line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
package Text::Summarize; |
2
|
1
|
|
|
1
|
|
392968
|
use strict; |
|
1
|
|
|
|
|
4
|
|
|
1
|
|
|
|
|
40
|
|
3
|
1
|
|
|
1
|
|
6
|
use warnings; |
|
1
|
|
|
|
|
1
|
|
|
1
|
|
|
|
|
30
|
|
4
|
1
|
|
|
1
|
|
1760
|
use Log::Log4perl; |
|
1
|
|
|
|
|
187936
|
|
|
1
|
|
|
|
|
6
|
|
5
|
1
|
|
|
1
|
|
1847
|
use Text::Categorize::Textrank; |
|
0
|
|
|
|
|
|
|
|
0
|
|
|
|
|
|
|
6
|
|
|
|
|
|
|
use Data::Dump qw(dump); |
7
|
|
|
|
|
|
|
|
8
|
|
|
|
|
|
|
BEGIN |
9
|
|
|
|
|
|
|
{ |
10
|
|
|
|
|
|
|
use Exporter (); |
11
|
|
|
|
|
|
|
use vars qw($VERSION @ISA @EXPORT @EXPORT_OK %EXPORT_TAGS); |
12
|
|
|
|
|
|
|
$VERSION = '0.50'; |
13
|
|
|
|
|
|
|
@ISA = qw(Exporter); |
14
|
|
|
|
|
|
|
@EXPORT = qw(getSumbasicRankingOfSentences); |
15
|
|
|
|
|
|
|
@EXPORT_OK = qw(getSumbasicRankingOfSentences); |
16
|
|
|
|
|
|
|
%EXPORT_TAGS = (); |
17
|
|
|
|
|
|
|
} |
18
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
#12345678901234567890123456789012345678901234 |
20
|
|
|
|
|
|
|
#Routine to compute summaries of text. |
21
|
|
|
|
|
|
|
|
22
|
|
|
|
|
|
|
=head1 NAME |
23
|
|
|
|
|
|
|
|
24
|
|
|
|
|
|
|
C - Routine to compute summaries of text. |
25
|
|
|
|
|
|
|
|
26
|
|
|
|
|
|
|
=head1 SYNOPSIS |
27
|
|
|
|
|
|
|
|
28
|
|
|
|
|
|
|
use strict; |
29
|
|
|
|
|
|
|
use warnings; |
30
|
|
|
|
|
|
|
use Text::Summarize; |
31
|
|
|
|
|
|
|
use Data::Dump qw(dump); |
32
|
|
|
|
|
|
|
my $listOfSentences = [ |
33
|
|
|
|
|
|
|
{ id => 0, listOfTokens => [qw(all people are equal)] }, |
34
|
|
|
|
|
|
|
{ id => 1, listOfTokens => [qw(all men are equal)] }, |
35
|
|
|
|
|
|
|
{ id => 2, listOfTokens => [qw(all are equal)] }, |
36
|
|
|
|
|
|
|
]; |
37
|
|
|
|
|
|
|
dump getSumbasicRankingOfSentences(listOfSentences => $listOfSentences); |
38
|
|
|
|
|
|
|
|
39
|
|
|
|
|
|
|
=head1 DESCRIPTION |
40
|
|
|
|
|
|
|
|
41
|
|
|
|
|
|
|
C contains a routine to score a list of sentences |
42
|
|
|
|
|
|
|
for inclusion in a summary of the text using the |
43
|
|
|
|
|
|
|
SumBasic algorithm from the report I |
44
|
|
|
|
|
|
|
by L. Vanderwendea, H. Suzukia, C. Brocketta, and A. Nenkovab. |
45
|
|
|
|
|
|
|
|
46
|
|
|
|
|
|
|
=head1 ROUTINES |
47
|
|
|
|
|
|
|
|
48
|
|
|
|
|
|
|
=head2 C |
49
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
use Text::Summarize; |
51
|
|
|
|
|
|
|
use Data::Dump qw(dump); |
52
|
|
|
|
|
|
|
my $listOfSentences = [ |
53
|
|
|
|
|
|
|
{ id => 0, listOfTokens => [qw(all people are equal)] }, |
54
|
|
|
|
|
|
|
{ id => 1, listOfTokens => [qw(all men are equal)] }, |
55
|
|
|
|
|
|
|
{ id => 2, listOfTokens => [qw(all are equal)] }, |
56
|
|
|
|
|
|
|
]; |
57
|
|
|
|
|
|
|
dump getSumbasicRankingOfSentences(listOfSentences => $listOfSentences); |
58
|
|
|
|
|
|
|
|
59
|
|
|
|
|
|
|
C computes the sumBasic score of the list of sentences |
60
|
|
|
|
|
|
|
provided. It returns an array reference containing the pairs C<[id, score]> sorted |
61
|
|
|
|
|
|
|
in descending order of score, where C is from C. |
62
|
|
|
|
|
|
|
|
63
|
|
|
|
|
|
|
=over |
64
|
|
|
|
|
|
|
|
65
|
|
|
|
|
|
|
=item C |
66
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
listOfSentences => [{id => '..', listOfTokens => [...]}, ..., {id => '..', listOfTokens => [...]}] |
68
|
|
|
|
|
|
|
|
69
|
|
|
|
|
|
|
C holds the list of sentences that are to be scored. Each |
70
|
|
|
|
|
|
|
item in the list is a hash reference of the form C<{id =E '..', listOfTokens =E [...]}> where |
71
|
|
|
|
|
|
|
C is a unique identifier for the sentence and C is an array |
72
|
|
|
|
|
|
|
reference of the list of tokens comprizing the sentence. |
73
|
|
|
|
|
|
|
|
74
|
|
|
|
|
|
|
=item C |
75
|
|
|
|
|
|
|
|
76
|
|
|
|
|
|
|
tokenWeight => {} |
77
|
|
|
|
|
|
|
|
78
|
|
|
|
|
|
|
C is a optional hash reference that provides the weight of the tokens defined |
79
|
|
|
|
|
|
|
in C. If C is defined, but undefined for a token in a sentence, |
80
|
|
|
|
|
|
|
then the tokens weight defaults to zero unless C is true, |
81
|
|
|
|
|
|
|
