blib/lib/WWW/Scraper/ISBN/Pearson_Driver.pm | |||
---|---|---|---|
Criterion | Covered | Total | % |
statement | 56 | 56 | 100.0 |
branch | 7 | 12 | 58.3 |
condition | 4 | 12 | 33.3 |
subroutine | 8 | 8 | 100.0 |
pod | 1 | 1 | 100.0 |
total | 76 | 89 | 85.3 |
line | stmt | bran | cond | sub | pod | time | code |
---|---|---|---|---|---|---|---|
1 | package WWW::Scraper::ISBN::Pearson_Driver; | ||||||
2 | |||||||
3 | 6 | 6 | 96794 | use strict; | |||
6 | 14 | ||||||
6 | 212 | ||||||
4 | 6 | 6 | 19 | use warnings; | |||
6 | 7 | ||||||
6 | 189 | ||||||
5 | |||||||
6 | 6 | 6 | 21 | use vars qw($VERSION @ISA); | |||
6 | 12 | ||||||
6 | 504 | ||||||
7 | $VERSION = '0.22'; | ||||||
8 | |||||||
9 | #-------------------------------------------------------------------------- | ||||||
10 | |||||||
11 | =head1 NAME | ||||||
12 | |||||||
13 | WWW::Scraper::ISBN::Pearson_Driver - Search driver for the Pearson Education online book catalog. | ||||||
14 | |||||||
15 | =head1 SYNOPSIS | ||||||
16 | |||||||
17 | See parent class documentation (L |
||||||
18 | |||||||
19 | =head1 DESCRIPTION | ||||||
20 | |||||||
21 | Searches for book information from the Pearson Education's online catalog. | ||||||
22 | |||||||
23 | =cut | ||||||
24 | |||||||
25 | #-------------------------------------------------------------------------- | ||||||
26 | |||||||
27 | ########################################################################### | ||||||
28 | # Inheritence | ||||||
29 | |||||||
30 | 6 | 6 | 26 | use base qw(WWW::Scraper::ISBN::Driver); | |||
6 | 7 | ||||||
6 | 2784 | ||||||
31 | |||||||
32 | ########################################################################### | ||||||
33 | # Modules | ||||||
34 | |||||||
35 | 6 | 6 | 8929 | use WWW::Mechanize; | |||
6 | 772507 | ||||||
6 | 246 | ||||||
36 | |||||||
37 | ########################################################################### | ||||||
38 | # Constants | ||||||
39 | |||||||
40 | 6 | 6 | 50 | use constant SEARCH => 'http://www.pearsoned.co.uk/Bookshop/'; | |||
6 | 8 | ||||||
6 | 402 | ||||||
41 | 6 | 6 | 28 | use constant DETAIL => 'http://www.pearsoned.co.uk/Bookshop/detail.asp?item='; | |||
6 | 7 | ||||||
6 | 3370 | ||||||
42 | |||||||
43 | #-------------------------------------------------------------------------- | ||||||
44 | |||||||
45 | ########################################################################### | ||||||
46 | # Public Interface | ||||||
47 | |||||||
48 | =head1 METHODS | ||||||
49 | |||||||
50 | =over 4 | ||||||
51 | |||||||
52 | =item C |
||||||
53 | |||||||
54 | Creates a query string, then passes the appropriate form fields to the Pearson | ||||||
55 | Education server. | ||||||
56 | |||||||
57 | The returned page should be the correct catalog page for that ISBN. If not the | ||||||
58 | function returns zero and allows the next driver in the chain to have a go. If | ||||||
59 | a valid page is returned, the following fields are returned via the book hash: | ||||||
60 | |||||||
61 | isbn (now returns isbn13) | ||||||
62 | isbn10 | ||||||
63 | isbn13 | ||||||
64 | ean13 (industry name) | ||||||
65 | author | ||||||
66 | title | ||||||
67 | book_link | ||||||
68 | image_link | ||||||
69 | description | ||||||
70 | pubdate | ||||||
71 | publisher | ||||||
72 | binding (if known) | ||||||
73 | pages (if known) | ||||||
74 | weight (if known) (in grammes) | ||||||
75 | width (if known) (in millimetres) | ||||||
76 | height (if known) (in millimetres) | ||||||
77 | |||||||
78 | The book_link and image_link refer back to the Pearson Education UK website. | ||||||
79 | |||||||
80 | =back | ||||||
81 | |||||||
82 | =cut | ||||||
83 | |||||||
84 | sub search { | ||||||
85 | 2 | 2 | 1 | 7328 | my $self = shift; | ||
86 | 2 | 4 | my $isbn = shift; | ||||
87 | 2 | 10 | $self->found(0); | ||||
88 | 2 | 28 | $self->book(undef); | ||||
89 | |||||||
90 | 2 | 24 | my $mech = WWW::Mechanize->new(); | ||||
91 | 2 | 12724 | $mech->agent_alias( 'Linux Mozilla' ); | ||||
92 | |||||||
