1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # Phind.pm -- the Phind classifier
|
---|
4 | #
|
---|
5 | # Copyright (C) 2000 Gordon W. Paynter
|
---|
6 | # Copyright (C) 2000 New Zealand Digital Library Project
|
---|
7 | #
|
---|
8 | #
|
---|
9 | # A component of the Greenstone digital library software
|
---|
10 | # from the New Zealand Digital Library Project at the
|
---|
11 | # University of Waikato, New Zealand.
|
---|
12 | #
|
---|
13 | # This program is free software; you can redistribute it and/or modify
|
---|
14 | # it under the terms of the GNU General Public License as published by
|
---|
15 | # the Free Software Foundation; either version 2 of the License, or
|
---|
16 | # (at your option) any later version.
|
---|
17 | #
|
---|
18 | # This program is distributed in the hope that it will be useful,
|
---|
19 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
20 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
21 | # GNU General Public License for more details.
|
---|
22 | #
|
---|
23 | # You should have received a copy of the GNU General Public License
|
---|
24 | # along with this program; if not, write to the Free Software
|
---|
25 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
26 | #
|
---|
27 | ###########################################################################
|
---|
28 |
|
---|
29 | # The Phind clasifier plugin.
|
---|
30 | # Type "classinfo.pl Phind" at the command line for a summary.
|
---|
31 |
|
---|
32 | # 12/05/02 Added usage datastructure - John Thompson
|
---|
33 |
|
---|
34 | package Phind;
|
---|
35 |
|
---|
36 | use BaseClassifier;
|
---|
37 | use util;
|
---|
38 | use ghtml;
|
---|
39 | use unicode;
|
---|
40 |
|
---|
41 | use strict;
|
---|
42 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
---|
43 |
|
---|
44 | my @removedirs = ();
|
---|
45 |
|
---|
46 | my %wanted_index_files = ('td'=>1,
|
---|
47 | 't'=>1,
|
---|
48 | 'ti'=>1,
|
---|
49 | 'tl'=>1,
|
---|
50 | 'tsd'=>1,
|
---|
51 | 'idb'=>1,
|
---|
52 | 'ib1'=>1,
|
---|
53 | 'ib2'=>1,
|
---|
54 | 'ib3'=>1,
|
---|
55 | 'i'=>1,
|
---|
56 | 'il'=>1,
|
---|
57 | 'w'=>1,
|
---|
58 | 'wa'=>1);
|
---|
59 |
|
---|
60 | sub BEGIN {
|
---|
61 | @Phind::ISA = ('BaseClassifier');
|
---|
62 | }
|
---|
63 |
|
---|
64 | sub END {
|
---|
65 |
|
---|
66 | # Tidy up stray files - we do this here as there's some weird problem
|
---|
67 | # preventing us from doing it in the get_classify_info() function (on
|
---|
68 | # windows at least) where the close() appears to fail on txthandle and
|
---|
69 | # dochandle, thus preventing us from deleting those files
|
---|
70 |
|
---|
71 | foreach my $dir (@removedirs) {
|
---|
72 | if (-d $dir && opendir (DIR, $dir)) {
|
---|
73 | my @files = readdir DIR;
|
---|
74 | closedir DIR;
|
---|
75 |
|
---|
76 | foreach my $file (@files) {
|
---|
77 | next if $file =~ /^\.\.?$/;
|
---|
78 | my ($suffix) = $file =~ /\.([^\.]+)$/;
|
---|
79 | if (!defined $suffix || !defined $wanted_index_files{$suffix}) {
|
---|
80 | # delete it!
|
---|
81 | &util::rm (&util::filename_cat ($dir, $file));
|
---|
82 | }
|
---|
83 | }
|
---|
84 | }
|
---|
85 | }
|
---|
86 | }
|
---|
87 |
|
---|
88 | my $arguments =
|
---|
89 | [ { 'name' => "text",
|
---|
90 | 'desc' => "{Phind.text}",
|
---|
91 | 'type' => "string",
|
---|
92 | 'deft' => "section:Title,section:text",
|
---|
93 | 'reqd' => "no" },
|
---|
94 | { 'name' => "title",
|
---|
95 | 'desc' => "{Phind.title}",
|
---|
96 | 'type' => "metadata",
|
---|
97 | 'deft' => "Title",
|
---|
98 | 'reqd' => "no" },
|
---|
99 | { 'name' => "buttonname",
|
---|
100 | 'desc' => "{BasClas.buttonname}",
|
---|
101 | 'type' => "string",
|
---|
102 | 'deft' => "Phrase",
|
---|
103 | 'reqd' => "no" },
|
---|
104 | { 'name' => "language",
|
---|
105 | 'desc' => "{Phind.language}",
|
---|
106 | 'type' => "string",
|
---|
107 | 'deft' => "en",
|
---|
108 | 'reqd' => "no" },
|
---|
109 | { 'name' => "savephrases",
|
---|
110 | 'desc' => "{Phind.savephrases}",
|
---|
111 | 'type' => "string",
|
---|
112 | 'deft' => "",
|
---|
113 | 'reqd' => "no" },
|
---|
114 | { 'name' => "suffixmode",
|
---|
115 | 'desc' => "{Phind.suffixmode}",
|
---|
116 | 'type' => "int",
|
---|
117 | 'deft' => "1",
|
---|
118 | 'range' => "0,1",
|
---|
119 | 'reqd' => "no" },
|
---|
120 | { 'name' => "min_occurs",
|
---|
121 | 'desc' => "{Phind.min_occurs}",
|
---|
122 | 'type' => "int",
|
---|
123 | 'deft' => "2",
|
---|
124 | 'range' => "1,",
|
---|
125 | 'reqd' => "no" },
|
---|
126 | { 'name' => "thesaurus",
|
---|
127 | 'desc' => "{Phind.thesaurus}",
|
---|
128 | 'type' => "string",
|
---|
129 | 'deft' => "",
|
---|
130 | 'reqd' => "no" },
|
---|
131 | { 'name' => "untidy",
|
---|
132 | 'desc' => "{Phind.untidy}",
|
---|
133 | 'type' => "flag",
|
---|
134 | 'reqd' => "no" } ];
|
---|
135 |
|
---|
136 | my $options = { 'name' => "Phind",
|
---|
137 | 'desc' => "{Phind.desc}",
|
---|
138 | 'abstract' => "no",
|
---|
139 | 'inherits' => "yes",
|
---|
140 | 'args' => $arguments };
|
---|
141 |
|
---|
142 |
|
---|
143 | # Phrase delimiter symbols - these should be abstracted out someplace
|
---|
144 |
|
---|
145 | my $colstart = "COLLECTIONSTART";
|
---|
146 | my $colend = "COLLECTIONEND";
|
---|
147 | my $doclimit = "DOCUMENTLIMIT";
|
---|
148 | my $senlimit = "SENTENCELIMIT";
|
---|
149 | my @delimiters = ($colstart, $colend, $doclimit, $senlimit);
|
---|
150 |
|
---|
151 |
|
---|
152 | # Create a new Phind browser based on collect.cfg
|
---|
153 |
|
---|
154 | sub new {
|
---|
155 | my ($class) = shift (@_);
|
---|
156 | my ($classifierslist,$inputargs,$hashArgOptLists) = @_;
|
---|
157 | push(@$classifierslist, $class);
|
---|
158 |
|
---|
159 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
---|
160 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
---|
161 |
|
---|
162 | my $self = new BaseClassifier($classifierslist, $inputargs, $hashArgOptLists);
|
---|
163 |
|
---|
164 | if ($self->{'info_only'}) {
|
---|
165 | # don't worry about any options etc
|
---|
166 | return bless $self, $class;
|
---|
167 | }
|
---|
168 |
|
---|
169 | # Ensure the Phind generate scripts are in place
|
---|
170 | my $file1 = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "suffix");
|
---|
171 | $file1 .= ".exe" if $ENV{'GSDLOS'} =~ /^windows$/;
|
---|
172 | my $src = &util::filename_cat($ENV{'GSDLHOME'}, "src", "phind", "generate");
|
---|
173 | if (!(-e $file1)) {
|
---|
174 | print STDERR "Phind.pm: ERROR: The Phind \"suffix\" program is not installed.\n\n";
|
---|
175 | exit(1);
|
---|
176 | }
|
---|
177 |
|
---|
178 | # things that may have ex. in them that need to be stripped off
|
---|
179 | $self->{'text'} = $self->strip_ex_from_metadata($self->{'text'});
|
---|
180 | $self->{'title'} = $self->strip_ex_from_metadata($self->{'title'});
|
---|
181 |
|
---|
182 | # Transfer value from Auto Parsing to the variable name that used in previous GreenStone.
