| line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
|
1
|
|
|
|
|
|
|
package Algorithm::RabinKarp; |
|
2
|
|
|
|
|
|
|
|
|
3
|
2
|
|
|
2
|
|
74829
|
use warnings; |
|
|
2
|
|
|
|
|
5
|
|
|
|
2
|
|
|
|
|
69
|
|
|
4
|
2
|
|
|
2
|
|
13
|
use strict; |
|
|
2
|
|
|
|
|
5
|
|
|
|
2
|
|
|
|
|
78
|
|
|
5
|
|
|
|
|
|
|
|
|
6
|
2
|
|
|
2
|
|
1448
|
use Algorithm::RabinKarp::Util qw(stream_fh stream_string); |
|
|
2
|
|
|
|
|
6
|
|
|
|
2
|
|
|
|
|
183
|
|
|
7
|
|
|
|
|
|
|
|
|
8
|
2
|
|
|
2
|
|
2180
|
use UNIVERSAL; |
|
|
2
|
|
|
|
|
30
|
|
|
|
2
|
|
|
|
|
11
|
|
|
9
|
|
|
|
|
|
|
|
|
10
|
2
|
|
|
2
|
|
69
|
use constant BASE => 2; |
|
|
2
|
|
|
|
|
3
|
|
|
|
2
|
|
|
|
|
1319
|
|
|
11
|
|
|
|
|
|
|
|
|
12
|
|
|
|
|
|
|
our $VERSION = "0.41"; |
|
13
|
|
|
|
|
|
|
|
|
14
|
|
|
|
|
|
|
=head1 NAME |
|
15
|
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
Algorithm::RabinKarp - Rabin-Karp streaming hash |
|
17
|
|
|
|
|
|
|
|
|
18
|
|
|
|
|
|
|
=head1 SYNOPSIS |
|
19
|
|
|
|
|
|
|
|
|
20
|
|
|
|
|
|
|
my $text = "A do run run run, a do run run"; |
|
21
|
|
|
|
|
|
|
my $kgram = Algorithm::RabinKarp->new($window, $text); |
|
22
|
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
or |
|
24
|
|
|
|
|
|
|
|
|
25
|
|
|
|
|
|
|
my $kgram2 = Algorithm::RabinKarp->new($window, $fh); |
|
26
|
|
|
|
|
|
|
|
|
27
|
|
|
|
|
|
|
or |
|
28
|
|
|
|
|
|
|
my $kgram3 = Algorithm::RabinKarp->new($window, sub { |
|
29
|
|
|
|
|
|
|
... |
|
30
|
|
|
|
|
|
|
return $num, $position; |
|
31
|
|
|
|
|
|
|
}); |
|
32
|
|
|
|
|
|
|
|
|
33
|
|
|
|
|
|
|
my ($hash, $start_position, $end_position) = $kgram->next; |
|
34
|
|
|
|
|
|
|
|
|
35
|
|
|
|
|
|
|
my @values = $kgram->values; |
|
36
|
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
my %occurances; # a dictionary of all kgrams. |
|
38
|
|
|
|
|
|
|
while (my ($hash, @pos) = @{shift @values}) { |
|
39
|
|
|
|
|
|
|
push @{$occurances{$hash}}, \@pos; |
|
40
|
|
|
|
|
|
|
} |
|
41
|
|
|
|
|
|
|
|
|
42
|
|
|
|
|
|
|
my $needle = Algorithm::RabinKarp->new(6, "needle"); |
|
43
|
|
|
|
|
|
|
open my $fh, '<', "haystack.txt"; |
|
44
|
|
|
|
|
|
|
my $haystack = Algorithm::RabinKarp->new(6, $fh); |
|
45
|
|
|
|
|
|
|
my $needle_hash = $needle->next; |
|
46
|
|
|
|
|
|
|
|
|
47
|
|
|
|
|
|
|
while (my ($hay_hash, @pos) = $haystack->next) { |
|
48
|
|
|
|
|
|
|
warn "Possible match for 'needle' at @pos" |
|
49
|
|
|
|
|
|
|
if $needle_hash eq $hay_hash; |
|
50
|
|
|
|
|
|
|
} |
|
51
|
|
|
|
|
|
|
|
|
52
|
|
|
|
|
|
|
|
|
53
|
|
|
|
|
|
|
=head1 DESCRIPTION |
|
54
|
|
|
|
|
|
|
|
|
55
|
|
|
|
|
|
|
This is an implementation of Rabin and Karp's streaming hash, as |
|
56
|
|
|
|
