File Coverage
| blib/lib/HTML/TableParser.pm |
|
| Criterion |
Covered |
Total |
% |
| statement |
132 |
141 |
93.6
|
| branch |
52 |
74 |
70.2
|
| condition |
36 |
52 |
69.2
|
| subroutine |
13 |
13 |
100.0
|
| pod |
1 |
1 |
100.0
|
| total |
234 |
281 |
83.2
|
| line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
|
1
|
|
|
|
|
|
|
package HTML::TableParser; |
|
2
|
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
# ABSTRACT: HTML::TableParser - Extract data from an HTML table |
|
4
|
|
|
|
|
|
|
|
|
5
|
|
|
|
|
|
|
require 5.8.1; |
|
6
|
7
|
|
|
7
|
|
374325
|
use strict; |
|
|
7
|
|
|
|
|
49
|
|
|
|
7
|
|
|
|
|
158
|
|
|
7
|
7
|
|
|
7
|
|
28
|
use warnings; |
|
|
7
|
|
|
|
|
7
|
|
|
|
7
|
|
|
|
|
202
|
|
|
8
|
|
|
|
|
|
|
|
|
9
|
|
|
|
|
|
|
our $VERSION = '0.43'; |
|
10
|
|
|
|
|
|
|
|
|
11
|
7
|
|
|
7
|
|
39
|
use Carp (); |
|
|
7
|
|
|
|
|
16
|
|
|
|
7
|
|
|
|
|
153
|
|
|
12
|
7
|
|
|
7
|
|
3012
|
use HTML::Parser; |
|
|
7
|
|
|
|
|
33501
|
|
|
|
7
|
|
|
|
|
197
|
|
|
13
|
|
|
|
|
|
|
|
|
14
|
7
|
|
|
7
|
|
2757
|
use HTML::TableParser::Table; |
|
|
7
|
|
|
|
|
17
|
|
|
|
7
|
|
|
|
|
11557
|
|
|
15
|
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
## no critic ( ProhibitAccessOfPrivateData ) |
|
17
|
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
our @ISA = qw(HTML::Parser); |
|
20
|
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
# Preloaded methods go here. |
|
22
|
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
our %Attr = ( Trim => 0, |
|
24
|
|
|
|
|
|
|
Decode => 1, |
|
25
|
|
|
|
|
|
|
Chomp => 0, |
|
26
|
|
|
|
|
|
|
MultiMatch => 0, |
|
27
|
|
|
|
|
|
|
DecodeNBSP => 0, |
|
28
|
|
|
|
|
|
|
); |
|
29
|
|
|
|
|
|
|
our @Attr = keys %Attr; |
|
30
|
|
|
|
|
|
|
|
|
31
|
|
|
|
|
|
|
our $Verbose = 0; |
|
32
|
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
sub new |
|
34
|
|
|
|
|
|
|
{ |
|
35
|
93
|
|
|
93
|
1
|
5720
|
my $class = shift; |
|
36
|
|
|
|
|
|
|
|
|
37
|
93
|
|
|
|
|
151
|
my $reqs = shift; |
|
38
|
|
|
|
|
|
|
|
|
39
|
93
|
|
|
|
|
374
|
my $self = $class->SUPER::new |
|
40
|
|
|
|
|
|
|
( |
|
41
|
|
|
|
|
|
|
api_version => 3, |
|
42
|
|
|
|
|
|
|
unbroken_text => 1, |
|
43
|
|
|
|
|
|
|
start_h => [ '_start', 'self, tagname, attr, line' ], |
|
44
|
|
|
|
|
|
|
end_h => [ '_end', 'self, tagname, attr, line' ], |
|
45
|
|
|
|
|
|
|
); |
|
46
|
|
|
|
|
|
|
|
|
47
|
93
|
50
|
33
|
|
|
5012
|
Carp::croak( __PACKAGE__, ": must specify a table request" ) |
|
48
|
|
|
|
|
|
|
unless defined $reqs and 'ARRAY' eq ref $reqs; |
|
49
|
|
|
|
|
|
|
|
|
50
|
93
|
|
100
|
|
|
238
|
my $attr = shift || {}; |
|
51
|
|
|
|
|
|
|
|
|
52
|
93
|
|
|
|
|
217
|
my @notvalid = grep { ! exists $Attr{$_} } keys %$attr; |
|
|
51
|
|
|
|
|
153
|
|
|
53
|
93
|
50
|
|
|
|
182
|
Carp::croak ( __PACKAGE__, ": Invalid attribute(s): '", |
|
54
|
|
|
|
|
|
|
join(" ,'", @notvalid ), "'" ) |
|
55
|
|
|
|
|
|
|
if @notvalid; |
|
56
|
|
|
|
|
|
|
|
|
57
|
93
|
|
|
|
|
340
|
my %attr = ( %Attr, %$attr ); |
|
58
|
|
|
|
|
|
|
|
|
59
|
93
|
|
|
|
|
245
|
$self->{reqs} = _tidy_reqs( $reqs, \%attr ); |
|
60
|
|
|
|
|
|
|
|
|
61
|
93
|
|
|
|
|
352
|
$self->{Tables} = [ HTML::TableParser::Table->new() ]; |
|
62
|
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
# by default we're not processing anything |
|
64
|
93
|
|
|
|
|
213
|
$self->_process(0); |
|
65
|
|
|
|
|
|
|
|
|
66
|
93
|
|
|
|
|
231
|
$self; |
|
67
|
|
|
|
|
|
|
} |
|
68
|
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
|
|
70
|
|
|
|
|
|
|
our @ReqAttr = ( qw( cols colre id idre class obj start end |
|
71
|
|
|
|
|
|
|
hdr row warn udata ), |
|
72
|
|
|
|
|
|
|
keys %Attr ); |
|
73
|
|
|
|
|
|
|
our %ReqAttr = map { $_ => 1 } @ReqAttr; |
|
74
|
|
|
|
|
|
|
|
|
75
|
|
|
|
|
|
|
# convert table requests into something that HTML::TableParser::Table can |
|
76
|
|
|
|
|
|
|
# handle |
|
77
|
|
|
|
|
|
|
sub _tidy_reqs |
|
78
|
|
|
|
|
|
|
{ |
|
79
|
93
|
|
|
93
|
|
166
|
my ( $reqs, $attr ) = @_; |
|
80
|
|
|
|
|
|
|
|
|
81
|
93
|
|
|
|
|
108
|
my @reqs; |
|
82
|
|
|
|
|
|
|
|
|
83
|
93
|
|
|
|
|
110
|
my $nreq = 0; |
|
84
|
93
|
|
|
|
|
170
|
for my $req ( @$reqs ) |
|
85
|
|
|
|
|
|
|
{ |
|
86
|
95
|
|
|
|
|
127
|
my %req; |
|
87
|
|
|
|
|
|
|
|
|
88
|
95
|
|
|
|
|
129
|
$nreq++; |
|
89
|
|
|
|
|
|
|
|
|
90
|
95
|
|
|
|
|
189
|
my @notvalid = grep { ! exists $ReqAttr{$_} } keys %$req; |
|
|
493
|
|
|
|
|
757
|
|
|
91
|
95
|
50
|
|
|
|
186
|
Carp::croak (__PACKAGE__, ": table request $nreq: invalid attribute(s): '", |
|
92
|
|
|
|
|
|
|
join(" ,'", @notvalid ), "'" ) |
|
93
|
|
|
|
|
|
|
if @notvalid; |
|
94
|
|
|
|
|
|
|
|
|
95
|
95
|
|
|
|
|
120
|
my $req_id = 0; |
|
96
|
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
|
|
98
|
|
|
|
|
|
|
# parse cols and id the same way |
|
99
|
95
|
|
|
|
|
158
|
for my $what ( qw( cols id ) ) |
|
100
|
|
|
|
|
|
|
{ |
|
101
|
190
|
|
|
|
|
341
|
$req{$what} = []; |
|
102
|
|
|
|
|
|
|
|
|
103
|
190
|
100
|
66
|
|
|
514
|
if ( exists $req->{$what} && defined $req->{$what} ) |
|
104
|
|
|
|
|
|
|
{ |
|
105
|
78
|
|
|
|
|
99
|
my @reqs; |
|
106
|
|
|
|
|
|
|
|
|
107
|
78
|
|
|
|
|
131
|
my $ref = ref $req->{$what}; |
|
108
|
|
|
|
|
|
|
|
|
109
|
78
|
100
|
100
|
|
|
295
|
if ( 'ARRAY' eq $ref ) |
