File Coverage

File:lib/CheckSpelling/UnknownWordSplitter.pm
Coverage:82.7%

linestmtbrancondsubtimecode
1#! -*-perl-*-
2
3# ~/bin/w
4# Search for potentially misspelled words
5# Output is:
6# misspellled
7# woord (WOORD, Woord, woord, woord's)
8package CheckSpelling::UnknownWordSplitter;
9
10
1
1
1
111537
1
use 5.022;
11
1
1
1
1
1
1
47
use feature 'unicode_strings';
12
1
1
1
1
1
5
7
use strict;
13
1
1
1
1
1
1
20
use warnings;
14
1
1
1
1
1
1
14
no warnings qw(experimental::vlb);
15
1
1
1
1
1
1
1
use utf8;
16
1
1
1
1
10
1
26
use Encode qw/decode_utf8 encode FB_DEFAULT/;
17
1
1
1
1
2
1
25
use File::Basename;
18
1
1
1
1
2
0
15
use Cwd 'abs_path';
19
1
1
1
1
1
1
17
use File::Spec;
20
1
1
1
1
1
1
16
use File::Temp qw/ tempfile tempdir /;
21
1
1
1
1
2
0
26
use File::Path qw/ make_path /;
22
1
1
1
1
306
1
18
use CheckSpelling::Util;
23
1
1
1
1
202
1308
940
use Digest::SHA;
24our $VERSION='0.1.0';
25
26my ($longest_word, $shortest_word, $word_match, $forbidden_re, $patterns_re, $candidates_re, $disable_word_collating, $check_file_names);
27my $homoglyph_re;
28my $begin_block_re = '';
29my @begin_block_list = ();
30my @end_block_list = ();
31my ($ignore_pattern, $upper_pattern, $lower_pattern, $not_lower_pattern, $not_upper_or_lower_pattern, $punctuation_pattern);
32my ($shortest, $longest) = (255, 0);
33my @forbidden_re_list;
34my %forbidden_re_descriptions;
35my @candidates_re_list;
36my $hunspell_dictionary_path;
37my @hunspell_dictionaries;
38my %dictionary = ();
39my $base_dict;
40my %unique;
41my %unique_unrecognized;
42my ($last_file, $words, $unrecognized) = ('', 0, 0);
43my ($ignore_next_line_pattern);
44my ($check_images, $ocr_directory);
45
46my $disable_flags;
47
48sub test_re {
49
32
32
21
  my ($expression) = @_;
50
32
32
17
209
  return eval { qr /$expression/ };
51}
52
53sub quote_re {
54
34
34
29
  my ($expression) = @_;
55
34
50
26
  return $expression if $expression =~ /\?\{/;
56
34
72
  $expression =~ s/
57   \G
58   (
59      (?:[^\\]|\\[^Q])*
60   )
61   (?:
62      \\Q
63      (?:[^\\]|\\[^E])*
64      (?:\\E)?
65   )?
66/
67
68
50
111
   $1 . (defined($2) ? quotemeta($2) : '')
68/xge;
69
34
42
  return $expression;
70}
71
72sub file_to_lists {
73
6
6
3
  my ($re) = @_;
74
6
7
  my @patterns;
75  my %hints;
76
6
0
  my $fh;
77
6
50
58
  if (open($fh, '<:utf8', $re)) {
78
6
8
    local $/=undef;
79
6
43
    my $file=<$fh>;
80
6
20
    close $fh;
81
6
2
    my $line_number = 0;
82
6
5
    my $hint = '';
83
6
22
    for (split /\R/, $file) {
84
32
21
      ++$line_number;
85
32
15
      chomp;
86
32
100
41
      if (/^#(?:\s(.+)|)/) {
87
12
100
67
25
        $hint = $1 if ($hint eq '' && defined $1);
88
12
9
        next;
89      }
90
20
100
21
      $hint = '' unless $_ ne '';
91
20
50
17
      next if $_ eq '$^';
92
20
14
      my $pattern = $_;
93
20
100
49
      next unless s/^(.+)/(?:$1)/;
94
13
15
      my $quoted = quote_re($1);
95
13
100
14
      unless (test_re $quoted) {
96
1
1
        my $error = $@;
97
1
64
        my $home = dirname(__FILE__);
98
1
27
        $error =~ s/$home.*?\.pm line \d+\./$re line $line_number (bad-regex)/;
99
1
14
        print STDERR $error;
100
1
4
        $_ = '(?:\$^ - skipped because bad-regex)';
101
1
1
        $hint = '';
102      }
103
13
100
20
      if (defined $hints{$_}) {
104
1
2
        my $pattern_length = length $pattern;
105
1
2
        my $wrapped = CheckSpelling::Util::wrap_in_backticks($pattern);
106
1
20
        print STDERR "$re:$line_number:1 ... $pattern_length, Warning - duplicate pattern: $wrapped (duplicate-pattern)\n";
107
1
2
        $_ = '(?:\$^ - skipped because duplicate-pattern on $line_number)';
108      } else {
109
12
15
        push @patterns, $_;
110
12
19
        $hints{$_} = $hint;
111      }
112
13
19
      $hint = '';
113    }
114  }
115
116  return {
117
6
26
    patterns => \@patterns,
118    hints => \%hints,
119  };
120}
121
122sub file_to_list {
123
5
5
1318
  my ($re) = @_;
124
5
8
  my $lists = file_to_lists($re);
125
126
5
5
3
18
  return @{$lists->{'patterns'}};
127}
128
129sub list_to_re {
130
5
5
6
  my (@list) = @_;
131
5
11
11
50
1
8
8
  @list = map { my $quoted = quote_re($_); test_re($quoted) ? $quoted : '' } @list;
132
5
11
4
12
  @list = grep { $_ ne '' } @list;
133
5
50
5
  return '$^' unless scalar @list;
134
5
13
  return join "|", (@list);
