| line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
|
1
|
|
|
|
|
|
|
package Algorithm::NGram; |
|
2
|
1
|
|
|
1
|
|
35568
|
use strict; |
|
|
1
|
|
|
|
|
3
|
|
|
|
1
|
|
|
|
|
41
|
|
|
3
|
1
|
|
|
1
|
|
6
|
use warnings; |
|
|
1
|
|
|
|
|
2
|
|
|
|
1
|
|
|
|
|
32
|
|
|
4
|
|
|
|
|
|
|
|
|
5
|
1
|
|
|
1
|
|
7
|
use Carp qw (croak); |
|
|
1
|
|
|
|
|
8
|
|
|
|
1
|
|
|
|
|
86
|
|
|
6
|
1
|
|
|
1
|
|
994
|
use Class::Accessor::Fast; |
|
|
1
|
|
|
|
|
3633
|
|
|
|
1
|
|
|
|
|
9
|
|
|
7
|
1
|
|
|
1
|
|
31
|
use List::Util qw (shuffle); |
|
|
1
|
|
|
|
|
2
|
|
|
|
1
|
|
|
|
|
131
|
|
|
8
|
1
|
|
|
1
|
|
1151
|
use Storable; |
|
|
1
|
|
|
|
|
3879
|
|
|
|
1
|
|
|
|
|
69
|
|
|
9
|
|
|
|
|
|
|
|
|
10
|
1
|
|
|
1
|
|
7
|
use base qw/Class::Accessor::Fast/; |
|
|
1
|
|
|
|
|
2
|
|
|
|
1
|
|
|
|
|
105
|
|
|
11
|
|
|
|
|
|
|
__PACKAGE__->mk_accessors(qw/ngram_width token_table tokens dirty/); |
|
12
|
|
|
|
|
|
|
|
|
13
|
|
|
|
|
|
|
use constant { |
|
14
|
1
|
|
|
|
|
1402
|
START_TOK => ':STARTTOK:', |
|
15
|
|
|
|
|
|
|
END_TOK => ':ENDTOK:', |
|
16
|
1
|
|
|
1
|
|
5
|
}; |
|
|
1
|
|
|
|
|
2
|
|
|
17
|
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
our $VERSION = '0.9'; |
|
19
|
|
|
|
|
|
|
|
|
20
|
|
|
|
|
|
|
=head1 NAME |
|
21
|
|
|
|
|
|
|
|
|
22
|
|
|
|
|
|
|
Algorithm::NGram |
|
23
|
|
|
|
|
|
|
|
|
24
|
|
|
|
|
|
|
=head1 SYNPOSIS |
|
25
|
|
|
|
|
|
|
|
|
26
|
|
|
|
|
|
|
use Algorithm::NGram; |
|
27
|
|
|
|
|
|
|
my $ng = Algorithm::NGram->new(ngram_width => 3); # use trigrams |
|
28
|
|
|
|
|
|
|
|
|
29
|
|
|
|
|
|
|
# feed in text |
|
30
|
|
|
|
|
|
|
$ng->add_text($text1); # analyze $text1 |
|
31
|
|
|
|
|
|
|
$ng->add_text($text2); # analyze $text2 |
|
32
|
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
# feed in arbitrary sequence of tokens |
|
34
|
|
|
|
|
|
|
$ng->add_start_token; |
|
35
|
|
|
|
|
|
|
$ng->add_tokens(qw/token1 token2 token3/); |
|
36
|
|
|
|
|
|
|
$ng->add_end_token; |
|
37
|
|
|
|
|
|
|
|
|
38
|
|
|
|
|
|
|
my $output = $ng->generate_text; |
|
39
|
|
|
|
|
|
|
|
|
40
|
|
|
|
|
|
|
=head1 DESCRIPTION |
|
41
|
|
|
|
|
|
|
|
|
42
|
|
|
|
|
|
|
This is a module for analyzing token sequences with n-grams. You can |
|
43
|
|
|
|
|
|
|
use it to parse a block of text, or feed in your own tokens. It can |
|
44
|
|
|
|
|
|
|
generate new sequences of tokens from what has been fed in. |
|
45
|
|
|
|
|
|
|
|
|
46
|
|
|
|
|
|
|
=head1 EXPORT |
|
47
|
|
|
|
|
|
|
|
|
48
|
|
|
|
|
|
|
None. |
|
49
|
|
|
|
|
|
|
|
|
50
|
|
|
|
|
|
|
=head1 METHODS |
|
51
|
|
|
|
|
|
|
|
|
52
|
|
|
|
|
|
|
=over 4 |