|
|
|
50
|
66
|
|
|
|
|
|
110
|
|
|
|
|
|
|
{ |
|
111
|
34
|
|
|
|
|
41
|
@reqs = @{$req->{$what}}; |
|
|
34
|
|
|
|
|
68
|
|
|
112
|
|
|
|
|
|
|
} |
|
113
|
|
|
|
|
|
|
elsif ( 'Regexp' eq $ref || |
|
114
|
|
|
|
|
|
|
'CODE' eq $ref || |
|
115
|
|
|
|
|
|
|
! $ref ) |
|
116
|
|
|
|
|
|
|
{ |
|
117
|
44
|
|
|
|
|
88
|
@reqs = ( $req->{$what} ); |
|
118
|
|
|
|
|
|
|
} |
|
119
|
|
|
|
|
|
|
else |
|
120
|
|
|
|
|
|
|
{ |
|
121
|
0
|
|
|
|
|
0
|
Carp::croak( __PACKAGE__, |
|
122
|
|
|
|
|
|
|
": table request $nreq: $what must be a scalar, arrayref, or coderef" ); |
|
123
|
|
|
|
|
|
|
} |
|
124
|
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
# now, check that we have legal things in there |
|
126
|
78
|
|
|
|
|
120
|
my %attr = (); |
|
127
|
|
|
|
|
|
|
|
|
128
|
78
|
|
|
|
|
122
|
for my $match ( @reqs ) |
|
129
|
|
|
|
|
|
|
{ |
|
130
|
81
|
|
|
|
|
110
|
my $ref = ref $match; |
|
131
|
81
|
50
|
66
|
|
|
318
|
Carp::croak( __PACKAGE__, |
|
|
|
|
100
|
|
|
|
|
|
|
|
|
66
|
|
|
|
|
|
132
|
|
|
|
|
|
|
": table request $nreq: illegal $what `$match': must be a scalar, regexp, or coderef" ) |
|
133
|
|
|
|
|
|
|
unless defined $match && ! $ref || 'Regexp' eq $ref |
|
134
|
|
|
|
|
|
|
|| 'CODE' eq $ref ; |
|
135
|
|
|
|
|
|
|
|
|
136
|
81
|
100
|
100
|
|
|
283
|
if ( ! $ref && $match eq '-' ) |
|
137
|
|
|
|
|
|
|
{ |
|
138
|
1
|
|
|
|
|
2
|
%attr = ( exclude => 1 ); |
|
139
|
1
|
|
|
|
|
2
|
next; |
|
140
|
|
|
|
|
|
|
} |
|
141
|
|
|
|
|
|
|
|
|
142
|
80
|
100
|
100
|
|
|
225
|
if ( ! $ref && $match eq '--' ) |
|
143
|
|
|
|
|
|
|
{ |
|
144
|
1
|
|
|
|
|
3
|
%attr = ( skip => 1 ); |
|
145
|
1
|
|
|
|
|
2
|
next; |
|
146
|
|
|
|
|
|
|
} |
|
147
|
|
|
|
|
|
|
|
|
148
|
79
|
50
|
66
|
|
|
208
|
if ( ! $ref && $match eq '+' ) |
|
149
|
|
|
|
|
|
|
{ |
|
150
|
0
|
|
|
|
|
0
|
%attr = (); |
|
151
|
0
|
|
|
|
|
0
|
next; |
|
152
|
|
|
|
|
|
|
} |
|
153
|
|
|
|
|
|
|
|
|
154
|
79
|
|
|
|
|
101
|
push @{$req{$what}}, { %attr, match => $match }; |
|
|
79
|
|
|
|
|
230
|
|
|
155
|
79
|
|
|
|
|
121
|
%attr = (); |
|
156
|
79
|
|
|
|
|
181
|
$req_id++; |
|
157
|
|
|
|
|
|
|
} |
|
158
|
|
|
|
|
|
|
} |
|
159
|
|
|
|
|
|
|
} |
|
160
|
|
|
|
|
|
|
|
|
161
|
|
|
|
|
|
|
# colre is now obsolete, but keep backwards compatibility |
|
162
|
|
|
|
|
|
|
# column regular expression match? |
|
163
|
95
|
100
|
|
|
|
195
|
if ( defined $req->{colre} ) |
|
164
|
|
|
|
|
|
|
{ |
|
165
|
17
|
|
|
|
|
25
|
my $colre; |
|
166
|
|
|
|
|
|
|
|
|
167
|
17
|
50
|
|
|
|
44
|
if ( 'ARRAY' eq ref $req->{colre} ) |
|
|
|
0
|
|
|
|
|
|
|
168
|
|
|
|
|
|
|
{ |
|
169
|
17
|
|
|
|
|
29
|
$colre = $req->{colre}; |
|
170
|
|
|
|
|
|
|
} |
|
171
|
|
|
|
|
|
|
elsif ( ! ref $req->{colre} ) |
|
172
|
|
|
|
|
|
|
{ |
|
173
|
0
|
|
|
|
|
0
|
$colre = [ $req->{colre} ]; |
|
174
|
|
|
|
|
|
|
} |
|
175
|
|
|
|
|
|
|
else |
|
176
|
|
|
|
|
|
|
{ |
|
177
|
0
|
|
|
|
|
0
|
Carp::croak( __PACKAGE__, |
|
178
|
|
|
|
|
|
|
": table request $nreq: colre must be a scalar or arrayref" ); |
|
179
|
|
|
|
|
|
|
} |
|
180
|
|
|
|
|
|
|
|
|
181
|
17
|
|
|
|
|
36
|
for my $re ( @$colre ) |
|
182
|
|
|
|
|
|
|
{ |
|
183
|
17
|
|
|
|
|
27
|
my $ref = ref $re; |
|
184
|
|
|
|
|
|
|
|
|
185
|
17
|
50
|
66
|
|
|
41
|
Carp::croak( __PACKAGE__, ": table request $nreq: colre must be a scalar" ) |
|
186
|
|
|
|
|
|
|
unless ! $ref or 'Regexp' eq $ref; |
|
187
|
17
|
100
|
|
|
|
20
|
push @{$req{cols}}, { include => 1, |
|
|
17
|
|
|
|
|
151
|
|
|
188
|
|
|
|
|
|
|
match => 'Regexp' eq $ref ? $re : qr/$re/ }; |
|
189
|
17
|
|
|
|
|
39
|
$req_id++; |
|
190
|
|
|
|
|
|
|
} |
|
191
|
|
|
|
|
|
|
} |
|
192
|
|
|
|
|
|
|
|
|
193
|
|
|
|
|
|
|
|
|
194
|
95
|
50
|
|
|
|
177
|
Carp::croak( __PACKAGE__, |
|
195
|
|
|
|
|
|
|
": table request $nreq: must specify at least one id method" ) |
|
196
|
|
|
|
|
|
|
unless $req_id; |
|
197
|
|
|
|
|
|
|
|
|
198
|
|
|
|
|
|
|
$req{obj} = $req->{obj} |
|
199
|
95
|
100
|
|
|
|
185
|
if exists $req->{obj}; |
|
200
|
|
|
|
|
|
|
|
|
201
|
|
|
|
|
|
|
$req{class} = $req->{class} |
|
202
|
95
|
100
|
|
|
|
170
|
if exists $req->{class}; |
|
203
|
|
|
|
|
|
|
|
|
204
|
95
|
|
|
|
|
148
|
for my $method ( qw( start end hdr row warn new ) ) |
|
205
|
|
|
|
|
|
|
{ |
|
206
|
570
|
100
|
100
|
|
|
1696
|
if ( exists $req->{$method} && 'CODE' eq ref $req->{$method} ) |
|
|
|
100
|
66
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
207
|
|
|
|
|
|
|
{ |
|
208
|
310
|
|
|
|
|
480
|
$req{$method} = $req->{$method}; |
|
209
|
|
|
|
|
|
|
} |
|
210
|
|
|
|
|
|
|
|
|
211
|
|
|
|
|
|
|
elsif ( exists $req{obj} || exists $req{class}) |
|
212
|
|
|
|
|
|
|
{ |
|
213
|
78
|
100
|
|
|
|
113
|
my $thing = exists $req{obj} ? $req{obj} : $req{class}; |
|
214
|
|
|
|
|
|
|
|
|
215
|
78
|
100
|
|
|
|
100
|
if ( exists $req->{$method} ) |
|
216
|
|
|
|
|
|
|
{ |
|
217
|
2
|
50
|
|
|
|
5
|
if ( defined $req->{$method} ) |
|
218
|
|
|
|
|
|
|
{ |
|
219
|
|
|
|
|
|
|
Carp::croak( __PACKAGE__, |
|
220
|
|
|
|
|
|
|
": table request $nreq: can't have object & non-scalar $method" ) |
|
221
|
0
|
0
|
|
|
|
0
|
if ref $req->{$method}; |
|
222
|
|
|
|
|
|
|
|
|
223
|
0
|
|
|
|
|
0
|
my $call = $req->{$method}; |
|
224
|
|
|
|
|
|
|
|
|
225
|
|
|
|
|
|
|
Carp::croak( __PACKAGE__, |
|