|
---|
183 |
|
---|
184 | $self->{"indexes"} = $self->{"text"};
|
---|
185 |
|
---|
186 | # Further setup
|
---|
187 | $self->{'collection'} = $ENV{'GSDLCOLLECTION'}; # classifier information
|
---|
188 | $self->{'collectiondir'} = $ENV{'GSDLCOLLECTDIR'}; # collection directories
|
---|
189 | if (! defined $self->{'builddir'}) {
|
---|
190 | $self->{'builddir'} = &util::filename_cat($ENV{'GSDLCOLLECTDIR'}, "building");
|
---|
191 | }
|
---|
192 | $self->{'total'} = 0;
|
---|
193 |
|
---|
194 | # Clean out the unused keys
|
---|
195 | delete $self->{"text"};
|
---|
196 |
|
---|
197 | return bless $self, $class;
|
---|
198 | }
|
---|
199 |
|
---|
200 |
|
---|
201 | # Initialise the Phind classifier
|
---|
202 |
|
---|
203 | sub init {
|
---|
204 | my $self = shift (@_);
|
---|
205 |
|
---|
206 | # ensure we have a build directory
|
---|
207 | my $builddir = $self->{'builddir'};
|
---|
208 | die unless (-e "$builddir");
|
---|
209 |
|
---|
210 | # create Phind directory
|
---|
211 | my $phnumber = 1;
|
---|
212 | my $phinddir = &util::filename_cat($builddir, "phind1");
|
---|
213 | while (-e "$phinddir") {
|
---|
214 | $phnumber++;
|
---|
215 | $phinddir = &util::filename_cat($builddir, "phind$phnumber");
|
---|
216 | }
|
---|
217 | &util::mk_dir("$phinddir");
|
---|
218 | $self->{'phinddir'} = $phinddir;
|
---|
219 | $self->{'phindnumber'} = $phnumber;
|
---|
220 |
|
---|
221 | push(@removedirs, $phinddir) unless $self->{'untidy'};
|
---|
222 |
|
---|
223 | # open filehandles for documents and text
|
---|
224 | my $clausefile = &util::filename_cat("$phinddir", "clauses");
|
---|
225 | &util::rm($clausefile) if (-e $clausefile);
|
---|
226 |
|
---|
227 | my $txthandle = 'TEXT' . $phnumber;
|
---|
228 | open($txthandle, ">$clausefile") || die "Cannot open $clausefile: $!";
|
---|
229 | $self->{'txthandle'} = $txthandle;
|
---|
230 |
|
---|
231 | my $docfile = &util::filename_cat("$phinddir", "docs.txt");
|
---|
232 | &util::rm($docfile) if (-e $docfile);
|
---|
233 |
|
---|
234 | my $dochandle = 'DOC' . $phnumber;
|
---|
235 | open($dochandle, ">$docfile") || die "Cannot open $docfile: $!";
|
---|
236 | $self->{'dochandle'} = $dochandle;
|
---|
237 |
|
---|
238 | }
|
---|
239 |
|
---|
240 |
|
---|
241 | # Classify each document.
|
---|
242 | #
|
---|
243 | # Each document is passed here in turn. The classifier extracts the
|
---|
244 | # text of each and stores it in the clauses file. Document details are
|
---|
245 | # stored in the docs.txt file.
|
---|
246 |
|
---|
247 | sub classify {
|
---|
248 | my $self = shift (@_);
|
---|
249 | my ($doc_obj,$edit_mode) = @_;
|
---|
250 |
|
---|
251 | my $verbosity = $self->{'verbosity'};
|
---|
252 | my $top_section = $doc_obj->get_top_section();
|
---|
253 |
|
---|
254 | my $titlefield = $self->{'title'};
|
---|
255 |
|
---|
256 | my $title = $doc_obj->get_metadata_element ($top_section, $titlefield);
|
---|
257 | if (!defined($title)) {
|
---|
258 | $title = "";
|
---|
259 | print STDERR "Phind: document has no title\n";
|
---|
260 | }
|
---|
261 | print "process: $title\n" if ($verbosity > 2);
|
---|
262 |
|
---|
263 | # Only consider the file if it is in the correct language
|
---|
264 | my $doclanguage = $doc_obj->get_metadata_element ($top_section, "Language");
|
---|
265 | my $phrlanguage = $self->{'language'};
|
---|
266 | return if ($doclanguage && ($doclanguage !~ /$phrlanguage/i));
|
---|
267 |
|
---|
268 | if ($edit_mode eq "delete") {
|
---|
269 | # This classifier works quite differently to the others
|
---|
270 | # Probably doesn't support incremental building anyway
|
---|
271 | return;
|
---|
272 | }
|
---|
273 |
|
---|
274 | # record this file
|
---|
275 | $self->{'total'} ++;
|
---|
276 | # what is $file ???
|
---|
277 | # print "file $self->{'total'}: $file\n" if ($self->{'$verbosity'});
|
---|
278 |
|
---|
279 | # Store document details
|
---|
280 | my $OID = $doc_obj->get_OID();
|
---|
281 | $OID = "NULL" unless defined $OID;
|
---|
282 | my $dochandle = $self->{'dochandle'};
|
---|
283 | print $dochandle "<Document>\t$OID\t$title\n";
|
---|
284 |
|
---|
285 | # Store the text occuring in this object
|
---|
286 |
|
---|
287 | # output the document delimiter
|
---|
288 | my $txthandle = $self->{'txthandle'};
|
---|
289 | print $txthandle "$doclimit\n";
|
---|
290 |
|
---|
291 | # iterate over the required indexes and store their text
|
---|
292 | my $indexes = $self->{'indexes'};
|
---|
293 | my $text = "";
|
---|
294 | my ($part, $level, $field, $section, $data, $dataref);
|
---|
295 |
|
---|
296 | foreach $part (split(/,/, $indexes)) {
|
---|
297 |
|
---|
298 | # Each field has a level and a data element ((e.g. document:Title)
|
---|
299 | ($level, $field) = split(/:/, $part);
|
---|
300 | die unless ($level && $field);
|
---|
301 |
|
---|
302 | # Extract the text from every section
|
---|
303 | # (In phind, document:text and section:text are equivalent)
|
---|
304 | if ($field eq "text") {
|
---|
305 | $data = "";
|
---|
306 | $section = $doc_obj->get_top_section();
|
---|
307 | while (defined($section)) {
|
---|
308 | $data .= $doc_obj->get_text($section) . "\n";
|
---|
309 | $section = $doc_obj->get_next_section($section);
|
---|
310 | }
|
---|
311 | $text .= convert_gml_to_tokens($phrlanguage, $data) . "\n";
|
---|
312 | }
|
---|
313 |
|
---|
314 | # Extract a metadata field from a document
|
---|
315 | # (If there is more than one element of the given type, get them all.)
|
---|
316 | elsif ($level eq "document") {
|
---|
317 | $dataref = $doc_obj->get_metadata($doc_obj->get_top_section(), $field);
|
---|
318 | foreach $data (@$dataref) {
|
---|
319 | $text .= convert_gml_to_tokens($phrlanguage, $data) . "\n";
|
---|
320 | }
|
---|
321 | }
|
---|
322 |
|
---|
323 | # Extract metadata from every section in a document
|
---|
324 | elsif ($level eq "section") {
|
---|
325 | $data = "";
|
---|
326 | $section = $doc_obj->get_top_section();
|
---|
327 | while (defined($section)) {
|
---|
328 | $dataref = $doc_obj->get_metadata($section, $field);
|
---|
329 | $data .= join("\n", @$dataref) . "\n";
|
---|
330 | $section = $doc_obj->get_next_section($section);
|
---|
331 | }
|
---|
332 | $text .= convert_gml_to_tokens($phrlanguage, $data) . "\n";
|
---|
333 | }
|
---|
334 |
|
---|
335 | # Some sort of specification which I don't understand
|
---|
336 | else {
|
---|
337 | die "Unknown level ($level) in Phind index ($part)\n";
|
---|
338 | }
|
---|
339 |
|
---|
340 | }
|
---|
341 |
|
---|
342 | # output the text
|
---|
343 | $text =~ tr/\n//s;
|
---|
344 | print $txthandle "$text";
|
---|
345 | }
|
---|
346 |
|
---|
347 |
|
---|
348 | # Construct the classifier from the information already gathered
|
---|
349 | #
|
---|
350 | # When get_classify_info is called, the clauses and docs.txt files have
|
---|
351 | # already been constructed in the Phind directory. This function will
|
---|
352 | # translate them into compressed, indexed MGPP files that can be read by
|
---|
353 | # the phindcgi script. It will also register our classifier so that it
|
---|
354 | # shows up in the navigation bar.
|
---|
355 |
|
---|
356 | sub get_classify_info {
|
---|
357 | my $self = shift (@_);
|
---|
358 | my ($gli) = @_;
|
---|
359 |
|
---|
360 | close $self->{'dochandle'};
|
---|
361 | close $self->{'txthandle'};
|
---|
362 | my $verbosity = $self->{'verbosity'};
|
---|
363 | my $out = $self->{'outhandle'};
|
---|
364 | my $phinddir = $self->{'phinddir'};
|
---|
365 |
|
---|
366 | my $osextra = "";
|
---|
367 | if ($ENV{'GSDLOS'} !~ /^windows$/i) {
|
---|
368 | $osextra = " -d /";
|
---|
369 | }
|
---|
370 |
|
---|
371 | print STDERR "</Stage>\n" if $gli;
|
---|
372 |
|
---|
373 | if ($verbosity) {
|
---|
374 | print $out "\n*** Phind.pm generating indexes for ", $self->{'indexes'}, "\n";
|
---|
375 | print $out "*** in ", $self->{'phinddir'}, "\n";
|
---|
376 | }
|
---|
377 |
|
---|
378 | print STDERR "<Stage name='Phind'>\n" if $gli;
|
---|
379 |
|
---|
380 | # Construct phind indexes
|
---|
381 | my $suffixmode = $self->{'suffixmode'};
|
---|
382 | my $min_occurs = $self->{'min_occurs'};
|
---|
383 | my ($command, $status);
|
---|
384 |
|
---|
385 | # Generate the vocabulary, symbol statistics, and numbers file
|
---|
386 | # from the clauses file
|
---|
387 | print $out "\nExtracting vocabulary and statistics\n" if $verbosity;
|
---|
388 | print STDERR "<Phase name='ExtractingVocab'/>\n" if $gli;
|
---|
389 | &extract_vocabulary($self);
|
---|
390 |
|
---|
391 | # Use the suffix program to generate the phind/phrases file
|
---|
392 | print $out "\nExtracting phrases from processed text (with suffix)\n" if $verbosity;
|
---|
393 | print STDERR "<Phase name='ExtractingPhrase'/>\n" if $gli;
|
---|
394 | &execute("suffix \"$phinddir\" $suffixmode $min_occurs $verbosity", $verbosity, $out);
|
---|
395 |
|
---|
396 | # check that we generated some files. It's not necessarily an error if
|
---|
397 | # we didn't (execute() would have quit on error), but we can't go on.