in which case the token is ignored and not used to compute the average weight |
82
|
|
|
|
|
|
|
of the sentences containing it. If C is undefined then the weights of the tokens |
83
|
|
|
|
|
|
|
are either their frequency of occurrence in the filtered text, or their textranks if C is defined. |
84
|
|
|
|
|
|
|
|
85
|
|
|
|
|
|
|
=item C |
86
|
|
|
|
|
|
|
|
87
|
|
|
|
|
|
|
ignoreUndefinedTokens => 0 |
88
|
|
|
|
|
|
|
|
89
|
|
|
|
|
|
|
If C is true, then any tokens for which C is |
90
|
|
|
|
|
|
|
undefined are ignored and not used to compute the average weight of a |
91
|
|
|
|
|
|
|
sentence; the default is false. |
92
|
|
|
|
|
|
|
|
93
|
|
|
|
|
|
|
=item C |
94
|
|
|
|
|
|
|
|
95
|
|
|
|
|
|
|
tokenWeightUpdateFunction => &subroutine (currentTokenWeight, initialTokenWeight, token, selectedSentenceId, selectedSentenceWeight) |
96
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
C is an optional parameter for defining the function that updates the |
98
|
|
|
|
|
|
|
weight of a token when it is contained in a selected sentence. Five parameters are passed to the |
99
|
|
|
|
|
|
|
subroutine: the token's current weight (float), the token's initial weight (float), the token (string), the C of the |
100
|
|
|
|
|
|
|
selected sentence (string), and the current average weight of the tokens in the selected sentence (float). |
101
|
|
|
|
|
|
|
The default is L. |
102
|
|
|
|
|
|
|
|
103
|
|
|
|
|
|
|
=item C |
104
|
|
|
|
|
|
|
|
105
|
|
|
|
|
|
|
textRankParameters => undef |
106
|
|
|
|
|
|
|
|
107
|
|
|
|
|
|
|
If C is defined, then the token weights |
108
|
|
|
|
|
|
|
are computed using L. The parameters to use for L, |
109
|
|
|
|
|
|
|
excluding the C parameters, can be set using the hash reference defined by C. |
110
|
|
|
|
|
|
|
For example, C {directedGraph =E 1}> would make the textrank weights |
111
|
|
|
|
|
|
|
be computed using a directed token graph. |
112
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
=back |
114
|
|
|
|
|
|
|
|
115
|
|
|
|
|
|
|
=cut |
116
|
|
|
|
|
|
|
|
117
|
|
|
|
|
|
|
sub getSumbasicRankingOfSentences |
118
|
|
|
|
|
|
|
{ |
119
|
|
|
|
|
|
|
my (%Parameters) = @_; |
120
|
|
|
|
|
|
|
|
121
|
|
|
|
|
|
|
# get the list of sentences. |
122
|
|
|
|
|
|
|
my $listOfSentences = $Parameters{listOfSentences} if exists $Parameters{listOfSentences}; |
123
|
|
|
|
|
|
|
return [] unless defined $listOfSentences; |
124
|
|
|
|
|
|
|
|
125
|
|
|
|
|
|
|
# get the original token weights. |
126
|
|
|
|
|
|
|
my $originalTokenWeights; |
127
|
|
|
|
|
|
|
$originalTokenWeights = $Parameters{tokenWeight} if (exists($Parameters{tokenWeight}) && defined($Parameters{tokenWeight})); |
128
|
|
|
|
|
|
|
|
129
|
|
|
|
|
|
|
# if textRankParameters is defined, compute the token weights via textrank. |
130
|
|
|
|
|
|
|
if (exists($Parameters{textRankParameters}) && defined($Parameters{textRankParameters})) |
131
|
|
|
|
|
|
|
{ |
132
|
|
|
|
|
|
|
$originalTokenWeights = _getTextRankWeightOfTokens(%Parameters, listOfSentences => $listOfSentences); |
133
|
|
|
|
|
|
|
} |
134
|
|
|
|
|
|
|
|
135
|
|
|
|
|
|
|
# if $originalTokenWeights is not defined, then use the frequency of the tokens as their weight. |
136
|
|
|
|
|
|
|
if (!defined ($originalTokenWeights)) |
137
|
|
|
|
|
|
|
{ |
138
|
|
|
|
|
|
|
$originalTokenWeights = _getFrequencyWeightOfTokens(listOfSentences => $listOfSentences); |
139
|
|
|
|
|
|
|
} |
140
|
|
|
|
|
|
|
|
141
|
|
|
|
|
|
|
# get the function to update the weights of the tokens. |
142
|
|
|
|
|
|
|
my $tokenWeightUpdateFunction = \&tokenWeightUpdateFunction_Squared; |
143
|
|
|
|
|
|
|
$tokenWeightUpdateFunction = $Parameters{tokenWeightUpdateFunction} if exists $Parameters{tokenWeightUpdateFunction}; |
144
|
|
|
|
|
|
|
|
145
|
|
|
|
|
|
|
# set the flag for ignoreUndefinedTokens. |
146
|
|
|
|
|
|
|
my $ignoreUndefinedTokens = exists $Parameters{ignoreUndefinedTokens} && $Parameters{ignoreUndefinedTokens}; |
147
|
|
|
|
|
|
|
|
148
|
|
|
|
|
|
|
# copy the weights of only the tokens that occur in the sentences. |
149
