[1643] | 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 | #
|
---|
| 31 | # options are:
|
---|
| 32 | # title=Title The title field for this classification
|
---|
| 33 | # text=fields The text used to build the phrase hierarchy
|
---|
| 34 | # phindexdir=directory Location of phind index files
|
---|
| 35 | # verbosity=num Control amount of output
|
---|
[1646] | 36 | # untidy=true Do not clean up intermediate files
|
---|
| 37 | # suffixmode=num Mode of suffix program (0 = all phrases, 1 = stopword)
|
---|
| 38 | # suffixsize=num Number of symbols available to suffix program
|
---|
[1643] | 39 |
|
---|
| 40 |
|
---|
| 41 | # How a classifier works.
|
---|
| 42 | #
|
---|
| 43 | # When a classifier is requested in the collect.cfg file, buildcol creates a
|
---|
| 44 | # new classifier object (such as the one defined in theis file) and later
|
---|
| 45 | # passes each document object to the classifier in turn. Four functions are
|
---|
| 46 | # used:
|
---|
| 47 | #
|
---|
| 48 | # 1. "new" is called before the documents are processed to set up the
|
---|
| 49 | # classifier.
|
---|
| 50 | #
|
---|
| 51 | # 2. "init" is called after buildcol.pl has created the indexes etc but
|
---|
| 52 | # before the documents are classified in order that the classifier might
|
---|
| 53 | # set any varioables it requiers, etc.
|
---|
| 54 | #
|
---|
| 55 | # 3. "classify" is called once for each document object. The classifier
|
---|
| 56 | # "classifies" each document and updates its local data accordingly.
|
---|
| 57 | #
|
---|
| 58 | # 4. "get_classify_info" is called after every document has been
|
---|
| 59 | # classified. It collates the information about the documents and
|
---|
| 60 | # stores a reference to the classifier so that Greenstone can later
|
---|
| 61 | # display it.
|
---|
| 62 |
|
---|
| 63 |
|
---|
| 64 | package phind;
|
---|
| 65 |
|
---|
| 66 | use BasClas;
|
---|
| 67 | use util;
|
---|
| 68 |
|
---|
| 69 | sub BEGIN {
|
---|
| 70 | @ISA = ('BasClas');
|
---|
| 71 | }
|
---|
| 72 |
|
---|
| 73 | # Define delimiter symbols - this should be abstracted out someplace
|
---|
| 74 | my $colstart = "COLLECTIONSTART";
|
---|
| 75 | my $colend = "COLLECTIONEND";
|
---|
| 76 | my $doclimit = "DOCUMENTLIMIT";
|
---|
| 77 | my $senlimit = "SENTENCELIMIT";
|
---|
| 78 | my @delimiters = ($colstart, $colend, $doclimit, $senlimit);
|
---|
| 79 |
|
---|
| 80 | # Create a new phind browser based on the options in collect.cfg
|
---|
| 81 |
|
---|
| 82 | sub new {
|
---|
| 83 | my ($class, @options) = @_;
|
---|
| 84 | my $self = new BasClas ($class, @_);
|
---|
| 85 | my $out = pop @options;
|
---|
| 86 |
|
---|
| 87 | # Phind installation check
|
---|
| 88 | # The phind phrase browser is research software and is not installed
|
---|
| 89 | # by defualt. If the user attepts to use it we warn them that it's a
|
---|
| 90 | # bit dodgy, then tell them how to install it. If they can do that
|
---|
| 91 | # and get all the files in place, then we let them proceed.
