| line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
|
1
|
|
|
|
|
|
|
package WARC::Record::Logical::Heuristics; # -*- CPerl -*- |
|
2
|
|
|
|
|
|
|
|
|
3
|
2
|
|
|
2
|
|
69458
|
use strict; |
|
|
2
|
|
|
|
|
12
|
|
|
|
2
|
|
|
|
|
53
|
|
|
4
|
2
|
|
|
2
|
|
9
|
use warnings; |
|
|
2
|
|
|
|
|
4
|
|
|
|
2
|
|
|
|
|
74
|
|
|
5
|
|
|
|
|
|
|
|
|
6
|
|
|
|
|
|
|
our @ISA = qw(); |
|
7
|
|
|
|
|
|
|
|
|
8
|
2
|
|
|
2
|
|
415
|
use WARC; *WARC::Record::Logical::Heuristics::VERSION = \$WARC::VERSION; |
|
|
2
|
|
|
|
|
5
|
|
|
|
2
|
|
|
|
|
70
|
|
|
9
|
|
|
|
|
|
|
|
|
10
|
2
|
|
|
2
|
|
10
|
use Carp; |
|
|
2
|
|
|
|
|
4
|
|
|
|
2
|
|
|
|
|
92
|
|
|
11
|
2
|
|
|
2
|
|
10
|
use File::Spec; |
|
|
2
|
|
|
|
|
4
|
|
|
|
2
|
|
|
|
|
4559
|
|
|
12
|
|
|
|
|
|
|
|
|
13
|
|
|
|
|
|
|
=head1 NAME |
|
14
|
|
|
|
|
|
|
|
|
15
|
|
|
|
|
|
|
WARC::Record::Logical::Heuristics - heuristics for locating record segments |
|
16
|
|
|
|
|
|
|
|
|
17
|
|
|
|
|
|
|
=head1 SYNOPSIS |
|
18
|
|
|
|
|
|
|
|
|
19
|
|
|
|
|
|
|
use WARC::Record::Logical::Heuristics; |
|
20
|
|
|
|
|
|
|
|
|
21
|
|
|
|
|
|
|
=head1 DESCRIPTION |
|
22
|
|
|
|
|
|
|
|
|
23
|
|
|
|
|
|
|
This is an internal module that provides functions for locating record |
|
24
|
|
|
|
|
|
|
segments when the needed information is not available from an index. |
|
25
|
|
|
|
|
|
|
|
|
26
|
|
|
|
|
|
|
These mostly assume that IIPC WARC guidelines have been followed, as |
|
27
|
|
|
|
|
|
|
otherwise there simply is no efficient solution. |
|
28
|
|
|
|
|
|
|
|
|
29
|
|
|
|
|
|
|
Implementations vary, however, with some using only an incrementing serial |
|
30
|
|
|
|
|
|
|
number and a constant timestamp from the initiation of the crawl job, while |
|
31
|
|
|
|
|
|
|
the guidelines and specification envision a timestamp reflecting the first |
|
32
|
|
|
|
|
|
|
write to that specific file rather than the start of the crawl. Constant |
|
33
|
|
|
|
|
|
|
timestamps are checked first, since the search is simpler. |
|
34
|
|
|
|
|
|
|
|
|
35
|
|
|
|
|
|
|
=over |
|
36
|
|
|
|
|
|
|
|
|
37
|
|
|
|
|
|
|
=item $WARC::Record::Logical::Heuristics::Patience |
|
38
|
|
|
|
|
|
|
|
|
39
|
|
|
|
|
|
|
This variable sets a threshold used to limit the reach of an unproductive |
|
40
|
|
|
|
|
|
|
search. This module tracks the "effort" expended (I/O performed) during a |
|
41
|
|
|
|
|
|
|
search and abandons the search if the threshold is exceeded. Finding |
|
42
|
|
|
|
|
|
|
results dynamically (and temporarily) increases this threshold during a |
|
43
|
|
|
|
|
|
|
search, such that this really sets how far the search will go between |
|
44
|
|
|
|
|
|
|
results before giving up and concluding that there are no more results. |
|
45
|
|
|
|
|
|
|
|
|
46
|
|
|
|
|
|
|
The search will reach farther if either the WARC files are not compressed, |
|
47
|
|
|
|
|
|
|
or the "sl" GZIP extension documented in L is used. |
|
48
|
|
|
|
|
|
|
Decompressing record data to find the next record is considerable effort |
|
49
|
|
|
|
|
|
|
for larger records, but is not counted for very small records that the |
|
50
|
|
|
|
|
|
|
system is likely to already have cached after the header has been read. |
|
51
|
|
|
|
|
|
|
|
|
52
|
|
|
|
|
|
|
=cut |
|
53
|
|
|
|
|
|
|
|
|
54
|
|
|
|
|
|
|
# These provide a simple mechanism to limit the scope of a search that is |
|
55
|
|
|
|
|
|
|
# not producing results. Both are localized in the top-level calls. |
|
56
|
|
|
|
|
|
|
|
|
57
|
|
|
|
|
|
|
our $Patience = 10000; # How much effort to put into a search? |
|
58
|
|
|
|
|
|
|
our $Effort = 0; # How much have we done so far during this search? |
|
59
|
|
|
|
|
|
|
|
|
60
|
|
|
|
|
|
|
# Most I/O incurs "effort", represented by incrementing $Effort, while |
|
61
|
|
|
|
|
|
|
# partial success (finding an interesting record) increases "patience", |
|
62
|
|
|
|
|
|
|
# represented by incrementing $Patience. The search stops when either |
|
63
|
|
|
|
|
|
|
# there are no more places to look or $Effort exceeds $Patience. |
|
64
|
|
|
|
|
|
|
|
|
65
|
|
|
|
|
|
|
=item %WARC::Record::Logical::Heuristics::Effort |
|
66
|
|
|
|
|
|
|
|
|
67
|
|
|
|
|
|
|
This internal hash indicates how costly certain operations should be |
|
68
|
|
|
|
|
|
|