226
|
|
|
|
|
|
|
": table request $nreq: class doesn't have method $call" ) |
|
227
|
0
|
0
|
0
|
|
|
0
|
if ( exists $req->{obj} && ! $req->{obj}->can( $call ) ) |
|
|
|
|
0
|
|
|
|
|
|
228
|
|
|
|
|
|
|
|| !UNIVERSAL::can( $thing, $call ); |
|
229
|
|
|
|
|
|
|
} |
|
230
|
|
|
|
|
|
|
|
|
231
|
|
|
|
|
|
|
# if $req->{$method} is undef, user must have explicitly |
|
232
|
|
|
|
|
|
|
# set it so, which is a signal to NOT call that method. |
|
233
|
|
|
|
|
|
|
} |
|
234
|
|
|
|
|
|
|
else |
|
235
|
|
|
|
|
|
|
{ |
|
236
|
76
|
100
|
|
|
|
221
|
$req{$method} = $method |
|
237
|
|
|
|
|
|
|
if UNIVERSAL::can( $thing, $method ); |
|
238
|
|
|
|
|
|
|
} |
|
239
|
|
|
|
|
|
|
} |
|
240
|
|
|
|
|
|
|
elsif( exists $req->{$method} ) |
|
241
|
|
|
|
|
|
|
{ |
|
242
|
0
|
|
|
|
|
0
|
Carp::croak( __PACKAGE__, ": invalid callback for $method" ); |
|
243
|
|
|
|
|
|
|
} |
|
244
|
|
|
|
|
|
|
} |
|
245
|
|
|
|
|
|
|
|
|
246
|
|
|
|
|
|
|
# last minute cleanups for things that don't fit in the above loop |
|
247
|
|
|
|
|
|
|
Carp::croak( __PACKAGE__, ": must specify valid constructor for class $req->{class}" ) |
|
248
|
95
|
50
|
66
|
|
|
200
|
if exists $req{class} && ! exists $req{new}; |
|
249
|
|
|
|
|
|
|
|
|
250
|
|
|
|
|
|
|
|
|
251
|
95
|
|
|
|
|
143
|
$req{udata} = undef; |
|
252
|
95
|
100
|
|
|
|
180
|
$req{udata} = exists $req->{udata} ? $req->{udata} : undef; |
|
253
|
|
|
|
|
|
|
|
|
254
|
95
|
|
|
|
|
192
|
$req{match} = 0; |
|
255
|
|
|
|
|
|
|
|
|
256
|
95
|
|
|
|
|
280
|
@req{@Attr} = @Attr{@Attr}; |
|
257
|
|
|
|
|
|
|
|
|
258
|
|
|
|
|
|
|
$req{$_} = $attr->{$_} |
|
259
|
95
|
|
|
|
|
169
|
foreach grep { defined $attr->{$_} } @Attr; |
|
|
475
|
|
|
|
|
859
|
|
|
260
|
|
|
|
|
|
|
|
|
261
|
|
|
|
|
|
|
$req{$_} = $req->{$_} |
|
262
|
95
|
|
|
|
|
133
|
foreach grep { defined $req->{$_} } @Attr; |
|
|
475
|
|
|
|
|
691
|
|
|
263
|
|
|
|
|
|
|
|
|
264
|
95
|
|
|
|
|
202
|
push @reqs, \%req; |
|
265
|
|
|
|
|
|
|
} |
|
266
|
|
|
|
|
|
|
|
|
267
|
93
|
|
|
|
|
191
|
\@reqs; |
|
268
|
|
|
|
|
|
|
} |
|
269
|
|
|
|
|
|
|
|
|
270
|
|
|
|
|
|
|
|
|
271
|
|
|
|
|
|
|
sub _process |
|
272
|
|
|
|
|
|
|
{ |
|
273
|
337
|
|
|
337
|
|
505
|
my ($self, $state) = @_; |
|
274
|
|
|
|
|
|
|
|
|
275
|
337
|
|
100
|
|
|
812
|
my $ostate = $self->{process} || 0; |
|
276
|
|
|
|
|
|
|
|
|
277
|
337
|
100
|
|
|
|
542
|
if ( $state ) |
|
278
|
|
|
|
|
|
|
{ |
|
279
|
123
|
|
|
|
|
569
|
$self->report_tags( qw( table th td tr ) ); |
|
280
|
123
|
|
|
|
|
495
|
$self->handler( 'text' => '_text', 'self, text, line' ); |
|
281
|
|
|
|
|
|
|
} |
|
282
|
|
|
|
|
|
|
|
|
283
|
|
|
|
|
|
|
else |
|
284
|
|
|
|
|
|
|
{ |
|
285
|
214
|
|
|
|
|
733
|
$self->report_tags( qw( table ) ); |
|
286
|
214
|
|
|
|
|
580
|
$self->handler( 'text' => '' ); |
|
287
|
|
|
|
|
|
|
} |
|
288
|
|
|
|
|
|
|
|
|
289
|
337
|
|
|
|
|
504
|
$self->{process} = $state; |
|
290
|
337
|
|
|
|
|
844
|
$ostate; |
|
291
|
|
|
|
|
|
|
} |
|
292
|
|
|
|
|
|
|
|
|
293
|
|
|
|
|
|
|
|
|
294
|
|
|
|
|
|
|
our %trans = ( tr => 'row', |
|
295
|
|
|
|
|
|
|
th => 'header', |
|
296
|
|
|
|
|
|
|
td => 'column' ); |
|
297
|
|
|
|
|
|
|
|
|
298
|
|
|
|
|
|
|
sub _start |
|
299
|
|
|
|
|
|
|
{ |
|
300
|
53261
|
|
|
53261
|
|
79654
|
my $self = shift; |
|
301
|
53261
|
|
|
|
|
58899
|
my $tagname = shift; |
|
302
|
|
|
|
|
|
|
|
|
303
|
53261
|
50
|
|
|
|
77121
|
print STDERR __PACKAGE__, "::start : $_[1] : $tagname \n" |
|
304
|
|
|
|
|
|
|
if $HTML::TableParser::Verbose; |
|
305
|
|
|
|
|
|
|
|
|
306
|
53261
|
100
|
|
|
|
75562
|
if ( 'table' eq $tagname ) |
|
307
|
|
|
|
|
|
|
{ |
|
308
|
122
|
|
|
|
|
232
|
$self->_start_table( @_ ); |
|
309
|
|
|
|
|
|
|
} |
|
310
|
|
|
|
|
|
|
|
|
311
|
|
|
|
|
|
|
else |
|
312
|
|
|
|
|
|
|
{ |
|
313
|
53139
|
|
|
|
|
74748
|
my $method = 'start_' . $trans{$tagname}; |
|
314
|
|
|
|
|
|
|
|
|
315
|
53139
|
|
|
|
|
101068
|
$self->{Tables}[-1]->$method(@_); |
|
316
|
|
|
|
|
|
|
} |
|
317
|
|
|
|
|
|
|
} |
|
318
|
|
|
|
|
|
|
|
|
319
|
|
|
|
|
|
|
|
|
320
|
|
|
|
|
|
|
sub _end |
|
321
|
|
|
|
|
|
|
{ |
|
322
|
9954
|
|
|
9954
|
|
14610
|
my $self = shift; |
|
323
|
9954
|
|
|
|
|
10744
|
my $tagname = shift; |
|
324
|
|
|
|
|
|
|
|
|
325
|
9954
|
50
|
|
|
|
14047
|
print STDERR __PACKAGE__, "::_end : $_[1]: $tagname \n" |
|
326
|
|
|
|
|
|
|
if $HTML::TableParser::Verbose; |
|
327
|
|
|
|
|
|
|
|
|
328
|
9954
|
100
|
|
|
|
14103
|
if ( 'table' eq $tagname ) |
|
329
|
|
|
|
|
|
|
{ |
|
330
|
123
|
|
|
|
|
250
|
$self->_end_table( @_ ); |
|
331
|
|
|
|
|
|
|
} |
|
332
|
|
|
|
|
|
|
|
|
333
|
|
|
|
|
|
|
else |
|
334
|
|
|
|
|
|
|
{ |
|
335
|
9831
|
|
|
|
|
13730
|
my $method = 'end_' . $trans{$tagname}; |
|
336
|
|
|
|
|
|
|
|
|
337
|
9831
|
|
|
|
|
18719
|
$self->{Tables}[-1]->$method(@_); |
|
338
|
|
|
|
|
|
|
} |
|
339
|
|
|
|
|
|
|
} |
|
340
|
|
|
|
|
|
|
|
|
341
|
|
|
|
|
|
|
|
|
342
|
|
|
|
|
|
|
sub _start_table |
|
343
|
|
|
|
|
|
|
{ |
|
344
|
122
|
|
|
122
|
|
209
|
my ( $self, $attr, $line ) = @_; |
|
345
|
|
|
|
|
|
|
|
|
346
|
|
|
|
|
|
|
my $tbl = HTML::TableParser::Table->new( $self, |
|
347
|
|
|
|
|
|
|
$self->{Tables}[-1]->ids, |
|
348
|
122
|
|
|
|
|
351
|
$self->{reqs}, $line ); |
|
349
|
|
|
|
|
|
|
|
|
350
|
122
|
50
|
|
|
|
272
|