|
|
|
|
|
|
|
# the default weight of a token is zero. |
150
|
|
|
|
|
|
|
my %tokenWeight; |
151
|
|
|
|
|
|
|
for (my $i = 0 ; $i < @$listOfSentences ; $i++) |
152
|
|
|
|
|
|
|
{ |
153
|
|
|
|
|
|
|
|
154
|
|
|
|
|
|
|
# if the sentence has no id, skip it. |
155
|
|
|
|
|
|
|
unless (exists $listOfSentences->[$i]->{id}) |
156
|
|
|
|
|
|
|
{ |
157
|
|
|
|
|
|
|
|
158
|
|
|
|
|
|
|
# get the list of tokens in the sentence as a string. |
159
|
|
|
|
|
|
|
my $stringOfTokens; |
160
|
|
|
|
|
|
|
if (exists($listOfSentences->[$i]->{listOfTokens})) |
161
|
|
|
|
|
|
|
{ |
162
|
|
|
|
|
|
|
$stringOfTokens = join(' ', @{ $listOfSentences->[$i]->{listOfTokens} }); |
163
|
|
|
|
|
|
|
} |
164
|
|
|
|
|
|
|
|
165
|
|
|
|
|
|
|
# create the message to log. |
166
|
|
|
|
|
|
|
my $logger = Log::Log4perl->get_logger(); |
167
|
|
|
|
|
|
|
my $message; |
168
|
|
|
|
|
|
|
if (defined $stringOfTokens) |
169
|
|
|
|
|
|
|
{ |
170
|
|
|
|
|
|
|
$message = "warning: skipping sentence number $i with tokens $stringOfTokens since it is missing an id.\n"; |
171
|
|
|
|
|
|
|
} |
172
|
|
|
|
|
|
|
else |
173
|
|
|
|
|
|
|
{ |
174
|
|
|
|
|
|
|
$message = "warning: skipping sentence number $i since it is missing an id and listOfTokens.\n"; |
175
|
|
|
|
|
|
|
} |
176
|
|
|
|
|
|
|
|
177
|
|
|
|
|
|
|
# log the message as a warning. |
178
|
|
|
|
|
|
|
$logger->logwarn($message); |
179
|
|
|
|
|
|
|
|
180
|
|
|
|
|
|
|
# skip processing the sentence. |
181
|
|
|
|
|
|
|
next; |
182
|
|
|
|
|
|
|
} |
183
|
|
|
|
|
|
|
|
184
|
|
|
|
|
|
|
# get the listOfTokens of the sentence. |
185
|
|
|
|
|
|
|
if ((exists $listOfSentences->[$i]->{listOfTokens}) && (@{ $listOfSentences->[$i]->{listOfTokens} })) |
186
|
|
|
|
|
|
|
{ |
187
|
|
|
|
|
|
|
my $listOfTokens = $listOfSentences->[$i]->{listOfTokens}; |
188
|
|
|
|
|
|
|
foreach my $token (@$listOfTokens) |
189
|
|
|
|
|
|
|
{ |
190
|
|
|
|
|
|
|
|
191
|
|
|
|
|
|
|
# if the weight is already defined for the token, skip it. |
192
|
|
|
|
|
|
|
next if exists $tokenWeight{$token}; |
193
|
|
|
|
|
|
|
|
194
|
|
|
|
|
|
|
# if weight for token not defined, it defaults to zero if ignoreUndefinedTokens is false. |
195
|
|
|
|
|
|
|
if (exists $originalTokenWeights->{$token}) |
196
|
|
|
|
|
|
|
{ |
197
|
|
|
|
|
|
|
$tokenWeight{$token} = $originalTokenWeights->{$token}; |
198
|
|
|
|
|
|
|
} |
199
|
|
|
|
|
|
|
elsif (!$ignoreUndefinedTokens) |
200
|
|
|
|
|
|
|
{ |
201
|
|
|
|
|
|
|
$tokenWeight{$token} = 0; |
202
|
|
|
|
|
|
|
} |
203
|
|
|
|
|
|
|
} |
204
|
|
|
|
|
|
|
} |
205
|
|
|
|
|
|
|
} |
206
|
|
|
|
|
|
|
|
207
|
|
|
|
|
|
|
# normalize the token weights to sum to one. |
208
|
|
|
|
|
|
|
my $sum = 0; |
209
|
|
|
|
|
|
|
while (my ($token, $weight) = each %tokenWeight) { $sum += $weight; } |
210
|
|
|
|
|
|
|
$sum = 1 if ($sum == 0); |
211
|
|
|
|
|
|
|
while (my ($token, $weight) = each %tokenWeight) { $tokenWeight{$token} /= $sum; } |
212
|
|
|
|
|
|
|
|
213
|
|
|
|
|
|
|
# keep a copy of the initial token weights. |
214
|
|
|
|
|
|
|
my %initialTokenWeight = %tokenWeight; |
215
|
|
|
|
|
|
|
|
216
|
|
|
|
|
|
|
# @listOfEmptySentenceIds will hold the list of empty sentence ids. |
217
|
|
|
|
|
|
|
my @listOfEmptySentenceIds; |
218
|
|
|
|
|
|
|
|
219
|
|
|
|
|
|
|
# make a copy of the list of sentences |
220
|
|
|
|
|
|
|
my @localListOfSentences; |
221
|
|
|
|
|
|
|
for (my $i = 0 ; $i < @$listOfSentences ; $i++) |
222
|
|
|
|
|
|
|
{ |
223
|
|
|
|
|
|
|
|
224
|
|
|
|
|
|
|
# if the sentence has no id, skip it. |
225
|
|
|
|
|
|
|
next unless exists $listOfSentences->[$i]->{id}; |
226
|
|
|
|
|
|
|
|
227
|
|
|
|
|
|
|
# copy the id of the sentence. |
228
|
|
|
|
|
|
|
my %sentence; |
229
|
|
|
|
|
|
|
$sentence{id} = $listOfSentences->[$i]->{id}; |
230
|
|
|
|
|
|
|
|
231
|
|
|
|
|
|
|
# convert the list of tokens in a sentence to a hash with the key as the token and the value its occurance in the sentence. |
232
|
|
|
|
|
|
|
if ((exists $listOfSentences->[$i]->{listOfTokens}) && (@{ $listOfSentences->[$i]->{listOfTokens} })) |