considered. The keys and their meanings are subject to change at whim, but |
|
69
|
|
|
|
|
|
|
this is available for quick tuning if needed. Generally, the better |
|
70
|
|
|
|
|
|
|
solution is to index your data rather than spend time tuning heuristics. |
|
71
|
|
|
|
|
|
|
|
|
72
|
|
|
|
|
|
|
=cut |
|
73
|
|
|
|
|
|
|
|
|
74
|
|
|
|
|
|
|
our %Effort = |
|
75
|
|
|
|
|
|
|
(# read_record: |
|
76
|
|
|
|
|
|
|
# effort incurred to read a record header, regardless of compression |
|
77
|
|
|
|
|
|
|
read_record => 5, |
|
78
|
|
|
|
|
|
|
|
|
79
|
|
|
|
|
|
|
# gzread_data_per_tick: |
|
80
|
|
|
|
|
|
|
# number of bytes to read while advancing past a compressed record to |
|
81
|
|
|
|
|
|
|
# incur one effort point; effort incurred rounds down, even to zero |
|
82
|
|
|
|
|
|
|
# |
|
83
|
|
|
|
|
|
|
# this value is a shot-in-the-dark estimate that gunzipping 320 KiB is |
|
84
|
|
|
|
|
|
|
# equivalent to the open/seek/read process for loading a record header |
|
85
|
|
|
|
|
|
|
gzread_data_per_tick => 64 * 1024, |
|
86
|
|
|
|
|
|
|
|
|
87
|
|
|
|
|
|
|
# readdir_files_per_tick: |
|
88
|
|
|
|
|
|
|
# number of file names to read and check while scanning a directory for |
|
89
|
|
|
|
|
|
|
# to incur one effort point; effort incurred rounds down, even to zero |
|
90
|
|
|
|
|
|
|
# |
|
91
|
|
|
|
|
|
|
# this value is a shot-in-the-dark estimate that reading/matching 1600 |
|
92
|
|
|
|
|
|
|
# file names is equivalent to loading a record header; this estimate |
|
93
|
|
|
|
|
|
|
# may be high or low depending on the number of axes used in the search |
|
94
|
|
|
|
|
|
|
readdir_files_per_tick => 320, |
|
95
|
|
|
|
|
|
|
); |
|
96
|
|
|
|
|
|
|
|
|
97
|
|
|
|
|
|
|
# Internal functions: |
|
98
|
|
|
|
|
|
|
|
|
99
|
|
|
|
|
|
|
## @axes = _split_digit_spans( $filename ) |
|
100
|
|
|
|
|
|
|
## |
|
101
|
|
|
|
|
|
|
## Extract possible sequence numbers from $filename and return list of |
|
102
|
|
|
|
|
|
|
## array references [PREFIX, NUMBER, SUFFIX] where NUMBER is a field that |
|
103
|
|
|
|
|
|
|
## can be adjusted to find "nearby" files if NUMBER turns out to actually |
|
104
|
|
|
|
|
|
|
## be a sequence number. Finds numerous false matches in normal use, but |
|
105
|
|
|
|
|
|
|
## broad searches cost only time while excessive narrowing causes failure. |
|
106
|
|
|
|
|
|
|
## |
|
107
|
|
|
|
|
|
|
## Does not perform I/O; does not increment $Effort. |
|
108
|
|
|
|
|
|
|
|
|
109
|
|
|
|
|
|
|
sub _split_digit_spans ($) { |
|
110
|
38
|
|
|
38
|
|
6464
|
my $name = shift; |
|
111
|
38
|
|
|
|
|
74
|
my @axes = (); |
|
112
|
|
|
|
|
|
|
|
|
113
|
|
|
|
|
|
|
# Split on zero-width boundaries between digits and non-digits. |
|
114
|
38
|
|
|
|
|
1061
|
my @pieces = split /(?=[0-9])(?<=[^0-9])|(?=[^0-9])(?<=[0-9])/, $name; |
|
115
|
|
|
|
|
|
|
# The @pieces array now contains alternating spans of digits and non-digits. |
|
116
|
|
|
|
|
|
|
|
|
117
|
38
|
|
|
|
|
124
|
for (my $i = 0; $i < @pieces; $i++) { |
|
118
|
524
|
100
|
100
|
|
|
1818
|
next unless ($pieces[$i] =~ /^[0-9]+$/ && length($pieces[$i]) < 9); |
|
119
|
|
|
|
|
|
|
# More than 8 digits is probably not a sequence number and may be |
|
120
|
|
|
|
|
|
|
# beyond the range of an integer anyway. Use indexes instead of |
|
121
|
|
|
|
|
|
|
# heuristics if you need to work with a billion WARC files. |
|
122
|
237
|
|
|
|
|
1339
|
push @axes, [join('', @pieces[0..($i-1)]), |
|
123
|
|
|
|
|
|
|
$pieces[$i], join('', @pieces[($i+1)..$#pieces])]; |
|
124
|
|
|
|
|
|
|
} |
|
125
|
|
|
|
|
|
|
|
|
126
|
38
|
|
|
|
|
200
|
return @axes; |
|
127
|
|
|
|
|
|
|
} |
|
128
|
|
|
|
|
|
|
|
|
129
|
|
|
|
|
|
|
## @found = _find_nearby_files( $direction, @axes ) |
|
130
|
|
|
|
|
|
|
## |
|
131
|
|
|
|
|
|
|
## Locate existing files that appear to be part of a contiguous sequence |
|
132
|
|
|
|
|
|
|
## along an axis in @axes. The $direction argument is either +1 to search |
|
133
|
|
|
|
|
|
|
## for higher numbers or -1 to search for lower numbers. A direction |
|
134
|
|
|
|
|
|
|
## value with a magnitude greater than 1 results in skipping possibilities |
|
135
|
|
|
|
|
|
|
## during the search. |
|
136
|
|
|
|
|
|
|
## |
|
137
|
|
|
|
|
|
|
## Returns a list of array references reflecting the files along each axis |
|
138
|
|
|
|
|
|
|