|
---|
398 | my $phrasesfile=&util::filename_cat($self->{'phinddir'}, 'phrases');
|
---|
399 | if (! -r $phrasesfile) {
|
---|
400 | print STDERR "<Warning name='NoPhrasesFound'/>\n" if $gli;
|
---|
401 | print $out "\nNo phrases found for Phind classifier!\n";
|
---|
402 | return;
|
---|
403 | }
|
---|
404 |
|
---|
405 | # Create the phrase file and put phrase numbers in phind/phrases
|
---|
406 | print $out "\nSorting and renumbering phrases for input to mgpp\n" if $verbosity;
|
---|
407 | print STDERR "<Phase name='SortAndRenumber'/>\n" if $gli;
|
---|
408 | &renumber_phrases($self);
|
---|
409 |
|
---|
410 | print $out "\nCreating phrase databases\n";
|
---|
411 | print STDERR "<Phase name='PhraseDatabases'/>\n" if $gli;
|
---|
412 | my $mg_input = &util::filename_cat($phinddir, "pdata.txt");
|
---|
413 | my $mg_stem = &util::filename_cat($phinddir, "pdata");
|
---|
414 |
|
---|
415 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T1 \"$mg_input\"", $verbosity, $out);
|
---|
416 | &execute("mgpp_compression_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
---|
417 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T2 \"$mg_input\"", $verbosity, $out);
|
---|
418 |
|
---|
419 | # create the mg index of words
|
---|
420 | print $out "\nCreating word-level search indexes\n";
|
---|
421 | print STDERR "<Phase name='WordLevelIndexes'/>\n" if $gli;
|
---|
422 | $mg_input = &util::filename_cat($phinddir, "pword.txt");
|
---|
423 | $mg_stem = &util::filename_cat($phinddir, "pword");
|
---|
424 |
|
---|
425 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T1 -I1 \"$mg_input\"", $verbosity, $out);
|
---|
426 | &execute("mgpp_compression_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
---|
427 | &execute("mgpp_perf_hash_build $osextra -f \"$mg_stem\"", $verbosity, $out);
|
---|
428 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T2 -I2 \"$mg_input\"", $verbosity, $out);
|
---|
429 | &execute("mgpp_weights_build $osextra -f \"$mg_stem\"", $verbosity, $out);
|
---|
430 | &execute("mgpp_invf_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
---|
431 |
|
---|
432 | &execute("mgpp_stem_idx $osextra -f \"$mg_stem\" -s 1", $verbosity, $out);
|
---|
433 | &execute("mgpp_stem_idx $osextra -f \"$mg_stem\" -s 2", $verbosity, $out);
|
---|
434 | &execute("mgpp_stem_idx $osextra -f \"$mg_stem\" -s 3", $verbosity, $out);
|
---|
435 |
|
---|
436 | # create the mg document information database
|
---|
437 | print $out "\nCreating document information databases\n";
|
---|
438 | print STDERR "<Phase name='DocInfoDatabases'/>\n" if $gli;
|
---|
439 | $mg_input = &util::filename_cat($phinddir, "docs.txt");
|
---|
440 | $mg_stem = &util::filename_cat($phinddir, "docs");
|
---|
441 |
|
---|
442 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T1 \"$mg_input\"", $verbosity, $out);
|
---|
443 | &execute("mgpp_compression_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
---|
444 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T2 \"$mg_input\"", $verbosity, $out);
|
---|
445 |
|
---|
446 | # Return the information about the classifier that we'll later want to
|
---|
447 | # use to create macros when the Phind classifier document is displayed.
|
---|
448 | my %classifyinfo = ('thistype'=>'Invisible',
|
---|
449 | 'childtype'=>'Phind',
|
---|
450 | 'Title'=>$self->{'buttonname'},
|
---|
451 | 'parameters'=>"phindnumber=$self->{'phindnumber'}",
|
---|
452 | 'contains'=>[]);
|
---|
453 |
|
---|
454 | my $collection = $self->{'collection'};
|
---|
455 | my $url = "library?a=p&p=phind&c=$collection";
|
---|
456 | push (@{$classifyinfo{'contains'}}, {'OID'=>$url});
|
---|
457 |
|
---|
458 | return \%classifyinfo;
|
---|
459 | }
|
---|
460 |
|
---|
461 |
|
---|
462 |
|
---|
463 | sub convert_gml_to_tokens {
|
---|
464 |
|
---|
465 | my ($language_exp, $text) = @_;
|
---|
466 |
|
---|
467 | # escape any magic words... - jrm21
|
---|
468 | foreach my $delim (@delimiters) {
|
---|
469 | my $replacement=lc($delim);
|
---|
470 | my $num= $text=~ s/$delim/$replacement/g;
|
---|
471 | if (!$num) {$num=0;}
|
---|
472 | }
|
---|
473 |
|
---|
474 | if ($language_exp =~ /en/) {
|
---|
475 | return &convert_gml_to_tokens_EN($text);
|
---|
476 | }
|
---|
477 |
|
---|
478 | if ($language_exp =~ /zh/) {
|
---|
479 | return &convert_gml_to_tokens_ZH($text);
|
---|
480 | }
|
---|
481 |
|
---|
482 | $_ = $text;
|
---|
483 |
|
---|
484 | # 1. remove GML tags
|
---|
485 |
|
---|
486 | # Remove everything that is in a tag
|
---|
487 | s/\s*<p>\s*/ PARAGRAPHBREAK /isgo;
|
---|
488 | s/\s*<br>\s*/ LINEBREAK /isgo;
|
---|
489 | s/<[^>]*>/ /sgo;
|
---|
490 |
|
---|
491 | # Now we have the text, but it may contain HTML
|
---|
492 | # elements coded as > etc. Remove these tags.
|
---|
493 | s/&/&/sgo;
|
---|
494 | s/</</sgo;
|
---|
495 | s/>/>/sgo;
|
---|
496 | s/\s*<p>\s*/ PARAGRAPHBREAK /isgo;
|
---|
497 | s/\s*<br>\s*/ LINEBREAK /isgo;
|
---|
498 | s/<[^>]*>/ /sgo;
|
---|
499 |
|
---|
500 | # replace<p> and <br> placeholders with clause break symbol (\n)
|
---|
501 | s/\s+/ /gso;
|
---|
502 | s/PARAGRAPHBREAK/\n/sgo;
|
---|
503 | s/LINEBREAK/\n/sgo;
|
---|
504 |
|
---|
505 |
|
---|
506 |
|
---|
507 |
|
---|
508 | # 2. Split the remaining text into space-delimited tokens
|
---|
509 |
|
---|
510 | # Convert any HTML special characters (like ") to their UTF8 equivalent
|
---|
511 | s/&([^;]+);/&unicode::ascii2utf8(\&ghtml::getcharequiv($1,1))/gse;
|
---|
512 |
|
---|
513 | # Split text at word boundaries
|
---|
514 | s/\b/ /go;
|
---|
515 |
|
---|
516 | # 3. Convert the remaining text to "clause format"
|
---|
517 |
|
---|
518 | # Insert newline if the end of a sentence is detected
|
---|
519 | # (delimter is: "[\.\?\!]\s")
|
---|
520 | # s/\s*[\.\?\!]\s+/\n/go;
|
---|
521 |
|
---|
522 | # remove unnecessary punctuation and replace with clause break symbol (\n)
|
---|
523 | s/[^\w ]/\n/go;
|
---|
524 |
|
---|
525 | # remove extraneous whitespace
|
---|
526 | s/ +/ /sgo;
|
---|
527 | s/^\s+//mgo;
|
---|
528 | s/\s*$/\n/mgo;
|
---|
529 |
|
---|
530 | # remove lines that contain one word or less
|
---|
531 | s/^\S*$//mgo;
|
---|
532 | s/^\s*$//mgo;
|
---|
533 | tr/\n//s;
|
---|
534 |
|
---|
535 | return $_;
|
---|
536 | }
|
---|
537 |
|
---|
538 | # a chinese version
|
---|
539 | sub convert_gml_to_tokens_ZH {
|
---|
540 |
|
---|
541 | $_ = shift @_;
|
---|
542 |
|
---|
543 | # Replace all whitespace with a simple space
|
---|
544 | s/\s+/ /gs;
|
---|
545 | # Remove everything that is in a tag
|
---|
546 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
---|
547 | s/\s*<br>\s*/ LINEBREAK /isg;
|
---|
548 | s/<[^>]*>/ /sg;
|
---|
549 |
|
---|
550 | # Now we have the text, but it may contain HTML
|
---|
551 | # elements coded as > etc. Remove these tags.