|
53
|
|
|
|
|
|
|
|
|
54
|
|
|
|
|
|
|
=item new |
|
55
|
|
|
|
|
|
|
|
|
56
|
|
|
|
|
|
|
Create a new n-gram analyzer instance. |
|
57
|
|
|
|
|
|
|
|
|
58
|
|
|
|
|
|
|
B |
|
59
|
|
|
|
|
|
|
|
|
60
|
|
|
|
|
|
|
=over 4 |
|
61
|
|
|
|
|
|
|
|
|
62
|
|
|
|
|
|
|
=item ngram_width |
|
63
|
|
|
|
|
|
|
|
|
64
|
|
|
|
|
|
|
This is the "window size" of how many tokens the analyzer will keep |
|
65
|
|
|
|
|
|
|
track of. A ngram_width of two will make a bigram, a ngram_width of |
|
66
|
|
|
|
|
|
|
three will make a trigram, etc... |
|
67
|
|
|
|
|
|
|
|
|
68
|
|
|
|
|
|
|
=back |
|
69
|
|
|
|
|
|
|
|
|
70
|
|
|
|
|
|
|
=cut |
|
71
|
|
|
|
|
|
|
|
|
72
|
|
|
|
|
|
|
sub new { |
|
73
|
2
|
|
|
2
|
1
|
15
|
my ($class, %opts) = @_; |
|
74
|
|
|
|
|
|
|
|
|
75
|
|
|
|
|
|
|
# trigram by default |
|
76
|
2
|
|
50
|
|
|
7
|
my $ngram_width = delete $opts{ngram_width} || 3; |
|
77
|
|
|
|
|
|
|
|
|
78
|
2
|
|
100
|
|
|
8
|
my $token_table = delete $opts{token_table} || {}; |
|
79
|
2
|
|
100
|
|
|
8
|
my $tokens = delete $opts{tokens} || []; |
|
80
|
|
|
|
|
|
|
|
|
81
|
2
|
|
|
|
|
6
|
my $self = { |
|
82
|
|
|
|
|
|
|
ngram_width => $ngram_width, |
|
83
|
|
|
|
|
|
|
token_table => $token_table, |
|
84
|
|
|
|
|
|
|
tokens => $tokens, |
|
85
|
|
|
|
|
|
|
}; |
|
86
|
|
|
|
|
|
|
|
|
87
|
2
|
|
|
|
|
4
|
bless $self, $class; |
|
88
|
2
|
|
|
|
|
8
|
$self->dirty(1); |
|
89
|
|
|
|
|
|
|
|
|
90
|
2
|
|
|
|
|
20
|
return $self; |
|
91
|
|
|
|
|
|
|
} |
|
92
|
|
|
|
|
|
|
|
|
93
|
|
|
|
|
|
|
=item ngram_width |
|
94
|
|
|
|
|
|
|
|
|
95
|
|
|
|
|
|
|
Returns token window size (e.g. the "n" in n-gram) |
|
96
|
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
=cut |
|
98
|
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
=item token_table |
|
100
|
|
|
|
|
|
|
|
|
101
|
|
|
|
|
|
|
Returns n-gram table |
|
102
|
|
|
|
|
|
|
|
|
103
|
|
|
|
|
|
|
=cut |
|
104
|
|
|
|
|
|
|
|
|
105
|
|
|
|
|
|
|
=item add_text |
|
106
|
|
|
|
|
|
|
|
|
107
|
|
|
|
|
|
|
Splits a block of text up by whitespace and processes each word as a |
|
108
|
|
|
|
|
|
|
token. Automatically calls C at the beginning of |
|
109
|
|
|
|
|
|
|
the text and C at the end. |
|
110
|
|
|
|
|
|
|
|
|
111
|
|
|
|
|
|
|
=cut |
|
112
|
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
# process a block of text, auto-tokenizing it |
|
114
|
|
|
|
|
|
|
sub add_text { |
|
115
|
2
|
|
|
2
|
1
|
12
|
my ($self, $text) = @_; |
|
116
|
|
|
|
|
|
|
|
|
117
|
2
|
|
|
|
|
5
|
$self->add_start_token; |
|
118
|
|
|
|
|
|
|
|
|
119
|
|
|
|
|
|
|
# tokenize text |
|
120