135}
136
137sub not_empty {
138
107
107
85
  my ($thing) = @_;
139
107
67
506
  return defined $thing && $thing ne '' && $thing =~ /^\d+$/;
140}
141
142sub parse_block_list {
143
3
3
2
  my ($re) = @_;
144
3
3
  my @file;
145
3
50
31
  return @file unless (open(my $file_fh, '<:utf8', $re));
146
147
3
4
  local $/=undef;
148
3
24
  my $file=<$file_fh>;
149
3
4
  my $last_line = $.;
150
3
9
  close $file_fh;
151
3
10
  for (split /\R/, $file) {
152
8
100
11
    next if /^#/;
153
5
3
    chomp;
154
5
4
    s/^\\#/#/;
155
5
50
6
    next unless /^./;
156
5
5
    push @file, $_;
157  }
158
159
3
3
  my $pairs = (0+@file) / 2;
160
3
4
  my $true_pairs = $pairs | 0;
161
3
100
3
  unless ($pairs == $true_pairs) {
162
1
1
    my $early_warnings = CheckSpelling::Util::get_file_from_env('early_warnings', '/dev/null');
163
1
10
    open my $early_warnings_fh, ">>:encoding(UTF-8)", $early_warnings;
164
1
28
    print $early_warnings_fh "$re:$last_line:Block delimiters must come in pairs (uneven-block-delimiters)\n";
165
1
19
    close $early_warnings_fh;
166
1
1
    my $i = 0;
167
1
2
    while ($i < $true_pairs) {
168
0
0
      print STDERR "block-delimiter $i S: $file[$i*2]\n";
169
0
0
      print STDERR "block-delimiter $i E: $file[$i*2+1]\n";
170
0
0
      $i++;
171    }
172
1
11
    print STDERR "block-delimiter unmatched S: `$file[$i*2]`\n";
173
1
2
    @file = ();
174  }
175
176
3
9
  return @file;
177}
178
179sub valid_word {
180  # shortest_word is an absolute
181
28
28
17
  our ($shortest, $longest, $shortest_word, $longest_word);
182
28
50
35
  $shortest = $shortest_word if $shortest_word;
183
28
100
100
27
  if ($longest_word) {
184    # longest_word is an absolute
185
26
21
    $longest = $longest_word;
186  } elsif (not_empty($longest)) {
187    # we allow for some sloppiness (a couple of stuck keys per word)
188    # it's possible that this should scale with word length
189
1
1
    $longest += 2;
190  }
191
28
18
  our ($upper_pattern, $lower_pattern, $punctuation_pattern);
192
28
84
100
31
210
  my $word_pattern = join '|', (grep { defined $_ && /./ } ($upper_pattern, $lower_pattern, $punctuation_pattern));
193
28
100
26
  $word_pattern = q<\\w|'> unless $word_pattern;
194
28
50
50
58
  if ((defined $shortest && not_empty($longest)) &&
195      ($shortest > $longest)) {
196
0
0
    $word_pattern = "(?:$word_pattern){3}";
197
0
0
    return qr/$word_pattern/;
198  }
199
28
100
35
  $shortest = 3 unless defined $shortest;
200
28
100
20
  $longest = '' unless not_empty($longest);
201
28
112
  $word_match = "(?:$word_pattern){$shortest,$longest}";
202
28
226
  return qr/\b$word_match\b/;
203}
204
205sub load_dictionary {
206
15
15
2011
  my ($dict) = @_;
207
15
7
  our ($word_match, $longest, $shortest, $longest_word, $shortest_word, %dictionary);
208
15
13
  $longest_word = CheckSpelling::Util::get_val_from_env('INPUT_LONGEST_WORD', undef);
209
15
14
  $shortest_word = CheckSpelling::Util::get_val_from_env('INPUT_SHORTEST_WORD', 0);
210
15
10
  our ($ignore_pattern, $upper_pattern, $lower_pattern, $not_lower_pattern, $not_upper_or_lower_pattern, $punctuation_pattern);
211
15
13
  $ignore_pattern = CheckSpelling::Util::get_file_from_env_utf8('INPUT_IGNORE_PATTERN', q<[^a-zA-Z']>);
212
15
55
  $upper_pattern = CheckSpelling::Util::get_file_from_env_utf8('INPUT_UPPER_PATTERN', '[A-Z]');
213
15
35
  $lower_pattern = CheckSpelling::Util::get_file_from_env_utf8('INPUT_LOWER_PATTERN', '[a-z]');
214
15
26
  $not_lower_pattern = CheckSpelling::Util::get_file_from_env_utf8('INPUT_NOT_LOWER_PATTERN', '[^a-z]');
215
15
28
  $not_upper_or_lower_pattern = CheckSpelling::Util::get_file_from_env_utf8('INPUT_NOT_UPPER_OR_LOWER_PATTERN', '[^A-Za-z]');
216
15
27
  $punctuation_pattern = CheckSpelling::Util::get_file_from_env_utf8('INPUT_PUNCTUATION_PATTERN', q<'>);
217
15
30
  my $homoglyph_list_path = CheckSpelling::Util::get_file_from_env_utf8('homoglyph_list_path', '/dev/null');
218
15
100
99
  if (-s $homoglyph_list_path) {
219
1
1
1
1
344
1
3384
    use CheckSpelling::Homoglyph;
220
14
21
    CheckSpelling::Homoglyph::init($homoglyph_list_path);
221
14
20
    my $homoglyphs = $CheckSpelling::Homoglyph::homoglyphs;
222    # problematic characters: `\\`, `-`, `]`
223
14
138
    $homoglyphs =~ s/([-\\\]])/\\$1/g;
224
14
36
    $homoglyphs = "[$homoglyphs]";
225
14
80