|
---|
552 | s/</</sg;
|
---|
553 | s/>/>/sg;
|
---|
554 |
|
---|
555 | s/\s+/ /sg;
|
---|
556 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
---|
557 | s/\s*<br>\s*/ LINEBREAK /isg;
|
---|
558 | s/<[^>]*>/ /sg;
|
---|
559 |
|
---|
560 | # remove & and other miscellaneous markup tags
|
---|
561 | s/&/&/sg;
|
---|
562 | s/</</sg;
|
---|
563 | s/>/>/sg;
|
---|
564 | s/&/&/sg;
|
---|
565 |
|
---|
566 | # replace<p> and <br> placeholders with carriage returns
|
---|
567 | s/PARAGRAPHBREAK/\n/sg;
|
---|
568 | s/LINEBREAK/\n/sg;
|
---|
569 |
|
---|
570 |
|
---|
571 | # print STDERR "text:$_\n";
|
---|
572 | return $_;
|
---|
573 | }
|
---|
574 |
|
---|
575 | # A version of convert_gml_to_tokens that is fine-tuned to the English language.
|
---|
576 |
|
---|
577 | sub convert_gml_to_tokens_EN {
|
---|
578 | $_ = shift @_;
|
---|
579 |
|
---|
580 | # FIRST, remove GML tags
|
---|
581 |
|
---|
582 | # Replace all whitespace with a simple space
|
---|
583 | s/\s+/ /gs;
|
---|
584 |
|
---|
585 | # Remove everything that is in a tag
|
---|
586 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
---|
587 | s/\s*<br>\s*/ LINEBREAK /isg;
|
---|
588 | s/<[^>]*>/ /sg;
|
---|
589 |
|
---|
590 | # Now we have the text, but it may contain HTML
|
---|
591 | # elements coded as > etc. Remove these tags.
|
---|
592 | s/</</sg;
|
---|
593 | s/>/>/sg;
|
---|
594 |
|
---|
595 | s/\s+/ /sg;
|
---|
596 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
---|
597 | s/\s*<br>\s*/ LINEBREAK /isg;
|
---|
598 | s/<[^>]*>/ /sg;
|
---|
599 |
|
---|
600 | # remove & and other miscellaneous markup tags
|
---|
601 | s/&/&/sg;
|
---|
602 | s/</</sg;
|
---|
603 | s/>/>/sg;
|
---|
604 | s/&/&/sg;
|
---|
605 |
|
---|
606 | # replace<p> and <br> placeholders with carriage returns
|
---|
607 | s/PARAGRAPHBREAK/\n/sg;
|
---|
608 | s/LINEBREAK/\n/sg;
|
---|
609 |
|
---|
610 |
|
---|
611 | # Exceptional punctuation
|
---|
612 | #
|
---|
613 | # We make special cases of some punctuation
|
---|
614 |
|
---|
615 | # remove any apostrophe that indicates omitted letters
|
---|
616 | s/(\w+)\'(\w*\s)/ $1$2 /g;
|
---|
617 |
|
---|
618 | # remove period that appears in a person's initals
|
---|
619 | s/\s([A-Z])\./ $1 /g;
|
---|
620 |
|
---|
621 | # replace hyphens in hypheanted words and names with a space
|
---|
622 | s/([A-Za-z])-\s*([A-Za-z])/$1 $2/g;
|
---|
623 |
|
---|
624 | # Convert the remaining text to "clause format",
|
---|
625 | # This means removing all excess punctuation and garbage text,
|
---|
626 | # normalising valid punctuation to fullstops and commas,
|
---|
627 | # then putting one cluse on each line.
|
---|
628 |
|
---|
629 | # Insert newline when the end of a sentence is detected
|
---|
630 | # (delimter is: "[\.\?\!]\s")
|
---|
631 | s/\s*[\.\?\!]\s+/\n/g;
|
---|
632 |
|
---|
633 | # split numbers after four digits
|
---|
634 | s/(\d\d\d\d)/$1 /g;
|
---|
635 |
|
---|
636 | # split words after 32 characters
|
---|
637 |
|
---|
638 | # squash repeated punctuation
|
---|
639 | tr/A-Za-z0-9 //cs;
|
---|
640 |
|
---|
641 | # save email addresses
|
---|
642 | # s/\w+@\w+\.[\w\.]+/EMAIL/g;
|
---|
643 |
|
---|
644 | # normalise clause breaks (mostly punctuation symbols) to commas
|
---|
645 | s/[^A-Za-z0-9 \n]+/ , /g;
|
---|
646 |
|
---|
647 | # Remove repeated commas, and replace with newline
|
---|
648 | s/\s*,[, ]+/\n/g;
|
---|
649 |
|
---|
650 | # remove extra whitespace
|
---|
651 | s/ +/ /sg;
|
---|
652 | s/^\s+//mg;
|
---|
653 | s/\s*$/\n/mg;
|
---|
654 |
|
---|
655 | # remove lines that contain one word or less
|
---|
656 | s/^\w*$//mg;
|
---|
657 | s/^\s*$//mg;
|
---|
658 | tr/\n//s;
|
---|
659 |
|
---|
660 | return $_;
|
---|
661 |
|
---|
662 | }
|
---|
663 |
|
---|
664 |
|
---|
665 |
|
---|
666 | # Execute a system command
|
---|
667 |
|
---|
668 | sub execute {
|
---|
669 | my ($command, $verbosity, $outhandle) = @_;
|
---|
670 | print $outhandle "Executing: $command\n" if ($verbosity > 2);
|
---|
671 | $! = 0;
|
---|
672 | my $status = system($command);
|
---|
673 | if ($status != 0) {
|
---|
674 | print STDERR "Phind - Error executing '$command': $!\n";
|
---|
675 | exit($status); # this causes the build to fail...
|
---|
676 | }
|
---|
677 | }
|
---|
678 |
|
---|
679 |
|
---|
680 | # Generate the vocabulary, symbol statistics, and numbers file from the
|
---|
681 | # clauses file. This is legacy code, so is a bit messy and probably wont
|
---|
682 | # run under windows.
|
---|
683 |
|
---|
684 | sub extract_vocabulary {
|
---|
685 | my ($self) = @_;
|
---|
686 |
|
---|
687 | my $verbosity = $self->{'verbosity'};
|
---|
688 | my $out = $self->{'outhandle'};
|
---|
689 |
|
---|
690 | my $collectiondir = $self->{'collectiondir'};
|
---|
691 | my $phinddir = $self->{'phinddir'};
|
---|
692 |
|
---|
693 | my $language_exp = $self->{'language'};
|
---|
694 |
|
---|
695 | my ($w, $l, $line, $word);
|
---|
696 |
|
---|
697 | my ($first_delimiter, $last_delimiter,
|
---|
698 | $first_stopword, $last_stopword,
|
---|
699 | $first_extractword, $last_extractword,
|
---|
700 | $first_contentword, $last_contentword,
|
---|
701 | $phrasedelimiter);
|
---|
702 |
|
---|
703 | my $thesaurus = $self->{'thesaurus'};
|
---|
704 | my ($thesaurus_links, $thesaurus_terms,
|
---|
705 | %thesaurus, $first_thesaurusword, $last_thesaurusword);
|
---|
706 |
|
---|
707 | my %symbol;
|
---|
708 | my (%freq);
|
---|
709 |
|
---|
710 | print $out "Calculating vocabulary\n" if ($verbosity > 1);
|
---|
711 |
|
---|
712 | # Read and store the stopwords
|
---|
713 | my $stopdir = &util::filename_cat($ENV{'GSDLHOME'}, "etc", "packages", "phind", "stopword");
|
---|
714 | my $stopword_files = ();
|
---|
715 | my ($language, $language_dir, $file, $file_name);
|
---|
716 | my %stopwords;
|
---|
717 |
|
---|
718 | # Examine each directory in the stopword directory
|
---|
719 | opendir(STOPDIR, $stopdir);
|
---|
720 | foreach $language (readdir STOPDIR) {
|
---|
721 |
|
---|
722 | # Ignore entries that do not match the classifier's language
|
---|
723 | next unless ($language =~ /$language_exp/);
|
---|
724 | $language_dir = &util::filename_cat($stopdir, $language);
|
---|
725 | next unless (-d "$language_dir");
|
---|
726 |
|
---|
727 | opendir(LANGDIR, $language_dir);
|
---|
728 | foreach $file (readdir LANGDIR) {
|
---|
729 |
|
---|
730 | # Ignore entries that are not stopword files
|
---|
731 | next unless ($file =~ /sw$/);
|
---|
732 | $file_name = &util::filename_cat($language_dir, $file);
|
---|
733 | next unless (-f "$file_name");
|
---|
734 |
|
---|
735 | # Read the stopwords
|
---|
736 | open(STOPFILE, "<$file_name");
|
---|
737 | while (<STOPFILE>) {
|
---|
738 | s/^\s+//;
|
---|
739 | s/\s.*//;
|
---|
740 | $word = $_;
|
---|
741 | $l = lc($word);
|
---|
742 | $stopwords{$l} = $word;
|
---|
743 | }
|
---|
744 | close STOPFILE;
|
---|
745 |
|
---|
746 | }
|
---|
747 | closedir LANGDIR;
|
---|
748 | }
|
---|
749 | closedir STOPDIR;
|
---|
750 |
|
---|
751 | # Read thesaurus information
|
---|
752 | if ($thesaurus) {