## from the argument list but omitting axes on which no files were found. |
|
139
|
|
|
|
|
|
|
## |
|
140
|
|
|
|
|
|
|
## Performs only directory lookups, which have highly unpredictable costs |
|
141
|
|
|
|
|
|
|
## and are usually cached by the system; does not increment $Effort. |
|
142
|
|
|
|
|
|
|
|
|
143
|
|
|
|
|
|
|
sub _find_nearby_files ($@) { |
|
144
|
20
|
|
|
20
|
|
3034
|
my $direction = shift; |
|
145
|
20
|
|
|
|
|
44
|
my @found = (); |
|
146
|
|
|
|
|
|
|
|
|
147
|
20
|
|
|
|
|
44
|
foreach my $axis (@_) { |
|
148
|
180
|
|
|
|
|
299
|
my @files = (); |
|
149
|
180
|
|
|
|
|
367
|
my $i = $axis->[1] + $direction; my $file; |
|
|
180
|
|
|
|
|
227
|
|
|
150
|
180
|
|
|
|
|
2559
|
while (-f ($file = join '', ($axis->[0], |
|
151
|
|
|
|
|
|
|
sprintf('%0*d', length $axis->[1], $i), |
|
152
|
|
|
|
|
|
|
$axis->[2]))) |
|
153
|
42
|
|
|
|
|
133
|
{ push @files, $file; $i += $direction } |
|
|
42
|
|
|
|
|
600
|
|
|
154
|
180
|
100
|
|
|
|
655
|
push @found, \@files if scalar @files; |
|
155
|
|
|
|
|
|
|
} |
|
156
|
|
|
|
|
|
|
|
|
157
|
20
|
|
|
|
|
64
|
return @found; |
|
158
|
|
|
|
|
|
|
} |
|
159
|
|
|
|
|
|
|
|
|
160
|
|
|
|
|
|
|
## @found = _scan_directory_for_axes( $dirname, @axes ) |
|
161
|
|
|
|
|
|
|
## |
|
162
|
|
|
|
|
|
|
## Locate existing files that may appear to be part of a sequence along an |
|
163
|
|
|
|
|
|
|
## axis in @axes, using wildcards for long digit spans. |
|
164
|
|
|
|
|
|
|
## |
|
165
|
|
|
|
|
|
|
## The $dirname argument specifies the name of a directory to search and |
|
166
|
|
|
|
|
|
|
## all @axes are interpreted relative to $dirname. This differs from |
|
167
|
|
|
|
|
|
|
## _find_nearby_files where each axis specifies full absolute filenames. |
|
168
|
|
|
|
|
|
|
## For this function, the axes are strictly filenames with no directory. |
|
169
|
|
|
|
|
|
|
## |
|
170
|
|
|
|
|
|
|
## Returns a list of array references reflecting the files along each axis |
|
171
|
|
|
|
|
|
|
## from the argument list but omitting axes on which no files were found. |
|
172
|
|
|
|
|
|
|
## |
|
173
|
|
|
|
|
|
|
## Performs directory reads; increments $Effort to count file names read. |
|
174
|
|
|
|
|
|
|
|
|
175
|
|
|
|
|
|
|
sub _scan_directory_for_axes ($@) { |
|
176
|
16
|
|
|
16
|
|
1217
|
my $dirname = shift; |
|
177
|
16
|
|
|
|
|
25
|
my $read_count = 0; |
|
178
|
|
|
|
|
|
|
|
|
179
|
|
|
|
|
|
|
my @re = map { |
|
180
|
16
|
|
|
|
|
38
|
my $pre = quotemeta $_->[0]; my $post = quotemeta $_->[2]; |
|
|
60
|
|
|
|
|
150
|
|
|
|
60
|
|
|
|
|
108
|
|
|
181
|
60
|
|
|
|
|
164
|
$pre =~ s/(?<=[^0-9])([0-9]{9,})(?=[^0-9])/'[0-9]{'.(length $1).'}'/eg; |
|
|
14
|
|
|
|
|
59
|
|
|
182
|
60
|
|
|
|
|
156
|
$post =~ s/(?<=[^0-9])([0-9]{9,})(?=[^0-9])/'[0-9]{'.(length $1).'}'/eg; |
|
|
2
|
|
|
|
|
16
|
|
|
183
|
60
|
|
|
|
|
92
|
my $midlen = length $_->[1]; qr/^$pre[0-9]{$midlen}$post/ } @_; |
|
|
60
|
|
|
|
|
993
|
|
|
184
|
|
|
|
|
|
|
|
|
185
|
16
|
|
|
|
|
31
|
my $filename; |
|
186
|
16
|
|
|
|
|
22
|
my @found = (); |
|
187
|
16
|
100
|
|
|
|
961
|
opendir my $dir, $dirname or croak "$dirname: $!"; |
|
188
|
15
|
|
|
|
|
567
|
while (defined ($filename = readdir $dir)) { |
|
189
|
435
|
|
|
|
|
853
|
foreach (0 .. $#re) { |
|
190
|
1740
|
100
|
|
|
|
4569
|
if ($filename =~ $re[$_]) |
|
191
|
148
|
|
|
|
|
196
|
{ push @{$found[$_]}, $filename } |
|
|
148
|
|
|
|
|
350
|
|
|
192
|
|
|
|
|
|
|
} |
|
193
|
435
|
|
|
|
|
1123
|
$read_count++; |
|
194
|
|
|
|
|
|
|
} |
|
195
|
15
|
|
|
|
|
235
|
closedir $dir; |
|
196
|
|
|
|
|
|
|
|
|
197
|
15
|
|
|
|
|
108
|
$Effort += int($read_count / $Effort{readdir_files_per_tick}); |
|
198
|
15
|
|
|
|
|
42
|
return grep {scalar @$_} @found; |
|
|
60
|
|
|
|
|
213
|
|
|
199
|
|
|
|
|
|
|
} |
|
200
|
|
|
|
|
|
|
|
|
201
|
|
|
|
|
|
|
## @similar = _find_similar_files( $seed ) |
|
202
|
|
|
|
|
|
|
## |
|
203
|
|
|
|
|
|
|
## Locate existing files that may appear to be part of a sequence involving |
|
204
|
|
|
|
|
|
|
## any digit span in $seed, using wildcards for long digit spans and |
|
205
|
|
|
|
|
|
|
## searching only the directory containing $seed. |
|
206
|
|
|
|
|
|
|
## |
|
207
|
|
|
|
|
|
|
## Returns a list of array references, each containing two array references |
|
208
|
|
|
|
|
|