|
|
|
described in "Winnowing: Local Algorithms for Document Fingerprinting" by |
|
57
|
|
|
|
|
|
|
Schleimer, Wilkerson, and Aiken. Following the suggestion of Schleimer, |
|
58
|
|
|
|
|
|
|
I am using their second equation: |
|
59
|
|
|
|
|
|
|
|
|
60
|
|
|
|
|
|
|
$H[ $c[2..$k + 1] ] = (( $H[ $c[1..$k] ] - $c[1] ** $k ) + $c[$k+1] ) * $k |
|
61
|
|
|
|
|
|
|
|
|
62
|
|
|
|
|
|
|
The results of this hash encodes information about the next k values in |
|
63
|
|
|
|
|
|
|
the stream (hense k-gram.) This means for any given stream of length n |
|
64
|
|
|
|
|
|
|
integer values (or characters), you will get back n - k + 1 hash |
|
65
|
|
|
|
|
|
|
values. |
|
66
|
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
For best results, you will want to create a code generator that filters |
|
68
|
|
|
|
|
|
|
your data to remove all unnecessary information. For example, in a large |
|
69
|
|
|
|
|
|
|
english document, you should probably remove all white space, as well |
|
70
|
|
|
|
|
|
|
as removing all capitalization. |
|
71
|
|
|
|
|
|
|
|
|
72
|
|
|
|
|
|
|
=head1 INTENT |
|
73
|
|
|
|
|
|
|
|
|
74
|
|
|
|
|
|
|
By preprocessing your document with the Rabin Karp hashing algorithm, |
|
75
|
|
|
|
|
|
|
it makes it possible to create a "fingerprint" of your document (or documents), |
|
76
|
|
|
|
|
|
|
and then perform multiple searches for fragments contained within your document |
|
77
|
|
|
|
|
|
|
database. |
|
78
|
|
|
|
|
|
|
|
|
79
|
|
|
|
|
|
|
Schleimer, Wilkerson, and Aiken suggest preproccessing to remove |
|
80
|
|
|
|
|
|
|
unnecessary information (like whitespace), as well as known redundent information |
|
81
|
|
|
|
|
|
|
(like, say, copyright notices or other boilerplate that is 'acceptable'.) |
|
82
|
|
|
|
|
|
|
|
|
83
|
|
|
|
|
|
|
They also suggest a post processing pass to reduce data volume, using a technique |
|
84
|
|
|
|
|
|
|
called winnowing (see the link at the end of this documentation.) |
|
85
|
|
|
|
|
|
|
|
|
86
|
|
|
|
|
|
|
=head1 METHODS |
|
87
|
|
|
|
|
|
|
|
|
88
|
|
|
|
|
|
|
=over |
|
89
|
|
|
|
|
|
|
|
|
90
|
|
|
|
|
|
|
=item new($k, [FileHandle|Scalar|Coderef] ) |
|
91
|
|
|
|
|
|
|
|
|
92
|
|
|
|
|
|
|
Creates a new hash generator. If you provide a callback function, it must |
|
93
|
|
|
|
|
|
|
return the next integer value in the stream. Additionally, you may |
|
94
|
|
|
|
|
|
|
return the original position of the value in the stream (ie, you may have been |
|
95
|
|
|
|
|
|
|
filtering characters out because they're redundant.) |
|
96
|
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
=cut |
|
98
|
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