|
---|
753 |
|
---|
754 | # link file exists
|
---|
755 | $thesaurus_links = &util::filename_cat($collectiondir, "etc", "$thesaurus.lnk");
|
---|
756 | die "Cannot find thesaurus link file" unless (-e "$thesaurus_links");
|
---|
757 |
|
---|
758 | # ensure term file exists in the correct language
|
---|
759 | if ($language_exp =~ /^([a-z][a-z])/) {
|
---|
760 | $language = $1;
|
---|
761 | } else {
|
---|
762 | $language = 'en';
|
---|
763 | }
|
---|
764 | $thesaurus_terms = &util::filename_cat($collectiondir, "etc", "$thesaurus.$language");
|
---|
765 | die "Cannot find thesaurus term file" unless (-e "$thesaurus_terms");
|
---|
766 |
|
---|
767 |
|
---|
768 | # Read the thesaurus terms
|
---|
769 | open(TH, "<$thesaurus_terms");
|
---|
770 | while(<TH>) {
|
---|
771 | s/^\d+ //;
|
---|
772 | s/\(.*\)//;
|
---|
773 | foreach $w (split(/\s+/, $_)) {
|
---|
774 | $thesaurus{lc($w)} = $w;
|
---|
775 | }
|
---|
776 | }
|
---|
777 | close TH;
|
---|
778 | }
|
---|
779 |
|
---|
780 | # Read words in the text and count occurences
|
---|
781 | open(TXT, "<$phinddir/clauses");
|
---|
782 |
|
---|
783 | my @words;
|
---|
784 | while(<TXT>) {
|
---|
785 | $line = $_;
|
---|
786 | next unless ($line =~ /./);
|
---|
787 |
|
---|
788 | @words = split(/\s+/, $line);
|
---|
789 | foreach $w (@words) {
|
---|
790 | $l = lc($w);
|
---|
791 | $w = $l if ((defined $stopwords{$l}) || (defined $thesaurus{$l}));
|
---|
792 | $freq{$w}++;
|
---|
793 | }
|
---|
794 | $freq{$senlimit}++;
|
---|
795 | }
|
---|
796 |
|
---|
797 | close TXT;
|
---|
798 |
|
---|
799 | # Calculate the "best" form of each word
|
---|
800 | my (%bestform, %totalfreq, %bestfreq);
|
---|
801 |
|
---|
802 | foreach $w (sort (keys %freq)) {
|
---|
803 | $l = lc($w);
|
---|
804 |
|
---|
805 | # totalfreq is the number of times a term appears in any form
|
---|
806 | $totalfreq{$l} += $freq{$w};
|
---|
807 |
|
---|
808 | if (defined $stopwords{$l}) {
|
---|
809 | $bestform{$l} = $stopwords{$l};
|
---|
810 |
|
---|
811 | } elsif (defined $thesaurus{$l}) {
|
---|
812 | $bestform{$l} = $thesaurus{$l};
|
---|
813 |
|
---|
814 | } elsif (!$bestform{$l} || ($freq{$w} > $bestfreq{$l})) {
|
---|
815 | $bestfreq{$l} = $freq{$w};
|
---|
816 | $bestform{$l} = $w;
|
---|
817 | }
|
---|
818 | }
|
---|
819 | undef %freq;
|
---|
820 | undef %bestfreq;
|
---|
821 |
|
---|
822 |
|
---|
823 | # Assign symbol numbers to tokens
|
---|
824 | my $nextsymbol = 1;
|
---|
825 | my (@vocab);
|
---|
826 |
|
---|
827 | # Delimiters
|
---|
828 | $first_delimiter = 1;
|
---|
829 |
|
---|
830 | foreach $word (@delimiters) {
|
---|
831 |
|
---|
832 | # $word = lc($word); # jrm21
|
---|
833 | $word = uc($word);
|
---|
834 | $bestform{$word} = $word;
|
---|
835 | $vocab[$nextsymbol] = $word;
|
---|
836 | $symbol{$word} = $nextsymbol;
|
---|
837 | $nextsymbol++;
|
---|
838 | }
|
---|
839 | $last_delimiter = $nextsymbol - 1;
|
---|
840 | # Stopwords
|
---|
841 | $first_stopword = $nextsymbol;
|
---|
842 |
|
---|
843 | foreach my $word (sort keys %stopwords) {
|
---|
844 | # don't include stopword unless it occurs in the text
|
---|
845 | $word = lc($word);
|
---|
846 | next unless ($totalfreq{$word});
|
---|
847 | next if ($symbol{$word});
|
---|
848 |
|
---|
849 | $vocab[$nextsymbol] = $word;
|
---|
850 | $symbol{$word} = $nextsymbol;
|
---|
851 | $nextsymbol++;
|
---|
852 | }
|
---|
853 | $last_stopword = $nextsymbol - 1;
|
---|
854 | $first_contentword = $nextsymbol;
|
---|
855 |
|
---|
856 | # Thesaurus terms
|
---|
857 | if ($thesaurus) {
|
---|
858 | $first_thesaurusword = $nextsymbol;
|
---|
859 |
|
---|
860 | foreach my $word (sort keys %thesaurus) {
|
---|
861 |
|
---|
862 | $word = lc($word);
|
---|
863 | next if ($symbol{$word});
|
---|
864 | $bestform{$word} = $thesaurus{$word};
|
---|
865 |
|
---|
866 | $vocab[$nextsymbol] = $word;
|
---|
867 | $symbol{$word} = $nextsymbol;
|
---|
868 | $nextsymbol++;
|
---|
869 |
|
---|
870 | }
|
---|
871 | $last_thesaurusword = $nextsymbol - 1;
|
---|
872 | }
|
---|
873 |
|
---|
874 | # Other content words
|
---|
875 | $first_extractword = $nextsymbol;
|
---|
876 |
|
---|
877 | foreach my $word (sort (keys %bestform)) {
|
---|
878 |
|
---|
879 | next if ($symbol{$word});
|
---|
880 |
|
---|
881 | $vocab[$nextsymbol] = $word;
|
---|
882 | $symbol{$word} = $nextsymbol;
|
---|
883 | $nextsymbol++;
|
---|
884 | }
|
---|
885 | $last_extractword = $nextsymbol - 1;
|
---|
886 | $last_contentword = $nextsymbol - 1;
|
---|
887 |
|
---|
888 | # Outut the words
|
---|
889 | print $out "Saving vocabulary in $phinddir/clauses.vocab\n" if ($verbosity > 1);
|
---|
890 | open(VOC, ">$phinddir/clauses.vocab");
|
---|
891 |
|
---|
892 | for (my $i = 1; $i < $nextsymbol; $i++) {
|
---|
893 | $w = $vocab[$i];
|
---|
894 |
|
---|
895 | print VOC "$bestform{$w}\n";
|
---|
896 | $totalfreq{$w} = 0 unless ($totalfreq{$w});
|
---|
897 | }
|
---|
898 | close VOC;
|
---|
899 |
|
---|
900 |
|
---|
901 | # Create statistics file
|
---|
902 | # Output statistics about the vocablary
|
---|
903 | print $out "Saving statistics in $phinddir/clauses.stats\n" if ($verbosity > 1);
|
---|
904 | &util::rm("$phinddir/clauses.stats") if (-e "$phinddir/clauses.stats");
|
---|
905 |
|
---|
906 | open(STAT, ">$phinddir/clauses.stats")
|
---|
907 | || die "Cannot open $phinddir/clauses.stats: $!";
|
---|
908 |
|
---|
909 | print STAT "first_delimiter $first_delimiter\n";
|
---|
910 | print STAT "last_delimiter $last_delimiter\n";
|
---|
911 | print STAT "first_stopword $first_stopword\n";
|
---|
912 | print STAT "last_stopword $last_stopword\n";
|
---|
913 | if ($thesaurus) {
|
---|
914 | print STAT "first_thesaurusword $first_thesaurusword\n";
|
---|
915 | print STAT "last_thesaurusword $last_thesaurusword\n";
|
---|
916 | }
|
---|
917 | print STAT "first_extractword $first_extractword\n";
|
---|
918 | print STAT "last_extractword $last_extractword\n";
|
---|
919 | print STAT "first_contentword $first_contentword\n";
|
---|
920 | print STAT "last_contentword $last_contentword\n";
|
---|
921 | print STAT "first_symbol $first_delimiter\n";
|
---|
922 | print STAT "last_symbol $last_contentword\n";
|
---|
923 | print STAT "first_word $first_stopword\n";
|
---|
924 | print STAT "last_word $last_contentword\n";
|
---|
925 | close STAT;
|
---|
926 |
|
---|
927 | undef @vocab;
|
---|
928 |
|
---|
929 |
|
---|
930 | # Create numbers file
|
---|
931 | # Save text as symbol numbers
|
---|
932 | print $out "Saving text as numbers in $phinddir/clauses.numbers\n" if ($verbosity > 1);
|
---|
933 |
|
---|
934 | open(TXT, "<$phinddir/clauses");
|
---|
935 | open(NUM, ">$phinddir/clauses.numbers");
|
---|
936 |
|
---|
937 | ## $phrasedelimiter = $symbol{lc($senlimit)}; # jrm21
|
---|
938 | ## print NUM "$symbol{lc($colstart)}\n"; # jrm21
|
---|
939 | $phrasedelimiter = $symbol{$senlimit};
|
---|
940 | print NUM "$symbol{$colstart}\n";
|
---|
941 |
|
---|
942 | # set up the special symbols that delimit documents and sentences
|
---|
943 | while(<TXT>) {
|