233
|
|
|
|
|
|
|
{ |
234
|
|
|
|
|
|
|
my %tokenCount; |
235
|
|
|
|
|
|
|
my $empty = 1; |
236
|
|
|
|
|
|
|
foreach my $token (@{ $listOfSentences->[$i]->{listOfTokens} }) |
237
|
|
|
|
|
|
|
{ |
238
|
|
|
|
|
|
|
|
239
|
|
|
|
|
|
|
# if the weight for the token is not defined, skip it. |
240
|
|
|
|
|
|
|
if (exists $tokenWeight{$token}) |
241
|
|
|
|
|
|
|
{ |
242
|
|
|
|
|
|
|
++$tokenCount{$token}; |
243
|
|
|
|
|
|
|
$empty = 0; |
244
|
|
|
|
|
|
|
} |
245
|
|
|
|
|
|
|
} |
246
|
|
|
|
|
|
|
|
247
|
|
|
|
|
|
|
# if the sentence has no defined tokens, store the id on @listOfEmptySentenceIds. |
248
|
|
|
|
|
|
|
if ($empty) |
249
|
|
|
|
|
|
|
{ |
250
|
|
|
|
|
|
|
push @listOfEmptySentenceIds, [ $listOfSentences->[$i]->{id}, scalar @{ $listOfSentences->[$i]->{listOfTokens} } ]; |
251
|
|
|
|
|
|
|
} |
252
|
|
|
|
|
|
|
else |
253
|
|
|
|
|
|
|
{ |
254
|
|
|
|
|
|
|
$sentence{tokenCounts} = \%tokenCount; |
255
|
|
|
|
|
|
|
|
256
|
|
|
|
|
|
|
# store the sentence in a list. |
257
|
|
|
|
|
|
|
push @localListOfSentences, \%sentence; |
258
|
|
|
|
|
|
|
} |
259
|
|
|
|
|
|
|
} |
260
|
|
|
|
|
|
|
else |
261
|
|
|
|
|
|
|
{ |
262
|
|
|
|
|
|
|
|
263
|
|
|
|
|
|
|
# if the sentence has no tokens, store the id on @listOfEmptySentenceIds. |
264
|
|
|
|
|
|
|
push @listOfEmptySentenceIds, [ $listOfSentences->[$i]->{id}, scalar @{ $listOfSentences->[$i]->{listOfTokens} } ]; |
265
|
|
|
|
|
|
|
} |
266
|
|
|
|
|
|
|
} |
267
|
|
|
|
|
|
|
|
268
|
|
|
|
|
|
|
# compute the average weight of each sentence and initialize its selected flag to false. |
269
|
|
|
|
|
|
|
for (my $i = 0 ; $i < @localListOfSentences ; $i++) |
270
|
|
|
|
|
|
|
{ |
271
|
|
|
|
|
|
|
|
272
|
|
|
|
|
|
|
# get the pointer to the sentence. |
273
|
|
|
|
|
|
|
my $sentence = $localListOfSentences[$i]; |
274
|
|
|
|
|
|
|
|
275
|
|
|
|
|
|
|
# compute the weight of the sentence. |
276
|
|
|
|
|
|
|
my $weight = 0; |
277
|
|
|
|
|
|
|
my $tokenCountSum = 0; |
278
|
|
|
|
|
|
|
while (my ($token, $count) = each %{ $sentence->{tokenCounts} }) |
279
|
|
|
|
|
|
|
{ |
280
|
|
|
|
|
|
|
$weight += $count * $tokenWeight{$token}; |
281
|
|
|
|
|
|
|
$tokenCountSum += $count; |
282
|
|
|
|
|
|
|
} |
283
|
|
|
|
|
|
|
$sentence->{size} = $tokenCountSum; |
284
|
|
|
|
|
|
|
$sentence->{weight} = $weight / $sentence->{size}; |
285
|
|
|
|
|
|
|
|
286
|
|
|
|
|
|
|
# initialize each sentence as not selected. |
287
|
|
|
|
|
|
|
$sentence->{selected} = 0; |
288
|
|
|
|
|
|
|
} |
289
|
|
|
|
|
|
|
|
290
|
|
|
|
|
|
|
# build the inverted index of the sentences and tokens, called tokenSentenceIndex. |
291
|
|
|
|
|
|
|
my %tokenSentenceIndex; |
292
|
|
|
|
|
|
|
for (my $i = 0 ; $i < @localListOfSentences ; $i++) |
293
|
|
|
|
|
|
|
{ |
294
|
|
|
|
|
|
|
|
295
|
|
|
|
|
|
|
# get the pointer to the sentence. |
296
|
|
|
|
|
|
|
my $sentence = $localListOfSentences[$i]; |
297
|
|
|
|
|
|
|
|
298
|
|
|
|
|
|
|
# get the list of tokens in the sentence. |
299
|
|
|
|
|
|
|
foreach my $token (keys %{ $sentence->{tokenCounts} }) |
300
|
|
|
|
|
|
|
{ |
301
|
|
|
|
|
|
|
|
302
|
|
|
|
|
|
|
# add the weightSentence pointer to the tokenSentenceIndex. |
303
|
|
|
|
|
|
|
$tokenSentenceIndex{$token} = [] unless exists $tokenSentenceIndex{$token}; |
304
|
|
|
|
|
|
|
|
305
|
|
|
|
|
|
|
# note we are storing the index of the sentence, not the pointer to the sentence. |
306
|
|
|
|
|
|
|
push @{ $tokenSentenceIndex{$token} }, $i; |
307
|
|
|
|
|
|
|
} |
308
|
|
|
|
|
|
|
} |
309
|
|
|
|
|
|
|
|
310
|
|
|
|
|
|
|
# make the list of just the tokens. |
311
|
|
|
|
|
|
|
my @listOfTokens = keys %tokenWeight; |
312
|
|
|
|
|
|
|
|
313
|
|
|
|
|
|
|
# @rankedListOfSentences will hold the sentences in sumbasic order. |
314
|
|
|
|
|
|
|
my @rankedListOfSentences; |
315
|
|
|
|
|
|
|
|
316
|
|
|
|
|
|
|
# loop over the sentences until they have all been selected. |
317
|
|
|
|
|
|
|
while (scalar(@rankedListOfSentences) < scalar(@localListOfSentences)) |
318
|
|
|
|
|
|
|
{ |
319
|
|
|
|
|
|
|
|
320
|
|
|
|