sub new { |
|
100
|
5
|
|
|
5
|
1
|
2276
|
my $class = shift; |
|
101
|
5
|
|
|
|
|
9
|
my $k = shift; |
|
102
|
5
|
|
|
|
|
17
|
my $stream = $class->make_stream(shift); |
|
103
|
4
|
|
|
|
|
8
|
my $rm_k = BASE; |
|
104
|
|
|
|
|
|
|
|
|
105
|
4
|
|
33
|
|
|
49
|
bless { |
|
106
|
|
|
|
|
|
|
k => $k % 32, |
|
107
|
|
|
|
|
|
|
vals => [], |
|
108
|
|
|
|
|
|
|
stream => $stream, |
|
109
|
|
|
|
|
|
|
}, ref $class || $class; |
|
110
|
|
|
|
|
|
|
} |
|
111
|
|
|
|
|
|
|
|
|
112
|
|
|
|
|
|
|
sub make_stream { |
|
113
|
5
|
|
|
5
|
0
|
8
|
my $class = shift; |
|
114
|
5
|
|
|
|
|
8
|
my $source = shift; |
|
115
|
|
|
|
|
|
|
|
|
116
|
5
|
100
|
|
|
|
19
|
return $source if ref $source eq 'CODE'; |
|
117
|
|
|
|
|
|
|
|
|
118
|
4
|
|
|
|
|
7
|
my $stream; |
|
119
|
4
|
100
|
66
|
|
|
33
|
if (defined $source && !ref $source) { |
|
|
|
50
|
33
|
|
|
|
|
|
120
|
3
|
|
|
|
|
17
|
$stream = stream_string($source); |
|
121
|
|
|
|
|
|
|
} elsif (UNIVERSAL::isa($source, "IO::Handle") |
|
122
|
|
|
|
|
|
|
|| UNIVERSAL::isa($source,"GLOB")) { |
|
123
|
0
|
|
|
|
|
0
|
require IO::Handle; |
|
124
|
|
|
|
|
|
|
# The simplest way of getting character position right now. |
|
125
|
0
|
|
|
|
|
0
|
$stream = stream_fh($source); |
|
126
|
|
|
|
|
|
|
} else { |
|
127
|
1
|
|
|
|
|
12
|
die __PACKAGE__." requires its source stream be one of the ". |
|
128
|
|
|
|
|
|
|
"following types: scalar, file handle, coderef, or IO::Handle"; |
|
129
|
|
|
|
|
|
|
} |
|
130
|
3
|
|
|
|
|
8
|
return $stream; |
|
131
|
|
|
|
|
|
|
} |
|
132
|
|
|
|
|
|
|
|
|
133
|
|
|
|
|
|
|
|
|
134
|
|
|
|
|
|
|
=item next() |
|
135
|
|
|
|
|
|
|
|
|
136
|
|
|
|
|
|
|
Returns an array containing (kgram hash value, start position , end position, start, end) |
|
137
|
|
|
|
|
|
|
for every call that can have a hash generated, or () when we have reached the end |
|
138
|
|
|
|
|
|
|
of the stream. |
|
139
|
|
|
|
|
|
|
|
|
140
|
|
|
|
|
|
|
C pulls the first $k from the stream on the first call. Each successive |
|
141
|
|
|
|
|
|
|
call to C has a complexity of O(1). |
|
142
|
|
|
|
|
|
|
|
|
143
|
|
|
|
|
|
|
=cut |
|
144
|
|
|
|
|
|
|
sub next { |
|
145
|
18
|
|
|
18
|
1
|
20
|
my $self = shift; |
|
146
|
|
|
|
|
|
|
|
|
147
|
|
|
|
|
|
|
# assume, for now, that each value is an integer, or can |
|
148
|
|
|
|
|
|
|
# auto cast to char |
|
149
|
18
|
|
|
|
|
21
|
my $values = $self->{vals}; #assume that @values always contains k values |
|
150
|
18
|
|
100
|
|
|
38
|
my $prev = shift @$values || [0, undef]; |
|
151
|
18
|
|
100
|
|
|
38
|
my $hash = $self->{hash} || 0; |
|
152
|
18
|
|
|
|
|
34
|
while (@$values < $self->{k}) { |
|
153
|
