C4/External/Syndetics.pm

   1 package C4::External::Syndetics;
   2 # Copyright (C) 2006 LibLime
   3 # <jmf at liblime dot com>
   4 #
   5 # This file is part of Koha.
   6 #
   7 # Koha is free software; you can redistribute it and/or modify it under the
   8 # terms of the GNU General Public License as published by the Free Software
   9 # Foundation; either version 2 of the License, or (at your option) any later
  10 # version.
  11 #
  12 # Koha is distributed in the hope that it will be useful, but WITHOUT ANY
  13 # WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
  14 # A PARTICULAR PURPOSE.  See the GNU General Public License for more details.
  15 #
  16 # You should have received a copy of the GNU General Public License along
  17 # with Koha; if not, write to the Free Software Foundation, Inc.,
  18 # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
  19
  20 use XML::Simple;
  21 use XML::LibXML;
  22 use LWP::Simple;
  23 use LWP::UserAgent;
  24 use HTTP::Request::Common;
  25
  26 use strict;
  27 use warnings;
  28
  29 use vars qw($VERSION @ISA @EXPORT);
  30
  31 BEGIN {
  32     require Exporter;
  33     $VERSION = 3.07.00.049;
  34     @ISA = qw(Exporter);
  35     @EXPORT = qw(
  36         &get_syndetics_index
  37         &get_syndetics_summary
  38         &get_syndetics_toc
  39         &get_syndetics_editions
  40         &get_syndetics_excerpt
  41         &get_syndetics_reviews
  42         &get_syndetics_anotes
  43     );
  44 }
  45
  46 # package-level variable
  47 my $parser = XML::LibXML->new();
  48
  49 =head1 NAME
  50
  51 C4::External::Syndetics - Functions for retrieving Syndetics content in Koha
  52
  53 =head1 FUNCTIONS
  54
  55 This module provides facilities for retrieving Syndetics.com content in Koha
  56
  57 =head2 get_syndetics_summary
  58
  59   my $syndetics_summary= &get_syndetics_summary( $isbn );
  60
  61 Get Summary data from Syndetics
  62
  63 =cut
  64
  65 sub get_syndetics_index {
  66     my ( $isbn, $upc, $oclc ) = @_;
  67
  68     return unless ( $isbn || $upc || $oclc );
  69
  70     my $response = _fetch_syndetics_content('INDEX.XML', $isbn, $upc, $oclc);
  71     unless ($response->content_type =~ /xml/) {
  72         return;
  73     }
  74
  75     my $content = $response->content;
  76     my $xmlsimple = XML::Simple->new();
  77     $response = $xmlsimple->XMLin(
  78         $content,
  79     ) unless !$content;
  80
  81     my $syndetics_elements;
  82     for my $available_type ('SUMMARY','TOC','FICTION','AWARDS1','SERIES1','SPSUMMARY','SPREVIEW', 'AVPROFILE', 'AVSUMMARY','DBCHAPTER','LJREVIEW','PWREVIEW','SLJREVIEW','CHREVIEW','BLREVIEW','HBREVIEW','KIREVIEW','CRITICASREVIEW','ANOTES') {
  83         if (exists $response->{$available_type} && $response->{$available_type} =~ /$available_type/) {
  84             $syndetics_elements->{$available_type} = $available_type;
  85         }
  86     }
  87
  88     return $syndetics_elements if $syndetics_elements;
  89 }
  90
  91 sub get_syndetics_summary {
  92     my ( $isbn, $upc, $oclc, $syndetics_elements ) = @_;
  93
  94     my $summary_type = exists($syndetics_elements->{'AVSUMMARY'}) ? 'AVSUMMARY.XML' : 'SUMMARY.XML';
  95     my $response = _fetch_syndetics_content($summary_type, $isbn, $upc, $oclc);
  96     unless ($response->content_type =~ /xml/) {
  97         return;
  98     }
  99
 100     my $content = $response->content;
 101
 102     my $summary;
 103     eval {
 104         my $doc = $parser->parse_string($content);
 105         $summary = $doc->findvalue('//Fld520');
 106     };
 107     if ($@) {