|
## for files sorting before and after $seed, reflecting the files along |
|
209
|
|
|
|
|
|
|
## each axis derived from $seed on which files other than $seed were found. |
|
210
|
|
|
|
|
|
|
## |
|
211
|
|
|
|
|
|
|
## Uses _scan_directory_for_axes; does not perform I/O directly. |
|
212
|
|
|
|
|
|
|
|
|
213
|
|
|
|
|
|
|
sub _find_similar_files ($) { |
|
214
|
12
|
|
|
12
|
|
6342
|
my $seedfile = shift; |
|
215
|
|
|
|
|
|
|
|
|
216
|
12
|
|
|
|
|
59
|
my $fs_volname; my $dirname; my $filename; |
|
|
12
|
|
|
|
|
0
|
|
|
217
|
12
|
|
|
|
|
140
|
($fs_volname, $dirname, $filename) = File::Spec->splitpath($seedfile); |
|
218
|
|
|
|
|
|
|
|
|
219
|
12
|
|
|
|
|
90
|
my @found = _scan_directory_for_axes |
|
220
|
|
|
|
|
|
|
(File::Spec->catpath($fs_volname, $dirname, ''), |
|
221
|
|
|
|
|
|
|
_split_digit_spans $filename); |
|
222
|
12
|
|
|
|
|
48
|
my @similar = (); |
|
223
|
12
|
|
|
|
|
29
|
foreach my $axis_files (@found) { |
|
224
|
48
|
|
|
|
|
74
|
my @before = (); my @after = (); |
|
|
48
|
|
|
|
|
62
|
|
|
225
|
48
|
|
|
|
|
69
|
foreach my $fname (@$axis_files) { |
|
226
|
114
|
100
|
|
|
|
270
|
if ($fname lt $filename) { |
|
|
|
100
|
|
|
|
|
|
|
227
|
32
|
|
|
|
|
178
|
push @before, File::Spec->catpath($fs_volname, $dirname, $fname); |
|
228
|
|
|
|
|
|
|
} elsif ($fname gt $filename) { |
|
229
|
34
|
|
|
|
|
176
|
push @after, File::Spec->catpath($fs_volname, $dirname, $fname); |
|
230
|
|
|
|
|
|
|
} |
|
231
|
|
|
|
|
|
|
} |
|
232
|
18
|
|
|
|
|
43
|
push @similar, [[sort {$a cmp $b} @before], |
|
233
|
48
|
100
|
|
|
|
182
|
[sort {$a cmp $b} @after]] if @before + @after; |
|
|
28
|
|
|
|
|
52
|
|
|
234
|
|
|
|
|
|
|
} |
|
235
|
|
|
|
|
|
|
|
|
236
|
12
|
|
|
|
|
59
|
return @similar; |
|
237
|
|
|
|
|
|
|
} |
|
238
|
|
|
|
|
|
|
|
|
239
|
|
|
|
|
|
|
## ($checkpoint, @records) = |
|
240
|
|
|
|
|
|
|
## _scan_volume( $volume, $start, $end, [$field, $value]... ) |
|
241
|
|
|
|
|
|
|
## |
|
242
|
|
|
|
|
|
|
## Search $volume for segment records where any $field matches $value |
|
243
|
|
|
|
|
|
|
## starting at offset $start and ending at or after offset $end. If $end |
|
244
|
|
|
|
|
|
|
## is an undefined value, searches until end-of-file. |
|
245
|
|
|
|
|
|
|
## |
|
246
|
|
|
|
|
|
|
## Only returns records that have a 'WARC-Segment-Number' header. |
|
247
|
|
|
|
|
|
|
## |
|
248
|
|
|
|
|
|
|
## The returned $checkpoint is the last record examined, regardless of |
|
249
|
|
|
|
|
|
|
## header values, and provides a valid offset for resuming a search. |
|
250
|
|
|
|
|
|
|
|
|
251
|
|
|
|
|
|
|
sub _scan_volume ($$$@) { |
|
252
|
97
|
|
|
97
|
|
5086
|
my $volume = shift; |
|
253
|
97
|
|
|
|
|
149
|
my $start = shift; |
|
254
|
97
|
|
|
|
|
136
|
my $end = shift; |
|
255
|
|
|
|
|
|
|
|
|
256
|
97
|
|
|
|
|
233
|
my $record = $volume->record_at($start); |
|
257
|
97
|
|
|
|
|
401
|
my @records = (); |
|
258
|
|
|
|
|
|
|
|
|
259
|
97
|
|
100
|
|
|
396
|
while ($record && (!defined $end || $record->offset <= $end)) { |
|
|
|
|
100
|
|
|
|
|
|
260
|
885
|
|
|
|
|
2203
|
$Effort += $Effort{read_record}; |
|
261
|
|
|
|
|
|
|
next unless (defined $record->field('WARC-Segment-Number') |
|
262
|
885
|
100
|
100
|
|
|
1817
|
&& grep {defined $record->field($_->[0])} @_); |
|
|
327
|
|
|
|
|
772
|
|
|
263
|
224
|
100
|
|
|
|
477
|
push @records, $record if grep {defined $record->field($_->[0]) |
|
|
286
|
100
|
|
|
|
633
|
|
|
264
|
|
|
|
|
|
|
&& $record->field($_->[0]) eq $_->[1]} @_; |
|
265
|
|
|
|
|
|
|
} continue { $Effort += int($record->field('Content-Length') |
|
266
|
|
|
|
|
|
|
/ $Effort{gzread_data_per_tick}) |
|
267
|
|
|
|
|
|
|
if (defined $record->{compression} |
|
268
|
885
|
100
|
100
|
|
|
2053
|
&& !defined $record->{sl_packed_size}); |
|
269
|
885
|
|
|
|
|
1783
|
$record = $record->next } |
|
270
|
|
|
|
|
|
|
|
|
271
|
97
|
|
|
|
|
390
|
return $record, @records; |
|
272
|
|
|
|
|
|
|
} |
|
273
|
|
|
|
|
|
|
|
|
274
|
|
|
|
|
|
|
=item ( $first_segment, @clues ) = find_first_segment( $record ) |
|
275
|
|
|
|
|
|
|
|
|
276
|
|
|
|
|
|
|
Attempt to locate the first segment of the logical record suggested by the |
|
277
|
|
|
|
|
|
|
given record without using indexes. Croaks if given a record that does not |
|
278
|
|
|
|
|
|
|