|
---|
| 92 |
|
---|
| 93 | print $out "The Phind classifier for Greenstone.\n";
|
---|
| 94 | print $out "Checking the phind phrase browser requirements...\n";
|
---|
| 95 |
|
---|
| 96 | # Make sure we're not in windows
|
---|
| 97 | if ($ENV{'GSDLOS'} =~ /windows/i) {
|
---|
| 98 | print STDERR "Phind currently only works under Unix";
|
---|
| 99 | exit(1);
|
---|
| 100 | }
|
---|
| 101 |
|
---|
| 102 | # Ensure the Phind generate scripts are in place
|
---|
| 103 | my $file1 = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "suffix");
|
---|
| 104 | my $src = &util::filename_cat($ENV{'GSDLHOME'}, "src", "phind", "generate");
|
---|
| 105 |
|
---|
| 106 | if (!(-e $file1)) {
|
---|
| 107 | print STDERR "The phind \"suffix\" program is not installed. ";
|
---|
| 108 | print STDERR "To install it, change to the directory\n";
|
---|
| 109 | print STDERR " $src\n";
|
---|
| 110 | print STDERR "and type \"make install-phind\".\n\n";
|
---|
| 111 | exit(1);
|
---|
| 112 | }
|
---|
| 113 |
|
---|
| 114 | # Ensure the Phind CGI script is in place
|
---|
| 115 | $file1 = &util::filename_cat($ENV{'GSDLHOME'}, "cgi-bin", "phindcgi");
|
---|
| 116 | $src = &util::filename_cat($ENV{'GSDLHOME'}, "src", "phind", "host");
|
---|
| 117 |
|
---|
| 118 | if (!(-e $file1)) {
|
---|
| 119 | print STDERR "The phind CGI program is not installed. ";
|
---|
| 120 | print STDERR "To install it, change to the directory\n";
|
---|
| 121 | print STDERR " $src\n";
|
---|
| 122 | print STDERR "and type \"make install-phind\".\n\n";
|
---|
| 123 | exit(1);
|
---|
| 124 | }
|
---|
| 125 |
|
---|
| 126 | # Ensure the Phind Java applet is in place
|
---|
| 127 | $src = &util::filename_cat($ENV{'GSDLHOME'}, "src", "phind", "client");
|
---|
| 128 | $file1 = &util::filename_cat($src, "Phind.class");
|
---|
| 129 |
|
---|
| 130 | if (!(-e $file1)) {
|
---|
| 131 | print STDERR "The phind Java classes are not compiled. ";
|
---|
| 132 | print STDERR "To compile them, change to the directory\n";
|
---|
| 133 | print STDERR " $src\n";
|
---|
| 134 | print STDERR "and use your Java compiler to compile Phind.java.\n";
|
---|
| 135 | print STDERR "(if you have Java 1.2 installed, type \"javac Phind.java\")\n\n";
|
---|
| 136 | exit(1);
|
---|
| 137 | }
|
---|
| 138 |
|
---|
| 139 |
|
---|
| 140 | # The installation appears OK - set up the classifier
|
---|
| 141 | my $collection = $ENV{'GSDLCOLLECTION'};
|
---|
| 142 | my $phindexdir = &util::filename_cat($ENV{'GSDLCOLLECTDIR'},"phindex");
|
---|
| 143 | my $language = "english";
|
---|
| 144 |
|
---|
| 145 | my $title = "Topic";
|
---|
| 146 | my $indexes = "section:Title,section:text";
|
---|
| 147 |
|
---|
[1646] | 148 | my $suffixmode = 1;
|
---|
| 149 | my $suffixsize = 40000000;
|
---|
| 150 |
|
---|
[1643] | 151 | my $verbosity = 2;
|
---|
| 152 | my $untidy = 0;
|
---|
| 153 |
|
---|
| 154 | # parse the options
|
---|
| 155 | foreach $option (@options) {
|
---|
| 156 |
|
---|
| 157 | if ($option =~ /^text=(.*)$/i) {
|
---|
| 158 | $indexes = $1;
|
---|
| 159 | } elsif ($option =~ /^title=(.*)$/i) {
|
---|
| 160 | $title = $1;
|
---|
| 161 | } elsif ($option =~ /^phindexdir=(.*)$/i) {
|
---|
| 162 | $phindexdir = $1;
|
---|
[1646] | 163 | } elsif ($option =~ /^suffixsize=(.*)$/i) {