print STDERR __PACKAGE__, "::_start_table : $tbl->{id}\n" |
|
351
|
|
|
|
|
|
|
if $HTML::TableParser::Verbose; |
|
352
|
|
|
|
|
|
|
|
|
353
|
122
|
|
|
|
|
253
|
$self->_process( $tbl->process ); |
|
354
|
|
|
|
|
|
|
|
|
355
|
122
|
|
|
|
|
144
|
push @{$self->{Tables}}, $tbl; |
|
|
122
|
|
|
|
|
937
|
|
|
356
|
|
|
|
|
|
|
} |
|
357
|
|
|
|
|
|
|
|
|
358
|
|
|
|
|
|
|
|
|
359
|
|
|
|
|
|
|
sub _end_table |
|
360
|
|
|
|
|
|
|
{ |
|
361
|
123
|
|
|
123
|
|
192
|
my ( $self, $attr, $line ) = @_; |
|
362
|
|
|
|
|
|
|
|
|
363
|
|
|
|
|
|
|
|
|
364
|
123
|
|
|
|
|
136
|
my $tbl = pop @{$self->{Tables}}; |
|
|
123
|
|
|
|
|
202
|
|
|
365
|
|
|
|
|
|
|
|
|
366
|
123
|
50
|
|
|
|
215
|
print STDERR __PACKAGE__, "::_end_table : $tbl->{id}\n" |
|
367
|
|
|
|
|
|
|
if $HTML::TableParser::Verbose; |
|
368
|
|
|
|
|
|
|
|
|
369
|
|
|
|
|
|
|
# the first table in the list is our sentinel table. if we're about |
|
370
|
|
|
|
|
|
|
# to delete it, it means that we've hit one too many |
tags
|
371
|
|
|
|
|
|
|
# we delay the croak until after the pop so that the verbose error |
|
372
|
|
|
|
|
|
|
# message prints something nice. no harm anyway as we're about to |
|
373
|
|
|
|
|
|
|
# keel over and croak. |
|
374
|
|
|
|
|
|
|
|
|
375
|
|
|
|
|
|
|
Carp::croak( __PACKAGE__, |
|
376
|
|
|
|
|
|
|
": $line: unbalanced tags; too many tags" ) |
|
377
|
123
|
100
|
|
|
|
142
|
if 0 == @{$self->{Tables}}; |
|
|
123
|
|
|
|
|
380
|
|
|
378
|
|
|
|
|
|
|
|
|
379
|
122
|
|
|
|
|
323
|
undef $tbl; |
|
380
|
|
|
|
|
|
|
|
|
381
|
122
|
|
|
|
|
4793
|
$self->_process( $self->{Tables}[-1]->process ); |
|
382
|
|
|
|
|
|
|
} |
|
383
|
|
|
|
|
|
|
|
|
384
|
|
|
|
|
|
|
|
|
385
|
|
|
|
|
|
|
sub _text |
|
386
|
|
|
|
|
|
|
{ |
|
387
|
62098
|
|
|
62098
|
|
145638
|
my ( $self, $text, $line ) = @_; |
|
388
|
|
|
|
|
|
|
|
|
389
|
62098
|
|
|
|
|
111474
|
$self->{Tables}[-1]->text( $text ); |
|
390
|
|
|
|
|
|
|
} |
|
391
|
|
|
|
|
|
|
|
|
392
|
|
|
|
|
|
|
1; |
|
393
|
|
|
|
|
|
|
|
|
394
|
|
|
|
|
|
|
# |
|
395
|
|
|
|
|
|
|
# This file is part of HTML-TableParser |
|
396
|
|
|
|
|
|
|
# |
|
397
|
|
|
|
|
|
|
# This software is Copyright (c) 2018 by Smithsonian Astrophysical Observatory. |
|
398
|
|
|
|
|
|
|
# |
|
399
|
|
|
|
|
|
|
# This is free software, licensed under: |
|
400
|
|
|
|
|
|
|
# |
|
401
|
|
|
|
|
|
|
# The GNU General Public License, Version 3, June 2007 |
|
402
|
|
|
|
|
|
|
# |
|
403
|
|
|
|
|
|
|
|
|
404
|
|
|
|
|
|
|
=pod |
|
405
|
|
|
|
|
|
|
|
|
406
|
|
|
|
|
|
|
=head1 NAME |
|
407
|
|
|
|
|
|
|
|
|
408
|
|
|
|
|
|
|
HTML::TableParser - HTML::TableParser - Extract data from an HTML table |
|
409
|
|
|
|
|
|
|
|
|
410
|
|
|
|
|
|
|
=head1 VERSION |
|
411
|
|
|
|
|
|
|
|
|
412
|
|
|
|
|
|
|
version 0.43 |
|
413
|
|
|
|
|
|
|
|
|
414
|
|
|
|
|
|
|
=head1 SYNOPSIS |
|
415
|
|
|
|
|
|
|
|
|
416
|
|
|
|
|
|
|
use HTML::TableParser; |
|
417
|
|
|
|
|
|
|
|
|
418
|
|
|
|
|
|
|
@reqs = ( |
|
419
|
|
|
|
|
|
|
{ |
|
420
|
|
|
|
|
|
|
id => 1.1, # id for embedded table |
|
421
|
|
|
|
|
|
|
hdr => \&header, # function callback |
|
422
|
|
|
|
|
|
|
row => \&row, # function callback |
|
423
|
|
|
|
|
|
|
start => \&start, # function callback |
|
424
|
|
|
|
|
|
|
end => \&end, # function callback |
|
425
|
|
|
|
|
|
|
udata => { Snack => 'Food' }, # arbitrary user data |
|
426
|
|
|
|
|
|
|
}, |
|
427
|
|
|
|
|
|
|
{ |
|
428
|
|
|
|
|
|
|
id => 1, # table id |
|
429
|
|
|
|
|
|
|
cols => [ 'Object Type', |
|
430
|
|
|
|
|
|
|
qr/object/ ], # column name matches |
|
431
|
|
|
|
|
|
|
obj => $obj, # method callbacks |
|
432
|
|
|
|
|
|
|
}, |
|
433
|
|
|
|
|
|
|
); |
|
434
|
|
|
|
|
|
|
|
|
435
|
|
|
|
|
|
|
# create parser object |
|
436
|
|
|
|
|
|
|
$p = HTML::TableParser->new( \@reqs, |
|
437
|
|
|
|
|
|
|
{ Decode => 1, Trim => 1, Chomp => 1 } ); |
|
438
|
|
|
|
|
|
|
$p->parse_file( 'foo.html' ); |
|
439
|
|
|
|
|
|
|
|
|
440
|
|
|
|
|
|
|
|
|
441
|
|
|
|
|
|
|
# function callbacks |
|
442
|
|
|
|
|
|
|
sub start { |
|
443
|
|
|
|
|
|
|
my ( $id, $line, $udata ) = @_; |
|
444
|
|
|
|
|
|
|
#... |
|
445
|
|
|
|
|
|
|
} |
|
446
|
|
|
|
|
|
|
|
|
447
|
|
|
|
|
|
|
sub end { |
|
448
|
|
|
|
|
|
|
my ( $id, $line, $udata ) = @_; |
|
449
|
|
|
|
|
|
|
#... |
|
450
|
|
|
|
|
|
|
} |
|
451
|
|
|
|
|
|
|
|
|
452
|
|
|
|
|
|
|
sub header { |
|
453
|
|
|
|
|
|
|
my ( $id, $line, $cols, $udata ) = @_; |
|
454
|
|
|
|
|
|
|
#... |
|
455
|
|
|
|
|
|
|
} |
|
456
|
|
|
|
|
|
|
|
|
457
|
|
|
|
|
|
|
sub row { |
|
458
|
|
|
|
|
|
|
my ( $id, $line, $cols, $udata ) = @_; |
|
459
|
|
|
|
|
|
|
#... |
|
460
|
|
|
|
|
|
|
} |
|
461
|
|
|
|
|
|
|
|
|
462
|
|
|
|
|
|
|
=head1 DESCRIPTION |
|
463
|
|
|
|
|
|
|
|
|
464
|
|
|
|
|
|
|
B uses B to extract data from an HTML |
|
465
|
|
|
|
|
|
|
table. The data is returned via a series of user defined callback |
|
466
|
|
|
|
|
|
|
functions or methods. Specific tables may be selected either by a |
|
467
|
|
|
|
|
|
|
matching a unique table id or by matching against the column names. |
|
468
|
|
|
|
|
|
|
Multiple (even nested) tables may be parsed in a document in one pass. |
|
469
|
|
|
|
|
|
|
|
|
470
|
|
|
|
|
|
|
=head2 Table Identification |
|
471
|
|
|
|
|
|
|
|
|
472
|
|
|
|
|
|
|