appear to have been written using WARC segmentation. Returns a |
|
279
|
|
|
|
|
|
|
C object for the first record and a list of other objects |
|
280
|
|
|
|
|
|
|
that may be useful for locating continuation records. Returns undef in the |
|
281
|
|
|
|
|
|
|
first slot if no clear first segment was found, but can still return other |
|
282
|
|
|
|
|
|
|
records encountered during the search even if the search was ultimately |
|
283
|
|
|
|
|
|
|
unsuccessful. |
|
284
|
|
|
|
|
|
|
|
|
285
|
|
|
|
|
|
|
=cut |
|
286
|
|
|
|
|
|
|
|
|
287
|
|
|
|
|
|
|
## Each "clue" can be a WARC::Record, or a hint in the form of [key => value]. |
|
288
|
|
|
|
|
|
|
## |
|
289
|
|
|
|
|
|
|
## The hint keys currently are: |
|
290
|
|
|
|
|
|
|
## |
|
291
|
|
|
|
|
|
|
## tail => $record |
|
292
|
|
|
|
|
|
|
## last record examined in initial volume |
|
293
|
|
|
|
|
|
|
## (a good starting point to search for more segments) |
|
294
|
|
|
|
|
|
|
## |
|
295
|
|
|
|
|
|
|
## files_on_axes => [$filename, ...]... |
|
296
|
|
|
|
|
|
|
## array of arrays from _find_nearby_files |
|
297
|
|
|
|
|
|
|
## files_from_dir => [[$filename...], [$filename...]]... |
|
298
|
|
|
|
|
|
|
## array of arrays from _find_similar_files |
|
299
|
|
|
|
|
|
|
## Note that the filenames are set to undef in these hints as the |
|
300
|
|
|
|
|
|
|
## corresponding WARC volumes are scanned, with any relevant records |
|
301
|
|
|
|
|
|
|
## added directly to the clue list as they are found. |
|
302
|
|
|
|
|
|
|
|
|
303
|
|
|
|
|
|
|
sub find_first_segment { |
|
304
|
11
|
|
|
11
|
1
|
899
|
local $Patience = $Patience; |
|
305
|
11
|
|
|
|
|
21
|
local $Effort = 0; |
|
306
|
|
|
|
|
|
|
|
|
307
|
11
|
|
|
|
|
20
|
my $initial = shift; |
|
308
|
|
|
|
|
|
|
|
|
309
|
11
|
100
|
|
|
|
37
|
croak 'searching for segments for unsegmented record' |
|
310
|
|
|
|
|
|
|
unless defined $initial->field('WARC-Segment-Number'); |
|
311
|
|
|
|
|
|
|
|
|
312
|
10
|
|
|
|
|
28
|
my $origin_id = $initial->field('WARC-Segment-Origin-ID'); |
|
313
|
10
|
|
|
|
|
23
|
my @clues = (); my $point; my @records; |
|
|
10
|
|
|
|
|
18
|
|
|
314
|
|
|
|
|
|
|
|
|
315
|
|
|
|
|
|
|
# First we search the volume containing the initial record, since |
|
316
|
|
|
|
|
|
|
# multiple WARC files may have been concatenated together after writing. |
|
317
|
10
|
|
|
|
|
29
|
($point, @records) = _scan_volume $initial->volume, 0, $initial->offset, |
|
318
|
|
|
|
|
|
|
[WARC_Segment_Origin_ID => $origin_id], [WARC_Record_ID => $origin_id]; |
|
319
|
|
|
|
|
|
|
# ... @records will always include $initial ... |
|
320
|
10
|
|
|
|
|
40
|
push @clues, @records, [tail => $point]; |
|
321
|
|
|
|
|
|
|
|
|
322
|
10
|
|
|
|
|
23
|
foreach my $record (@records) { |
|
323
|
10
|
100
|
|
|
|
28
|
return $record, @clues if $record->field('WARC-Record-ID') eq $origin_id; |
|
324
|
|
|
|
|
|
|
} |
|
325
|
7
|
|
|
|
|
18
|
$Patience += $Effort * ((scalar @records) - 1); |
|
326
|
7
|
100
|
|
|
|
27
|
return undef, @clues if $Effort > $Patience; |
|
327
|
|
|
|
|
|
|
|
|
328
|
|
|
|
|
|
|
# If we get this far, the first segment must be in another volume. |
|
329
|
|
|
|
|
|
|
{ |
|
330
|
6
|
|
|
|
|
18
|
my @simple_axes = _split_digit_spans $initial->volume->filename; |
|
331
|
6
|
|
|
|
|
18
|
my @nearby = _find_nearby_files -1, @simple_axes; |
|
332
|
|
|
|
|
|
|
|
|
333
|
|
|
|
|
|
|
# A simple sequence number may be in use; we can check these volumes |
|
334
|
|
|
|
|
|
|
# before reading the directory to handle varying timestamps. |
|
335
|
6
|
100
|
|
|
|
24
|
push @clues, [files_on_axes => @nearby] if scalar @nearby; |
|
336
|
6
|
|
|
|
|
29
|
foreach my $axis_files (reverse @nearby) { |
|
337
|
|
|
|
|
|
|
# Work backwards on the assumption that sequence numbers are nearer |
|
338
|
|
|
|
|
|
|
# to the end of the filename. (Correct for Wget and Wpull.) |
|
339
|
4
|
|
|
|
|
23
|
foreach my $name (@$axis_files) { |
|
340
|
6
|
|
|
|
|
11
|
my $previousEffort = $Effort; |
|
341
|
6
|
|
|
|
|
18
|
my $volume = mount WARC::Volume ($name); |
|
342
|
6
|
|
|
|
|
121
|
(undef, @records) = _scan_volume $volume, 0, undef, |
|
343
|
|
|
|
|
|
|
[WARC_Segment_Origin_ID => $origin_id], |
|
344
|
|
|
|
|
|
|
[WARC_Record_ID => $origin_id]; |
|
345
|
6
|
|
|
|
|
18
|
