blib/lib/WWW/Scraper/ISBN/TheNile_Driver.pm | |||
---|---|---|---|
Criterion | Covered | Total | % |
statement | 60 | 60 | 100.0 |
branch | 10 | 18 | 55.5 |
condition | 8 | 18 | 44.4 |
subroutine | 7 | 7 | 100.0 |
pod | 1 | 1 | 100.0 |
total | 86 | 104 | 82.6 |
line | stmt | bran | cond | sub | pod | time | code |
---|---|---|---|---|---|---|---|
1 | package WWW::Scraper::ISBN::TheNile_Driver; | ||||||
2 | |||||||
3 | 6 | 6 | 219242 | use strict; | |||
6 | 25 | ||||||
6 | 282 | ||||||
4 | 6 | 6 | 36 | use warnings; | |||
6 | 10 | ||||||
6 | 259 | ||||||
5 | |||||||
6 | 6 | 6 | 32 | use vars qw($VERSION @ISA); | |||
6 | 15 | ||||||
6 | 617 | ||||||
7 | $VERSION = '0.13'; | ||||||
8 | |||||||
9 | #-------------------------------------------------------------------------- | ||||||
10 | |||||||
11 | =head1 NAME | ||||||
12 | |||||||
13 | WWW::Scraper::ISBN::TheNile_Driver - Search driver for TheNile 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 TheNile online book catalog | ||||||
22 | |||||||
23 | =cut | ||||||
24 | |||||||
25 | #-------------------------------------------------------------------------- | ||||||
26 | |||||||
27 | ########################################################################### | ||||||
28 | # Inheritence | ||||||
29 | |||||||
30 | 6 | 6 | 37 | use base qw(WWW::Scraper::ISBN::Driver); | |||
6 | 14 | ||||||
6 | 333146 | ||||||
31 | |||||||
32 | ########################################################################### | ||||||
33 | # Modules | ||||||
34 | |||||||
35 | 6 | 6 | 20867 | use WWW::Mechanize; | |||
6 | 1464460 | ||||||
6 | 330 | ||||||
36 | |||||||
37 | ########################################################################### | ||||||
38 | # Constants | ||||||
39 | |||||||
40 | 6 | 6 | 191 | use constant SEARCH => 'http://www.thenile.com.au/search.php?s='; | |||
6 | 14 | ||||||
6 | 8347 | ||||||
41 | |||||||
42 | #-------------------------------------------------------------------------- | ||||||
43 | |||||||
44 | ########################################################################### | ||||||
45 | # Public Interface | ||||||
46 | |||||||
47 | =head1 METHODS | ||||||
48 | |||||||
49 | =over 4 | ||||||
50 | |||||||
51 | =item C |
||||||
52 | |||||||
53 | Creates a query string, then passes the appropriate form fields to the | ||||||
54 | TheNile server. | ||||||
55 | |||||||
56 | The returned page should be the correct catalog page for that ISBN. If not the | ||||||
57 | function returns zero and allows the next driver in the chain to have a go. If | ||||||
58 | a valid page is returned, the following fields are returned via the book hash: | ||||||
59 | |||||||
60 | isbn (now returns isbn13) | ||||||
61 | isbn10 | ||||||
62 | isbn13 | ||||||
63 | ean13 (industry name) | ||||||
64 | author | ||||||
65 | title | ||||||
66 | book_link | ||||||
67 | image_link | ||||||
68 | description | ||||||
69 | pubdate | ||||||
70 | publisher | ||||||
71 | binding (if known) | ||||||
72 | pages (if known) | ||||||
73 | weight (if known) (in grammes) | ||||||
74 | width (if known) (in millimetres) | ||||||
75 | height (if known) (in millimetres) | ||||||
76 | |||||||
77 | The book_link and image_link refer back to the TheNile website. | ||||||
78 | |||||||
79 | =back | ||||||
80 | |||||||
81 | =cut | ||||||
82 | |||||||
83 | sub search { | ||||||
84 | 4 | 4 | 1 | 35146 | my $self = shift; | ||
85 | 4 | 384 | my $isbn = shift; | ||||
86 | 4 | 27 | $self->found(0); | ||||
87 | 4 | 70 | $self->book(undef); | ||||
88 | |||||||
89 | # validate and convert into EAN13 format | ||||||
90 | 4 | 54 | my $ean = $self->convert_to_ean13($isbn); | ||||
91 | 4 | 50 | 66 | 267 | return $self->handler("Invalid ISBN specified [$isbn]") | ||
33 | |||||||
66 | |||||||
33 | |||||||
92 | if(!$ean || (length $isbn == 13 && $isbn ne $ean) | ||||||
93 | || (length $isbn == 10 && $isbn ne $self->convert_to_isbn10($ean))); | ||||||
