| 1 | package MARC::Indexer; | ||
| 2 | |||
| 3 | use strict; | ||
| 4 | use warnings; | ||
| 5 | |||
| 6 | $MARC::Indexer::VERSION = '0.04'; | ||
| 7 | |||
| 8 | use MARC::Indexer::Config; | ||
| 9 | use MARC::Loop qw(marcparse TAG VALREF); | ||
| 10 | use Unicode::Normalize; | ||
| 11 | use POSIX qw(strftime); | ||
| 12 | |||
| 13 | sub new { | ||
| 14 | my $cls = shift; | ||
| 15 | bless { @_ }, $cls; | ||
| 16 | } | ||
| 17 | |||
| 18 | sub compile { | ||
| 19 | my ($self) = @_; | ||
| 20 | if (!$self->{'is_compiled'}) { | ||
| 21 | my %source2eval; | ||
| 22 | $self->{'source2eval'} = \%source2eval; | ||
| 23 | foreach my $term (values %{ $self->{'terms'} }) { | ||
| 24 | my ($name, $source, $derived, $norm, $match, $repeat) | ||
| 25 | = @$term{qw(name source derived norm match repeat)}; | ||
| 26 | my $eval; | ||
| 27 | if ($derived) { | ||
| 28 | $eval = derived2eval($term); | ||
| 29 | } | ||
| 30 | else { | ||
| 31 | $eval = $source2eval{$source} ||= source2eval($source); | ||
| 32 | $self->{'want'}{source2tag($source)} = 1; | ||
| 33 | } | ||
| 34 | $term->{'eval'} = $eval; | ||
| 35 | $term->{'norm'} = mknorm(@$norm); | ||
| 36 | } | ||
| 37 | } | ||
| 38 | $self->{'is_compiled'} = 1; | ||
| 39 | } | ||
| 40 | |||
| 41 | sub mknorm { | ||
| 42 | return sub { $_ } if !@_ || !defined $_[0]; | ||
| 43 | my @subs = map { __PACKAGE__->can('norm_'.trim($_)) || die "Unknown normalizer: $_" } @_; | ||
| 44 | return sub { | ||
| 45 | my ($v) = @_; | ||
| 46 | return if !defined $v; | ||
| 47 | foreach my $sub (@subs) { | ||
| 48 | $v = $sub->($v); | ||
| 49 | return if !defined $v; | ||
| 50 | } | ||
| 51 | return $v; | ||
| 52 | }; | ||
| 53 | } | ||
| 54 | |||
| 55 | sub derived2eval { | ||
| 56 | my ($term) = @_; | ||
| 57 | if ($term->{'permuted'}) { | ||
| 58 | |||
| 59 | } | ||
| 60 | else { | ||
| 61 | |||
| 62 | } | ||
| 63 | } | ||
| 64 | |||
| 65 | sub source2tag { | ||
| 66 | return 'M' if $_[0] =~ /^::/; | ||
| 67 | return 'L' if $_[0] =~ /^(?:L|leader)\b/; | ||
| 68 | return $1 if $_[0] =~ /^raw:(...)/; | ||
| 69 | return substr($_[0], 0, 3); | ||
| 70 | } | ||
| 71 | |||
| 72 | sub index { | ||
| 73 | my ($self, $marc) = @_; | ||
| 74 | my $marcref = ref($marc) ? $marc : \$marc; | ||
| 75 | # 1. Compile the indexing code if needed | ||
| 76 | $self->compile if !$self->{'is_compiled'}; | ||
| 77 | # 2. Gather all fields used for indexing | ||
| 78 | my $data = marcdata($marcref, $self->{'want'}); | ||
| 79 | # 3. Index each term | ||
| 80 | my %index; | ||
| 81 | while (my ($name, $term) = each %{ $self->{'terms'} }) { | ||
| 82 | next if $term->{'derived'}; | ||
| 83 | my $tag = source2tag($term->{'source'}); | ||
| 84 | my @vals = @{ $data->{$tag} or next }; | ||
| 85 | my $eval = $term->{'eval'} or next; | ||
| 86 | my $norm = $term->{'norm'}; | ||
| 87 | @vals = map { $eval->($_) } @vals; | ||
| 88 | @vals = map { $norm->($_) } @vals if $norm; | ||
| 89 | next if !@vals; | ||