93 | 2 | 121 | eval { $mech->get( SEARCH ) }; | ||||
2 | 7 | ||||||
94 | 2 | 50 | 33 | 428350 | return $self->handler("Pearson Education website appears to be unavailable.") | ||
33 | |||||||
95 | if($@ || !$mech->success() || !$mech->content()); | ||||||
96 | |||||||
97 | #print STDERR "\n# content=[\n".$mech->content()."\n]\n"; | ||||||
98 | |||||||
99 | 2 | 144 | $mech->form_id('hipGlobalSearchForm'); | ||||
100 | 2 | 17905 | $mech->set_fields( 'txtSearch' => $isbn ); | ||||
101 | |||||||
102 | 2 | 166 | eval { $mech->submit() }; | ||||
2 | 12 | ||||||
103 | 2 | 50 | 33 | 1293616 | return $self->handler("Failed to find that book on Pearson Education website.") | ||
33 | |||||||
104 | if($@ || !$mech->success() || !$mech->content()); | ||||||
105 | |||||||
106 | # The Book page | ||||||
107 | 2 | 106 | my $html = $mech->content(); | ||||
108 | #print STDERR "\n# html=[\n$html\n]\n"; | ||||||
109 | |||||||
110 | 2 | 50 | 1583 | return $self->handler("Failed to find that book on Pearson Education website.") | |||
111 | if($html =~ m! Your search for \d+ returned 0 results. Please search again. !si); |
||||||
112 | |||||||
113 | 2 | 4 | my $data; | ||||
114 | 2 | 196 | ($data->{image}) = $html =~ m!"(http://images.pearsoned-ema.com/jpeg/large/\d+\.jpg)"!i; | ||||
115 | 2 | 190 | ($data->{thumb}) = $html =~ m!"(http://images.pearsoned-ema.com/jpeg/small/\d+\.jpg)"!i; | ||||
116 | 2 | 1384 | ($data->{title}, | ||||
117 | $data->{author}, | ||||||
118 | $data->{pubdate}, | ||||||
119 | $data->{binding}, | ||||||
120 | $data->{pages}, | ||||||
121 | $data->{isbn13}, | ||||||
122 | $data->{isbn10}) = $html =~ m! \s* |
||||||
123 | (.*?)\s*(?:.*?\s*)? |
||||||
124 | ]+>(.*?)\s*\s* |
||||||
125 | ([^,]+),\s*([^,]+)(?:,\s*(\d+)\s+pages)? \s* |
||||||
126 | ISBN13:\s*(\d+)\s* \s* |
||||||
127 | ISBN10:\s*([\dX]+)\s* !ix; |
||||||
128 | 2 | 1245 | ($data->{description}) = $html =~ m! ([^<]+)!is; |
||||
129 | 2 | 838 | ($data->{bookid}) = $html =~ m!recommend.asp\?item=(\d+)!i; | ||||
130 | |||||||
131 | #use Data::Dumper; | ||||||
132 | #print STDERR "\n# " . Dumper($data); | ||||||
133 | |||||||
134 | 2 | 50 | 10 | return $self->handler("Could not extract data from Pearson Education result page.") | |||
135 | unless(defined $data); | ||||||
136 | |||||||
137 | # remove HTML tags | ||||||
138 | 2 | 6 | for(qw(author binding)) { | ||||
139 | 4 | 50 | 11 | next unless(defined $data->{$_}); | |||
140 | 4 | 37 | $data->{$_} =~ s!<[^>]+>!!g; | ||||
141 | } | ||||||
142 | |||||||
143 | # trim top and tail | ||||||
144 | 2 | 12 | for(keys %$data) { | ||||
145 | 22 | 100 | 36 | next unless(defined $data->{$_}); | |||
146 | 21 | 36 | $data->{$_} =~ s/^\s+//; | ||||
147 | 21 | 69 | $data->{$_} =~ s/\s+$//; | ||||
148 | } | ||||||
149 | |||||||
150 | 2 | 15 | my $uri = $mech->uri(); #DETAIL . $data->{bookid}; | ||||
151 | |||||||
152 | 2 | 61 | my $bk = { | ||||
153 | 'ean13' => $data->{isbn13}, | ||||||
154 | 'isbn13' => $data->{isbn13}, | ||||||
155 | 'isbn10' => $data->{isbn10}, | ||||||
156 | 'isbn' => $data->{isbn13}, | ||||||
157 | 'author' => $data->{author}, | ||||||
158 | 'title' => $data->{title}, | ||||||
159 | 'book_link' => "$uri", | ||||||
160 | 'image_link' => $data->{image}, | ||||||
161 | 'thumb_link' => $data->{thumb}, | ||||||
162 | 'description' => $data->{description}, | ||||||
163 | 'pubdate' => $data->{pubdate}, | ||||||
164 | 'publisher' => q!Pearson Education!, | ||||||
165 | 'binding' => $data->{binding}, | ||||||
166 | 'pages' => $data->{pages}, | ||||||
167 | 'weight' => $data->{weight}, | ||||||
168 | 'width' => $data->{width}, | ||||||
169 | 'height' => $data->{height}, | ||||||
170 | 'html' => $html | ||||||
171 | }; | ||||||
172 | |||||||
173 | #use Data::Dumper; | ||||||
174 | #print STDERR "\n# book=".Dumper($bk); | ||||||
175 | |||||||
176 | 2 | 47 | $self->book($bk); | ||||
177 | 2 | 32 | $self->found(1); | ||||
178 | 2 | 15 | return $self->book; | ||||
179 | } | ||||||
180 | |||||||
181 | 1; | ||||||
182 | __END__ |