|
|
|
# if there are no tokens left, exit the loop. |
321
|
|
|
|
|
|
|
last unless @listOfTokens > 0; |
322
|
|
|
|
|
|
|
|
323
|
|
|
|
|
|
|
# get the token with the greatest (weight, length, -order). |
324
|
|
|
|
|
|
|
my $maxIndex = 0; |
325
|
|
|
|
|
|
|
my $maxToken = $listOfTokens[$maxIndex]; |
326
|
|
|
|
|
|
|
my $maxTokenWeight = $tokenWeight{$maxToken}; |
327
|
|
|
|
|
|
|
for (my $i = 1 ; $i < scalar(@listOfTokens) ; $i++) |
328
|
|
|
|
|
|
|
{ |
329
|
|
|
|
|
|
|
my $cmp; |
330
|
|
|
|
|
|
|
if ($maxTokenWeight < $tokenWeight{ $listOfTokens[$i] }) |
331
|
|
|
|
|
|
|
{ |
332
|
|
|
|
|
|
|
|
333
|
|
|
|
|
|
|
# $maxTokenWeight is smaller. |
334
|
|
|
|
|
|
|
$cmp = -1; |
335
|
|
|
|
|
|
|
} |
336
|
|
|
|
|
|
|
elsif ($maxTokenWeight > $tokenWeight{ $listOfTokens[$i] }) |
337
|
|
|
|
|
|
|
{ |
338
|
|
|
|
|
|
|
|
339
|
|
|
|
|
|
|
# $maxTokenWeight is larger. |
340
|
|
|
|
|
|
|
$cmp = 1; |
341
|
|
|
|
|
|
|
} |
342
|
|
|
|
|
|
|
else |
343
|
|
|
|
|
|
|
{ |
344
|
|
|
|
|
|
|
|
345
|
|
|
|
|
|
|
# weights are equal, compare token lengths, choose the longer one. |
346
|
|
|
|
|
|
|
$cmp = length($maxToken) <=> length($listOfTokens[$i]); |
347
|
|
|
|
|
|
|
|
348
|
|
|
|
|
|
|
# if tokens have equal length, choose the one lexically smaller. |
349
|
|
|
|
|
|
|
if ($cmp == 0) { $cmp = $listOfTokens[$i] cmp $maxToken; } |
350
|
|
|
|
|
|
|
} |
351
|
|
|
|
|
|
|
|
352
|
|
|
|
|
|
|
# if the current max is smaller, replace it. |
353
|
|
|
|
|
|
|
if ($cmp == -1) |
354
|
|
|
|
|
|
|
{ |
355
|
|
|
|
|
|
|
$maxIndex = $i; |
356
|
|
|
|
|
|
|
$maxToken = $listOfTokens[$maxIndex]; |
357
|
|
|
|
|
|
|
$maxTokenWeight = $tokenWeight{$maxToken}; |
358
|
|
|
|
|
|
|
} |
359
|
|
|
|
|
|
|
} |
360
|
|
|
|
|
|
|
|
361
|
|
|
|
|
|
|
# copy the last token to where the max was, it may be popped off if there are no |
362
|
|
|
|
|
|
|
# sentences left containing it. |
363
|
|
|
|
|
|
|
$listOfTokens[$maxIndex] = $listOfTokens[-1]; |
364
|
|
|
|
|
|
|
$listOfTokens[-1] = $maxToken; |
365
|
|
|
|
|
|
|
|
366
|
|
|
|
|
|
|
# if there are no sentences remaining with the token, move on to the next token. |
367
|
|
|
|
|
|
|
unless (exists $tokenSentenceIndex{$maxToken}) |
368
|
|
|
|
|
|
|
{ |
369
|
|
|
|
|
|
|
pop @listOfTokens; |
370
|
|
|
|
|
|
|
next; |
371
|
|
|
|
|
|
|
} |
372
|
|
|
|
|
|
|
|
373
|
|
|
|
|
|
|
# get the list of sentences that have the token. |
374
|
|
|
|
|
|
|
my $listOfSentencesWithToken = $tokenSentenceIndex{$maxToken}; |
375
|
|
|
|
|
|
|
|
376
|
|
|
|
|
|
|
# if there are no sentences remaining with the token, move on to the next token. |
377
|
|
|
|
|
|
|
unless (scalar(@$listOfSentencesWithToken) > 0) |
378
|
|
|
|
|
|
|
{ |
379
|
|
|
|
|
|
|
pop @listOfTokens; |
380
|
|
|
|
|
|
|
delete $tokenSentenceIndex{$maxToken}; |
381
|
|
|
|
|
|
|
next; |
382
|
|
|
|
|
|
|
} |
383
|
|
|
|
|
|
|
|
384
|
|
|
|
|
|
|
# find the sentence having the token with the highest weight not yet selected. |
385
|
|
|
|
|
|
|
my $maxSentenceIndex; |
386
|
|
|
|
|
|
|
my $maxSentence; |
387
|
|
|
|
|
|
|
my @remainingListOfSentencesWithToken; |
388
|
|
|
|
|
|
|
foreach my $sentenceIndex (@$listOfSentencesWithToken) |
389
|
|
|
|
|
|
|
{ |
390
|
|
|
|
|
|
|
|
391
|
|
|
|
|
|
|
# get the pointer to the sentence. |
392
|
|
|
|
|
|
|
my $sentence = $localListOfSentences[$sentenceIndex]; |
393
|
|
|
|
|
|
|
|
394
|
|
|
|
|
|
|
# skip the sentence if already selected. |
395
|
|
|
|
|
|
|
next if $sentence->{selected}; |
396
|
|
|
|
|
|
|
|
397
|
|
|
|
|
|
|
# if no sentence has been selected, just take the first valid sentence. |
398
|
|
|
|
|
|
|
unless (defined($maxSentence)) |
399
|
|
|
|
|
|
|
{ |
400
|
|
|
|
|
|
|
$maxSentence = $sentence; |
401
|
|
|
|
|
|
|
$maxSentenceIndex = $sentenceIndex; |
402
|
|
|
|
|
|
|
next; |
403
|
|
|
|
|
|
|
} |
404
|
|
|
|
|
|
|
|
405
|
|
|
|
|
|
|
# choose the sentence with the greater weight, or the greater size, or the lesser id. |
406
|
|
|
|
|
|
|
my $cmp = |
407
|
|
|
|
|
|
|
($sentence->{weight} <=> $maxSentence->{weight}) |
408