 108         warn "Error parsing Syndetics $summary_type";
 109     }
 110     return $summary if $summary;
 111 }
 112
 113 sub get_syndetics_toc {
 114     my ( $isbn,$upc,$oclc ) = @_;
 115
 116     my $response = _fetch_syndetics_content('TOC.XML', $isbn, $upc, $oclc);
 117     unless ($response->content_type =~ /xml/) {
 118         return;
 119     }
 120
 121     my $content = $response->content;
 122     my $xmlsimple = XML::Simple->new();
 123     $response = $xmlsimple->XMLin(
 124         $content,
 125         forcearray => [ qw(Fld970) ],
 126     ) unless !$content;
 127     # manipulate response USMARC VarFlds VarDFlds Notes Fld520 a
 128     my $toc;
 129     $toc = \@{$response->{VarFlds}->{VarDFlds}->{SSIFlds}->{Fld970}} if $response;
 130     return $toc if $toc;
 131 }
 132
 133 sub get_syndetics_excerpt {
 134     my ( $isbn,$upc,$oclc ) = @_;
 135
 136     my $response = _fetch_syndetics_content('DBCHAPTER.XML', $isbn, $upc, $oclc);
 137     unless ($response->content_type =~ /xml/) {
 138         return;
 139     }
 140
 141     my $content = $response->content;
 142     my $xmlsimple = XML::Simple->new();
 143     $response = $xmlsimple->XMLin(
 144         $content,
 145         forcearray => [ qw(Fld520) ],
 146     ) unless !$content;
 147     # manipulate response USMARC VarFlds VarDFlds Notes Fld520 a
 148     my $excerpt;
 149     $excerpt = \@{$response->{VarFlds}->{VarDFlds}->{Notes}->{Fld520}} if $response;
 150     return XMLout($excerpt, NoEscape => 1) if $excerpt;
 151 }
 152
 153 sub get_syndetics_reviews {
 154     my ( $isbn,$upc,$oclc,$syndetics_elements ) = @_;
 155
 156     my @reviews;
 157     my $review_sources = [
 158     {title => 'Library Journal Review', file => 'LJREVIEW.XML', element => 'LJREVIEW'},
 159     {title => 'Publishers Weekly Review', file => 'PWREVIEW.XML', element => 'PWREVIEW'},
 160     {title => 'School Library Journal Review', file => 'SLJREVIEW.XML', element => 'SLJREVIEW'},
 161     {title => 'CHOICE Review', file => 'CHREVIEW.XML', element => 'CHREVIEW'},
 162     {title => 'Booklist Review', file => 'BLREVIEW.XML', element => 'BLREVIEW'},
 163     {title => 'Horn Book Review', file => 'HBREVIEW.XML', element => 'HBREVIEW'},
 164     {title => 'Kirkus Book Review', file => 'KIREVIEW.XML', element => 'KIREVIEW'},
 165     {title => 'Criticas Review', file => 'CRITICASREVIEW.XML', element => 'CRITICASREVIEW'},
 166     {title => 'Spanish Review', file => 'SPREVIEW.XML', element => 'SPREVIEW'},
 167     ];
 168
 169     for my $source (@$review_sources) {
 170         if ($syndetics_elements->{$source->{element}} and $source->{element} =~ $syndetics_elements->{$source->{element}}) {
 171
 172         } else {
 173             #warn "Skipping $source->{element} doesn't match $syndetics_elements->{$source->{element}} \n";
 174             next;
 175         }
 176         my $response = _fetch_syndetics_content($source->{file}, $isbn, $upc, $oclc);
 177         unless ($response->content_type =~ /xml/) {
 178             next;
 179         }
 180
 181         my $content = $response->content;
 182
 183         eval {
 184             my $doc = $parser->parse_string($content);
 185
 186             # note that using findvalue strips any HTML elements embedded
 187             # in that review.  That helps us handle slight differences
 188             # in the output provided by Syndetics 'old' and 'new' versions
 189             # of their service and cleans any questionable HTML that
 190             # may be present in the reviews, but does mean that any
 191             # <B> and <I> tags used to format the review are also gone.