Each table is given a unique id, relative to its parent, based upon its |
|
473
|
|
|
|
|
|
|
order and nesting. The first top level table has id C<1>, the second |
|
474
|
|
|
|
|
|
|
C<2>, etc. The first table nested in table C<1> has id C<1.1>, the |
|
475
|
|
|
|
|
|
|
second C<1.2>, etc. The first table nested in table C<1.1> has id |
|
476
|
|
|
|
|
|
|
C<1.1.1>, etc. These, as well as the tables' column names, may |
|
477
|
|
|
|
|
|
|
be used to identify which tables to parse. |
|
478
|
|
|
|
|
|
|
|
|
479
|
|
|
|
|
|
|
=head2 Data Extraction |
|
480
|
|
|
|
|
|
|
|
|
481
|
|
|
|
|
|
|
As the parser traverses a selected table, it will pass data to user |
|
482
|
|
|
|
|
|
|
provided callback functions or methods after it has digested |
|
483
|
|
|
|
|
|
|
particular structures in the table. All functions are passed the |
|
484
|
|
|
|
|
|
|
table id (as described above), the line number in the HTML source |
|
485
|
|
|
|
|
|
|
where the table was found, and a reference to any table specific user |
|
486
|
|
|
|
|
|
|
provided data. |
|
487
|
|
|
|
|
|
|
|
|
488
|
|
|
|
|
|
|
=over 8 |
|
489
|
|
|
|
|
|
|
|
|
490
|
|
|
|
|
|
|
=item Table Start |
|
491
|
|
|
|
|
|
|
|
|
492
|
|
|
|
|
|
|
The B callback is invoked when a matched table has been found. |
|
493
|
|
|
|
|
|
|
|
|
494
|
|
|
|
|
|
|
=item Table End |
|
495
|
|
|
|
|
|
|
|
|
496
|
|
|
|
|
|
|
The B callback is invoked after a matched table has been parsed. |
|
497
|
|
|
|
|
|
|
|
|
498
|
|
|
|
|
|
|
=item Header |
|
499
|
|
|
|
|
|
|
|
|
500
|
|
|
|
|
|
|
The B callback is invoked after the table header has been read in. |
|
501
|
|
|
|
|
|
|
Some tables do not use the BthE> tag to indicate a header, so this |
|
502
|
|
|
|
|
|
|
function may not be called. It is passed the column names. |
|
503
|
|
|
|
|
|
|
|
|
504
|
|
|
|
|
|
|
=item Row |
|
505
|
|
|
|
|
|
|
|
|
506
|
|
|
|
|
|
|
The B callback is invoked after a row in the table has been read. |
|
507
|
|
|
|
|
|
|
It is passed the column data. |
|
508
|
|
|
|
|
|
|
|
|
509
|
|
|
|
|
|
|
=item Warn |
|
510
|
|
|
|
|
|
|
|
|
511
|
|
|
|
|
|
|
The B callback is invoked when a non-fatal error occurs during |
|
512
|
|
|
|
|
|
|
parsing. Fatal errors croak. |
|
513
|
|
|
|
|
|
|
|
|
514
|
|
|
|
|
|
|
=item New |
|
515
|
|
|
|
|
|
|
|
|
516
|
|
|
|
|
|
|
This is the class method to call to create a new object when |
|
517
|
|
|
|
|
|
|
B is supposed to create new objects upon table |
|
518
|
|
|
|
|
|
|
start. |
|
519
|
|
|
|
|
|
|
|
|
520
|
|
|
|
|
|
|
=back |
|
521
|
|
|
|
|
|
|
|
|
522
|
|
|
|
|
|
|
=head2 Callback API |
|
523
|
|
|
|
|
|
|
|
|
524
|
|
|
|
|
|
|
Callbacks may be functions or methods or a mixture of both. |
|
525
|
|
|
|
|
|
|
In the latter case, an object must be passed to the constructor. |
|
526
|
|
|
|
|
|
|
(More on that later.) |
|
527
|
|
|
|
|
|
|
|
|
528
|
|
|
|
|
|
|
The callbacks are invoked as follows: |
|
529
|
|
|
|
|
|
|
|
|
530
|
|
|
|
|
|
|
start( $tbl_id, $line_no, $udata ); |
|
531
|
|
|
|
|
|
|
|
|
532
|
|
|
|
|
|
|
end( $tbl_id, $line_no, $udata ); |
|
533
|
|
|
|
|
|
|
|
|
534
|
|
|
|
|
|
|
hdr( $tbl_id, $line_no, \@col_names, $udata ); |
|
535
|
|
|
|
|
|
|
|
|
536
|
|
|
|
|
|
|
row( $tbl_id, $line_no, \@data, $udata ); |
|
537
|
|
|
|
|
|
|
|
|
538
|
|
|
|
|
|
|
warn( $tbl_id, $line_no, $message, $udata ); |
|
539
|
|
|
|
|
|
|
|
|
540
|
|
|
|
|
|
|
new( $tbl_id, $udata ); |
|
541
|
|
|
|
|
|
|
|
|
542
|
|
|
|
|
|
|
=head2 Data Cleanup |
|
543
|
|
|
|
|
|
|
|
|
544
|
|
|
|
|
|
|
There are several cleanup operations that may be performed automatically: |
|
545
|
|
|
|
|
|
|
|
|
546
|
|
|
|
|
|
|
=over 8 |
|
547
|
|
|
|
|
|
|
|
|
548
|
|
|
|
|
|
|
=item Chomp |
|
549
|
|
|
|
|
|
|
|
|
550
|
|
|
|
|
|
|
B the data |
|
551
|
|
|
|
|
|
|
|
|
552
|
|
|
|
|
|
|
=item Decode |
|
553
|
|
|
|
|
|
|
|
|
554
|
|
|
|
|
|
|
Run the data through B. |
|
555
|
|
|
|
|
|
|
|
|
556
|
|
|
|
|
|
|
=item DecodeNBSP |
|
557
|
|
|
|
|
|
|
|
|
558
|
|
|
|
|
|
|
Normally B changes a non-breaking space into |
|
559
|
|
|
|
|
|
|
a character which doesn't seem to be matched by Perl's whitespace |
|
560
|
|
|
|
|
|
|
regexp. Setting this attribute changes the HTML C character to |
|
561
|
|
|
|
|
|
|
a plain 'ol blank. |
|
562
|
|
|
|
|
|
|
|
|
563
|
|
|
|
|
|
|
=item Trim |
|
564
|
|
|
|
|
|
|
|
|
565
|
|
|
|
|
|
|
remove leading and trailing white space. |
|
566
|
|
|
|
|
|
|
|
|
567
|
|
|
|
|
|
|
=back |
|
568
|
|
|
|
|
|
|
|
|
569
|
|
|
|
|
|
|
=head2 Data Organization |
|
570
|
|
|
|
|
|
|
|
|
571
|
|
|
|
|
|
|
Column names are derived from cells delimited by the BthE> and |
|
572
|
|
|
|
|
|
|
B/thE> tags. Some tables have header cells which span one or |
|
573
|
|
|
|
|
|
|
more columns or rows to make things look nice. B |
|
574
|
|
|
|
|
|
|
determines the actual number of columns used and provides column |
|
575
|
|
|
|
|
|
|
names for each column, repeating names for spanned columns and |
|
576
|
|
|
|
|
|
|
concatenating spanned rows and columns. For example, if the |
|
577
|
|
|
|
|
|
|
table header looks like this: |
|
578
|
|
|
|
|
|
|
|
|
579
|
|
|
|
|
|
|
+----+--------+----------+-------------+-------------------+ |
|
580
|
|
|
|
|
|
|
| | | Eq J2000 | | Velocity/Redshift | |
|
581
|
|
|
|
|
|
|