|
---|
| 164 | $suffixsize = $1;
|
---|
| 165 | } elsif ($option =~ /^suffixmode=(.*)$/i) {
|
---|
| 166 | $suffixmode = $1;
|
---|
[1643] | 167 | } elsif ($option =~ /^verbosity=(.*)$/i) {
|
---|
| 168 | $verbosity = $1;
|
---|
| 169 | } elsif ($option =~ /^untidy/i) {
|
---|
| 170 | $untidy = 1;
|
---|
| 171 | }
|
---|
| 172 | }
|
---|
| 173 |
|
---|
| 174 |
|
---|
| 175 | $self->{'collection'} = $collection;
|
---|
| 176 | $self->{'title'} = $title;
|
---|
| 177 | $self->{'indexes'} = $indexes;
|
---|
[1646] | 178 |
|
---|
| 179 | $self->{'suffixmode'} = $suffixmode;
|
---|
| 180 | $self->{'suffixsize'} = $suffixsize;
|
---|
| 181 |
|
---|
[1643] | 182 | $self->{'verbosity'} = $verbosity;
|
---|
| 183 | $self->{'untidy'} = $untidy;
|
---|
| 184 |
|
---|
| 185 | # limit languages
|
---|
| 186 | $language =~ s/,/\|/g;
|
---|
| 187 | $self->{'language_exp'} = $language;
|
---|
| 188 | $self->{'delimiter'} = $delimiter;
|
---|
| 189 |
|
---|
| 190 | # reset phindex directory
|
---|
| 191 | if (-e "$phindexdir") {
|
---|
| 192 | &util::rm_r("$phindexdir");
|
---|
| 193 | }
|
---|
| 194 | &util::mk_dir("$phindexdir");
|
---|
| 195 | $self->{'phindexdir'} = $phindexdir;
|
---|
| 196 |
|
---|
| 197 | return bless $self, $class;
|
---|
| 198 | }
|
---|
| 199 |
|
---|
| 200 |
|
---|
| 201 | # Initialise the phind classifier
|
---|
| 202 |
|
---|
| 203 | sub init {
|
---|
| 204 | my $self = shift (@_);
|
---|
| 205 |
|
---|
| 206 | # open filehandles for documents and text
|
---|
| 207 | my $phindexdir = $self->{'phindexdir'};
|
---|
| 208 |
|
---|
| 209 | my $clausefile = &util::filename_cat("$phindexdir", "clauses");
|
---|
| 210 | &util::rm($clausefile) if (-e $clausefile);
|
---|
| 211 | open(TEXT, ">$clausefile") || die "Cannot open $clausefile: $!";
|
---|
| 212 | $self->{'txthandle'} = TEXT;
|
---|
| 213 |
|
---|
| 214 | my $docfile = &util::filename_cat("$phindexdir", "docs.txt");
|
---|
| 215 | &util::rm($docfile) if (-e $docfile);
|
---|
| 216 | open(DOCS, ">$docfile") || die "Cannot open $docfile: $!";
|
---|
| 217 | $self->{'dochandle'} = DOCS;
|
---|
| 218 |
|
---|
| 219 | }
|
---|
| 220 |
|
---|
| 221 | # Classify each document.
|
---|
| 222 | #
|
---|
| 223 | # Each document is passed here in turn. The classifier extracts the
|
---|
| 224 | # text of each and stores it in the clauses file. Document details are
|
---|
| 225 | # stored in the docs.txt file.
|
---|
| 226 |
|
---|
| 227 | sub classify {
|
---|
| 228 | my $self = shift (@_);
|
---|
| 229 | my ($doc_obj) = @_;
|
---|
| 230 |
|
---|
| 231 | my $verbosity = $self->{'verbosity'};
|
---|
| 232 | my $top_section = $doc_obj->get_top_section();
|
---|
| 233 |
|
---|
| 234 | my $title = $doc_obj->get_metadata_element ($top_section, "Title");
|
---|
| 235 | print "process: $title\n" if ($verbosity > 2);
|
---|
| 236 |
|
---|
| 237 |
|
---|
| 238 | # only consider english-language files
|
---|
| 239 | my $doclanguage = $doc_obj->get_metadata_element ($top_section, "Language");
|
---|
| 240 | my $phrlanguage = $self->{'language_exp'};
|
---|
| 241 | return if ($doclanguage && ($doclanguage !~ /$phrlanguage/i));
|
---|
| 242 |
|
---|
| 243 | # record this file
|
---|
| 244 | my $total++;
|
---|
| 245 | print "file $total: $file\n" if ($self->{'$verbosity'});