---|
944 |
|
---|
945 | # split sentence into a list of tokens
|
---|
946 | $line = $_;
|
---|
947 | next unless ($line =~ /./);
|
---|
948 | @words = split(/\s+/, $line);
|
---|
949 |
|
---|
950 | # output one token at a time
|
---|
951 | foreach $word (@words) {
|
---|
952 | # don't lower-case special delimiters - jrm21
|
---|
953 | if (!map {if ($word eq $_) {1} else {()}} @delimiters) {
|
---|
954 | $word = lc($word);
|
---|
955 | }
|
---|
956 | print NUM "$symbol{$word}\n";
|
---|
957 | }
|
---|
958 |
|
---|
959 | # output phrase delimiter
|
---|
960 | print NUM "$phrasedelimiter\n";
|
---|
961 | }
|
---|
962 |
|
---|
963 | close TXT;
|
---|
964 | # print NUM "$symbol{lc($colend)}\n";# jrm21
|
---|
965 | print NUM "$symbol{$colend}\n";
|
---|
966 | close NUM;
|
---|
967 |
|
---|
968 | # Save thesaurus data in one convienient file
|
---|
969 | if ($thesaurus) {
|
---|
970 |
|
---|
971 | my $thesaurusfile = &util::filename_cat($phinddir, "$thesaurus.numbers");
|
---|
972 |
|
---|
973 |
|
---|
974 | print $out "Saving thesaurus as numbers in $thesaurusfile\n"
|
---|
975 | if ($verbosity > 1);
|
---|
976 |
|
---|
977 | # Read the thesaurus terms
|
---|
978 | my ($num, $text, %thes_symbols);
|
---|
979 |
|
---|
980 | open(TH, "<$thesaurus_terms");
|
---|
981 | while(<TH>) {
|
---|
982 | chomp;
|
---|
983 | @words = split(/\s+/, $_);
|
---|
984 | $num = shift @words;
|
---|
985 | $text = "";
|
---|
986 |
|
---|
987 | # translate words into symbol numbers
|
---|
988 | foreach $word (@words) {
|
---|
989 | $word = lc($word);
|
---|
990 | if ($symbol{$word}) {
|
---|
991 | $text .= "s$symbol{$word} ";
|
---|
992 | } elsif ($verbosity) {
|
---|
993 | print $out "Phind: No thesaurus symbol, ignoring \"$word\"\n";
|
---|
994 | }
|
---|
995 | }
|
---|
996 | $text =~ s/ $//;
|
---|
997 | $thes_symbols{$num} = $text;
|
---|
998 | }
|
---|
999 | close TH;
|
---|
1000 |
|
---|
1001 | # Read the thesaurus links and write the corresponding data
|
---|
1002 | open(TH, "<$thesaurus_links");
|
---|
1003 | open(THOUT, ">$thesaurusfile");
|
---|
1004 |
|
---|
1005 | while(<TH>) {
|
---|
1006 | chomp;
|
---|
1007 | ($num, $text) = split(/:/, $_);
|
---|
1008 |
|
---|
1009 | if (defined($thes_symbols{$num})) {
|
---|
1010 | print THOUT "$num:$thes_symbols{$num}:$text\n";
|
---|
1011 | } else {
|
---|
1012 | print THOUT "$num:untranslated:$text\n";
|
---|
1013 | }
|
---|
1014 | }
|
---|
1015 | close TH;
|
---|
1016 | close THOUT;
|
---|
1017 | }
|
---|
1018 |
|
---|
1019 |
|
---|
1020 |
|
---|
1021 |
|
---|
1022 | }
|
---|
1023 |
|
---|
1024 |
|
---|
1025 | # renumber_phrases
|
---|
1026 | #
|
---|
1027 | # Prepare the phrases file to be input to mgpp. The biggest problem is
|
---|
1028 | # reconciling the phrase identifiers used by the suffix program (which
|
---|
1029 | # we'll call suffix-id numbers) with the numbers used in the thesaurus
|
---|
1030 | # (theesaurus-id) to create a ciommon set of phind id numbers (phind-id).
|
---|
1031 | # Phind-id numbers must be sorted by frequency of occurance.
|
---|
1032 | #
|
---|
1033 | # Start creating a set of phind-id numbers from the sorted suffix-id
|
---|
1034 | # numbers and (if required) the thesaurus-id numbers. Then add any other
|
---|
1035 | # phrases occuring in the thesaurus.
|
---|
1036 | #
|
---|
1037 | # The last thing we have to do is restore the vocabulary information to the
|
---|
1038 | # phrase file so that the phrases are stored as words, not as symbol
|
---|
1039 | # numbers.
|
---|
1040 |
|
---|
1041 | # The original phrases file looks something like this:
|
---|
1042 | # 159396-1:s5175:4:1:116149-2:3:d2240,2;d2253;d2254
|
---|
1043 | # 159409-1:s5263:6:1:159410-2:6:d2122;d2128;d2129;d2130;d2215;d2380
|
---|
1044 | # 159415-1:s5267:9:1:159418-2:8:d3,2;d632;d633;d668;d1934;d2010;d2281;d2374
|
---|
1045 | # 159426-1:s5273:5:2:159429-2,115168-17:5:d252;d815;d938;d939;d2361
|
---|
1046 |
|
---|
1047 |
|
---|
1048 | sub renumber_phrases {
|
---|
1049 | my ($self) = @_;
|
---|
1050 |
|
---|
1051 | renumber_suffix_data($self);
|
---|
1052 | renumber_thesaurus_data($self);
|
---|
1053 | restore_vocabulary_data($self);
|
---|
1054 |
|
---|
1055 | }
|
---|
1056 |
|
---|
1057 |
|
---|
1058 |
|
---|
1059 | # renumber_suffix_data
|
---|
1060 | #
|
---|
1061 | # Translate phrases file to phrases.2 using phind keys instead
|
---|
1062 | # of suffix keys and sorting the expansion data.
|
---|
1063 |
|
---|
1064 | sub renumber_suffix_data {
|
---|
1065 | my ($self) = @_;
|
---|
1066 |
|
---|
1067 | my $verbosity = $self->{'verbosity'};
|
---|
1068 | my $out = $self->{'outhandle'};
|
---|
1069 | print $out "Translate phrases: suffix-ids become phind-id's\n"
|
---|
1070 | if ($verbosity);
|
---|
1071 |
|
---|
1072 | my $phinddir = $self->{'phinddir'};
|
---|
1073 | my $infile = &util::filename_cat($phinddir, 'phrases');
|
---|
1074 | my $outfile = &util::filename_cat($phinddir, 'phrases.2');
|
---|
1075 |
|
---|
1076 | # Read the phrase file. Calculate initial set of phind-id
|
---|
1077 | # numbers and store (suffixid -> frequency) relation.
|
---|
1078 |
|
---|
1079 | my %suffixtophind;
|
---|
1080 | my @totalfrequency;
|
---|
1081 | my (@fields, $suffixid);
|
---|
1082 | my $nextphind = 1;
|
---|
1083 |
|
---|
1084 | open(IN, "<$infile");
|
---|
1085 | while(<IN>) {
|
---|
1086 |
|
---|
1087 | chomp;
|
---|
1088 | @fields = split(/:/, $_);
|
---|
1089 |
|
---|
1090 | # get next suffixid and phindid
|
---|
1091 | $suffixid = shift @fields;
|
---|
1092 | $suffixtophind{$suffixid} = $nextphind;
|
---|
1093 |
|
---|
1094 | # store total frequency
|
---|
1095 | shift @fields;
|
---|
1096 | $totalfrequency[$nextphind] = shift @fields;
|
---|
1097 |
|
---|
1098 | $nextphind++;
|
---|
1099 | }
|
---|
1100 | close IN;
|
---|
1101 |
|
---|
1102 |
|
---|
1103 | # Translate phrases file to phrases.2. Use phind keys (not suffix
|
---|
1104 | # keys), sort expansion and document occurance data in order of
|
---|
1105 | # descending frequency..
|
---|
1106 | open(IN, "<$infile");
|
---|
1107 | open(OUT, ">$outfile");
|
---|
1108 |
|
---|
1109 | my ($phindid, $text, $tf, $countexp, $expansions, $countdocs, $documents);
|
---|
1110 | my (@documwents, @newexp, $k, $n);
|
---|
1111 | my $linenumber = 0;
|
---|
1112 |
|
---|
1113 | while(<IN>) {
|
---|
1114 |
|
---|
1115 | # read the line
|
---|
1116 | chomp;
|
---|
1117 | @fields = split(/:/, $_);
|
---|
1118 |
|
---|
1119 | # get a phrase number for this line
|
---|
1120 | $suffixid = shift @fields;
|
---|
1121 | die unless (defined($suffixtophind{$suffixid}));
|
---|
1122 | $phindid = $suffixtophind{$suffixid};
|
---|
1123 |
|
---|
1124 | # get the symbols in the phrase
|
---|
1125 | $text = shift @fields;
|
---|
1126 |
|
---|
1127 | # output status information
|
---|
1128 | $linenumber++;
|
---|
1129 | if ($verbosity > 2) {
|
---|
1130 | if ($linenumber % 1000 == 0) {
|
---|
1131 | print $out "line $linenumber:\t$phindid\t$suffixid\t($text)\n";
|
---|
1132 | }
|
---|
1133 | # what are $num and $key??