|
2
|
|
|
|
|
17
|
foreach my $tok (split(/ /, $text)) { |
|
121
|
16
|
|
|
|
|
63
|
$tok =~ s/ +//g; # remove spaces |
|
122
|
|
|
|
|
|
|
|
|
123
|
16
|
50
|
|
|
|
22
|
next unless $tok; |
|
124
|
|
|
|
|
|
|
|
|
125
|
16
|
|
|
|
|
26
|
$self->add_token($tok); |
|
126
|
|
|
|
|
|
|
} |
|
127
|
|
|
|
|
|
|
|
|
128
|
2
|
|
|
|
|
15
|
$self->add_end_token; |
|
129
|
|
|
|
|
|
|
} |
|
130
|
|
|
|
|
|
|
|
|
131
|
|
|
|
|
|
|
=item add_tokens |
|
132
|
|
|
|
|
|
|
|
|
133
|
|
|
|
|
|
|
Adds an arbitrary list of tokens. |
|
134
|
|
|
|
|
|
|
|
|
135
|
|
|
|
|
|
|
=cut |
|
136
|
|
|
|
|
|
|
|
|
137
|
|
|
|
|
|
|
*add_token = \&add_tokens; |
|
138
|
|
|
|
|
|
|
sub add_tokens { |
|
139
|
20
|
|
|
20
|
1
|
25
|
my ($self, @tokens) = @_; |
|
140
|
20
|
|
|
|
|
17
|
push @{$self->{tokens}}, @tokens; |
|
|
20
|
|
|
|
|
33
|
|
|
141
|
20
|
|
|
|
|
40
|
$self->dirty(1); |
|
142
|
|
|
|
|
|
|
} |
|
143
|
|
|
|
|
|
|
|
|
144
|
|
|
|
|
|
|
=item add_start_token |
|
145
|
|
|
|
|
|
|
|
|
146
|
|
|
|
|
|
|
Adds the "start token." This is useful because you often will want to |
|
147
|
|
|
|
|
|
|
mark the beginnings and ends of a token sequence so that when |
|
148
|
|
|
|
|
|
|
generating your output the generator will know what tokens start a |
|
149
|
|
|
|
|
|
|
sequence and when to end. |
|
150
|
|
|
|
|
|
|
|
|
151
|
|
|
|
|
|
|
=cut |
|
152
|
|
|
|
|
|
|
|
|
153
|
|
|
|
|
|
|
sub add_start_token { |
|
154
|
2
|
|
|
2
|
1
|
3
|
my ($self) = @_; |
|
155
|
2
|
|
|
|
|
5
|
$self->add_token(START_TOK); |
|
156
|
|
|
|
|
|
|
} |
|
157
|
|
|
|
|
|
|
|
|
158
|
|
|
|
|
|
|
=item add_end_token |
|
159
|
|
|
|
|
|
|
|
|
160
|
|
|
|
|
|
|
Adds the "end token." See C. |
|
161
|
|
|
|
|
|
|
|
|
162
|
|
|
|
|
|
|
=cut |
|
163
|
|
|
|
|
|
|
|
|
164
|
|
|
|
|
|
|
sub add_end_token { |
|
165
|
2
|
|
|
2
|
1
|
2
|
my ($self) = @_; |
|
166
|
2
|
|
|
|
|
4
|
$self->add_token(END_TOK); |
|
167
|
|
|
|
|
|
|
} |
|
168
|
|
|
|
|
|
|
|
|
169
|
|
|
|
|
|
|
=item analyze |
|
170
|
|
|
|
|
|
|
|
|
171
|
|
|
|
|
|
|
Generates an n-gram frequency table. Returns a hashref of |
|
172
|
|
|
|
|
|
|
I<< N => tokens => count >>, where N is the number of tokens (will be from 2 |
|
173
|
|
|
|
|
|
|
to ngram_width). You will not normally need to call this unless you |
|
174
|
|
|
|
|
|
|
want to get the n-gram frequency table. |
|
175
|
|
|
|
|
|
|
|
|
176
|
|
|
|
|
|
|
=cut |
|
177
|
|
|
|
|
|
|
|
|
178
|
|
|
|
|
|
|
sub analyze { |
|
179
|
1
|
|
|
1
|
1
|
19
|
my $self = shift; |
|
180
|
|
|
|
|
|
|
|
|
181
|
1
|
|
|
|
|
3
|
$self->{token_table} = {}; |
|
182
|
|
|
|
|