    qr/$homoglyphs/;
226
14
18
    my $any_char = "(?:$upper_pattern|$lower_pattern)";
227
14
28
    qr/$any_char/;
228
14
15
    my $any_char_or_punctuation = "(?:$upper_pattern|$lower_pattern|$punctuation_pattern)";
229
14
25
    my $homoglyphs_or_any_char_or_punctuation = "$homoglyphs|$any_char_or_punctuation";
230
14
50
18
    my $longest_word_string = defined $longest_word ? $longest_word : '';
231
14
17
    my $homoglyphs_or_any_char_or_punctuation_short_to_long = "(?:$homoglyphs_or_any_char_or_punctuation){$shortest_word,$longest_word_string}";
232
14
77
    my $not_upper_or_lower_pattern_or_end = "(?:$not_upper_or_lower_pattern|$))";
233
14
74
    our $homoglyph_re = "(?=$homoglyphs_or_any_char_or_punctuation_short_to_long$not_upper_or_lower_pattern_or_end)($any_char+$homoglyphs(?:$any_char_or_punctuation|$homoglyphs)*|$homoglyphs+$any_char(?:$any_char_or_punctuation|$homoglyphs)*)";
234  }
235
15
19
  %dictionary = ();
236
237
15
547
  open(my $dict_fh, '<:utf8', $dict);
238
15
67
  while (!eof($dict_fh)) {
239
53
59
    my $word = <$dict_fh>;
240
53
45
    chomp $word;
241
53
100
128
    next unless $word =~ $word_match;
242
50
44
    my $l = length $word;
243
50
100
36
    $longest = -1 unless not_empty($longest);
244
50
100
49
    $longest = $l if $l > $longest;
245
50
100
41
    $shortest = $l if $l < $shortest;
246
50
100
    $dictionary{$word}=1;
247  }
248
15
37
  close $dict_fh;
249
250
15
14
  $word_match = valid_word();
251}
252
253sub hunspell_dictionary {
254
3
3
4
  my ($dict) = @_;
255
3
6
  my $name = $dict;
256
3
7
  $name =~ s{/src/index/hunspell/index\.dic$}{};
257
3
18
  $name =~ s{.*/}{};
258
3
3
  my $aff = $dict;
259
3
3
  my $encoding;
260
3
9
  $aff =~ s/\.dic$/.aff/;
261
3
50
40
  if (open my $aff_fh, '<', $aff) {
262
3
21
    while (<$aff_fh>) {
263
0
0
0
      next unless /^SET\s+(\S+)/;
264
0
0
0
      $encoding = $1 if ($1 !~ /utf-8/i);
265
0
0
      last;
266    }
267
3
57
    close $aff_fh;
268  }
269  return {
270
3
334
    name => $name,
271    dict => $dict,
272    aff => $aff,
273    encoding => $encoding,
274    engine => Text::Hunspell->new($aff, $dict),
275  }
276}
277
278sub init {
279
12
12
15715
  my ($configuration) = @_;
280
12
9
  our ($word_match, %unique, $patterns_re, @forbidden_re_list, $forbidden_re, @candidates_re_list, $candidates_re);
281
12
9
  our ($begin_block_re, @begin_block_list, @end_block_list);
282
12
26
  our $sandbox = CheckSpelling::Util::get_file_from_env('sandbox', '');
283
12
13
  our $hunspell_dictionary_path = CheckSpelling::Util::get_file_from_env('hunspell_dictionary_path', '');
284
12
15
  our $timeout = CheckSpelling::Util::get_val_from_env('splitter_timeout', 30);
285
12
10
  our %forbidden_re_descriptions;
286
12
100
13
  if ($hunspell_dictionary_path) {
287
3
42
    our @hunspell_dictionaries = ();
288
1
1
1
1
1
1
1
1
1
3
50
1
1
1
290
1106
21
6
2
14
9
2
17
187
    if (eval 'use Text::Hunspell; 1') {
289
3
131
      my @hunspell_dictionaries_list = glob("$hunspell_dictionary_path/*.dic");
290
3
8
      for my $hunspell_dictionary_file (@hunspell_dictionaries_list) {
291
3
5
        push @hunspell_dictionaries, hunspell_dictionary($hunspell_dictionary_file);
292      }
293    } else {
294
0
0
      print STDERR "Could not load Text::Hunspell for dictionaries (hunspell-unavailable)\n";
295    }
296  }
297
298
12
100
70
  if (-e "$configuration/block-delimiters.list") {
299
3
4
    my @block_delimiters = parse_block_list "$configuration/block-delimiters.list";
300
3
100
4
    if (@block_delimiters) {
301
2
2
      @begin_block_list = ();
302
2
1
      @end_block_list = ();
303
304
2
2
      while (@block_delimiters) {
305
2
3
        my ($begin, $end) = splice @block_delimiters, 0, 2;
306
2
2
        push @begin_block_list, $begin;
307
2
3
        push @end_block_list, $end;
308      }
309
310
2
2
1
3
      $begin_block_re = join '|', (map { '('.quote_re("\Q$_\E").')' } @begin_block_list);
311    }
312  }
313
314
12
20
  my (@patterns_re_list, %in_patterns_re_list);
315
12
50
48
  if (-e "$configuration/patterns.txt") {
316
0
0
    @patterns_re_list = file_to_list "$configuration/patterns.txt";
317
0
0
    $patterns_re = list_to_re @patterns_re_list;
318
0
0
0
0
    %in_patterns_re_list = map {$_ => 1} @patterns_re_list;
319  } else {
320
12
10
    $patterns_re = undef;
321  }