 192             my $result = $doc->findvalue('//Fld520');
 193             push @reviews, {title => $source->{title}, reviews => [ { content => $result } ]} if $result;
 194         };
 195         if ($@) {
 196             warn "Error parsing Syndetics $source->{title} review";
 197         }
 198     }
 199     return \@reviews;
 200 }
 201
 202 sub get_syndetics_editions {
 203     my ( $isbn,$upc,$oclc ) = @_;
 204
 205     my $response = _fetch_syndetics_content('FICTION.XML', $isbn, $upc, $oclc);
 206     unless ($response->content_type =~ /xml/) {
 207         return;
 208     }
 209
 210     my $content = $response->content;
 211
 212     my $xmlsimple = XML::Simple->new();
 213     $response = $xmlsimple->XMLin(
 214         $content,
 215         forcearray => [ qw(Fld020) ],
 216     ) unless !$content;
 217     # manipulate response USMARC VarFlds VarDFlds Notes Fld520 a
 218     my $similar_items;
 219     $similar_items = \@{$response->{VarFlds}->{VarDFlds}->{NumbCode}->{Fld020}} if $response;
 220     return $similar_items if $similar_items;
 221 }
 222
 223 sub get_syndetics_anotes {
 224     my ( $isbn,$upc,$oclc) = @_;
 225
 226     my $response = _fetch_syndetics_content('ANOTES.XML', $isbn, $upc, $oclc);
 227     unless ($response->content_type =~ /xml/) {
 228         return;
 229     }
 230
 231     my $content = $response->content;
 232
 233     my $xmlsimple = XML::Simple->new();
 234     $response = $xmlsimple->XMLin(
 235         $content,
 236         forcearray => [ qw(Fld980) ],
 237         ForceContent => 1,
 238     ) unless !$content;
 239     my @anotes;
 240     for my $fld980 (@{$response->{VarFlds}->{VarDFlds}->{SSIFlds}->{Fld980}}) {
 241         # this is absurd, but sometimes this data serializes differently
 242         if(ref($fld980->{a}->{content}) eq 'ARRAY') {
 243             for my $content (@{$fld980->{a}->{content}}) {
 244                 push @anotes, {content => $content};
 245
 246             }
 247         }
 248         else {
 249             push @anotes, {content => $fld980->{a}->{content}};
 250         }
 251     }
 252     return \@anotes;
 253 }
 254
 255 sub _fetch_syndetics_content {
 256     my ( $element, $isbn, $upc, $oclc ) = @_;
 257
 258     $isbn = '' unless defined $isbn;
 259     $upc  = '' unless defined $upc;
 260     $oclc = '' unless defined $oclc;
 261
 262     my $syndetics_client_code = C4::Context->preference('SyndeticsClientCode');
 263
 264     my $url = "http://www.syndetics.com/index.aspx?isbn=$isbn/$element&client=$syndetics_client_code&type=xw10&upc=$upc&oclc=$oclc";
 265     my $ua = LWP::UserAgent->new;
 266     $ua->timeout(10);
 267     $ua->env_proxy;
 268     my $response = $ua->get($url);
 269
 270     warn "could not retrieve $url" unless $response->content;
 271     return $response;
 272
 273 }
 274 1;
 275 __END__
 276
 277 =head1 NOTES
 278
 279 =cut
 280
 281 =head1 AUTHOR
 282
 283 Joshua Ferraro <jmf@liblime.com>
 284
 285 =cut