|
---|
1134 | #print $out "$num: $key\t($text)\n" if ($verbosity > 3);
|
---|
1135 | }
|
---|
1136 |
|
---|
1137 | # get the phrase frequency
|
---|
1138 | $tf = shift @fields;
|
---|
1139 |
|
---|
1140 | # get the number of expansions
|
---|
1141 | $countexp = shift @fields;
|
---|
1142 |
|
---|
1143 | # get the expansions, convert them into phind-id numbers, and sort them
|
---|
1144 | $expansions = shift @fields;
|
---|
1145 | @newexp = ();
|
---|
1146 | foreach $k (split(/,/, $expansions)) {
|
---|
1147 | die "ERROR - no phindid for: $k" unless (defined($suffixtophind{$k}));
|
---|
1148 | $n = $suffixtophind{$k};
|
---|
1149 | push @newexp, $n;
|
---|
1150 | }
|
---|
1151 | @newexp = sort {$totalfrequency[$b] <=> $totalfrequency[$a]} @newexp;
|
---|
1152 |
|
---|
1153 | # get the number of documents
|
---|
1154 | $countdocs = shift @fields;
|
---|
1155 |
|
---|
1156 | # get the documents and sort them
|
---|
1157 | $documents = shift @fields;
|
---|
1158 | $documents =~ s/d//g;
|
---|
1159 | my @documents = split(/;/, $documents);
|
---|
1160 | @documents = sort by_doc_frequency @documents;
|
---|
1161 |
|
---|
1162 | # output the phrase data
|
---|
1163 | print OUT "$phindid:$text:$tf:$countexp:$countdocs:";
|
---|
1164 | print OUT join(",", @newexp), ",:", join(";", @documents), ";\n";
|
---|
1165 |
|
---|
1166 | }
|
---|
1167 |
|
---|
1168 | close IN;
|
---|
1169 | close OUT;
|
---|
1170 | }
|
---|
1171 |
|
---|
1172 |
|
---|
1173 | # renumber_thesaurus_data
|
---|
1174 | #
|
---|
1175 | # Translate phrases.2 to phrases.3, adding thesaurus data if available.
|
---|
1176 |
|
---|
1177 | sub renumber_thesaurus_data {
|
---|
1178 | my ($self) = @_;
|
---|
1179 |
|
---|
1180 | my $out = $self->{'outhandle'};
|
---|
1181 | my $verbosity = $self->{'verbosity'};
|
---|
1182 | my $thesaurus = $self->{'thesaurus'};
|
---|
1183 |
|
---|
1184 | my $phinddir = $self->{'phinddir'};
|
---|
1185 | my $infile = &util::filename_cat($phinddir, "phrases.2");
|
---|
1186 | my $outfile = &util::filename_cat($phinddir, "phrases.3");
|
---|
1187 |
|
---|
1188 |
|
---|
1189 | # If no thesaurus is defined, simply move the phrases file.
|
---|
1190 | if (!$thesaurus) {
|
---|
1191 | print $out "Translate phrases.2: no thesaurus data\n"
|
---|
1192 | if ($verbosity);
|
---|
1193 | &util::mv($infile, $outfile);
|
---|
1194 | return;
|
---|
1195 | }
|
---|
1196 |
|
---|
1197 | print $out "Translate phrases.2: add thesaurus data\n"
|
---|
1198 | if ($verbosity);
|
---|
1199 |
|
---|
1200 | # 1.
|
---|
1201 | # Read thesaurus file and store (symbols->thesaurusid) mapping
|
---|
1202 | my $thesaurusfile = &util::filename_cat($phinddir, "$thesaurus.numbers");
|
---|
1203 | my %symbolstothesid;
|
---|
1204 | my (@fields, $thesid, $symbols);
|
---|
1205 |
|
---|
1206 | open(TH, "<$thesaurusfile");
|
---|
1207 |
|
---|
1208 | while (<TH>) {
|
---|
1209 |
|
---|
1210 | chomp;
|
---|
1211 | @fields = split(/:/, $_);
|
---|
1212 |
|
---|
1213 | # get id and text
|
---|
1214 | $thesid = shift @fields;
|
---|
1215 | $symbols = shift @fields;
|
---|
1216 | $symbolstothesid{$symbols} = $thesid;
|
---|
1217 | }
|
---|
1218 | close TH;
|
---|
1219 |
|
---|
1220 | # 2.
|
---|
1221 | # Read phrases file to find thesaurus entries that already
|
---|
1222 | # have a phindid. Store their phind-ids for later translation,
|
---|
1223 | # and store their frequency for later sorting.
|
---|
1224 | my %thesaurustophindid;
|
---|
1225 | my %phindidtofrequency;
|
---|
1226 | my ($phindid, $freq);
|
---|
1227 |
|
---|
1228 | open(IN, "<$infile");
|
---|
1229 |
|
---|
1230 | while(<IN>) {
|
---|
1231 |
|
---|
1232 | chomp;
|
---|
1233 | @fields = split(/:/, $_);
|
---|
1234 |
|
---|
1235 | # phindid and symbols for this line
|
---|
1236 | $phindid = shift @fields;
|
---|
1237 | $symbols = shift @fields;
|
---|
1238 | $freq = shift @fields;
|
---|
1239 |
|
---|
1240 | # do we have a thesaurus id corresponding to this phrase?
|
---|
1241 | if (defined($symbolstothesid{$symbols})) {
|
---|
1242 | $thesid = $symbolstothesid{$symbols};
|
---|
1243 | $thesaurustophindid{$thesid} = $phindid;
|
---|
1244 | $phindidtofrequency{$phindid} = $freq;
|
---|
1245 | }
|
---|
1246 | }
|
---|
1247 | close IN;
|
---|
1248 |
|
---|
1249 | undef %symbolstothesid;
|
---|
1250 |
|
---|
1251 | # 3.
|
---|
1252 | # Create phind-id numbers for remaining thesaurus entries,
|
---|
1253 | # and note that their frequency is 0 for later sorting.
|
---|
1254 | my $nextphindid = $phindid + 1;
|
---|
1255 |
|
---|
1256 | open(TH, "<$thesaurusfile");
|
---|
1257 | while(<TH>) {
|
---|
1258 |
|
---|
1259 | chomp;
|
---|
1260 | @fields = split(/:/, $_);
|
---|
1261 |
|
---|
1262 | # read thesaurus-id and ensure it has a corresponding phind-id
|
---|
1263 | $thesid = shift @fields;
|
---|
1264 | if (!defined($thesaurustophindid{$thesid})) {
|
---|
1265 | $thesaurustophindid{$thesid} = $nextphindid;
|
---|
1266 | $phindidtofrequency{$nextphindid} = 0;
|
---|
1267 | $nextphindid++;
|
---|
1268 | }
|
---|
1269 | }
|
---|
1270 | close TH;
|
---|
1271 |
|
---|
1272 | # 4.
|
---|
1273 | # Translate thesaurus file, replacing thesaurus-id numbers with
|
---|
1274 | # phind-id numbers.
|
---|
1275 | my $newthesaurusfile = &util::filename_cat($phinddir, "$thesaurus.phindid");
|
---|
1276 | my ($relations, $linkcounter, $linktext, $linktype, @linkdata);
|
---|
1277 | my (@links, $linkid, %linkidtotype, $newrelation);
|
---|
1278 |
|
---|
1279 | open(TH, "<$thesaurusfile");
|
---|
1280 | open(TO, ">$newthesaurusfile");
|
---|
1281 | while(<TH>) {
|
---|
1282 |
|
---|
1283 | chomp;
|
---|
1284 | @fields = split(/:/, $_);
|
---|
1285 |
|
---|
1286 | # phindid and symbols for this line
|
---|
1287 | ($thesid, $symbols, $relations) = @fields;
|
---|
1288 |
|
---|
1289 | die unless ($thesid && $symbols);
|
---|
1290 | die unless $thesaurustophindid{$thesid};
|
---|
1291 | $phindid = $thesaurustophindid{$thesid};
|
---|
1292 |
|
---|
1293 | # convert each part of the relation string to use phind-id numbers
|
---|
1294 | # at the same time, we want to sort the list by frequency.
|
---|
1295 | undef %linkidtotype;
|
---|
1296 |
|
---|
1297 | foreach $linktext (split(/;/, $relations)) {
|
---|
1298 | @linkdata = split(/,/, $linktext);
|
---|
1299 |
|
---|
1300 | # remember the linktype (e.g. BT, NT)
|
---|
1301 | $linktype = shift @linkdata;
|
---|
1302 |
|
---|
1303 | # store the type of each link
|
---|
1304 | foreach $thesid (@linkdata) {
|
---|
1305 | die unless (defined($thesaurustophindid{$thesid}));
|
---|
1306 | $linkidtotype{$thesaurustophindid{$thesid}} = $linktype;
|
---|
1307 | }
|
---|
1308 | }
|
---|
1309 |
|
---|
1310 | # sort the list of links, first by frequency, then by type.