322
323
12
100
38
  if (-e "$configuration/forbidden.txt") {
324
1
2
    my $forbidden_re_info = file_to_lists "$configuration/forbidden.txt";
325
1
1
1
2
    @forbidden_re_list = @{$forbidden_re_info->{'patterns'}};
326
1
1
1
3
    %forbidden_re_descriptions = %{$forbidden_re_info->{'hints'}};
327
1
1
    $forbidden_re = list_to_re @forbidden_re_list;
328  } else {
329
11
12
    $forbidden_re = undef;
330  }
331
332
12
100
46
  if (-e "$configuration/candidates.txt") {
333
4
6
    @candidates_re_list = file_to_list "$configuration/candidates.txt";
334
4
8
8
50
33
5
8
17
    @candidates_re_list = map { my $quoted = quote_re($_); $in_patterns_re_list{$_} || !test_re($quoted) ? '' : $quoted } @candidates_re_list;
335
4
5
    $candidates_re = list_to_re @candidates_re_list;
336  } else {
337
8
6
    $candidates_re = undef;
338  }
339
340
12
23
  our $largest_file = CheckSpelling::Util::get_val_from_env('INPUT_LARGEST_FILE', 1024*1024);
341
342
12
12
  my $disable_flags = CheckSpelling::Util::get_file_from_env('INPUT_DISABLE_CHECKS', '');
343
12
10
  our $disable_word_collating = $disable_flags =~ /(?:^|,|\s)word-collating(?:,|\s|$)/;
344
12
6
  our $disable_minified_file = $disable_flags =~ /(?:^|,|\s)minified-file(?:,|\s|$)/;
345
12
9
  our $disable_single_line_file = $disable_flags =~ /(?:^|,|\s)single-line-file(?:,|\s|$)/;
346
347
12
10
  our $ignore_next_line_pattern = CheckSpelling::Util::get_file_from_env('INPUT_IGNORE_NEXT_LINE', '');
348
12
10
  $ignore_next_line_pattern =~ s/\s+/|/g;
349
350
12
9
  our $check_images = CheckSpelling::Util::get_val_from_env('INPUT_CHECK_IMAGES', '');
351
12
6
  $check_images = $check_images =~ /^(?:1|true)$/i;
352
12
50
9
  if ($check_images) {
353
0
0
    our $ocr_directory = CheckSpelling::Util::get_file_from_env('ocr_directory', '/tmp/ocr');
354
0
0
0
    $ocr_directory = $1 if ($ocr_directory =~ /^(.*)$/);
355  }
356
357
12
9
  our $check_file_names = CheckSpelling::Util::get_file_from_env('check_file_names', '');
358
359
12
9
  our $use_magic_file = CheckSpelling::Util::get_val_from_env('INPUT_USE_MAGIC_FILE', '');
360
361
12
17
  $word_match = valid_word();
362
363
12
24
  our $base_dict = CheckSpelling::Util::get_file_from_env('dict', "$configuration/words");
364
12
100
58
  $base_dict = '/usr/share/dict/words' unless -e $base_dict;
365
12
16
  load_dictionary($base_dict);
366}
367
368sub split_line {
369
1161
1161
514
  our (%dictionary, $word_match, $disable_word_collating);
370
1161
522
  our ($ignore_pattern, $upper_pattern, $lower_pattern, $not_lower_pattern, $not_upper_or_lower_pattern, $punctuation_pattern);
371
1161
481
  our @hunspell_dictionaries;
372
1161
445
  our $shortest;
373
1161
882
  my $shortest_threshold = $shortest + 2;
374
1161
624
  my $pattern = '.';
375  # $pattern = "(?:$upper_pattern){$shortest,}|$upper_pattern(?:$lower_pattern){2,}\n";
376
377  # https://www.fileformat.info/info/unicode/char/2019/
378
1161
646
  my $rsqm = "\xE2\x80\x99";
379
380
1161
640
  my ($words, $unrecognized) = (0, 0);
381
1161
850
  my ($line, $unique_ref, $unique_unrecognized_ref, $unrecognized_line_items_ref) = @_;
382
1161
5425
    $line =~ s/(?:$rsqm|&apos;|&#39;|\%27|&#8217;|&#x2019;|&rsquo;|\\u2019|\x{2019}|')+/'/g;
383
1161
2253
    $line =~ s/(?:$ignore_pattern)+/ /g;
384
1161
1736
    while ($line =~ s/($upper_pattern{2,})($upper_pattern$lower_pattern{2,})/ $1 $2 /g) {}
385
1161
3324
    while ($line =~ s/((?:$lower_pattern|$punctuation_pattern)+)($upper_pattern)/$1 $2/g) {}
386
1161
1411
    for my $token (split /\s+/, $line) {
387
3645
100
3421
      next unless $token =~ /$pattern/;
388
2485
1901
      $token =~ s/^(?:'|$rsqm)+//g;
389
2485
2477
      $token =~ s/(?:'|$rsqm)+s?$//g;
390
2485
1692
      my $raw_token = $token;
391
2485
1383
      $token =~ s/^[^Ii]?'+(.*)/$1/;
392
2485
1353
      $token =~ s/(.*?)'+$/$1/;
393
2485
100
3858
      next unless $token =~ $word_match;
394
2318
100
2239
      if (defined $dictionary{$token}) {
395
1038
440
        ++$words;
396
1038
632
        $unique_ref->{$token}=1;
397
1038
854
        next;
398      }
399
1280
100
1027
      if (@hunspell_dictionaries) {
400
1254
658
        my $found = 0;
401
1254
769
        for my $hunspell_dictionary (@hunspell_dictionaries) {
402          my $token_encoded = defined $hunspell_dictionary->{'encoding'} ?