| 90 | splice @vals, 1 if !$term->{'repeat'}; | ||
| 91 | push @{ $index{$name} ||= [] }, @vals; | ||
| 92 | } | ||
| 93 | return \%index; | ||
| 94 | } | ||
| 95 | |||
| 96 | sub marcdata { | ||
| 97 | my ($marcref, $want) = @_; | ||
| 98 | my %data; | ||
| 99 | $data{'M'} = [$marcref]; | ||
| 100 | $data{'L'} = [substr($$marcref, 0, 24)] if $want->{'L'}; | ||
| 101 | my $baseaddr = substr($$marcref, 12, 5) + 0; | ||
| 102 | pos($$marcref) = 24; | ||
| 103 | while ($$marcref =~ /\G([0-9A-Za-z]{3})([0-9]{4})([0-9]{5})/gc) { | ||
| 104 | push @{ $data{$1} ||= [] }, substr($$marcref, $baseaddr + $3, $2 - 1) if $want->{$1}; | ||
| 105 | } | ||
| 106 | return \%data; | ||
| 107 | } | ||
| 108 | |||
| 109 | sub source2eval { | ||
| 110 | my ($source) = @_; | ||
| 111 | if ($source =~ m{^(?:L|leader)/(\d+)$}) { | ||
| 112 | my $pos = $1; | ||
| 113 | return sub { substr($_, $pos, 1) }; | ||
| 114 | } | ||
| 115 | elsif ($source =~ m{^(00[0-9A-Za-z])(?:/(\d+)(?:-(\d+))?)?$}) { | ||
| 116 | my ($tag, $b1, $bn) = ($1, $2, $3); | ||
| 117 | return sub { substr($_, $b1,$bn-$b1+1) } if defined $bn; | ||
| 118 | return sub { substr($_, $b1, 1) } if defined $b1; | ||
| 119 | return sub { $_ }; | ||
| 120 | } | ||
| 121 | elsif ($source =~ m{^::(\w+)}) { | ||
| 122 | my $sub = __PACKAGE__->can('src_'.$1) || die "Unknown source function: $1"; | ||
| 123 | return sub { | ||
| 124 | my ($marcref) = @_; | ||
| 125 | return $sub->($marcref); | ||
| 126 | } | ||
| 127 | } | ||
| 128 | elsif ($source =~ m{^(raw:)?([0-9A-Za-z]{3})\$(.+)$}) { | ||
| 129 | my ($raw, $tag, @subs) = ($1, $2, split //, $3); | ||
| 130 | my %want = map { $_ => 1 } @subs; | ||
| 131 | if ($raw) { | ||
| 132 | return sub { | ||
| 133 | my @pieces = ( substr($_, 0, 2) ); | ||
| 134 | pos($_) = 2; | ||
| 135 | while (/(\x1f([^\x1f])[^\x1f]+)/gc) { | ||
| 136 | push @pieces, $1 if $want{$2}; | ||
| 137 | } | ||
| 138 | return if @pieces < 2; | ||
| 139 | return join('', @pieces); | ||
| 140 | }; | ||
| 141 | } | ||
| 142 | else { | ||
| 143 | return sub { | ||
| 144 | pos($_) = 2; | ||
| 145 | my @subs; | ||
| 146 | while (/\x1f([^\x1f])([^\x1f]+)/gc) { | ||
| 147 | push @subs, $2 if $want{$1}; | ||
| 148 | } | ||
| 149 | return if !@subs; | ||
| 150 | return join(' ', @subs); | ||
| 151 | }; | ||
| 152 | } | ||
| 153 | } | ||
| 154 | } | ||
| 155 | |||
| 156 | sub trim { | ||
| 157 | local $_ = shift; | ||
| 158 | s/^\s+|\s+$//g; | ||
| 159 | return $_; | ||
| 160 | } | ||
| 161 | |||
| 162 | sub norm_blank2hash { | ||
| 163 | local $_ = shift; | ||
| 164 | tr/ /#/; | ||
| 165 | return $_; | ||
| 166 | } | ||
| 167 | |||
| 168 | sub norm_date2unix { | ||
| 169 | local $_ = shift; | ||
| 170 | /^([0-9]{4})-?([0-9]{2})-?([0-9]{2})T?([0-9]{2}):?([0-9]{2}):?([0-9]{2})/ or return 0; | ||
| 171 | return strftime('%s', $6, $5, $4, $3, $2-1, $1-1900); | ||
| 172 | } | ||
| 173 | |||
| 174 | sub norm_nfc { | ||
| 175 | NFC(shift); | ||
| 176 | } | ||
| 177 | |||
| 178 | sub norm_nfd { | ||