|
---|
| 246 |
|
---|
| 247 |
|
---|
| 248 | # Store document details
|
---|
| 249 | my $OID = $doc_obj->get_OID();
|
---|
| 250 | $OID = "NULL" unless defined $OID;
|
---|
| 251 | my $dochandle = $self->{'dochandle'};
|
---|
| 252 | print $dochandle "<Document>\t$OID\t$title\n";
|
---|
| 253 |
|
---|
| 254 | # Store the text occuring in this object
|
---|
| 255 |
|
---|
| 256 | # output the document delimiter
|
---|
| 257 | my $txthandle = $self->{'txthandle'};
|
---|
| 258 | print $txthandle "$doclimit\n";
|
---|
| 259 |
|
---|
| 260 | # iterarate over the required indexes and store their text
|
---|
| 261 | my $indexes = $self->{'indexes'};
|
---|
| 262 | my $text = "";
|
---|
| 263 | my ($part, $level, $field, $section, $data);
|
---|
| 264 |
|
---|
| 265 | foreach $part (split(/,/, $indexes)) {
|
---|
| 266 |
|
---|
| 267 | # Each field has a level and a data element ((e.g. document:Title)
|
---|
| 268 | ($level, $field) = split(/:/, $part);
|
---|
| 269 | die unless ($level && $field);
|
---|
| 270 |
|
---|
| 271 | # Extract the text from every section
|
---|
| 272 | # (In phind, document:text and section:text are equivalent)
|
---|
| 273 | if ($field eq "text") {
|
---|
| 274 | $data = "";
|
---|
| 275 | $section = $doc_obj->get_top_section();
|
---|
| 276 | while (defined($section)) {
|
---|
| 277 | $data .= $doc_obj->get_text($section) . "\n";
|
---|
| 278 | $section = $doc_obj->get_next_section($section);
|
---|
| 279 | }
|
---|
| 280 | $text .= convert_gml_to_tokens($data) . "\n";
|
---|
| 281 | }
|
---|
| 282 |
|
---|
| 283 | # Extract a metadata field from a document
|
---|
| 284 | elsif ($level eq "document") {
|
---|
| 285 | $data = $doc_obj->get_metadata_element($doc_obj->get_top_section(), $field);
|
---|
| 286 | $text .= convert_gml_to_tokens($data) . "\n";
|
---|
| 287 | }
|
---|
| 288 |
|
---|
| 289 | # Extract metadata from every section in a document
|
---|
| 290 | elsif ($level eq "section") {
|
---|
| 291 | $data = "";
|
---|
| 292 | $section = $doc_obj->get_top_section();
|
---|
| 293 | while (defined($section)) {
|
---|
| 294 | $data .= $doc_obj->get_metadata_element($section, $field) . "\n";
|
---|
| 295 | $section = $doc_obj->get_next_section($section);
|
---|
| 296 | }
|
---|
| 297 | $text .= convert_gml_to_tokens($data) . "\n";
|
---|
| 298 | }
|
---|
| 299 |
|
---|
| 300 | # Some sort of specification which I don't understand
|
---|
| 301 | else {
|
---|
| 302 | die "Unknown level ($level) in phind index ($part)\n";
|
---|
| 303 | }
|
---|
| 304 |
|
---|
| 305 | }
|
---|
| 306 |
|
---|
| 307 | # output the text
|
---|
| 308 | $text =~ tr/\n//s;
|
---|
| 309 | print $txthandle "$text";
|
---|
| 310 |
|
---|
| 311 | }
|
---|
| 312 |
|
---|
| 313 |
|
---|
| 314 |
|
---|
| 315 | # Construct the classifier from the information already gathered
|
---|
| 316 | #
|
---|
| 317 | # When get_classify_info is called, the clauses and docs.txt files have
|
---|
| 318 | # already been constructed in the phindex directory. This function will
|
---|
| 319 | # translate them into compressed, indexed MGPP files that can be read by
|
---|
| 320 | # the phindcgi script. It will also register our classifier so that it
|
---|
| 321 | # shows up in thenavigation bar.