403
1254
50
1135
            encode($hunspell_dictionary->{'encoding'}, $token) : $token;
404
1254
50
2986
          next unless ($hunspell_dictionary->{'engine'}->check($token_encoded));
405
0
0
          ++$words;
406
0
0
          $dictionary{$token} = 1;
407
0
0
          $unique_ref->{$token}=1;
408
0
0
          $found = 1;
409
0
0
          last;
410        }
411
1254
50
1019
        next if $found;
412      }
413
1280
1034
      my $key = lc $token;
414
1280
100
1121
      if (defined $dictionary{$key}) {
415
6
4
        ++$words;
416
6
6
        $unique_ref->{$key}=1;
417
6
8
        next;
418      }
419
1274
50
929
      unless ($disable_word_collating) {
420
1274
689
        $key =~ s/''+/'/g;
421
1274
100
1331
        $key =~ s/'[sd]$// unless length $key >= $shortest_threshold;
422      }
423
1274
50
1127
      if (defined $dictionary{$key}) {
424
0
0
        ++$words;
425
0
0
        $unique_ref->{$key}=1;
426
0
0
        next;
427      }
428
1274
727
      ++$unrecognized;
429
1274
796
      $unique_unrecognized_ref->{$raw_token}=1;
430
1274
1670
      $unrecognized_line_items_ref->{$raw_token}=1;
431    }
432
1161
1694
    return ($words, $unrecognized);
433}
434
435sub skip_file {
436
7
7
27
  my ($temp_dir, $reason) = @_;
437
7
227
  open(my $skipped_fh, '>:utf8', "$temp_dir/skipped");
438
7
35
  print $skipped_fh $reason;
439
7
120
  close $skipped_fh;
440}
441
442sub maybe_ocr_file {
443
0
0
0
  my ($file) = @_;
444
0
0
  our $ocr_directory;
445
0
0
  my $ocr_file = "$ocr_directory/$file";
446
0
0
  $ocr_file =~ /^(.*)$/;
447
0
0
  $ocr_file = $1;
448
0
0
  my $ocr_source_sha = "$ocr_file.sha1";
449
0
0
  $ocr_file = "$ocr_file.txt";
450
0
0
  my $sha = Digest::SHA->new(1)->addfile($file, 'b')->hexdigest;
451
0
0
0
0
  if (-e $ocr_file &&
452      -e $ocr_source_sha &&
453      open my $source_sha, '<', $ocr_source_sha) {
454
0
0
    my $last_sha = <$source_sha>;
455
0
0
    close $source_sha;
456
0
0
0
    if ($last_sha =~ /(.*)/) {
457
0
0
0
      return ($ocr_file, 1) if ($1 eq $sha);
458    }
459  }
460
0
0
  my $tesseract = dirname(dirname(dirname(__FILE__)))."/wrappers/run-tesseract";
461
0
0
  $ENV{'input'} = $file;
462
0
0
  my $text_file = `"$tesseract"`;
463
0
0
  delete $ENV{'input'};
464
0
0
0
  return ($file, 0) unless defined $text_file;
465
0
0
  my $file_converted = 0;
466
0
0
  chomp $text_file;
467
0
0
0
  if ($text_file =~ /^(.*)$/) {
468
0
0
    $text_file = $1;
469
0
0
    my $file_size = -s $text_file;
470
0
0
0
    if ($file_size > 20) {
471
0
0
      $file_converted = 1;
472
0
0
      make_path(dirname($ocr_source_sha));
473
0
0
      open my $source_sha, '>', $ocr_source_sha;
474
0
0
      print $source_sha $sha;
475
0
0
      close $source_sha;
476
0
0
      rename($text_file, $ocr_file);
477
0
0
      $file = $ocr_file;
478    } else {
479
0
0
      unlink($text_file);
480    }
481  }
482
0
0
  return ($file, $file_converted);
483}
484
485sub split_file {
486
18
18
12308
  my ($file) = @_;
487  our (
488
18
17
    $unrecognized, $shortest, $largest_file, $words,
489    $word_match, %unique, %unique_unrecognized, $forbidden_re,
490    @forbidden_re_list, $patterns_re, %dictionary,
491    $begin_block_re, @begin_block_list, @end_block_list,
492    $candidates_re, @candidates_re_list, $check_file_names, $use_magic_file, $disable_minified_file,
493    $disable_single_line_file,
494    $ignore_next_line_pattern,
495    $sandbox,
496    $check_images,
497  );
498
18
100
35
  $ignore_next_line_pattern = '$^' unless $ignore_next_line_pattern =~ /./;
499
500
18
12
  our %forbidden_re_descriptions;
501
18
7
  our ($ignore_pattern, $upper_pattern, $lower_pattern, $not_lower_pattern, $not_upper_or_lower_pattern, $punctuation_pattern);
502
503  # https://www.fileformat.info/info/unicode/char/2019/
504
18
13
  my $rsqm = "\xE2\x80\x99";
505
506
18
22
  my @candidates_re_hits = (0) x scalar @candidates_re_list;
507
18
15
  my @candidates_re_lines = (0) x scalar @candidates_re_list;
508
18
14
  my @forbidden_re_hits = (0) x scalar @forbidden_re_list;
509
18
30
  my @forbidden_re_lines = (0) x scalar @forbidden_re_list;
510
18
42
  my $temp_dir = tempdir(DIR=>$sandbox);
511
18
50
3048
  print STDERR "checking file: $file\n" if defined $ENV{'DEBUG'};
512
18
460
  open(my $name_fh, '>', "$temp_dir/name");
513
18
40
    print $name_fh $file;
514
18
250
  close $name_fh;
515
18
100
67
211
  if (defined readlink($file) &&
516      rindex(File::Spec->abs2rel(abs_path($file)), '../', 0) == 0) {
517
1
2
    skip_file($temp_dir, "file only has a single line (out-of-bounds-symbolic-link)\n");
518
1
4
    return $temp_dir;
519  }
520
17
100
50
38
  if ($use_magic_file) {
521
8
50
13196
    if (open(my $file_fh, '-|',
522              '/usr/bin/file',
523              '-b',
524              '--mime',
525              '-e', 'cdf',
526              '-e', 'compress',
527              '-e', 'csv',