|
|
|
|
183
|
1
|
|
|
|
|
3
|
my @all_tokens = @{$self->tokens}; |
|
|
1
|
|
|
|
|
4
|
|
|
184
|
|
|
|
|
|
|
|
|
185
|
1
|
|
|
|
|
14
|
for (my $i = 1; $i <= $self->ngram_width; $i++) { |
|
186
|
3
|
|
|
|
|
23
|
for (my $tok_idx = 0; $tok_idx < @all_tokens; $tok_idx++) { |
|
187
|
60
|
|
|
|
|
63
|
my $tok_idx_end = $tok_idx + $i - 1; |
|
188
|
60
|
100
|
|
|
|
95
|
$tok_idx_end = @all_tokens if $tok_idx_end > @all_tokens; |
|
189
|
|
|
|
|
|
|
|
|
190
|
|
|
|
|
|
|
# get tokens |
|
191
|
60
|
|
|
|
|
101
|
my @tokens = @all_tokens[$tok_idx ... $tok_idx_end]; |
|
192
|
|
|
|
|
|
|
|
|
193
|
|
|
|
|
|
|
# get the token that follows this ngram |
|
194
|
60
|
|
|
|
|
75
|
my $next_tok = $all_tokens[$tok_idx_end + 1]; |
|
195
|
60
|
100
|
66
|
|
|
197
|
next unless $next_tok && @tokens; |
|
196
|
|
|
|
|
|
|
|
|
197
|
|
|
|
|
|
|
# don't care about what follows END_TOK |
|
198
|
54
|
100
|
|
|
|
92
|
next if $tokens[0] eq END_TOK; |
|
199
|
|
|
|
|
|
|
|
|
200
|
51
|
|
|
|
|
78
|
my $token_key = $self->token_key(@tokens); |
|
201
|
|
|
|
|
|
|
|
|
202
|
|
|
|
|
|
|
# increment the count of $next_tok after this ngram |
|
203
|
51
|
|
|
|
|
228
|
$self->{token_table}->{$i}->{$token_key}->{$next_tok}++; |
|
204
|
|
|
|
|
|
|
} |
|
205
|
|
|
|
|
|
|
} |
|
206
|
|
|
|
|
|
|
|
|
207
|
1
|
|
|
|
|
6
|
$self->dirty(0); |
|
208
|
|
|
|
|
|
|
|
|
209
|
1
|
|
|
|
|
6
|
return $self->{token_table}; |
|
210
|
|
|
|
|
|
|
} |
|
211
|
|
|
|
|
|
|
|
|
212
|
|
|
|
|
|
|
=item generate_text |
|
213
|
|
|
|
|
|
|
|
|
214
|
|
|
|
|
|
|
After feeding in text tokens, this will return a new block of text |
|
215
|
|
|
|
|
|
|
based on whatever text was added. |
|
216
|
|
|
|
|
|
|
|
|
217
|
|
|
|
|
|
|
=cut |
|
218
|
|
|
|
|
|
|
|
|
219
|
|
|
|
|
|
|
sub generate_text { |
|
220
|
1
|
|
|
1
|
1
|
10
|
my ($self, %opts) = @_; |
|
221
|
|
|
|
|
|
|
|
|
222
|
1
|
|
|
|
|
6
|
my @toks = $self->generate(%opts); |
|
223
|
1
|
|
|
|
|
11
|
return join(' ', @toks); |
|
224
|
|
|
|
|
|
|
} |
|
225
|
|
|
|
|
|
|
|
|
226
|
|
|
|
|
|
|
=item generate |
|
227
|
|
|
|
|
|
|
|
|
228
|
|
|
|
|
|
|
Generates a new sequence of tokens based on whatever tokens have |
|
229
|
|
|
|
|
|
|
previously been fed in. |
|
230
|
|
|
|
|
|
|
|
|
231
|
|
|
|
|
|
|
=cut |
|
232
|
|
|
|
|
|
|
|
|
233
|
|
|
|
|
|
|
sub generate { |
|
234
|
1
|
|
|
1
|
1
|
2
|
my ($self, %opts) = @_; |
|
235
|
|
|
|
|
|
|
|
|
236
|
|
|
|
|
|
|
# update n-gram if things have changed |
|
237
|
1
|
50
|
33
|
|
|
4
|
$self->analyze |
|
238
|
|
|
|
|
|
|
if $self->dirty && ! $opts{no_analyze}; |
|
239
|
|
|