push @clues, @records; $name = undef; |
|
|
6
|
|
|
|
|
10
|
|
|
346
|
6
|
|
|
|
|
13
|
foreach my $record (@records) { |
|
347
|
3
|
100
|
|
|
|
11
|
return $record, @clues |
|
348
|
|
|
|
|
|
|
if $record->field('WARC-Record-ID') eq $origin_id; |
|
349
|
|
|
|
|
|
|
} |
|
350
|
5
|
|
|
|
|
13
|
$Patience += ($Effort - $previousEffort) * scalar @records; |
|
351
|
5
|
100
|
|
|
|
26
|
return undef, @clues if $Effort > $Patience; |
|
352
|
|
|
|
|
|
|
} |
|
353
|
|
|
|
|
|
|
} |
|
354
|
|
|
|
|
|
|
} |
|
355
|
|
|
|
|
|
|
|
|
356
|
|
|
|
|
|
|
# If we get this far, the first segment is in another volume and multiple |
|
357
|
|
|
|
|
|
|
# numbers must change to find that other volume. Assume that timestamps |
|
358
|
|
|
|
|
|
|
# are in use in the file names, confounding the simple sequence search. |
|
359
|
|
|
|
|
|
|
{ |
|
360
|
6
|
|
|
|
|
14
|
my @nearby = _find_similar_files $initial->volume->filename; |
|
|
4
|
|
|
|
|
10
|
|
|
|
4
|
|
|
|
|
17
|
|
|
361
|
|
|
|
|
|
|
|
|
362
|
4
|
100
|
|
|
|
16
|
push @clues, [files_from_dir => @nearby] if scalar @nearby; |
|
363
|
|
|
|
|
|
|
# Work forwards on the assumption that sequence numbers are nearer to |
|
364
|
|
|
|
|
|
|
# the beginning of the filename. (Correct in Internet Archive samples.) |
|
365
|
4
|
|
|
|
|
14
|
foreach my $fname ((map {reverse @{$_->[0]}} @nearby), |
|
|
8
|
|
|
|
|
11
|
|
|
|
8
|
|
|
|
|
19
|
|
|
366
|
|
|
|
|
|
|
# work backwards within the "before" list on each axis |
|
367
|
|
|
|
|
|
|
# ... and forwards within the "after" list on each axis |
|
368
|
8
|
|
|
|
|
12
|
(map {@{$_->[1]}} @nearby)) { |
|
|
8
|
|
|
|
|
17
|
|
|
369
|
9
|
|
|
|
|
16
|
my $previousEffort = $Effort; |
|
370
|
9
|
|
|
|
|
33
|
my $volume = mount WARC::Volume ($fname); |
|
371
|
9
|
|
|
|
|
167
|
(undef, @records) = _scan_volume $volume, 0, undef, |
|
372
|
|
|
|
|
|
|
[WARC_Segment_Origin_ID => $origin_id], |
|
373
|
|
|
|
|
|
|
[WARC_Record_ID => $origin_id]; |
|
374
|
9
|
|
|
|
|
26
|
push @clues, @records; $fname = undef; |
|
|
9
|
|
|
|
|
17
|
|
|
375
|
9
|
|
|
|
|
21
|
foreach my $record (@records) { |
|
376
|
4
|
100
|
|
|
|
12
|
return $record, @clues |
|
377
|
|
|
|
|
|
|
if $record->field('WARC-Record-ID') eq $origin_id; |
|
378
|
|
|
|
|
|
|
} |
|
379
|
7
|
|
|
|
|
18
|
$Patience += ($Effort - $previousEffort) * scalar @records; |
|
380
|
7
|
100
|
|
|
|
25
|
return undef, @clues if $Effort > $Patience; |
|
381
|
|
|
|
|
|
|
} |
|
382
|
|
|
|
|
|
|
} |
|
383
|
|
|
|
|
|
|
|
|
384
|
|
|
|
|
|
|
# If we get this far, we have run out of places to look and the user will |
|
385
|
|
|
|
|
|
|
# need to build an index instead of relying on heuristics. |
|
386
|
1
|
|
|
|
|
7
|
return undef, @clues; |
|
387
|
|
|
|
|
|
|
} |
|
388
|
|
|
|
|
|
|
|
|
389
|
|
|
|
|
|
|
=item ( @segments ) = find_continuation( $first_segment, @clues ) |
|
390
|
|
|
|
|
|
|
|
|
391
|
|
|
|
|
|
|
Attempt to locate the continuation segments of a logical record without |
|
392
|
|
|
|
|
|
|
using indexes. Uses the clues returned from C to aid |
|
393
|
|
|
|
|
|
|
in the search and returns a list of continuation records found that appear |
|
394
|
|
|
|
|
|
|
to be part of the same logical record as the given first segment. |
|
395
|
|
|
|
|
|
|
|
|
396
|
|
|
|
|
|
|
=cut |
|
397
|
|
|
|
|
|
|
|
|
398
|
|
|
|
|
|
|
sub _add_segments (\$\@\@) { |
|
399
|
78
|
|
|
78
|
|
117
|
my $total_segment_count_ref = shift; |
|
400
|
78
|
|
|
|
|
109
|
my $have_segments_ref = shift; |
|
401
|
78
|
|
|
|
|
100
|
my $new_segments_ref = shift; |
|
402
|
|
|
|
|
|
|
|
|
403
|
78
|
|
|
|
|
177
|
foreach (@$new_segments_ref) { |
|
404
|
38
|
|
|
|
|
87
|
$have_segments_ref->[$_->field('WARC-Segment-Number')]++; |
|
405
|
38
|
100
|
|
|
|
107
|
$$total_segment_count_ref = $_->field('WARC-Segment-Number') |
|
406
|
|
|
|
|
|
|
if defined $_->field('WARC-Segment-Total-Length'); |
|
407
|
|
|
|
|
|
|
} |
|
408
|
|
|
|
|
|
|
} |
|
409
|
|
|
|
|
|
|
sub _have_all_segments_p ($@) { |
|
410
|
86
|
|
|
86
|
|
135
|
my $total_segment_count = shift; |
|
411
|
|
|
|
|
|
|
|
|
412
|
|
|
|
|
|
|
# We cannot have all segments if we have not seen the last segment yet. |
|
413
|
86
|
100
|
|
|
|
380
|
return 0 unless defined $total_segment_count; |
|
414
|
|
|
|
|
|
|
|
|
415
|
|
|
|
|
|