528              '-e', 'elf',
529              '-e', 'json',
530              '-e', 'tar',
531              $file)) {
532
8
31073
      my $file_kind = <$file_fh>;
533
8
5014
      close $file_fh;
534
8
18
      my $file_converted = 0;
535
8
50
33
26
      if ($check_images && $file_kind =~ m<^image/>) {
536
0
0
        ($file, $file_converted) = maybe_ocr_file($file);
537      }
538
8
100
67
199
      if ($file_converted == 0 && $file_kind =~ /^(.*?); charset=binary/) {
539
2
35
        skip_file($temp_dir, "it appears to be a binary file (`$1`) (binary-file)\n");
540
2
47
        return $temp_dir;
541      }
542    }
543  } elsif ($file =~ /\.(?:png|jpe?g|gif)$/) {
544
0
0
    my $file_converted = 0;
545
0
0
    ($file, $file_converted) = maybe_ocr_file($file);
546  }
547
15
102
  my $file_size = -s $file;
548
15
50
23
  if (defined $largest_file) {
549
15
50
17
    unless ($check_file_names eq $file) {
550
15
100
19
      if ($file_size > $largest_file) {
551
1
3
        skip_file($temp_dir, "size `$file_size` exceeds limit `$largest_file` (large-file)\n");
552
1
3
        return $temp_dir;
553      }
554    }
555  }
556
14
165
  open my $file_fh, '<', $file;
557
14
13
  binmode $file_fh;
558
14
10
  my $head;
559
14
130
  read($file_fh, $head, 4096);
560
14
822
  $head =~ s/(?:\r|\n)+$//;
561
14
59
  my $dos_new_lines = () = $head =~ /\r\n/gi;
562
14
35
  my $unix_new_lines = () = $head =~ /\n/gi;
563
14
117
  my $mac_new_lines = () = $head =~ /\r/gi;
564
14
58
  local $/;
565
14
100
100
100
100
100
85
  if ($unix_new_lines == 0 && $mac_new_lines == 0) {
566
3
6
    $/ = "\n";
567  } elsif ($dos_new_lines >= $unix_new_lines && $dos_new_lines >= $mac_new_lines) {
568
1
7
    $/ = "\r\n";
569  } elsif ($mac_new_lines > $unix_new_lines) {
570
2
8
    $/ = "\r";
571  } else {
572
8
9
    $/ = "\n";
573  }
574
14
33
  seek($file_fh, 0, 0);
575
14
21
  ($words, $unrecognized) = (0, 0);
576
14
39
  %unique = ();
577
14
33
  %unique_unrecognized = ();
578
579  local $SIG{__WARN__} = sub {
580
0
0
0
    my $message = shift;
581
0
0
    $message =~ s/> line/> in $file - line/;
582
0
0
    chomp $message;
583
0
0
    print STDERR "$message\n";
584
14
128
  };
585
586
14
394
  open(my $warnings_fh, '>:utf8', "$temp_dir/warnings");
587
14
25
  our $timeout;
588
14
10
  eval {
589
14
0
0
111
0
    local $SIG{ALRM} = sub { die "alarm\n" }; # NB: \n required
590
14
42
    alarm $timeout;
591
592
14
30
    my $ignore_next_line = 0;
593
14
34
    my ($current_begin_marker, $next_end_marker, $start_marker_line) = ('', '', '');
594
14
19
    my $offset = 0;
595
14
108
    LINE: while (<$file_fh>) {
596
1170
100
1154
      if ($. == 1) {
597
14
50
16
        unless ($disable_minified_file) {
598
14
100
100
56
          if ($file_size >= 512 && length($_) == $file_size) {
599
1
13
            skip_file($temp_dir, "file only has a single line (single-line-file)\n");
600
1
5
            last;
601          }
602        }
603      }
604
1169
2498
      $_ = decode_utf8($_, FB_DEFAULT);
605
1169
50
2949
      if (/[\x{D800}-\x{DFFF}]/) {
606
0
0
        skip_file($temp_dir, "file contains a UTF-16 surrogate -- UTF-16 surrogates are not supported (utf16-surrogate-file)\n");
607
0
0
        last;
608      }
609
1169
1527
      s/\R$//;
610
1169
100
1029
      s/^\x{FEFF}// if $. == 1;
611
1169
100
1170
      next unless /./;
612
1168
810
      my $raw_line = $_;
613
1168
567
      my $parsed_block_markers;
614
615      # hook for custom multiline based text exclusions:
616
1168
100
804
      if ($begin_block_re) {
617
1148
553
        FIND_END_MARKER: while (1) {
618
1150
981
          while ($next_end_marker ne '') {
619
6
100
25
            next LINE unless /\Q$next_end_marker\E/;
620
1
5
            s/.*?\Q$next_end_marker\E//;
621
1
1
            ($current_begin_marker, $next_end_marker, $start_marker_line) = ('', '', '');
622
1
2
            $parsed_block_markers = 1;
623          }
624
1145
1236
          my @captured = (/^.*?$begin_block_re/);
625
1145
100
1104
          last unless (@captured);
626
2
2
          for my $capture (0 .. $#captured) {
627
2
50
3
            if ($captured[$capture]) {
628
2
5
              ($current_begin_marker, $next_end_marker, $start_marker_line) = ($begin_block_list[$capture], $end_block_list[$capture], "$.:1 ... 1");
629
2
12
              s/^.*?\Q$begin_block_list[$capture]\E//;
630
2
2
              $parsed_block_markers = 1;
631
2
3
              next FIND_END_MARKER;
632            }
633          }
634        }
635
1143
100
848
        next if $parsed_block_markers;
636      }
637
638
1162
626
      my $ignore_this_line = $ignore_next_line;
639
1162
993
      $ignore_next_line = ($_ =~ /$ignore_next_line_pattern/);
640
1162
100
859
      next if $ignore_this_line;
641
642      # hook for custom line based text exclusions:
643
1161
100
833
      if (defined $patterns_re) {
644
2
6
11
8