|
|
|
|
|
|
240
|
1
|
|
|
|
|
1
|
my @ret_toks; |
|
241
|
1
|
|
|
|
|
2
|
my $tok = START_TOK; |
|
242
|
|
|
|
|
|
|
|
|
243
|
1
|
|
|
|
|
1
|
my @cur_toks = (); |
|
244
|
|
|
|
|
|
|
|
|
245
|
1
|
|
66
|
|
|
2
|
do { |
|
246
|
9
|
100
|
|
|
|
18
|
push @ret_toks, $tok if $tok ne START_TOK; |
|
247
|
|
|
|
|
|
|
|
|
248
|
9
|
|
|
|
|
15
|
push @cur_toks, $tok; |
|
249
|
9
|
|
|
|
|
19
|
shift @cur_toks while @cur_toks > $self->ngram_width; |
|
250
|
|
|
|
|
|
|
|
|
251
|
9
|
|
|
|
|
68
|
$tok = $self->next_tok(@cur_toks); |
|
252
|
|
|
|
|
|
|
} while $tok && $tok ne END_TOK; |
|
253
|
|
|
|
|
|
|
|
|
254
|
1
|
|
|
|
|
14
|
return @ret_toks; |
|
255
|
|
|
|
|
|
|
} |
|
256
|
|
|
|
|
|
|
|
|
257
|
|
|
|
|
|
|
=item next_tok |
|
258
|
|
|
|
|
|
|
|
|
259
|
|
|
|
|
|
|
Given a list of tokens, will pick a possible token to come next. |
|
260
|
|
|
|
|
|
|
|
|
261
|
|
|
|
|
|
|
=cut |
|
262
|
|
|
|
|
|
|
|
|
263
|
|
|
|
|
|
|
sub next_tok { |
|
264
|
9
|
|
|
9
|
1
|
13
|
my ($self, @toks) = @_; |
|
265
|
|
|
|
|
|
|
|
|
266
|
9
|
50
|
|
|
|
7
|
return undef unless %{$self->token_table}; |
|
|
9
|
|
|
|
|
17
|
|
|
267
|
|
|
|
|
|
|
|
|
268
|
9
|
|
|
|
|
45
|
my $tok_next = $self->token_lookup(@toks); |
|
269
|
9
|
50
|
|
|
|
16
|
croak "No next tokens defined for tokens " . $self->token_key(@toks) |
|
270
|
|
|
|
|
|
|
unless defined $tok_next; |
|
271
|
|
|
|
|
|
|
|
|
272
|
9
|
|
|
|
|
7
|
my @possible_toks; |
|
273
|
|
|
|
|
|
|
|
|
274
|
9
|
|
|
|
|
24
|
while (my ($next_tok, $count) = each %$tok_next) { |
|
275
|
11
|
|
|
|
|
44
|
push @possible_toks, $next_tok for 1 .. $count; |
|
276
|
|
|
|
|
|
|
} |
|
277
|
|
|
|
|
|
|
|
|
278
|
9
|
|
|
|
|
63
|
@possible_toks = shuffle @possible_toks; |
|
279
|
9
|
|
|
|
|
11
|
my $tok = shift @possible_toks; |
|
280
|
|
|
|
|
|
|
|
|
281
|
9
|
|
|
|
|
40
|
return $tok; |
|
282
|
|
|
|
|
|
|
} |
|
283
|
|
|
|
|
|
|
|
|
284
|
|
|
|
|
|
|
=item token_lookup |
|
285
|
|
|
|
|
|
|
|
|
286
|
|
|
|
|
|
|
Returns a hashref of the counts of tokens that follow a sequence of tokens. |
|
287
|
|
|
|
|
|
|
|
|
288
|
|
|
|
|
|
|
=cut |
|
289
|
|
|
|
|
|
|
|
|
290
|
|
|
|
|
|
|
sub token_lookup { |
|
291
|
9
|
|
|
9
|
1
|
14
|
my ($self, @toks) = @_; |
|
292
|
|
|
|
|
|
|
|
|
293
|
9
|
|
|
|
|
7
|
my $tok_count = @toks; |
|
294
|
9
|
50
|
|
|
|
18
|
croak "token_lookup passed more than ngram_width tokens" |
|
295
|
|
|
|
|
|
|
unless $tok_count <= $self->ngram_width; |
|
296
|
|
|
|
|
|
|
|
|
297
|
9
|
|
|
|
|
38
|
my $tok_key = $self->token_key(@toks); |
|
298
|
|
|
|
|
|
|
|
|
299
|
9
|
|
50
|
|
|
29