|
# We have seen the last segment, do we have all of the others? |
|
416
|
24
|
100
|
|
|
|
59
|
for (my $i = 2; $i < $total_segment_count; $i++) { return 0 unless $_[$i] } |
|
|
59
|
|
|
|
|
187
|
|
|
417
|
|
|
|
|
|
|
# Start the search at 2 because offsets 0 and 1 are not used here. |
|
418
|
|
|
|
|
|
|
|
|
419
|
9
|
|
|
|
|
83
|
return 1; |
|
420
|
|
|
|
|
|
|
} |
|
421
|
|
|
|
|
|
|
|
|
422
|
|
|
|
|
|
|
sub find_continuation { |
|
423
|
16
|
|
|
16
|
1
|
6355
|
local $Patience = $Patience; |
|
424
|
16
|
|
|
|
|
33
|
local $Effort = 0; |
|
425
|
|
|
|
|
|
|
|
|
426
|
16
|
|
|
|
|
27
|
my $first_segment = shift; my $origin_id = $first_segment->id; |
|
|
16
|
|
|
|
|
51
|
|
|
427
|
|
|
|
|
|
|
|
|
428
|
|
|
|
|
|
|
# First we unpack the clues and check if all segments were found while |
|
429
|
|
|
|
|
|
|
# searching for the first segment. |
|
430
|
16
|
|
|
|
|
37
|
my @segments = (); my @nearby_volume_files = (); |
|
|
16
|
|
|
|
|
26
|
|
|
431
|
16
|
|
|
|
|
27
|
my $have_tail = 0; my $point = undef; |
|
|
16
|
|
|
|
|
26
|
|
|
432
|
16
|
|
|
|
|
24
|
my @similar_volume_files_before = (); my @similar_volume_files_after = (); |
|
|
16
|
|
|
|
|
30
|
|
|
433
|
16
|
|
|
|
|
39
|
foreach my $clue (@_) { |
|
434
|
35
|
100
|
|
|
|
107
|
if (UNIVERSAL::isa($clue, 'WARC::Record')) { |
|
|
|
100
|
|
|
|
|
|
|
435
|
21
|
100
|
|
|
|
58
|
push @segments, $clue unless $clue == $first_segment; |
|
436
|
|
|
|
|
|
|
} elsif (ref $clue eq 'ARRAY') { |
|
437
|
13
|
|
|
|
|
30
|
my $tag = shift @$clue; |
|
438
|
13
|
100
|
|
|
|
40
|
if ($tag eq 'tail') { |
|
|
|
100
|
|
|
|
|
|
|
|
|
100
|
|
|
|
|
|
|
439
|
7
|
|
|
|
|
11
|
$have_tail = 1; |
|
440
|
7
|
|
|
|
|
14
|
$point = shift @$clue; |
|
441
|
|
|
|
|
|
|
} elsif ($tag eq 'files_on_axes') { |
|
442
|
3
|
|
|
|
|
9
|
push @nearby_volume_files, map {[grep defined, @$_]} @$clue; |
|
|
6
|
|
|
|
|
19
|
|
|
443
|
|
|
|
|
|
|
} elsif ($tag eq 'files_from_dir') { |
|
444
|
2
|
|
|
|
|
8
|
foreach (@$clue) { |
|
445
|
6
|
|
|
|
|
8
|
push @similar_volume_files_before, [grep defined, @{$_->[0]}]; |
|
|
6
|
|
|
|
|
19
|
|
|
446
|
6
|
|
|
|
|
10
|
push @similar_volume_files_after, [grep defined, @{$_->[1]}]; |
|
|
6
|
|
|
|
|
17
|
|
|
447
|
|
|
|
|
|
|
} |
|
448
|
1
|
|
|
|
|
25
|
} else { die "unrecognized hint tag: $tag" } |
|
449
|
1
|
|
|
|
|
10
|
} else { die "unrecognized clue" } |
|
450
|
33
|
|
|
|
|
76
|
$clue = undef; |
|
451
|
|
|
|
|
|
|
} |
|
452
|
14
|
|
|
|
|
36
|
@similar_volume_files_before = grep {scalar @$_} @similar_volume_files_before; |
|
|
6
|
|
|
|
|
12
|
|
|
453
|
14
|
|
|
|
|
25
|
@similar_volume_files_after = grep {scalar @$_} @similar_volume_files_after; |
|
|
6
|
|
|
|
|
13
|
|
|
454
|
|
|
|
|
|
|
|
|
455
|
14
|
|
|
|
|
17
|
my @have_segments = (); my $total_segment_count = undef; |
|
|
14
|
|
|
|
|
23
|
|
|
456
|
14
|
|
|
|
|
44
|
_add_segments $total_segment_count, @have_segments, @segments; |
|
457
|
|
|
|
|
|
|
|
|
458
|
14
|
100
|
|
|
|
37
|
return @segments if _have_all_segments_p $total_segment_count, @have_segments; |
|
459
|
|
|
|
|
|
|
|
|
460
|
|
|
|
|
|
|
# If we get to here, at least one segment was not found while searching |
|
461
|
|
|
|
|
|
|
# for the first segment, so we will need to search too. |
|
462
|
12
|
|
|
|
|
24
|
my @records = (); |
|
463
|
|
|
|
|
|
|
|
|
464
|
|
|
|
|
|
|
# Pick up where find_first_segment left off... |
|
465
|
12
|
100
|
|
|
|
35
|
if ($point) { |
|
|
|
100
|
|
|
|
|
|
|
466
|
4
|
|
|
|
|
13
|
(undef, @records) = _scan_volume $point->volume, $point->offset, undef, |
|
467
|
|
|
|
|
|
|
[WARC_Segment_Origin_ID => $origin_id]; |
|
468
|
4
|
|
|
|
|
17
|
_add_segments $total_segment_count, @have_segments, @records; |
|
469
|
4
|
|
|
|
|
9
|
push @segments, @records; |
|
470
|
|
|
|
|
|
|
} elsif (!$have_tail) { |
|
471
|
|
|
|
|
|
|
# The search may have begun with the first segment directly; ensure |
|
472
|
|
|
|
|
|
|
# that we scan the entire volume containing the first segment later. |
|
473
|
7
|
|
|
|
|
23
|
push @nearby_volume_files, [$first_segment->volume->filename]; |
|
474
|
|
|
|
|
|
|
} |
|
475
|
12
|
|
|
|
|
137
|
$Patience += $Effort * scalar @records; |
|
476
|
|
|
|
|
|
|
return @segments |
|
477
|
12
|
100
|
100
|
|
|
26
|