|
---|
1311 | @links = sort { ($phindidtofrequency{$b} <=> $phindidtofrequency{$a})
|
---|
1312 | or ($linkidtotype{$a} cmp $linkidtotype{$b}) } (keys %linkidtotype);
|
---|
1313 | $linkcounter = (scalar @links);
|
---|
1314 |
|
---|
1315 | # create a string describing the link information
|
---|
1316 | $linktype = $linkidtotype{$links[0]};
|
---|
1317 | $newrelation = $linktype;
|
---|
1318 | foreach $linkid (@links) {
|
---|
1319 | if ($linkidtotype{$linkid} ne $linktype) {
|
---|
1320 | $linktype = $linkidtotype{$linkid};
|
---|
1321 | $newrelation .= ";" . $linktype;
|
---|
1322 | }
|
---|
1323 | $newrelation .= "," . $linkid;
|
---|
1324 | }
|
---|
1325 | $newrelation .= ";";
|
---|
1326 |
|
---|
1327 |
|
---|
1328 | # output the new line
|
---|
1329 | print TO "$phindid:$symbols:$linkcounter:$newrelation:\n";
|
---|
1330 | }
|
---|
1331 | close TH;
|
---|
1332 | close TO;
|
---|
1333 |
|
---|
1334 | undef %thesaurustophindid;
|
---|
1335 | undef %linkidtotype;
|
---|
1336 | undef %phindidtofrequency;
|
---|
1337 |
|
---|
1338 | # 5.
|
---|
1339 | # Read thesaurus data (in phind-id format) into memory
|
---|
1340 | my %thesaurusdata;
|
---|
1341 |
|
---|
1342 | open(TH, "<$newthesaurusfile");
|
---|
1343 | while(<TH>) {
|
---|
1344 | chomp;
|
---|
1345 | ($phindid, $symbols, $linkcounter, $relations) = split(/:/, $_);
|
---|
1346 | die unless ($phindid && $symbols);
|
---|
1347 | $thesaurusdata{$phindid} = "$symbols:$linkcounter:$relations";
|
---|
1348 | }
|
---|
1349 | close TH;
|
---|
1350 |
|
---|
1351 | # 6.
|
---|
1352 | # Add thesaurus data to phrases file
|
---|
1353 | my ($text, $tf, $countexp, $expansions, $countdocs, $documents);
|
---|
1354 | my (@documwents, @newexp, $k, $n);
|
---|
1355 | my $linenumber = 0;
|
---|
1356 |
|
---|
1357 | open(IN, "<$infile");
|
---|
1358 | open(OUT, ">$outfile");
|
---|
1359 |
|
---|
1360 | # Update existing phrases
|
---|
1361 | while(<IN>) {
|
---|
1362 |
|
---|
1363 | chomp;
|
---|
1364 | @fields = split(/:/, $_);
|
---|
1365 |
|
---|
1366 | # get data for this line
|
---|
1367 | $phindid = shift @fields;
|
---|
1368 |
|
---|
1369 | # output the phrase data, with thesaurus information
|
---|
1370 | print OUT "$phindid:", join(":", @fields);
|
---|
1371 |
|
---|
1372 | # add thesaurus data
|
---|
1373 | if (defined($thesaurusdata{$phindid})) {
|
---|
1374 | @fields = split(/:/, $thesaurusdata{$phindid});
|
---|
1375 | shift @fields;
|
---|
1376 | $linkcounter = shift @fields;
|
---|
1377 | $relations = shift @fields;
|
---|
1378 |
|
---|
1379 | print OUT ":$linkcounter:$relations";
|
---|
1380 | $thesaurusdata{$phindid} = "";
|
---|
1381 | }
|
---|
1382 | print OUT "\n";
|
---|
1383 | }
|
---|
1384 | close IN;
|
---|
1385 |
|
---|
1386 | # Add phrases that aren't already in the file
|
---|
1387 | foreach $phindid (sort numerically keys %thesaurusdata) {
|
---|
1388 | next unless ($thesaurusdata{$phindid});
|
---|
1389 |
|
---|
1390 | @fields = split(/:/, $thesaurusdata{$phindid});
|
---|
1391 | $symbols = shift @fields;
|
---|
1392 | $linkcounter = shift @fields;
|
---|
1393 | $relations = shift @fields;
|
---|
1394 |
|
---|
1395 | print OUT "$phindid:$symbols:0:0:0:::$linkcounter:$relations\n";
|
---|
1396 | }
|
---|
1397 | close OUT;
|
---|
1398 |
|
---|
1399 | }
|
---|
1400 |
|
---|
1401 | # restore_vocabulary_data
|
---|
1402 | #
|
---|
1403 | # Read phrases.3 and restore vocabulary information. Then write
|
---|
1404 | # this data to the MGPP input files (pword.txt and pdata.txt) and
|
---|
1405 | # (if requested) to the saved phrases file.
|
---|
1406 |
|
---|
1407 | sub restore_vocabulary_data {
|
---|
1408 | my ($self) = @_;
|
---|
1409 |
|
---|
1410 | my $out = $self->{'outhandle'};
|
---|
1411 | my $verbosity = $self->{'verbosity'};
|
---|
1412 | print $out "Translate phrases.3: restore vocabulary\n" if ($verbosity);
|
---|
1413 |
|
---|
1414 | my $phinddir = $self->{'phinddir'};
|
---|
1415 | my $infile = &util::filename_cat($phinddir, 'phrases.3');
|
---|
1416 | my $vocabfile = &util::filename_cat($phinddir, 'clauses.vocab');
|
---|
1417 | my $datafile = &util::filename_cat($phinddir, 'pdata.txt');
|
---|
1418 | my $wordfile = &util::filename_cat($phinddir, 'pword.txt');
|
---|
1419 |
|
---|
1420 | my $savephrases = $self->{'savephrases'};
|
---|
1421 |
|
---|
1422 | # 1.
|
---|
1423 | # Read the vocabulary file
|
---|
1424 | open(V, "<$vocabfile")
|
---|
1425 | || die "Cannot open $vocabfile: $!";
|
---|
1426 | my @symbol;
|
---|
1427 | my $i = 1;
|
---|
1428 | while(<V>) {
|
---|
1429 | chomp;
|
---|
1430 | $symbol[$i++] = $_;
|
---|
1431 | }
|
---|
1432 | close V;
|
---|
1433 |
|
---|
1434 | # 2.
|
---|
1435 | # Translate phrases.3 to MGPP input files
|
---|
1436 | my ($key, $text, $word, $isThesaurus, $line);
|
---|
1437 | my @fields;
|
---|
1438 | my $linenumber = 0;
|
---|
1439 |
|
---|
1440 | open(IN, "<$infile");
|
---|
1441 | open(DATA, ">$datafile");
|
---|
1442 | open(WORD, ">$wordfile");
|
---|
1443 |
|
---|
1444 | # Save the phrases in a separate text file
|
---|
1445 | if ($savephrases) {
|
---|
1446 | print $out "Saving phrases in $savephrases\n" if ($verbosity);
|
---|
1447 | open(SAVE, ">$savephrases");
|
---|
1448 | }
|
---|
1449 |
|
---|
1450 | while(<IN>) {
|
---|
1451 |
|
---|
1452 | # read the line
|
---|
1453 | chomp;
|
---|
1454 | $line = $_;
|
---|
1455 | @fields = split(/:/, $line);
|
---|
1456 |
|
---|
1457 | # get a phrase number for this line
|
---|
1458 | $key = shift @fields;
|
---|
1459 |
|
---|
1460 | # restore the text of the phrase
|
---|
1461 | $text = shift @fields;
|
---|
1462 | $text =~ s/s(\d+)/$symbol[$1]/g;
|
---|
1463 | if ($text =~ / /) {
|
---|
1464 | $word = "";
|
---|
1465 | } elsif ($text ne 'untranslated') {
|
---|
1466 | $word = $text;
|
---|
1467 | }
|
---|
1468 |
|
---|
1469 | # output the phrase data
|
---|
1470 | print DATA "<Document>";
|
---|
1471 | print DATA "$key:$text:", join(":", @fields), ":\n";
|
---|
1472 |
|
---|
1473 | # output the word index search data
|
---|
1474 | print WORD "<Document>$word\n";
|
---|
1475 |
|
---|
1476 | # output the phrases to a text file
|
---|
1477 | if ($savephrases) {
|
---|
1478 | if ((scalar @fields) == 7) {
|
---|
1479 | $isThesaurus = 1;
|
---|
1480 | } else {
|
---|
1481 | $isThesaurus = 0;
|
---|
1482 | }
|
---|
1483 | print SAVE $fields[0], "\t", $fields[2], "\t$isThesaurus\t$text\n";
|
---|
1484 | }
|
---|
1485 | }
|
---|
1486 | close IN;
|
---|
1487 | close WORD;
|
---|
1488 | close DATA;
|
---|
1489 | close SAVE if ($savephrases);
|
---|
1490 |
|
---|
1491 | }
|
---|
1492 |
|
---|
1493 |
|
---|
1494 |
|
---|
1495 | # sort routines used to renumber phrases
|
---|
1496 |
|
---|
1497 | sub numerically { $a <=> $b }
|
---|
1498 |
|
---|
1499 | sub by_doc_frequency {
|
---|
1500 | my $fa = 1;
|
---|
1501 | if ($a =~ /,/) {
|
---|
1502 | $fa = $a;
|
---|
1503 | $fa =~ s/\d+,//;
|
---|
1504 | }
|
---|
1505 | my $fb = 1;
|
---|
1506 | if ($b =~ /,/) {
|
---|
1507 | $fb = $b;
|
---|
1508 | $fb =~ s/\d+,//;
|
---|
1509 | }
|
---|
1510 |
|
---|
1511 | return ($fb <=> $fa);
|
---|
1512 | }
|
---|
1513 |
|
---|
1514 | 1;
|
---|