X-Git-Url: http://git.equinoxoli.org/?p=migration-tools.git;a=blobdiff_plain;f=Equinox-Migration%2Flib%2FEquinox%2FMigration%2FMapDrivenMARCXMLProc.pm;h=194a593415b20a25e88603e9636a8c3a6f152c72;hp=805d70ca7db351dba75bd96c870490a9f53b6dfe;hb=3bd5dd6ddf3857cf8c54cfe10df9fef27e2478d0;hpb=92480d2b668d7ddd2237cca8d4d6c20f5f2e43f9 diff --git a/Equinox-Migration/lib/Equinox/Migration/MapDrivenMARCXMLProc.pm b/Equinox-Migration/lib/Equinox/Migration/MapDrivenMARCXMLProc.pm index 805d70c..194a593 100644 --- a/Equinox-Migration/lib/Equinox/Migration/MapDrivenMARCXMLProc.pm +++ b/Equinox-Migration/lib/Equinox/Migration/MapDrivenMARCXMLProc.pm @@ -4,7 +4,8 @@ use warnings; use strict; use XML::Twig; -use Equinox::Migration::SubfieldMapper 1.003; +use DBM::Deep; +use Equinox::Migration::SubfieldMapper 1.004; =head1 NAME @@ -13,11 +14,18 @@ Equinox::Migration::MapDrivenMARCXMLProc =head1 VERSION -Version 1.000 +Version 1.003 =cut -our $VERSION = '1.000'; +our $VERSION = '1.003'; + +my $dstore; +my $sfmap; +my @mods = qw( multi bib required ); +my $multis = {}; +my $reccount; +my $verbose = 0; =head1 SYNOPSIS @@ -44,31 +52,33 @@ and C (the MARC data to be processed). sub new { my ($class, %args) = @_; - my $self = bless { mods => { multi => {}, - bib => {}, - required => {}, - }, - data => { recs => undef, # X::T record objects - rptr => 0, # next record pointer - crec => undef, # parsed record storage - tmap => undef, # tag_id-to-tag array map - }, + $verbose = 1 if $args{verbose}; + + my $self = bless { multis => \$multis, }, $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->{data}{tags} = $self->{map}->tags; + die "Argument 'mapfile' must be specified\n" unless ($args{mapfile}); + $sfmap = Equinox::Migration::SubfieldMapper->new( file => $args{mapfile}, + mods => \@mods ); + + # initialize datastore + die "Datastore file 'EMMXSSTORAGE.dbmd' already exists. Exiting.\n" + if (-e "EMMXSSTORAGE.dbmd"); + $dstore = DBM::Deep->new( file => "EMMXSSTORAGE.dbmd", + max_buckets => 64, + #data_sector_size => 256, + autoflush => 0, + ); + $reccount = 0; # next record ptr + $dstore->{tags} = $sfmap->tags; # list of all tags + $self->{data} = $dstore; # initialize twig - die "Argument 'marcfile' must be specified\n" unless (defined $args{marcfile}); + die "Argument 'marcfile' must be specified\n" unless ($args{marcfile}); if (-r $args{marcfile}) { - $self->{twig} = XML::Twig->new; - $self->{twig}->parsefile($args{marcfile}); - my @records = $self->{twig}->root->children; - $self->{data}{recs} = \@records; + my $xmltwig = XML::Twig->new( twig_handlers => { record => \&parse_record } ); + $xmltwig->parsefile( $args{marcfile} ); } else { die "Can't open marc file: $!\n"; } @@ -76,48 +86,37 @@ sub new { return $self; } +sub DESTROY { unlink "EMMXSSTORAGE.dbmd" } =head2 parse_record -Extracts data from the next record, per the mapping file. Returns a -normalized datastructure (see L for details) on -success; returns 0 otherwise. - - while (my $rec = $m->parse_record) { - # handle extracted record data - } +Extracts data from the next record, per the mapping file. =cut sub parse_record { - my ($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, tags => undef }; - $self->{data}{tmap} = {}; + my ($twig, $record) = @_; + my $crec = {}; # current record my @fields = $record->children; for my $f (@fields) - { $self->process_field($f) } + { process_field($f, $crec) } # cleanup memory and increment pointer $record->purge; - $self->{data}{rptr}++; + $reccount++; # check for required fields - $self->check_required; + check_required(); + push @{ $dstore->{recs} }, $crec; - return $self->{data}{crec}; + print STDERR "$reccount\n" + if ($verbose and !($reccount % 1000)); } sub process_field { - my ($self, $field) = @_; - my $map = $self->{map}; + my ($field, $crec) = @_; my $tag = $field->{'att'}->{'tag'}; - my $crec = $self->{data}{crec}; - my $tmap = $self->{data}{tmap}; # leader unless (defined $tag) { @@ -131,50 +130,56 @@ sub process_field { $crec->{egid} = $sub->text; return; } - if ($map->has($tag)) { + if ($sfmap->has($tag)) { push @{$crec->{tags}}, { tag => $tag, uni => undef, multi => undef }; - push @{$tmap->{$tag}}, (@{$crec->{tags}} - 1); + push @{$crec->{tmap}{$tag}}, (@{$crec->{tags}} - 1); 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) } ) { + { process_subs($tag, $sub, $crec) } + + # check map to ensure all declared tags and subs have a value + for my $mappedsub ( @{ $sfmap->subfields($tag) } ) { + my $fieldname = $sfmap->field($tag, $mappedsub); + my $mods = $sfmap->mods($fieldname); next if $mods->{multi}; $crec->{tags}[-1]{uni}{$mappedsub} = '' unless defined $crec->{tags}[-1]{uni}{$mappedsub}; } + for my $mappedtag ( @{ $sfmap->tags }) { + $crec->{tmap}{$mappedtag} = undef + unless defined $crec->{tmap}{$mappedtag}; + } } } sub process_subs { - my ($self, $tag, $sub) = @_; - my $map = $self->{map}; + my ($tag, $sub, $crec) = @_; my $code = $sub->{'att'}->{'code'}; # handle unmapped tag/subs - return unless ($map->has($tag, $code)); + return unless ($sfmap->has($tag, $code)); - # fetch our datafield struct and fieldname - my $dataf = $self->{data}{crec}{tags}[-1]; - my $field = $map->field($tag, $code); + # fetch our datafield struct and fiel + my $dataf = $crec->{tags}[-1]; + my $field = $sfmap->field($tag, $code); # test filters - for my $filter ( @{$map->filters($field)} ) { + for my $filter ( @{$sfmap->filters($field)} ) { return if ($sub->text =~ /$filter/i); } + # handle multi modifier - if (my $mods = $map->mods($field)) { + if (my $mods = $sfmap->mods($field)) { if ($mods->{multi}) { - my $name = $tag . $code; - push @{$dataf->{multi}{$name}}, $sub->text; + push @{$dataf->{multi}{$code}}, $sub->text; + $multis->{$tag}{$code} = 1; return; } } # 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}); + ($reccount + 1),"\n" if (defined $dataf->{uni}{$code}); # everything seems okay $dataf->{uni}{$code} = $sub->text; @@ -182,9 +187,8 @@ sub process_subs { sub check_required { - my ($self) = @_; - my $mods = $self->{map}->mods; - my $crec = $self->{data}{crec}; + my $mods = $sfmap->mods; + my $crec = $dstore->{crec}; for my $tag_id (keys %{$mods->{required}}) { for my $code (@{$mods->{required}{$tag_id}}) { @@ -195,13 +199,42 @@ sub check_required { $found = 1 if ($tag->{uni}{$code}); } - die "Required mapping $tag_id$code not found in rec ",$self->{data}{rptr},"\n" + die "Required mapping $tag_id$code not found in rec ",$reccount,"\n" unless ($found); } } } +=head2 recno + +Returns current record number (starting from zero) + +=cut + +sub recno { my ($self) = @_; return $self->{data}{rcnt} } + +=head2 name + +Returns mapped fieldname when passed a tag, and code + + my $name = $m->name(999,'a'); + +=cut + +sub name { my ($self, $t, $c) = @_; return $sfmap->field($t, $c) } + +=head2 get_multis + +Returns hashref of C<{tag}{code}> for all mapped multi fields + +=cut + +sub get_multis { + my ($self) = @_; + return $multis; +} + =head1 MODIFIERS MapDrivenMARCXMLProc implements the following modifiers, and passes @@ -240,15 +273,17 @@ Then C<$rec> will look like: tags => [ { tag => tag_id, - multi => { (tag_id . sub_code) => [ val1, val2, ... ] }, + multi => { code => [ val1, val2, ... ] }, uni => { code => value, code2 => value2, ... }, }, ... - ] + ], + tmap => { tag_id => [ INDEX_LIST ], tag_id2 => [ INDEX_LIST ], ... } } That is, there is an C key which points to the Evergreen ID of -that record, and a C key which points to an arrayref. +that record, a C key which points to an arrayref, and a C +key which points to a hashref. =head3 C @@ -259,9 +294,8 @@ Each tag hash holds its own id (e.g. C<998>), and two references to two more hashrefs, C and C. The C hash holds the extracted data for tag/sub mappings which -have the C modifier on them. The keys in C are -composed of the tag id and subfield code, catenated -(e.g. C<901c>). The values are arrayrefs containing the content of all +have the C modifier on them. The keys in C subfield +codes. The values are arrayrefs containing the content of all instances of that subfield in that instance of that tag. If no tags are defined as C, it will be C. @@ -276,6 +310,22 @@ 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. +=head3 tmap + +A hashref, where each key (a tag id like "650") points to a listref +containing the index (or indices) of C where that tag has +extracted data. + +The intended use of this is to simplify the processing of data from +tags which can appear more than once in a MARC record, like +holdings. If your holdings data is in 852, C{852}> will be a +listref with the indices of C which hold the data from the 852 +datafields. + +Complimentarily, C prevents data from singular datafields from +having to be copied for every instance of a multiple datafield, as it +lets you get the data from that record's one instance of whichever +field you're looking for. =head1 AUTHOR