|
---|
| 322 |
|
---|
| 323 | sub get_classify_info {
|
---|
| 324 | my $self = shift (@_);
|
---|
| 325 |
|
---|
| 326 | my $verbosity = $self->{'verbosity'};
|
---|
| 327 | my $phindexdir = $self->{'phindexdir'};
|
---|
| 328 | my $language = "english";
|
---|
| 329 |
|
---|
[1645] | 330 | if ($verbosity) {
|
---|
| 331 | print STDERR "\n*** phind.pm generating indexes for ", $self->{'indexes'}, "\n";
|
---|
| 332 | }
|
---|
| 333 |
|
---|
[1643] | 334 | # Construct phind indexes
|
---|
[1646] | 335 | my $suffixmode = $self->{'suffixmode'};
|
---|
| 336 | my $suffixsize = $self->{'suffixsize'};
|
---|
[1643] | 337 | my ($command, $status);
|
---|
| 338 |
|
---|
| 339 | # Generate the vocabulary, symbol statistics, and numbers file
|
---|
| 340 | # from the clauses file
|
---|
| 341 | print "\nExtracting vocabulary and statistics\n" if $verbosity;
|
---|
| 342 | &extract_vocabulary($phindexdir, $language, $verbosity);
|
---|
| 343 |
|
---|
| 344 | # Use the suffix program to generate the phindex/phrases file
|
---|
[1645] | 345 | print "\nExtracting phrases from processed text (with suffix)\n" if $verbosity;
|
---|
[1646] | 346 | &execute("suffix $phindexdir $suffixsize $suffixmode", $verbosity);
|
---|
[1643] | 347 |
|
---|
| 348 | # Create the phrase file and put phrase numbers in phindex/phrases
|
---|
| 349 | print "\nSorting and Renumbering phrases for input to mgpp\n" if $verbosity;
|
---|
| 350 | &renumber_phrases("$phindexdir", $verbosity);
|
---|
| 351 |
|
---|
| 352 | # Create the mg phrase database
|
---|
| 353 | my $mgpp = &util::filename_cat($ENV{'GSDLHOME'}, "src", "mgpp");
|
---|
| 354 | my $mg_passes = &util::filename_cat($mgpp, "text", "mg_passes");
|
---|
| 355 | my $mg_compression_dict = &util::filename_cat($mgpp, "text", "mg_compression_dict");
|
---|
| 356 |
|
---|
| 357 | my $mg_perf_hash_build = &util::filename_cat($mgpp, "text", "mg_perf_hash_build");
|
---|
| 358 | my $mg_weights_build = &util::filename_cat($mgpp, "text", "mg_weights_build");
|
---|
| 359 | my $mg_invf_dict = &util::filename_cat($mgpp, "text", "mg_invf_dict");
|
---|
| 360 | my $mg_stem_idx = &util::filename_cat($mgpp, "text", "mg_stem_idx");
|
---|
| 361 |
|
---|
| 362 | print "\nCreating phrase databases\n";
|
---|
| 363 | my $mg_input = &util::filename_cat($phindexdir, "pdata.txt");
|
---|
| 364 | my $mg_stem = "pdata";
|
---|
| 365 |
|
---|
| 366 | &execute("$mg_passes -d $phindexdir -f $mg_stem -T1 $mg_input", $verbosity);
|
---|
| 367 | &execute("$mg_compression_dict -d $phindexdir -f $mg_stem", $verbosity);
|
---|
| 368 | &execute("$mg_passes -d $phindexdir -f $mg_stem -T2 $mg_input", $verbosity);
|
---|
| 369 |
|
---|
| 370 | # create the mg index of words
|
---|
| 371 | print "\nCreating word-level search indexes\n";
|
---|
| 372 | $mg_input = &util::filename_cat($phindexdir, "pword.txt");
|
---|
| 373 | $mg_stem = "pword";
|
---|
| 374 |
|
---|
| 375 | &execute("$mg_passes -d $phindexdir -f $mg_stem -T1 -I1 $mg_input", $verbosity);
|
---|