        s/($patterns_re)/"="x length($1)/ge;
645      }
646
1161
710
      my $initial_line_state = $_;
647
1161
695
      my $previous_line_state = $_;
648
1161
616
      my $line_flagged;
649
1161
100
792
      if ($forbidden_re) {
650
9
5
58
12
        while (s/($forbidden_re)/"="x length($1)/e) {
651
5
4
          $line_flagged = 1;
652
5
9
          my ($begin, $end, $match) = ($-[0] + 1, $+[0] + 1, $1);
653
5
3
          my $found_trigger_re;
654
5
7
          for my $i (0 .. $#forbidden_re_list) {
655
7
6
            my $forbidden_re_singleton = $forbidden_re_list[$i];
656
7
5
            my $test_line = $previous_line_state;
657
7
4
100
55
7
            if ($test_line =~ s/($forbidden_re_singleton)/"="x length($1)/e) {
658
4
50
4
              next unless $test_line eq $_;
659
4
7
              my ($begin_test, $end_test, $match_test) = ($-[0] + 1, $+[0] + 1, $1);
660
4
50
4
              next unless $begin == $begin_test;
661
4
50
2
              next unless $end == $end_test;
662
4
50
5
              next unless $match eq $match_test;
663
4
2
              $found_trigger_re = $forbidden_re_singleton;
664
4
8
              my $hit = "$.:$begin:$end";
665
4
2
              $forbidden_re_hits[$i]++;
666
4
100
4
              $forbidden_re_lines[$i] = $hit unless $forbidden_re_lines[$i];
667
4
8
              last;
668            }
669          }
670
5
4
          my $wrapped = CheckSpelling::Util::wrap_in_backticks($match);
671
5
100
5
          if ($found_trigger_re) {
672
4
100
9
            my $description = $forbidden_re_descriptions{$found_trigger_re} || '';
673
4
8
            $found_trigger_re =~ s/^\(\?:(.*)\)$/$1/;
674
4
4
            my $quoted_trigger_re = CheckSpelling::Util::truncate_with_ellipsis(CheckSpelling::Util::wrap_in_backticks($found_trigger_re), 99);
675
4
100
4
            if ($description ne '') {
676
3
13
              print $warnings_fh ":$.:$begin ... $end, Warning - $wrapped matches a line_forbidden.patterns rule: $description - $quoted_trigger_re (forbidden-pattern)\n";
677            } else {
678
1
5
              print $warnings_fh ":$.:$begin ... $end, Warning - $wrapped matches a line_forbidden.patterns entry: $quoted_trigger_re (forbidden-pattern)\n";
679            }
680          } else {
681
1
3
            print $warnings_fh ":$.:$begin ... $end, Warning - $wrapped matches a line_forbidden.patterns entry (forbidden-pattern)\n";
682          }
683
5
24
          $previous_line_state = $_;
684        }
685
9
9
        $_ = $initial_line_state;
686      }
687      # This is to make it easier to deal w/ rules:
688
1161
1126
      s/^/ /;
689
1161
636
      my %unrecognized_line_items = ();
690
1161
467
      our $homoglyph_re;
691
1161
50
806
      if (defined $homoglyph_re) {
692
1161
668
        my $check_line_for_homoglyphs = $_;
693
1161
653
        my $homoglyphs = $CheckSpelling::Homoglyph::homoglyphs;
694
1161
10094
        while ($check_line_for_homoglyphs =~ /($homoglyph_re)/g) {
695
1
3
          my ($token, $token_raw, $begin, $end) = ($1, $1, $-[0] + 1, $+[0] + 1);
696
1
54
          $token =~ s/([$homoglyphs])/$CheckSpelling::Homoglyph::homoglyph_to_glyph{$1}/g;
697
1
50
2
          if (defined $dictionary{$token}) {
698
1
1
            my $token_raw = CheckSpelling::Util::wrap_in_backticks($token_raw);
699
1
1
            my $token = CheckSpelling::Util::wrap_in_backticks($token);
700
1
1
            my $wrapped = "check $token_raw should probably be $token (homoglyph-word)";
701
1
7
            print $warnings_fh ":$.:$begin ... $end: $wrapped\n";
702          }
703        }
704      }
705
1161
1216
      my ($new_words, $new_unrecognized) = split_line($_, \%unique, \%unique_unrecognized, \%unrecognized_line_items);
706
1161
682
      $words += $new_words;
707
1161
555
      $unrecognized += $new_unrecognized;
708
1161
813
      my $line_length = length($raw_line);
709
1161
1675
      for my $token (sort CheckSpelling::Util::case_biased keys %unrecognized_line_items) {
710
1021
483
        my $found_token = 0;
711
1021
536
        my $raw_token = $token;
712
1021
510
        $token =~ s/'/(?:'|\x{2019}|\&apos;|\&#39;)+/g;
713
1021
445
        my $before;
714
1021
100
50
1546
        if ($token =~ /^$upper_pattern$lower_pattern/) {
715
5
3
          $before = '(?<=.)';
716        } elsif ($token =~ /^$upper_pattern/) {
717
0
0
          $before = "(?<!$upper_pattern)";
718        } else {
719
1016
661
          $before = "(?<=$not_lower_pattern)";
720        }
721
1021
50
1219
        my $after = ($token =~ /$upper_pattern$/) ? "(?=$not_upper_or_lower_pattern)|(?=$upper_pattern$lower_pattern)" : "(?=$not_lower_pattern)";
722
1021
2212
        while ($raw_line =~ /(?:\b|$before)($token)(?:\b|$after)/g) {
723
1271
571
          $line_flagged = 1;
724
1271
526
          $found_token = 1;
725
1271
1793
          my ($begin, $end, $match) = ($-[0] + 1, $+[0] + 1, $1);
726
1271
50