|
|
|
|
|
|
|
|| ($sentence->{size} <=> $maxSentence->{size}) |
409
|
|
|
|
|
|
|
|| ($sentence->{id} cmp $maxSentence->{id}); |
410
|
|
|
|
|
|
|
|
411
|
|
|
|
|
|
|
# store the new maximum sentence. |
412
|
|
|
|
|
|
|
if ($cmp == 1) |
413
|
|
|
|
|
|
|
{ |
414
|
|
|
|
|
|
|
|
415
|
|
|
|
|
|
|
# store the previous maximum as an unselected sentence. |
416
|
|
|
|
|
|
|
push @remainingListOfSentencesWithToken, $maxSentenceIndex; |
417
|
|
|
|
|
|
|
$maxSentence = $sentence; |
418
|
|
|
|
|
|
|
$maxSentenceIndex = $sentenceIndex; |
419
|
|
|
|
|
|
|
} |
420
|
|
|
|
|
|
|
else |
421
|
|
|
|
|
|
|
{ |
422
|
|
|
|
|
|
|
|
423
|
|
|
|
|
|
|
# store the current sentence as unselected. |
424
|
|
|
|
|
|
|
push @remainingListOfSentencesWithToken, $sentenceIndex; |
425
|
|
|
|
|
|
|
} |
426
|
|
|
|
|
|
|
} |
427
|
|
|
|
|
|
|
|
428
|
|
|
|
|
|
|
# update the list of sentences with the token that were not selected for the summary. |
429
|
|
|
|
|
|
|
if (@remainingListOfSentencesWithToken == 0) |
430
|
|
|
|
|
|
|
{ |
431
|
|
|
|
|
|
|
delete $tokenSentenceIndex{$maxToken}; |
432
|
|
|
|
|
|
|
} |
433
|
|
|
|
|
|
|
else |
434
|
|
|
|
|
|
|
{ |
435
|
|
|
|
|
|
|
|
436
|
|
|
|
|
|
|
# update the list of sentences that the token is contained in. |
437
|
|
|
|
|
|
|
$tokenSentenceIndex{$maxToken} = \@remainingListOfSentencesWithToken; |
438
|
|
|
|
|
|
|
} |
439
|
|
|
|
|
|
|
|
440
|
|
|
|
|
|
|
# if no sentence selected, then there are no unselected sentences with the |
441
|
|
|
|
|
|
|
# token, so move on to the next token. |
442
|
|
|
|
|
|
|
unless (defined $maxSentence) |
443
|
|
|
|
|
|
|
{ |
444
|
|
|
|
|
|
|
pop @listOfTokens; |
445
|
|
|
|
|
|
|
delete $tokenSentenceIndex{$maxToken}; |
446
|
|
|
|
|
|
|
next; |
447
|
|
|
|
|
|
|
} |
448
|
|
|
|
|
|
|
|
449
|
|
|
|
|
|
|
# store the sentence selected and its weight. |
450
|
|
|
|
|
|
|
$maxSentence->{selected} = 1; |
451
|
|
|
|
|
|
|
push @rankedListOfSentences, [ $maxSentence, $maxSentence->{weight} ]; |
452
|
|
|
|
|
|
|
|
453
|
|
|
|
|
|
|
# update the weight of all the tokens in the max sentence. |
454
|
|
|
|
|
|
|
my @sentenceTokens = keys %{ $maxSentence->{tokenCounts} }; |
455
|
|
|
|
|
|
|
foreach my $token (@sentenceTokens) |
456
|
|
|
|
|
|
|
{ |
457
|
|
|
|
|
|
|
|
458
|
|
|
|
|
|
|
# (currentTokenWeight, initialTokenWeight, token, selectedSentenceId, selectedSentenceWeight) |
459
|
|
|
|
|
|
|
$tokenWeight{$token} = |
460
|
|
|
|
|
|
|
&$tokenWeightUpdateFunction($tokenWeight{$token}, $initialTokenWeight{$token}, $token, $maxSentence->{id}, $maxSentence->{weight}); |
461
|
|
|
|
|
|
|
} |
462
|
|
|
|
|
|
|
|
463
|
|
|
|
|
|
|
# get all of the sentences that share tokens with the max sentence. |
464
|
|
|
|
|
|
|
my %sentencesToUpdate; |
465
|
|
|
|
|
|
|
foreach my $token (@sentenceTokens) |
466
|
|
|
|
|
|
|
{ |
467
|
|
|
|
|
|
|
next unless exists $tokenSentenceIndex{$token}; |
468
|
|
|
|
|
|
|
foreach my $sentenceIndex (@{ $tokenSentenceIndex{$token} }) |
469
|
|
|
|
|
|
|
{ |
470
|
|
|
|
|
|
|
$sentencesToUpdate{$sentenceIndex} = 1; |
471
|
|
|
|
|
|
|
} |
472
|
|
|
|
|
|
|
} |
473
|
|
|
|
|
|
|
my @listOfSentencesToUpdate = keys %sentencesToUpdate; |
474
|
|
|
|
|
|
|
|
475
|
|
|
|
|
|
|
# recompute the weight of the sentences that have tokens whose weight changed. |
476
|
|
|
|
|
|
|
# floating point calculations will become unstable due to rounding errors if the |
477
|
|
|
|
|
|
|
# old weights are subtracted and the new weights added. slower, but best to |
478
|
|
|
|
|
|
|
# recompute the average weights by summing. |
479
|
|
|
|
|
|
|
foreach my $sentenceIndex (@listOfSentencesToUpdate) |
480
|
|
|
|
|
|
|
{ |
481
|
|
|
|
|
|
|
|
482
|
|
|
|
|
|
|
# get the pointer to the sentence. |
483
|
|
|
|
|
|
|
my $sentence = $localListOfSentences[$sentenceIndex]; |
484
|
|
|
|
|
|
|
|
485
|
|
|
|
|
|
|
# skip the sentence if it was already selected. |
486
|
|
|
|
|
|
|
next if $sentence->{selected}; |
487
|
|
|
|
|
|
|
|
488
|
|
|
|
|
|
|
# compute the weight of the sentence. |
489
|
|
|
|
|
|
|