94 | 4 | 55 | $isbn = $ean; | ||||
95 | |||||||
96 | 4 | 37 | my $mech = WWW::Mechanize->new(); | ||||
97 | 4 | 18441 | $mech->agent_alias( 'Linux Mozilla' ); | ||||
98 | |||||||
99 | 4 | 253 | eval { $mech->get( SEARCH . $isbn ) }; | ||||
4 | 24 | ||||||
100 | 4 | 50 | 33 | 3565673 | return $self->handler("TheNile website appears to be unavailable.") | ||
33 | |||||||
101 | if($@ || !$mech->success() || !$mech->content()); | ||||||
102 | |||||||
103 | # The Book page | ||||||
104 | 4 | 282 | my $html = $mech->content(); | ||||
105 | |||||||
106 | 4 | 50 | 6326 | return $self->handler("Failed to find that book on TheNile website.") | |||
107 | if($html =~ m!Sorry your search.*?did not return any results|This book is currently unavailable!si); | ||||||
108 | |||||||
109 | #print STDERR "\n# html=[\n$html\n]\n"; | ||||||
110 | |||||||
111 | 4 | 101 | my $data; | ||||
112 | 4 | 227 | ($data->{image}) = $html =~ m!(http://c\d+\.mrcdn\.net/[\w\.\/]+\.jpg)!si; | ||||
113 | 4 | 183 | ($data->{thumb}) = $html =~ m!(http://c\d+\.mrcdn\.net/[\w\.\/]+\.jpg)!si; | ||||
114 | 4 | 3577 | ($data->{isbn13},$data->{isbn10}) = $html =~ m! ]+>\s*(\d+)\s*/\s*(\d+)\s* |
||||
115 | 4 | 3454 | ($data->{author}) = $html =~ m! |
||||
116 | 4 | 50 | 23 | ($data->{author}) = $html =~ m! |
|||
117 | 4 | 1444 | ($data->{title}) = $html =~ m! |
||||
118 | 4 | 4061 | ($data->{publisher}) = $html =~ m! ]+>\s* ]+>\s*([^<]+) |
||||
119 | 4 | 50 | 26 | ($data->{publisher}) = $html =~ m! |
|||
120 | 4 | 1639 | ($data->{pubdate}) = $html =~ m! |
||||
121 | 4 | 3636 | ($data->{binding}) = $html =~ m! ]+>\s*([^<]+) |
||||
122 | 4 | 1661 | ($data->{pages}) = $html =~ m! |
||||
123 | 4 | 4644 | ($data->{weight}) = $html =~ m! |
||||
124 | 4 | 1652 | ($data->{width},$data->{height}) = $html =~ m! |
||||
125 | 4 | 1919 | ($data->{description}) = $html =~ m!Annotation\s*\s*(.*?) !si; |
||||
126 | 4 | 50 | 23 | ($data->{description}) = $html =~ m!Publisher Description\s*\s*(.*?) !si unless($data->{description}); |
|||
127 | |||||||
128 | 4 | 18 | for(qw(author publisher description)) { | ||||
129 | 12 | 50 | 346 | $data->{$_} =~ s!<[^>]+>!!g if($data->{$_}); | |||
130 | } | ||||||
131 | |||||||
132 | #use Data::Dumper; | ||||||
133 | #print STDERR "\n# " . Dumper($data); | ||||||
134 | |||||||
135 | 4 | 50 | 17 | return $self->handler("Could not extract data from TheNile result page.") | |||
136 | unless(defined $data); | ||||||
137 | |||||||
138 | # trim top and tail | ||||||
139 | 4 | 100 | 39 | foreach (keys %$data) { next unless(defined $data->{$_});$data->{$_} =~ s/^\s+//;$data->{$_} =~ s/\s+$//; } | |||
56 | 124 | ||||||
52 | 118 | ||||||
52 | 164 | ||||||
140 | |||||||
141 | 4 | 54 | my $url = $mech->uri(); | ||||
142 | |||||||
143 | 4 | 212 | my $bk = { | ||||
144 | 'ean13' => $data->{isbn13}, | ||||||
145 | 'isbn13' => $data->{isbn13}, | ||||||
146 | 'isbn10' => $data->{isbn10}, | ||||||
147 | 'isbn' => $data->{isbn13}, | ||||||
148 | 'author' => $data->{author}, | ||||||
149 | 'title' => $data->{title}, | ||||||
150 | 'book_link' => "$url", | ||||||
151 | 'image_link' => $data->{image}, | ||||||
152 | 'thumb_link' => $data->{thumb}, | ||||||
153 | 'description' => $data->{description}, | ||||||
154 | 'pubdate' => $data->{pubdate}, | ||||||
155 | 'publisher' => $data->{publisher}, | ||||||
156 | 'binding' => $data->{binding}, | ||||||
157 | 'pages' => $data->{pages}, | ||||||
158 | 'weight' => $data->{weight}, | ||||||
159 | 'width' => $data->{width}, | ||||||
160 | 'height' => $data->{height}, | ||||||
161 | 'depth' => $data->{depth}, | ||||||
162 | 'html' => $html | ||||||
163 | }; | ||||||
164 | |||||||
165 | #use Data::Dumper; | ||||||
166 | #print STDERR "\n# book=".Dumper($bk); | ||||||
167 | |||||||
168 | 4 | 278 | $self->book($bk); | ||||
169 | 4 | 66 | $self->found(1); | ||||
170 | 4 | 45 | return $self->book; | ||||
171 | } | ||||||
172 | |||||||
173 | 1; | ||||||
174 | |||||||
175 | __END__ |