| 376 | &execute("$mg_compression_dict -d $phindexdir -f $mg_stem", $verbosity);
|
---|
| 377 | &execute("$mg_perf_hash_build -d $phindexdir -f $mg_stem", $verbosity);
|
---|
| 378 | &execute("$mg_passes -d $phindexdir -f $mg_stem -T2 -I2 $mg_input", $verbosity);
|
---|
| 379 | &execute("$mg_weights_build -d $phindexdir -f $mg_stem", $verbosity);
|
---|
| 380 | &execute("$mg_invf_dict -d $phindexdir -f $mg_stem", $verbosity);
|
---|
| 381 |
|
---|
| 382 | &execute("$mg_stem_idx -d $phindexdir -f $mg_stem -s 1", $verbosity);
|
---|
| 383 | &execute("$mg_stem_idx -d $phindexdir -f $mg_stem -s 2", $verbosity);
|
---|
| 384 | &execute("$mg_stem_idx -d $phindexdir -f $mg_stem -s 3", $verbosity);
|
---|
| 385 |
|
---|
| 386 | # create the mg document information database
|
---|
| 387 | print "\nCreating document information databases\n";
|
---|
| 388 | $mg_input = &util::filename_cat($phindexdir, "docs.txt");
|
---|
| 389 | $mg_stem = "docs";
|
---|
| 390 |
|
---|
| 391 | &execute("$mg_passes -d $phindexdir -f $mg_stem -T1 $mg_input", $verbosity);
|
---|
| 392 | &execute("$mg_compression_dict -d $phindexdir -f $mg_stem", $verbosity);
|
---|
| 393 | &execute("$mg_passes -d $phindexdir -f $mg_stem -T2 $mg_input", $verbosity);
|
---|
| 394 |
|
---|
| 395 |
|
---|
| 396 | # Tidy up stray files
|
---|
| 397 | if (!$untidy) {
|
---|
[1645] | 398 | print "\nCleaning up\n" if ($verbosity > 2);
|
---|
[1643] | 399 | &util::rm("$phindexdir/clauses", "$phindexdir/clauses.numbers",
|
---|
| 400 | "$phindexdir/clauses.vocab", "$phindexdir/clauses.stats",
|
---|
| 401 | "$phindexdir/phrases", "$phindexdir/docs.txt",
|
---|
| 402 | "$phindexdir/pdata.txt", "$phindexdir/pword.txt");
|
---|
| 403 | my $outfile = 1;
|
---|
| 404 | while (-e "$phindexdir/outPhrase.$outfile") {
|
---|
| 405 | &util::rm("$phindexdir/outPhrase.$outfile");
|
---|
| 406 | $outfile++;
|
---|
| 407 | }
|
---|
| 408 | }
|
---|
| 409 |
|
---|
| 410 |
|
---|
| 411 | # Insert the classifier into.... what?
|
---|
| 412 | my $collection = $self->{'collection'};
|
---|
| 413 | my $url = "library?a=p&p=phind&c=$collection";
|
---|
| 414 |
|
---|
| 415 | my %classifyinfo = ('thistype'=>'Invisible',
|
---|
| 416 | 'childtype'=>'Phind',
|
---|
| 417 | 'Title'=>$self->{'title'},
|
---|
| 418 | 'contains'=>[]);
|
---|
| 419 |
|
---|
| 420 | push (@{$classifyinfo{'contains'}}, {'OID'=>$url});
|
---|
| 421 | return \%classifyinfo;
|
---|
| 422 | }
|
---|
| 423 |
|
---|
| 424 |
|
---|
| 425 |
|
---|
| 426 | sub convert_gml_to_tokens {
|
---|
| 427 |
|
---|
| 428 | $_ = shift @_;
|
---|
| 429 |
|
---|
| 430 | # FIRST, remove GML tags
|
---|
| 431 |
|
---|
| 432 | # Replace all whitespace with a simple space
|
---|
| 433 | s/\s+/ /gs;
|
---|
| 434 |
|
---|
| 435 | # Remove everything that is in a tag
|
---|
| 436 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
---|
| 437 | s/\s*<br>\s*/ LINEBREAK /isg;
|
---|
| 438 | s/<[^>]*>/ /sg;
|
---|
| 439 |
|
---|
| 440 | # Now we have the text, but it may contain HTML
|
---|
| 441 | # elements coded as > etc. Remove these tags.