| No | Object |----------| Object Type |-------------------| |
|
582
|
|
|
|
|
|
|
| | | RA | Dec | | km/s | z | Qual | |
|
583
|
|
|
|
|
|
|
+----+--------+----------+-------------+-------------------+ |
|
584
|
|
|
|
|
|
|
|
|
585
|
|
|
|
|
|
|
The columns will be: |
|
586
|
|
|
|
|
|
|
|
|
587
|
|
|
|
|
|
|
No |
|
588
|
|
|
|
|
|
|
Object |
|
589
|
|
|
|
|
|
|
Eq J2000 RA |
|
590
|
|
|
|
|
|
|
Eq J2000 Dec |
|
591
|
|
|
|
|
|
|
Object Type |
|
592
|
|
|
|
|
|
|
Velocity/Redshift km/s |
|
593
|
|
|
|
|
|
|
Velocity/Redshift z |
|
594
|
|
|
|
|
|
|
Velocity/Redshift Qual |
|
595
|
|
|
|
|
|
|
|
|
596
|
|
|
|
|
|
|
Row data are derived from cells delimited by the BtdE> and |
|
597
|
|
|
|
|
|
|
B/tdE> tags. Cells which span more than one column or row are |
|
598
|
|
|
|
|
|
|
handled correctly, i.e. the values are duplicated in the appropriate |
|
599
|
|
|
|
|
|
|
places. |
|
600
|
|
|
|
|
|
|
|
|
601
|
|
|
|
|
|
|
=head1 METHODS |
|
602
|
|
|
|
|
|
|
|
|
603
|
|
|
|
|
|
|
=over 8 |
|
604
|
|
|
|
|
|
|
|
|
605
|
|
|
|
|
|
|
=item new |
|
606
|
|
|
|
|
|
|
|
|
607
|
|
|
|
|
|
|
$p = HTML::TableParser->new( \@reqs, \%attr ); |
|
608
|
|
|
|
|
|
|
|
|
609
|
|
|
|
|
|
|
This is the class constructor. It is passed a list of table requests |
|
610
|
|
|
|
|
|
|
as well as attributes which specify defaults for common operations. |
|
611
|
|
|
|
|
|
|
Table requests are documented in L. |
|
612
|
|
|
|
|
|
|
|
|
613
|
|
|
|
|
|
|
The C<%attr> hash provides default values for some of the table |
|
614
|
|
|
|
|
|
|
request attributes, namely the data cleanup operations ( C, |
|
615
|
|
|
|
|
|
|
C, C ), and the multi match attribute C, |
|
616
|
|
|
|
|
|
|
i.e., |
|
617
|
|
|
|
|
|
|
|
|
618
|
|
|
|
|
|
|
$p = HTML::TableParser->new( \@reqs, { Chomp => 1 } ); |
|
619
|
|
|
|
|
|
|
|
|
620
|
|
|
|
|
|
|
will set B on for all of the table requests, unless overridden |
|
621
|
|
|
|
|
|
|
by them. The data cleanup operations are documented above; C |
|
622
|
|
|
|
|
|
|
is documented in L. |
|
623
|
|
|
|
|
|
|
|
|
624
|
|
|
|
|
|
|
B defaults to on; all of the others default to off. |
|
625
|
|
|
|
|
|
|
|
|
626
|
|
|
|
|
|
|
=item parse_file |
|
627
|
|
|
|
|
|
|
|
|
628
|
|
|
|
|
|
|
This is the same function as in B. |
|
629
|
|
|
|
|
|
|
|
|
630
|
|
|
|
|
|
|
=item parse |
|
631
|
|
|
|
|
|
|
|
|
632
|
|
|
|
|
|
|
This is the same function as in B. |
|
633
|
|
|
|
|
|
|
|
|
634
|
|
|
|
|
|
|
=back |
|
635
|
|
|
|
|
|
|
|
|
636
|
|
|
|
|
|
|
=head1 Table Requests |
|
637
|
|
|
|
|
|
|
|
|
638
|
|
|
|
|
|
|
A table request is a hash used by B to determine |
|
639
|
|
|
|
|
|
|
which tables are to be parsed, the callbacks to be invoked, and any |
|
640
|
|
|
|
|
|
|
data cleanup. There may be multiple requests processed by one call to |
|
641
|
|
|
|
|
|
|
the parser; each table is associated with a single request (even if |
|
642
|
|
|
|
|
|
|
several requests match the table). |
|
643
|
|
|
|
|
|
|
|
|
644
|
|
|
|
|
|
|
A single request may match several tables, however unless the |
|
645
|
|
|
|
|
|
|
B attribute is specified for that request, it will be used |
|
646
|
|
|
|
|
|
|
for the first matching table only. |
|
647
|
|
|
|
|
|
|
|
|
648
|
|
|
|
|
|
|
A table request which matches a table id of C will be used as |
|
649
|
|
|
|
|
|
|
a catch-all request, and will match all tables not matched by other |
|
650
|
|
|
|
|
|
|
requests. Please note that tables are compared to the requests in the |
|
651
|
|
|
|
|
|
|
order that the latter are passed to the B method; place the |
|
652
|
|
|
|
|
|
|
B method last for proper behavior. |
|
653
|
|
|
|
|
|
|
|
|
654
|
|
|
|
|
|
|
=head2 Identifying tables to parse |
|
655
|
|
|
|
|
|
|
|
|
656
|
|
|
|
|
|
|
B needs to be told which tables to parse. This can |
|
657
|
|
|
|
|
|
|
be done by matching table ids or column names, or a combination of |
|
658
|
|
|
|
|
|
|
both. The table request hash elements dedicated to this are: |
|
659
|
|
|
|
|
|
|
|
|
660
|
|
|
|
|
|
|
=over 8 |
|
661
|
|
|
|
|
|
|
|
|
662
|
|
|
|
|
|
|
=item id |
|
663
|
|
|
|
|
|
|
|
|
664
|
|
|
|
|
|
|
This indicates a match on table id. It can take one of these forms: |
|
665
|
|
|
|
|
|
|
|
|
666
|
|
|
|
|
|
|
=over 8 |
|
667
|
|
|
|
|
|
|
|
|
668
|
|
|
|
|
|
|
=item exact match |
|
669
|
|
|
|
|
|
|
|
|
670
|
|
|
|
|
|
|
id => $match |
|
671
|
|
|
|
|
|
|
id => '1.2' |
|
672
|
|
|
|
|
|
|
|
|
673
|
|
|
|
|
|
|
Here C<$match> is a scalar which is compared directly to the table id. |
|
674
|
|
|
|
|
|
|
|
|
675
|
|
|
|
|
|
|
=item regular expression |
|
676
|
|
|
|
|
|
|
|
|
677
|
|
|
|
|
|
|
id => $re |
|
678
|
|
|
|
|
|
|
id => qr/1\.\d+\.2/ |
|
679
|
|
|
|
|
|
|
|
|
680
|
|
|
|
|
|
|
C<$re> is a regular expression, which must be constructed with the |
|
681
|
|
|
|
|
|
|
C operator. |
|
682
|
|
|
|
|
|
|
|
|
683
|
|
|
|
|
|
|
=item subroutine |
|
684
|
|
|
|
|
|
|
|
|
685
|
|
|
|
|
|
|
id => \&my_match_subroutine |
|
686
|
|
|
|
|
|
|
id => sub { my ( $id, $oids ) = @_ ; |
|
687
|
|
|
|
|
|
|
$oids[0] > 3 && $oids[1] < 2 } |
|
688
|
|
|
|
|
|
|
|
|
689
|
|
|
|
|
|
|
Here C is assigned a coderef to a subroutine which returns |
|
690
|
|
|
|
|
|
|
true if the table matches, false if not. The subroutine is passed |
|
691
|
|