22
|
|
|
|
|
51
|
my $nextval = [$self->{stream}->()]; |
|
154
|
22
|
100
|
|
|
|
46
|
return unless @$nextval; |
|
155
|
21
|
|
|
|
|
25
|
push @$values, $nextval; |
|
156
|
21
|
|
|
|
|
19
|
$hash <<= 1; |
|
157
|
21
|
|
|
|
|
27
|
$hash -= $prev->[0] << $self->{k}; |
|
158
|
21
|
|
|
|
|
44
|
$hash += $nextval->[0]; |
|
159
|
|
|
|
|
|
|
|
|
160
|
|
|
|
|
|
|
} |
|
161
|
|
|
|
|
|
|
|
|
162
|
17
|
|
|
|
|
21
|
$self->{hash} = $hash; |
|
163
|
|
|
|
|
|
|
|
|
164
|
17
|
|
|
|
|
49
|
return $hash, $values->[0][1], $values->[-1][1], @{ $values }[0, -1]; |
|
|
17
|
|
|
|
|
67
|
|
|
165
|
|
|
|
|
|
|
} |
|
166
|
|
|
|
|
|
|
|
|
167
|
|
|
|
|
|
|
=item values |
|
168
|
|
|
|
|
|
|
|
|
169
|
|
|
|
|
|
|
Returns an array containing all C hash values contained |
|
170
|
|
|
|
|
|
|
within the data stream, and the positions associated with them (in the same |
|
171
|
|
|
|
|
|
|
format as yielded by L.) |
|
172
|
|
|
|
|
|
|
|
|
173
|
|
|
|
|
|
|
After calling C the stream will be completely exhausted, causing |
|
174
|
|
|
|
|
|
|
subsequent calls to C and C to return C. |
|
175
|
|
|
|
|
|
|
|
|
176
|
|
|
|
|
|
|
NOTE: You should use C if your source stream is infinite, as values |
|
177
|
|
|
|
|
|
|
will greedily attempt to consume all values. |
|
178
|
|
|
|
|
|
|
|
|
179
|
|
|
|
|
|
|
=cut |
|
180
|
|
|
|
|
|
|
|
|
181
|
|
|
|
|
|
|
sub values { |
|
182
|
1
|
|
|
1
|
1
|
5
|
my $self = shift; |
|
183
|
|
|
|
|
|
|
|
|
184
|
1
|
|
|
|
|
2
|
my @values; |
|
185
|
1
|
|
|
|
|
5
|
while (my @next = $self->next()) { |
|
186
|
17
|
|
|
|
|
131
|
push @values, \@next; |
|
187
|
|
|
|
|
|
|
} |
|
188
|
1
|
|
|
|
|
23
|
return @values; |
|
189
|
|
|
|
|
|
|
} |
|
190
|
|
|
|
|
|
|
|
|
191
|
|
|
|
|
|
|
=back |
|
192
|
|
|
|
|
|
|
|
|
193
|
|
|
|
|
|
|
=cut |
|
194
|
|
|
|
|
|
|
|
|
195
|
|
|
|
|
|
|
=head1 BUGS |
|
196
|
|
|
|
|
|
|
|
|
197
|
|
|
|
|
|
|
The current multipliers and modulus lead to very poor hash |
|
198
|
|
|
|
|
|
|
distributions. I'll investigate methods of improving this |
|
199
|
|
|
|
|
|
|
in future versions. |
|
200
|
|
|
|
|
|
|
|
|
201
|
|
|
|
|
|
|
=head1 SEE ALSO |
|
202
|
|
|
|
|
|
|
|
|
203
|
|
|
|
|
|
|
"Winnowing: Local Algorithms for Document Fingerprinting" |
|
204
|
|
|
|
|
|
|
L |
|
205
|
|
|
|
|
|
|
|
|
206
|
|
|
|
|
|
|
Wikipedia: Rabin-Karp string search algorithm |
|
207
|
|
|
|
|
|
|
L |
|
208
|
|
|
|
|
|
|
|
|
209
|
|
|
|
|
|
|
=head1 AUTHOR |
|
210
|
|
|
|
|
|
|
|
|
211
|
|
|
|
|
|
|
Norman Nunley Ennunley@gmail.comE |
|
212
|
|
|
|
|
|
|
Nicholas Clark (Who paired with me) |
|
213
|
|
|
|
|
|
|
|
|
214
|
|
|
|
|
|
|
=cut |
|
215
|
|
|
|
|
|
|
|
|
216
|
|
|
|
|
|
|
1; |