if (_have_all_segments_p $total_segment_count, @have_segments |
|
478
|
|
|
|
|
|
|
or $Effort > $Patience); |
|
479
|
|
|
|
|
|
|
|
|
480
|
|
|
|
|
|
|
# Search for more volumes in a simple sequence... |
|
481
|
|
|
|
|
|
|
{ |
|
482
|
10
|
|
|
|
|
23
|
my @simple_axes = _split_digit_spans $first_segment->volume->filename; |
|
483
|
10
|
|
|
|
|
28
|
my @nearby = _find_nearby_files 1, @simple_axes; |
|
484
|
|
|
|
|
|
|
|
|
485
|
|
|
|
|
|
|
# Were more volumes found in the simple sequence search now or previously? |
|
486
|
10
|
|
|
|
|
33
|
foreach my $axis_files ((reverse @nearby), |
|
487
|
|
|
|
|
|
|
(reverse @nearby_volume_files)) { |
|
488
|
|
|
|
|
|
|
# Work backwards on the assumption that sequence numbers are nearer |
|
489
|
|
|
|
|
|
|
# to the end of the filename. (Correct for Wget and Wpull.) |
|
490
|
17
|
|
|
|
|
34
|
foreach my $name (@$axis_files) { |
|
491
|
27
|
|
|
|
|
46
|
my $previousEffort = $Effort; |
|
492
|
27
|
|
|
|
|
423
|
my $volume = mount WARC::Volume ($name); |
|
493
|
27
|
|
|
|
|
507
|
(undef, @records) = _scan_volume $volume, 0, undef, |
|
494
|
|
|
|
|
|
|
[WARC_Segment_Origin_ID => $origin_id]; |
|
495
|
27
|
|
|
|
|
93
|
_add_segments $total_segment_count, @have_segments, @records; |
|
496
|
27
|
|
|
|
|
50
|
push @segments, @records; |
|
497
|
27
|
|
|
|
|
57
|
$Patience += ($Effort - $previousEffort) * scalar @records; |
|
498
|
|
|
|
|
|
|
return @segments |
|
499
|
27
|
100
|
100
|
|
|
61
|
if (_have_all_segments_p $total_segment_count, @have_segments |
|
500
|
|
|
|
|
|
|
or $Effort > $Patience); |
|
501
|
|
|
|
|
|
|
} |
|
502
|
|
|
|
|
|
|
} |
|
503
|
|
|
|
|
|
|
} |
|
504
|
|
|
|
|
|
|
|
|
505
|
|
|
|
|
|
|
# Search for more volumes by directory scan... |
|
506
|
|
|
|
|
|
|
{ |
|
507
|
10
|
100
|
|
|
|
21
|
unless (@similar_volume_files_before + @similar_volume_files_after) { |
|
|
6
|
|
|
|
|
18
|
|
|
|
6
|
|
|
|
|
18
|
|
|
508
|
|
|
|
|
|
|
# Unlike the simple sequence search, the directory scan finds files |
|
509
|
|
|
|
|
|
|
# in both directions from the starting point on all axes, but it may |
|
510
|
|
|
|
|
|
|
# not have been needed to find the first segment. Do it now if not. |
|
511
|
5
|
|
|
|
|
18
|
my @nearby = _find_similar_files $first_segment->volume->filename; |
|
512
|
5
|
|
|
|
|
11
|
foreach (@nearby) { |
|
513
|
13
|
|
|
|
|
19
|
push @similar_volume_files_before, $_->[0]; |
|
514
|
13
|
|
|
|
|
25
|
push @similar_volume_files_after, $_->[1]; |
|
515
|
|
|
|
|
|
|
} |
|
516
|
|
|
|
|
|
|
} |
|
517
|
|
|
|
|
|
|
# Any interesting records in volumes before the volume containing the |
|
518
|
|
|
|
|
|
|
# initial record were probably found while locating the first segment. |
|
519
|
6
|
|
|
|
|
13
|
foreach my $axis_files (@similar_volume_files_after, |
|
520
|
|
|
|
|
|
|
reverse @similar_volume_files_before) { |
|
521
|
|
|
|
|
|
|
# Work forwards on the assumption that sequence numbers are nearer to |
|
522
|
|
|
|
|
|
|
# the beginning of the filename. (Correct in Internet Archive samples.) |
|
523
|
23
|
|
|
|
|
42
|
foreach my $fname (@$axis_files) { |
|
524
|
33
|
|
|
|
|
52
|
my $previousEffort = $Effort; |
|
525
|
33
|
|
|
|
|
96
|
my $volume = mount WARC::Volume ($fname); |
|
526
|
33
|
|
|
|
|
574
|
(undef, @records) = _scan_volume $volume, 0, undef, |
|
527
|
|
|
|
|
|
|
[WARC_Segment_Origin_ID => $origin_id]; |
|
528
|
33
|
|
|
|
|
116
|
_add_segments $total_segment_count, @have_segments, @records; |
|
529
|
33
|
|
|
|
|
64
|
push @segments, @records; |
|
530
|
33
|
|
|
|
|
64
|
$Patience += ($Effort - $previousEffort) * scalar @records; |
|
531
|
|
|
|
|
|
|
return @segments |
|
532
|
33
|
100
|
100
|
|
|
69
|
if (_have_all_segments_p $total_segment_count, @have_segments |
|
533
|
|
|
|
|
|
|
or $Effort > $Patience); |
|
534
|
|
|
|
|
|
|
} |
|
535
|
|
|
|
|
|
|
} |
|
536
|
|
|
|
|
|
|
} |
|
537
|
|
|
|
|
|
|
|
|
538
|
|
|
|
|
|
|
# If we get to here, we have run out of places to look and the user will |
|
539
|
|
|
|
|
|
|
# need to build an index instead of relying on heuristics. |
|
540
|
2
|
|
|
|
|
16
|
return @segments; |
|
541
|
|
|
|
|
|
|
} |
|
542
|
|
|
|
|
|
|
|
|
543
|
|
|
|
|
|
|
=back |
|
544
|
|
|
|
|
|
|
|
|
545
|
|
|
|
|
|
|
=cut |
|
546
|
|
|
|
|
|
|
|
|
547
|
|
|
|
|
|
|
1; |
|
548
|
|
|
|
|
|
|
__END__ |