| 179 | NFD(shift); | ||
| 180 | } | ||
| 181 | |||
| 182 | sub norm_alpha { | ||
| 183 | local $_ = shift; | ||
| 184 | tr{'}{}d; | ||
| 185 | tr{-=./()[],:;"?!}{ }; | ||
| 186 | s/[^[:alpha] ]//g; | ||
| 187 | tr{ }{}s; | ||
| 188 | s/^ | $//g; | ||
| 189 | return $_ | ||
| 190 | } | ||
| 191 | |||
| 192 | sub norm_uppercase { | ||
| 193 | uc shift; | ||
| 194 | } | ||
| 195 | |||
| 196 | sub norm_lowercase { | ||
| 197 | lc shift; | ||
| 198 | } | ||
| 199 | |||
| 200 | sub norm_numeric { | ||
| 201 | local $_ = shift; | ||
| 202 | die if !/^[0-9]+$/; | ||
| 203 | $_ + 0; # This strips leading zeroes | ||
| 204 | } | ||
| 205 | |||
| 206 | sub norm_trim { | ||
| 207 | local $_ = shift; | ||
| 208 | tr/ / /s; | ||
| 209 | s/^ | $//g; | ||
| 210 | return $_; | ||
| 211 | } | ||
| 212 | |||
| 213 | sub norm_isbn { | ||
| 214 | local $_ = shift; | ||
| 215 | tr/-//d; | ||
| 216 | my $n = length; | ||
| 217 | die if $n != 10 && $n != 13; | ||
| 218 | norm_trim(uc $_); | ||
| 219 | } | ||
| 220 | |||
| 221 | sub norm_delnf1 { unshift @_, 1; goto &_remove_non_filing_chars } | ||
| 222 | sub norm_delnf2 { unshift @_, 2; goto &_remove_non_filing_chars } | ||
| 223 | sub norm_cook { | ||
| 224 | local $_ = shift; | ||
| 225 | s/^..\x1f.//; | ||
| 226 | s/\x1f./ /g; | ||
| 227 | return $_; | ||
| 228 | } | ||
| 229 | |||
| 230 | sub src_pub_date { | ||
| 231 | my ($marcref) = @_; | ||
| 232 | my ($leader, $fields) = marcparse($marcref); | ||
| 233 | my ($f008) = grep { $_->[TAG] eq '008' } @$fields; | ||
| 234 | return ($1) if defined($f008) && ${ $f008->[VALREF] } =~ /^.{7}([0-9]{4})/; | ||
| 235 | my ($f260) = grep { $_->[TAG] eq '260' } @$fields; | ||
| 236 | return ($1) if defined($f260) && ${ $f260->[VALREF] } =~ /\x1fc([0-9]{4})/; | ||
| 237 | return (); | ||
| 238 | } | ||
| 239 | |||
| 240 | sub _remove_non_filing_chars { | ||
| 241 | my $ind = shift; | ||
| 242 | local $_ = shift; | ||
| 243 | my $n = substr($_, $ind-1, 1); | ||
| 244 | if ($n =~ /[1-9]/) { | ||
| 245 | s/(?<=\x1fa)([^\x1f]+)/length($1) <= $n ? $1 : substr($1, $n)/e; | ||
| 246 | } | ||
| 247 | return $_; | ||
| 248 | } | ||
| 249 | |||
| 250 | 1; | ||
| 251 | |||
| 252 | =pod | ||
| 253 | |||
| 254 | =head1 NAME | ||
| 255 | |||
| 256 | MARC::Indexer - index MARC records | ||
| 257 | |||
| 258 | =head1 SYNOPSIS | ||
| 259 | |||
| 260 | $indexer = MARC::Indexer->new( | ||
| 261 | 'index_points' => [ | ||
| 262 | { 'name' => 'ctrlnum', | ||
| 263 | 'tag' => '001' }, | ||
| 264 | { 'name' => 'title', | ||
| 265 | 'tag' => '245', | ||
| 266 | 'extract' => 'subfields:abcdnp', | ||
| 267 | 'normalize' => [qw(text nfd uc)], | ||
| 268 | }, | ||
| 269 | { 'name' => 'subjtopic', | ||
| 270 | 'tag' => '650', | ||
| 271 | 'repeatable' => 1, | ||
| 272 | ... | ||
| 273 | }, | ||
| 274 | ... | ||
| 275 | ], | ||
| 276 | ); | ||
| 277 | $idx = $indexer->index($marc); | ||
| 278 | $idx_title = $idx->{'title'}; | ||
| 279 | |||
| 280 | =cut | ||
| 281 |