|
|
|
|
|
two arguments: the table id as a scalar string ( e.g. C<1.2.3>) and the |
|
692
|
|
|
|
|
|
|
table id as an arrayref (e.g. C<$oids = [ 1, 2, 3]>). |
|
693
|
|
|
|
|
|
|
|
|
694
|
|
|
|
|
|
|
=back |
|
695
|
|
|
|
|
|
|
|
|
696
|
|
|
|
|
|
|
C may be passed an array containing any combination of the |
|
697
|
|
|
|
|
|
|
above: |
|
698
|
|
|
|
|
|
|
|
|
699
|
|
|
|
|
|
|
id => [ '1.2', qr/1\.\d+\.2/, sub { ... } ] |
|
700
|
|
|
|
|
|
|
|
|
701
|
|
|
|
|
|
|
Elements in the array may be preceded by a modifier indicating |
|
702
|
|
|
|
|
|
|
the action to be taken if the table matches on that element. |
|
703
|
|
|
|
|
|
|
The modifiers and their meanings are: |
|
704
|
|
|
|
|
|
|
|
|
705
|
|
|
|
|
|
|
=over 8 |
|
706
|
|
|
|
|
|
|
|
|
707
|
|
|
|
|
|
|
=item C<-> |
|
708
|
|
|
|
|
|
|
|
|
709
|
|
|
|
|
|
|
If the id matches, it is explicitly excluded from being processed |
|
710
|
|
|
|
|
|
|
by this request. |
|
711
|
|
|
|
|
|
|
|
|
712
|
|
|
|
|
|
|
=item C<--> |
|
713
|
|
|
|
|
|
|
|
|
714
|
|
|
|
|
|
|
If the id matches, it is skipped by B requests. |
|
715
|
|
|
|
|
|
|
|
|
716
|
|
|
|
|
|
|
=item C<+> |
|
717
|
|
|
|
|
|
|
|
|
718
|
|
|
|
|
|
|
If the id matches, it will be processed by this request. This |
|
719
|
|
|
|
|
|
|
is the default action. |
|
720
|
|
|
|
|
|
|
|
|
721
|
|
|
|
|
|
|
=back |
|
722
|
|
|
|
|
|
|
|
|
723
|
|
|
|
|
|
|
An example: |
|
724
|
|
|
|
|
|
|
|
|
725
|
|
|
|
|
|
|
id => [ '-', '1.2', 'DEFAULT' ] |
|
726
|
|
|
|
|
|
|
|
|
727
|
|
|
|
|
|
|
indicates that this request should be used for all tables, |
|
728
|
|
|
|
|
|
|
except for table 1.2. |
|
729
|
|
|
|
|
|
|
|
|
730
|
|
|
|
|
|
|
id => [ '--', '1.2' ] |
|
731
|
|
|
|
|
|
|
|
|
732
|
|
|
|
|
|
|
Table 2 is just plain skipped altogether. |
|
733
|
|
|
|
|
|
|
|
|
734
|
|
|
|
|
|
|
=item cols |
|
735
|
|
|
|
|
|
|
|
|
736
|
|
|
|
|
|
|
This indicates a match on column names. It can take one of these forms: |
|
737
|
|
|
|
|
|
|
|
|
738
|
|
|
|
|
|
|
=over 8 |
|
739
|
|
|
|
|
|
|
|
|
740
|
|
|
|
|
|
|
=item exact match |
|
741
|
|
|
|
|
|
|
|
|
742
|
|
|
|
|
|
|
cols => $match |
|
743
|
|
|
|
|
|
|
cols => 'Snacks01' |
|
744
|
|
|
|
|
|
|
|
|
745
|
|
|
|
|
|
|
Here C<$match> is a scalar which is compared directly to the column names. |
|
746
|
|
|
|
|
|
|
If any column matches, the table is processed. |
|
747
|
|
|
|
|
|
|
|
|
748
|
|
|
|
|
|
|
=item regular expression |
|
749
|
|
|
|
|
|
|
|
|
750
|
|
|
|
|
|
|
cols => $re |
|
751
|
|
|
|
|
|
|
cols => qr/Snacks\d+/ |
|
752
|
|
|
|
|
|
|
|
|
753
|
|
|
|
|
|
|
C<$re> is a regular expression, which must be constructed with the |
|
754
|
|
|
|
|
|
|
C operator. Again, a successful match against any column name |
|
755
|
|
|
|
|
|
|
causes the table to be processed. |
|
756
|
|
|
|
|
|
|
|
|
757
|
|
|
|
|
|
|
=item subroutine |
|
758
|
|
|
|
|
|
|
|
|
759
|
|
|
|
|
|
|
cols => \&my_match_subroutine |
|
760
|
|
|
|
|
|
|
cols => sub { my ( $id, $oids, $cols ) = @_ ; |
|
761
|
|
|
|
|
|
|
... } |
|
762
|
|
|
|
|
|
|
|
|
763
|
|
|
|
|
|
|
Here C is assigned a coderef to a subroutine which returns |
|
764
|
|
|
|
|
|
|
true if the table matches, false if not. The subroutine is passed |
|
765
|
|
|
|
|
|
|
three arguments: the table id as a scalar string ( e.g. C<1.2.3>), the |
|
766
|
|
|
|
|
|
|
table id as an arrayref (e.g. C<$oids = [ 1, 2, 3]>), and the column |
|
767
|
|
|
|
|
|
|
names, as an arrayref (e.g. C<$cols = [ 'col1', 'col2' ]>). This |
|
768
|
|
|
|
|
|
|
option gives the calling routine the ability to make arbitrary |
|
769
|
|
|
|
|
|
|
selections based upon table id and columns. |
|
770
|
|
|
|
|
|
|
|
|
771
|
|
|
|
|
|
|
=back |
|
772
|
|
|
|
|
|
|
|
|
773
|
|
|
|
|
|
|
C may be passed an arrayref containing any combination of the |
|
774
|
|
|
|
|
|
|
above: |
|
775
|
|
|
|
|
|
|
|
|
776
|
|
|
|
|
|
|
cols => [ 'Snacks01', qr/Snacks\d+/, sub { ... } ] |
|
777
|
|
|
|
|
|
|
|
|
778
|
|
|
|
|
|
|
Elements in the array may be preceded by a modifier indicating |
|
779
|
|
|
|
|
|
|
the action to be taken if the table matches on that element. |
|
780
|
|
|
|
|
|
|
They are the same as the table id modifiers mentioned above. |
|
781
|
|
|
|
|
|
|
|
|
782
|
|
|
|
|
|
|
=item colre |
|
783
|
|
|
|
|
|
|
|
|
784
|
|
|
|
|
|
|
B |
|
785
|
|
|
|
|
|
|
An arrayref containing the regular expressions to match, or a scalar |
|
786
|
|
|
|
|
|
|
containing a single reqular expression |
|
787
|
|
|
|
|
|
|
|
|
788
|
|
|
|
|
|
|
=back |
|
789
|
|
|
|
|
|
|
|
|
790
|
|
|
|
|
|
|
More than one of these may be used for a single table request. A |
|
791
|
|
|
|
|
|
|
request may match more than one table. By default a request is used |
|
792
|
|
|
|
|
|
|
only once (even the C id match!). Set the C |
|
793
|
|
|
|
|
|
|
attribute to enable multiple matches per request. |
|
794
|
|
|
|
|
|
|
|
|
795
|
|
|
|
|
|
|
When attempting to match a table, the following steps are taken: |
|
796
|
|
|
|
|
|
|
|
|
797
|
|
|
|
|
|
|
=over 8 |
|
798
|
|
|
|
|
|
|
|
|
799
|
|
|
|
|
|
|
=item 1 |
|
800
|
|
|
|
|
|
|
|
|
801
|
|
|
|
|
|
|
The table id is compared to the requests which contain an id match. |
|
802
|
|
|
|
|
|
|
The first such match is used (in the order given in the passed array). |
|
803