my $weight = 0; |
490
|
|
|
|
|
|
|
while (my ($token, $count) = each %{ $sentence->{tokenCounts} }) |
491
|
|
|
|
|
|
|
{ |
492
|
|
|
|
|
|
|
$weight += $count * $tokenWeight{$token}; |
493
|
|
|
|
|
|
|
} |
494
|
|
|
|
|
|
|
$sentence->{weight} = $weight / $sentence->{size}; |
495
|
|
|
|
|
|
|
} |
496
|
|
|
|
|
|
|
} |
497
|
|
|
|
|
|
|
|
498
|
|
|
|
|
|
|
# normalize the sentence weights so they sum to one. |
499
|
|
|
|
|
|
|
my $totalSentenceWeight = 0; |
500
|
|
|
|
|
|
|
foreach my $sentenceWeight (@rankedListOfSentences) |
501
|
|
|
|
|
|
|
{ |
502
|
|
|
|
|
|
|
$totalSentenceWeight += $sentenceWeight->[1]; |
503
|
|
|
|
|
|
|
} |
504
|
|
|
|
|
|
|
$totalSentenceWeight = 1 if ($totalSentenceWeight == 0); |
505
|
|
|
|
|
|
|
|
506
|
|
|
|
|
|
|
foreach my $sentenceWeight (@rankedListOfSentences) |
507
|
|
|
|
|
|
|
{ |
508
|
|
|
|
|
|
|
|
509
|
|
|
|
|
|
|
# normalize the sentence weight. |
510
|
|
|
|
|
|
|
$sentenceWeight = [ $sentenceWeight->[0]->{id}, $sentenceWeight->[1] / $totalSentenceWeight ]; |
511
|
|
|
|
|
|
|
} |
512
|
|
|
|
|
|
|
|
513
|
|
|
|
|
|
|
# add the empty sentences to the list. |
514
|
|
|
|
|
|
|
push @rankedListOfSentences, map { [ $_->[0], 0 ] } sort { ($a->[1] <=> $b->[1]) || ($a->[0] cmp $b->[0]) } @listOfEmptySentenceIds; |
515
|
|
|
|
|
|
|
|
516
|
|
|
|
|
|
|
# adjust the weights to be descending (a kludge). |
517
|
|
|
|
|
|
|
if (@rankedListOfSentences) |
518
|
|
|
|
|
|
|
{ |
519
|
|
|
|
|
|
|
my $totalSentenceWeight = 0; |
520
|
|
|
|
|
|
|
my $runningSum = 0; |
521
|
|
|
|
|
|
|
for (my $i = @rankedListOfSentences - 1; $i > -1; $i--) |
522
|
|
|
|
|
|
|
{ |
523
|
|
|
|
|
|
|
$runningSum += $rankedListOfSentences[$i]->[1]; |
524
|
|
|
|
|
|
|
$rankedListOfSentences[$i]->[1] = $runningSum; |
525
|
|
|
|
|
|
|
$totalSentenceWeight += $rankedListOfSentences[$i]->[1]; |
526
|
|
|
|
|
|
|
} |
527
|
|
|
|
|
|
|
$totalSentenceWeight = 1 if ($totalSentenceWeight <= 0); |
528
|
|
|
|
|
|
|
foreach my $idWeight (@rankedListOfSentences) |
529
|
|
|
|
|
|
|
{ |
530
|
|
|
|
|
|
|
$idWeight->[1] = abs ($idWeight->[1]/ $totalSentenceWeight); |
531
|
|
|
|
|
|
|
} |
532
|
|
|
|
|
|
|
} |
533
|
|
|
|
|
|
|
|
534
|
|
|
|
|
|
|
return \@rankedListOfSentences; |
535
|
|
|
|
|
|
|
} |
536
|
|
|
|
|
|
|
|
537
|
|
|
|
|
|
|
=head2 C |
538
|
|
|
|
|
|
|
|
539
|
|
|
|
|
|
|
Returns the tokens current weight squared. |
540
|
|
|
|
|
|
|
|
541
|
|
|
|
|
|
|
=cut |
542
|
|
|
|
|
|
|
|
543
|
|
|
|
|
|
|
sub tokenWeightUpdateFunction_Squared # (currentTokenWeight, initialTokenWeight, token, selectedSentenceId, selectedSentenceWeight) |
544
|
|
|
|
|
|
|
{ |
545
|
|
|
|
|
|
|
return $_[0] * $_[0]; |
546
|
|
|
|
|
|
|
} |
547
|
|
|
|
|
|
|
|
548
|
|
|
|
|
|
|
=head2 C |
549
|
|
|
|
|
|
|
|
550
|
|
|
|
|
|
|
Returns the tokens current weight times its intial weight. |
551
|
|
|
|
|
|
|
|
552
|
|
|
|
|
|
|
=cut |
553
|
|
|
|
|
|
|
|
554
|
|
|
|
|
|
|
sub tokenWeightUpdateFunction_Multiplicative # (currentTokenWeight, initialTokenWeight, token, selectedSentenceId, selectedSentenceWeight) |
555
|
|
|
|
|
|
|
{ |
556
|
|
|
|
|
|
|
return $_[0] * $_[1]; |
557
|
|
|
|
|
|
|
} |
558
|
|
|
|
|
|
|
|
559
|
|
|
|
|
|
|
=head2 C |
560
|
|
|
|
|
|
|
|
561
|
|
|
|
|
|
|
Returns the tokens current weight times its the average weight of the tokens in the selected sentence. |
562
|
|
|
|
|
|
|
|
563
|
|
|
|
|
|
|
=cut |
564
|
|
|
|
|
|
|
|
565
|
|
|
|
|
|
|
sub tokenWeightUpdateFunction_Sentence # (currentTokenWeight, initialTokenWeight, token, selectedSentenceId, selectedSentenceWeight) |
566
|
|
|
|
|
|
|
{ |
567
|
|
|
|
|
|
|
return $_[0] * $_[4]; |
568
|
|
|
|
|
|
|
} |
569
|
|
|
|
|
|
|
|
570
|
|
|
|
|
|
|
# computes the textrank of the tokens. |
571
|
|
|
|
|
|
|
sub _getTextRankWeightOfTokens |
572
|
|
|
|
|
|
|
{ |
573
|
|
|
|
|
|
|
my %Parameters = @_; |
574
|
|
|
|
|
|
|
|
575
|
|
|
|
|
|
|
# use any textrank parameters if defined. |
576
|
|
|
|
|
|
|
my %textRankParameters; |
577
|
|
|
|
|
|
|