|
return $self->{token_table}->{$tok_count}->{$tok_key} || undef; |
|
300
|
|
|
|
|
|
|
} |
|
301
|
|
|
|
|
|
|
|
|
302
|
|
|
|
|
|
|
=item token_key |
|
303
|
|
|
|
|
|
|
|
|
304
|
|
|
|
|
|
|
Serializes a sequence of tokens for use as a key into the n-gram |
|
305
|
|
|
|
|
|
|
table. You will not normally need to call this. |
|
306
|
|
|
|
|
|
|
|
|
307
|
|
|
|
|
|
|
=cut |
|
308
|
|
|
|
|
|
|
|
|
309
|
|
|
|
|
|
|
sub token_key { |
|
310
|
60
|
|
|
60
|
1
|
88
|
my ($self, @toks) = @_; |
|
311
|
60
|
|
|
|
|
128
|
return join('-', @toks); |
|
312
|
|
|
|
|
|
|
} |
|
313
|
|
|
|
|
|
|
|
|
314
|
|
|
|
|
|
|
=item serialize |
|
315
|
|
|
|
|
|
|
|
|
316
|
|
|
|
|
|
|
Returns the tokens and n-gram (if one has been generated) in a string |
|
317
|
|
|
|
|
|
|
|
|
318
|
|
|
|
|
|
|
=cut |
|
319
|
|
|
|
|
|
|
|
|
320
|
|
|
|
|
|
|
sub serialize { |
|
321
|
1
|
|
|
1
|
1
|
2
|
my ($self) = @_; |
|
322
|
|
|
|
|
|
|
|
|
323
|
1
|
|
|
|
|
3
|
my $save = { |
|
324
|
|
|
|
|
|
|
ngram_width => $self->ngram_width, |
|
325
|
|
|
|
|
|
|
tokens => $self->tokens, |
|
326
|
|
|
|
|
|
|
ngram => $self->token_table, |
|
327
|
|
|
|
|
|
|
}; |
|
328
|
|
|
|
|
|
|
|
|
329
|
1
|
|
|
|
|
13
|
return Storable::nfreeze($save); |
|
330
|
|
|
|
|
|
|
} |
|
331
|
|
|
|
|
|
|
|
|
332
|
|
|
|
|
|
|
=item deserialize($string) |
|
333
|
|
|
|
|
|
|
|
|
334
|
|
|
|
|
|
|
Deserializes a string and returns an C instance |
|
335
|
|
|
|
|
|
|
|
|
336
|
|
|
|
|
|
|
=cut |
|
337
|
|
|
|
|
|
|
|
|
338
|
|
|
|
|
|
|
sub deserialize { |
|
339
|
1
|
|
|
1
|
1
|
131
|
my ($class, $str) = @_; |
|
340
|
1
|
50
|
|
|
|
3
|
croak "Empty string passed to deserialize" unless $str; |
|
341
|
|
|
|
|
|
|
|
|
342
|
1
|
50
|
|
|
|
4
|
my $save = Storable::thaw($str) |
|
343
|
|
|
|
|
|
|
or croak "Invalid serialized data passed to deserialize"; |
|
344
|
|
|
|
|
|
|
|
|
345
|
1
|
|
50
|
|
|
68
|
my $tokens = $save->{tokens} || []; |
|
346
|
1
|
|
|
|
|
2
|
my $ngram = $save->{ngram}; |
|
347
|
1
|
50
|
|
|
|
4
|
my $width = $save->{ngram_width} or croak "No n-gram width saved"; |
|
348
|
|
|
|
|
|
|
|
|
349
|
1
|
|
|
|
|
4
|
my $instance = $class->new( |
|
350
|
|
|
|
|
|
|
ngram_width => $width, |
|
351
|
|
|
|
|
|
|
token_table => $ngram, |
|
352
|
|
|
|
|
|
|
tokens => $tokens, |
|
353
|
|
|
|
|
|
|
); |
|
354
|
|
|
|
|
|
|
|
|
355
|
1
|
50
|
|
|
|
4
|
$instance->dirty(0) if $ngram; |
|
356
|
|
|
|
|
|
|
|
|
357
|
1
|
|
|
|
|
6
|
return $instance; |
|
358
|
|
|
|
|
|
|
} |
|
359
|
|
|
|
|
|
|
|
|
360
|
|
|
|
|
|
|
1; |
|
361
|
|
|
|
|
|
|
|
|
362
|
|
|
|
|
|
|
__END__ |