|
|
|
|
|
|
|
|
|
804
|
|
|
|
|
|
|
=item 2 |
|
805
|
|
|
|
|
|
|
|
|
806
|
|
|
|
|
|
|
If no explicit id match is found, column name matches are attempted. |
|
807
|
|
|
|
|
|
|
The first such match is used (in the order given in the passed array) |
|
808
|
|
|
|
|
|
|
|
|
809
|
|
|
|
|
|
|
=item 3 |
|
810
|
|
|
|
|
|
|
|
|
811
|
|
|
|
|
|
|
If no column name match is found (or there were none requested), |
|
812
|
|
|
|
|
|
|
the first request which matches an B of C is used. |
|
813
|
|
|
|
|
|
|
|
|
814
|
|
|
|
|
|
|
=back |
|
815
|
|
|
|
|
|
|
|
|
816
|
|
|
|
|
|
|
=head2 Specifying the data callbacks |
|
817
|
|
|
|
|
|
|
|
|
818
|
|
|
|
|
|
|
Callback functions are specified with the callback attributes |
|
819
|
|
|
|
|
|
|
C, C, C, C, and C. They should be set to |
|
820
|
|
|
|
|
|
|
code references, i.e. |
|
821
|
|
|
|
|
|
|
|
|
822
|
|
|
|
|
|
|
%table_req = ( ..., start => \&start_func, end => \&end_func ) |
|
823
|
|
|
|
|
|
|
|
|
824
|
|
|
|
|
|
|
To use methods, specify the object with the C key, and |
|
825
|
|
|
|
|
|
|
the method names via the callback attributes, which should be set |
|
826
|
|
|
|
|
|
|
to strings. If you don't specify method names they will default to (you |
|
827
|
|
|
|
|
|
|
guessed it) C, C, C, C, and C. |
|
828
|
|
|
|
|
|
|
|
|
829
|
|
|
|
|
|
|
$obj = SomeClass->new(); |
|
830
|
|
|
|
|
|
|
# ... |
|
831
|
|
|
|
|
|
|
%table_req_1 = ( ..., obj => $obj ); |
|
832
|
|
|
|
|
|
|
%table_req_2 = ( ..., obj => $obj, start => 'start', |
|
833
|
|
|
|
|
|
|
end => 'end' ); |
|
834
|
|
|
|
|
|
|
|
|
835
|
|
|
|
|
|
|
You can also have B create a new object for you |
|
836
|
|
|
|
|
|
|
for each table by specifying the C attribute. By default |
|
837
|
|
|
|
|
|
|
the constructor is assumed to be the class B method; if not, |
|
838
|
|
|
|
|
|
|
specify it using the C attribute: |
|
839
|
|
|
|
|
|
|
|
|
840
|
|
|
|
|
|
|
use MyClass; |
|
841
|
|
|
|
|
|
|
%table_req = ( ..., class => 'MyClass', new => 'mynew' ); |
|
842
|
|
|
|
|
|
|
|
|
843
|
|
|
|
|
|
|
To use a function instead of a method for a particular callback, |
|
844
|
|
|
|
|
|
|
set the callback attribute to a code reference: |
|
845
|
|
|
|
|
|
|
|
|
846
|
|
|
|
|
|
|
%table_req = ( ..., obj => $obj, end => \&end_func ); |
|
847
|
|
|
|
|
|
|
|
|
848
|
|
|
|
|
|
|
You don't have to provide all the callbacks. You should not use both |
|
849
|
|
|
|
|
|
|
C and C in the same table request. |
|
850
|
|
|
|
|
|
|
|
|
851
|
|
|
|
|
|
|
B automatically determines if your object |
|
852
|
|
|
|
|
|
|
or class has one of the required methods. If you wish it I |
|
853
|
|
|
|
|
|
|
to use a particular method, set it equal to C. For example |
|
854
|
|
|
|
|
|
|
|
|
855
|
|
|
|
|
|
|
%table_req = ( ..., obj => $obj, end => undef ) |
|
856
|
|
|
|
|
|
|
|
|
857
|
|
|
|
|
|
|
indicates the object's B method should not be called, even |
|
858
|
|
|
|
|
|
|
if it exists. |
|
859
|
|
|
|
|
|
|
|
|
860
|
|
|
|
|
|
|
You can specify arbitrary data to be passed to the callback functions |
|
861
|
|
|
|
|
|
|
via the C attribute: |
|
862
|
|
|
|
|
|
|
|
|
863
|
|
|
|
|
|
|
%table_req = ( ..., udata => \%hash_of_my_special_stuff ) |
|
864
|
|
|
|
|
|
|
|
|
865
|
|
|
|
|
|
|
=head2 Specifying Data cleanup operations |
|
866
|
|
|
|
|
|
|
|
|
867
|
|
|
|
|
|
|
Data cleanup operations may be specified uniquely for each table. The |
|
868
|
|
|
|
|
|
|
available keys are C, C, C. They should be |
|
869
|
|
|
|
|
|
|
set to a non-zero value if the operation is to be performed. |
|
870
|
|
|
|
|
|
|
|
|
871
|
|
|
|
|
|
|
=head2 Other Attributes |
|
872
|
|
|
|
|
|
|
|
|
873
|
|
|
|
|
|
|
The C key is used when a request is capable of handling |
|
874
|
|
|
|
|
|
|
multiple tables in the document. Ordinarily, a request will process |
|
875
|
|
|
|
|
|
|
a single table only (even C requests). |
|
876
|
|
|
|
|
|
|
Set it to a non-zero value to allow the request to handle more than |
|
877
|
|
|
|
|
|
|
one table. |
|
878
|
|
|
|
|
|
|
|
|
879
|
|
|
|
|
|
|
=head1 BUGS |
|
880
|
|
|
|
|
|
|
|
|
881
|
|
|
|
|
|
|
Please report any bugs or feature requests on the bugtracker website |
|
882
|
|
|
|
|
|
|
L or by |
|
883
|
|
|
|
|
|
|
email to |
|
884
|
|
|
|
|
|
|
L. |
|
885
|
|
|
|
|
|
|
|
|
886
|
|
|
|
|
|
|
When submitting a bug or request, please include a test-file or a |
|
887
|
|
|
|
|
|
|
patch to an existing test-file that illustrates the bug or desired |
|
888
|
|
|
|
|
|
|
feature. |
|
889
|
|
|
|
|
|
|
|
|
890
|
|
|
|
|
|
|
=head1 SOURCE |
|
891
|
|
|
|
|
|
|
|
|
892
|
|
|
|
|
|
|
The development version is on github at L |
|
893
|
|
|
|
|
|
|
and may be cloned from L |
|
894
|
|
|
|
|
|
|
|
|
895
|
|
|
|
|
|
|
=head1 AUTHOR |
|
896
|
|
|
|
|
|
|
|
|
897
|
|
|
|
|
|
|
Diab Jerius |
|
898
|
|
|
|
|
|
|
|
|
899
|
|
|
|
|
|
|
=head1 COPYRIGHT AND LICENSE |
|
900
|
|
|
|
|
|
|
|
|
901
|
|
|
|
|
|
|
This software is Copyright (c) 2018 by Smithsonian Astrophysical Observatory. |
|
902
|
|
|
|
|
|
|
|
|
903
|
|
|
|
|
|
|
This is free software, licensed under: |
|
904
|
|
|
|
|
|
|
|
|
905
|
|
|
|
|
|
|
The GNU General Public License, Version 3, June 2007 |
|
906
|
|
|
|
|
|
|
|
|
907
|
|
|
|
|
|
|
=cut |
|
908
|
|
|
|
|
|
|
|
|
909
|
|
|
|
|
|
|
__END__ |