%textRankParameters = %{ $Parameters{textRankParameters} } if ((exists $Parameters{textRankParameters}) && (defined $Parameters{textRankParameters})); |
578
|
|
|
|
|
|
|
|
579
|
|
|
|
|
|
|
# if no sentences, return now. |
580
|
|
|
|
|
|
|
return {} unless exists $Parameters{listOfSentences}; |
581
|
|
|
|
|
|
|
my $listOfSentences = $Parameters{listOfSentences}; |
582
|
|
|
|
|
|
|
|
583
|
|
|
|
|
|
|
# build the list of tokens. |
584
|
|
|
|
|
|
|
my @listOfTokens = map { ($_->{listOfTokens}) } @$listOfSentences; |
585
|
|
|
|
|
|
|
|
586
|
|
|
|
|
|
|
# return the textrank of each token. |
587
|
|
|
|
|
|
|
return getTextrankOfListOfTokens(%textRankParameters, listOfTokens => \@listOfTokens); |
588
|
|
|
|
|
|
|
} |
589
|
|
|
|
|
|
|
|
590
|
|
|
|
|
|
|
# computes the frequency of the tokens. |
591
|
|
|
|
|
|
|
sub _getFrequencyWeightOfTokens |
592
|
|
|
|
|
|
|
{ |
593
|
|
|
|
|
|
|
my %Parameters = @_; |
594
|
|
|
|
|
|
|
|
595
|
|
|
|
|
|
|
# if no sentences, return now. |
596
|
|
|
|
|
|
|
return {} unless exists $Parameters{listOfSentences}; |
597
|
|
|
|
|
|
|
my $listOfSentences = $Parameters{listOfSentences}; |
598
|
|
|
|
|
|
|
|
599
|
|
|
|
|
|
|
# compute total occurrence and frequency of the tokens. |
600
|
|
|
|
|
|
|
my $totalOccurrence = 0; |
601
|
|
|
|
|
|
|
my %tokenFrequency; |
602
|
|
|
|
|
|
|
|
603
|
|
|
|
|
|
|
foreach my $sentence (@$listOfSentences) |
604
|
|
|
|
|
|
|
{ |
605
|
|
|
|
|
|
|
foreach my $token (@{$sentence->{listOfTokens}}) |
606
|
|
|
|
|
|
|
{ |
607
|
|
|
|
|
|
|
++$tokenFrequency{$token}; |
608
|
|
|
|
|
|
|
++$totalOccurrence; |
609
|
|
|
|
|
|
|
} |
610
|
|
|
|
|
|
|
} |
611
|
|
|
|
|
|
|
$totalOccurrence = 1 if $totalOccurrence < 1; |
612
|
|
|
|
|
|
|
|
613
|
|
|
|
|
|
|
while (my ($token, undef) = each %tokenFrequency) |
614
|
|
|
|
|
|
|
{ |
615
|
|
|
|
|
|
|
$tokenFrequency{$token} /= $totalOccurrence; |
616
|
|
|
|
|
|
|
} |
617
|
|
|
|
|
|
|
|
618
|
|
|
|
|
|
|
# return the frequency of each token. |
619
|
|
|
|
|
|
|
return \%tokenFrequency; |
620
|
|
|
|
|
|
|
} |
621
|
|
|
|
|
|
|
|
622
|
|
|
|
|
|
|
=head1 INSTALLATION |
623
|
|
|
|
|
|
|
|
624
|
|
|
|
|
|
|
Use L to install the module and all its prerequisites: |
625
|
|
|
|
|
|
|
|
626
|
|
|
|
|
|
|
perl -MCPAN -e shell |
627
|
|
|
|
|
|
|
>install Text::Summarize |
628
|
|
|
|
|
|
|
|
629
|
|
|
|
|
|
|
=head1 BUGS |
630
|
|
|
|
|
|
|
|
631
|
|
|
|
|
|
|
Please email bugs reports or feature requests to C, or through |
632
|
|
|
|
|
|
|
the web interface at L. The author |
633
|
|
|
|
|
|
|
will be notified and you can be automatically notified of progress on the bug fix or feature request. |
634
|
|
|
|
|
|
|
|
635
|
|
|
|
|
|
|
=head1 AUTHOR |
636
|
|
|
|
|
|
|
|
637
|
|
|
|
|
|
|
Jeff Kubina |
638
|
|
|
|
|
|
|
|
639
|
|
|
|
|
|
|
=head1 COPYRIGHT |
640
|
|
|
|
|
|
|
|
641
|
|
|
|
|
|
|
Copyright (c) 2009 Jeff Kubina. All rights reserved. |
642
|
|
|
|
|
|
|
This program is free software; you can redistribute |
643
|
|
|
|
|
|
|
it and/or modify it under the same terms as Perl itself. |
644
|
|
|
|
|
|
|
|
645
|
|
|
|
|
|
|
The full text of the license can be found in the |
646
|
|
|
|
|
|
|
LICENSE file included with this module. |
647
|
|
|
|
|
|
|
|
648
|
|
|
|
|
|
|
=head1 KEYWORDS |
649
|
|
|
|
|
|
|
|
650
|
|
|
|
|
|
|
information processing, summary, summaries, summarization, summarize, sumbasic, textrank |
651
|
|
|
|
|
|
|
|
652
|
|
|
|
|
|
|
=head1 SEE ALSO |
653
|
|
|
|
|
|
|
|
654
|
|
|
|
|
|
|
L, L, L |
655
|
|
|
|
|
|
|
|
656
|
|
|
|
|
|
|
=begin html |
657
|
|
|
|
|
|
|
|
658
|
|
|
|
|
|
|
The SumBasic algorithm for ranking sentences is from |
659
|
|
|
|
|
|
|
Beyond SumBasic: Task-Focused Summarization with Sentence Simplification and Lexical Expansion |
660
|
|
|
|
|
|
|
by L. Vanderwendea, H. Suzukia, C. Brocketta, and A. Nenkovab. |
661
|
|
|
|
|
|
|
|
662
|
|
|
|
|
|
|
=end html |
663
|
|
|
|
|
|
|
|
664
|
|
|
|
|
|
|
=cut |
665
|
|
|
|
|
|
|
|
666
|
|
|
|
|
|
|
1; |
667
|
|
|
|
|
|
|
|
668
|
|
|
|
|
|
|
# The preceding line will help the module return a true value |
669
|
|
|
|
|
|
|
|