1194
          next unless $match =~ /./;
727
1271
901
          my $wrapped = CheckSpelling::Util::wrap_in_backticks($match);
728
1271
4935
          print $warnings_fh ":$.:$begin ... $end: $wrapped\n";
729        }
730
1021
100
1245
        unless ($found_token) {
731
3
50
33
29
          if ($raw_line !~ /$token.*$token/ && $raw_line =~ /($token)/) {
732
3
6
            my ($begin, $end, $match) = ($-[0] + 1, $+[0] + 1, $1);
733
3
3
            my $wrapped = CheckSpelling::Util::wrap_in_backticks($raw_token);
734
3
11
            print $warnings_fh ":$.:$begin ... $end: $wrapped\n";
735          } else {
736
0
0
            my $offset = $line_length + 1;
737
0
0
            my $wrapped = CheckSpelling::Util::wrap_in_backticks($raw_token);
738
0
0
            print $warnings_fh ":$.:1 ... $offset, Warning - Could not identify whole word $wrapped in line (token-is-substring)\n";
739          }
740        }
741      }
742
1161
100
100
1927
      if ($line_flagged && $candidates_re) {
743
2
2
        $_ = $previous_line_state = $initial_line_state;
744
2
2
21
5
        s/($candidates_re)/"="x length($1)/ge;
745
2
50
2
        if ($_ ne $initial_line_state) {
746
2
1
          $_ = $previous_line_state;
747
2
3
          for my $i (0 .. $#candidates_re_list) {
748
4
4
            my $candidate_re = $candidates_re_list[$i];
749
4
100
67
25
            next unless $candidate_re =~ /./ && $raw_line =~ /$candidate_re/;
750
2
2
50
7
4
            if (($_ =~ s/($candidate_re)/"="x length($1)/e)) {
751
2
3
              my ($begin, $end) = ($-[0] + 1, $+[0] + 1);
752
2
4
              my $hit = "$.:$begin:$end";
753
2
2
              $_ = $previous_line_state;
754
2
2
7
2
              my $replacements = ($_ =~ s/($candidate_re)/"="x length($1)/ge);
755
2
2
              $candidates_re_hits[$i] += $replacements;
756
2
50
7
              $candidates_re_lines[$i] = $hit unless $candidates_re_lines[$i];
757
2
5
              $_ = $previous_line_state;
758            }
759          }
760        }
761      }
762
1161
50
916
      unless ($disable_minified_file) {
763
1161
903
        s/={3,}//g;
764
1161
806
        $offset += length;
765
1161
1043
        my $ratio = int($offset / $.);
766
1161
617
        my $ratio_threshold = 1000;
767
1161
100
3364
        if ($ratio > $ratio_threshold) {
768
2
8
          skip_file($temp_dir, "average line width ($ratio) exceeds the threshold ($ratio_threshold) (minified-file)\n");
769
2
7
          last;
770        }
771      }
772    }
773
14
100
29
    if ($next_end_marker) {
774
1
50
2
      if ($start_marker_line) {
775
1
1
        my $wrapped = CheckSpelling::Util::wrap_in_backticks($current_begin_marker);
776
1
4
        print $warnings_fh ":$start_marker_line, Warning - Failed to find matching end marker for $wrapped (unclosed-block-ignore-begin)\n";
777      }
778
1
1
      my $wrapped = CheckSpelling::Util::wrap_in_backticks($next_end_marker);
779
1
3
      print $warnings_fh ":$.:1 ... 1, Warning - Expected to find end block marker $wrapped (unclosed-block-ignore-end)\n";
780    }
781
782
14
110
    alarm 0;
783  };
784
14
50
12
  if ($@) {
785
0
0
0
    die unless $@ eq "alarm\n";
786
0
0
    print $warnings_fh ":$.:1 ... 1, Warning - Could not parse file within time limit (slow-file)\n";
787
0
0
    skip_file($temp_dir, "it could not be parsed file within time limit (slow-file)\n");
788
0
0
    return $temp_dir;
789  }
790
791
14
58
  close $file_fh;
792
14
178
  close $warnings_fh;
793
794
14
100
75
32
  if ($unrecognized || @candidates_re_hits || @forbidden_re_hits) {
795
13
384
    open(my $stats_fh, '>:utf8', "$temp_dir/stats");
796
13
100
100
100
100
197
      print $stats_fh "{words: $words, unrecognized: $unrecognized, unknown: ".(keys %unique_unrecognized).
797      ", unique: ".(keys %unique).
798      (@candidates_re_hits ? ", candidates: [".(join ',', @candidates_re_hits)."]" : "").
799      (@candidates_re_lines ? ", candidate_lines: [".(join ',', @candidates_re_lines)."]" : "").
800      (@forbidden_re_hits ? ", forbidden: [".(join ',', @forbidden_re_hits)."]" : "").
801      (@forbidden_re_lines ? ", forbidden_lines: [".(join ',', @forbidden_re_lines)."]" : "").
802      "}";
803
13
176
    close $stats_fh;
804
13
349
    open(my $unknown_fh, '>:utf8', "$temp_dir/unknown");
805
13
20
56
36
      print $unknown_fh map { "$_\n" } sort CheckSpelling::Util::case_biased keys %unique_unrecognized;
806
13
145
    close $unknown_fh;
807  }
808
809
14
157
  return $temp_dir;
810}
811
812sub main {
813
4
4
402
  my ($configuration, @ARGV) = @_;
814
4
0
  our %dictionary;
815
4
100
4
  unless (%dictionary) {
816
1
1
    init($configuration);
817  }
818
819  # read all input
820
4
5
  my @reports;
821
822
4
1
  for my $file (@ARGV) {
823
4
4
    my $temp_dir = split_file($file);
824
4
8
    push @reports, "$temp_dir\n";
825  }
826
4
10
  print join '', @reports;
827}
828
8291;