>
-This arrayref holds anonymous hashrefs, one for each instance of each
-tag which occurs in the map. Each tag hashref holds its own id
-(e.g. C<998>), and two more hashrefs, C<multi> and C<uni>.
+A reference to a list of anonymous hashes, one for each instance of
+each tag which occurs in the map.
-The C<multi> hashref holds the extracted data for tag/sub mappings
-which have the C<multiple> modifier on them. The keys in C<multi> are
+Each tag hash holds its own id (e.g. C<998>), and two references to
+two more hashrefs, C<multi> and C<uni>.
+
+The C<multi> hash holds the extracted data for tag/sub mappings which
+have the C<multiple> modifier on them. The keys in C<multi> are
composed of the tag id and subfield code, catenated
(e.g. C<901c>). The values are arrayrefs containing the content of all
-instances of that subfield in that instance of that tag.
+instances of that subfield in that instance of that tag. If no tags
+are defined as C<multi>, it will be C<undef>.
+
+The C<uni> hash holds data for tag/sub mappings which occur only once
+per instance of a tag (but may occur multiple times in a record due to
+there being multiple instances of that tag in a record). Keys are
+subfield codes and values are subfield content.
-The C<uni> hashref holds data for tag/sub mappings which occur only
-once per instance of a tag (but may occur multiple times in a record
-due to there being multiple instances of that tag in a record). Keys
-are subfield codes and values are subfield content.
+All C<uni> subfields occuring in the map are guaranteed to be
+defined. Sufields which are mapped but do not occur in a particular
+datafield will be given a value of '' (the null string) in the current
+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 },
- sub_code2 => { value => VALUE, count => COUNT },
+ 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
Shawn Boyette, C<< <sboyette at esilibrary.com> >>
projects / migration-tools.git / blobdiff