|
---|
| 442 | s/</</sg;
|
---|
| 443 | s/>/>/sg;
|
---|
| 444 |
|
---|
| 445 | s/\s+/ /sg;
|
---|
| 446 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
---|
| 447 | s/\s*<br>\s*/ LINEBREAK /isg;
|
---|
| 448 | s/<[^>]*>/ /sg;
|
---|
| 449 |
|
---|
| 450 | # remove & and other miscellaneous markup tags
|
---|
| 451 | s/&/&/sg;
|
---|
| 452 | s/</</sg;
|
---|
| 453 | s/>/>/sg;
|
---|
| 454 | s/&/&/sg;
|
---|
| 455 |
|
---|
| 456 | # replace<p> and <br> placeholders with carriage returns
|
---|
| 457 | s/PARAGRAPHBREAK/\n/sg;
|
---|
| 458 | s/LINEBREAK/\n/sg;
|
---|
| 459 |
|
---|
| 460 |
|
---|
| 461 | # Exceptional punctuation
|
---|
| 462 | #
|
---|
| 463 | # We make special cases of some punctuation
|
---|
| 464 |
|
---|
| 465 | # remove any apostrophe that indicates omitted letters
|
---|
| 466 | s/(\w+)\'(\w*\s)/ $1$2 /g;
|
---|
| 467 |
|
---|
| 468 | # remove period that appears in a person's initals
|
---|
| 469 | s/\s([A-Z])\./ $1 /g;
|
---|
| 470 |
|
---|
| 471 | # replace hyphens in hypheanted words and names with a space
|
---|
| 472 | s/([A-Za-z])-\s*([A-Za-z])/$1 $2/g;
|
---|
| 473 |
|
---|
| 474 |
|
---|
| 475 | # Convert the remaining text to "clause format",
|
---|
| 476 | # This means removing all excess punctuation and garbage text,
|
---|
| 477 | # normalising valid punctuation to fullstops and commas,
|
---|
| 478 | # then putting one cluse on each line.
|
---|
| 479 |
|
---|
| 480 | # Insert newline when the end of a sentence is detected
|
---|
| 481 | # (delimter is: "[\.\?\!]\s")
|
---|
| 482 | s/\s*[\.\?\!]\s+/\n/g;
|
---|
| 483 |
|
---|
| 484 | # split numbers after four digits
|
---|
| 485 | s/(\d\d\d\d)/$1 /g;
|
---|
| 486 |
|
---|
| 487 | # split words after 32 characters
|
---|
| 488 |
|
---|
| 489 | # squash repeated punctuation
|
---|
| 490 | tr/A-Za-z0-9 //cs;
|
---|
| 491 |
|
---|
| 492 | # save email addresses
|
---|
| 493 | # s/\w+@\w+\.[\w\.]+/EMAIL/g;
|
---|
| 494 |
|
---|
| 495 | # normalise clause breaks (mostly punctuation symbols) to commas
|
---|
| 496 | s/[^A-Za-z0-9 \n]+/ , /g;
|
---|
| 497 |
|
---|
| 498 | # Remove repeated commas, and replace with newline
|
---|
| 499 | s/\s*,[, ]+/\n/g;
|
---|
| 500 |
|
---|
| 501 | # remove extra whitespace
|
---|
| 502 | s/ +/ /sg;
|
---|
| 503 | s/^\s+//mg;
|
---|
| 504 | s/\s*$/\n/mg;
|
---|
| 505 |
|
---|
| 506 | # remove lines that contain one word or less
|
---|
| 507 | s/^\w*$//mg;
|
---|
| 508 | s/^\s*$//mg;
|
---|
| 509 | tr/\n//s;
|
---|
| 510 |
|
---|
| 511 | return $_;
|
---|
| 512 | }
|
---|
| 513 |
|
---|
| 514 |
|
---|
| 515 | # Execute a system command
|
---|
| 516 |
|
---|
| 517 | sub execute {
|
---|
| 518 | my ($command, $verbosity) = @_;
|
---|
[1645] | 519 | print "Executing: $command\n" if ($verbosity > 2);
|
---|
[1643] | 520 | my $status = system($command);
|
---|
| 521 | if ($status != 0) {
|
---|
| 522 | print STDERR "phindgen.pl - Error executing $command: $!\n";
|
---|
| 523 | exit($status);
|
---|
| 524 | }
|
---|
| 525 | }
|
---|
| 526 |
|
---|
| 527 |
|
---|
| 528 | # Generate the vocabulary, symbol statistics, and numbers file from the
|
---|
| 529 | # clauses file. This is legacy code, so is a bit messy and probably wont
|
---|
| 530 | # run under windows.
|
---|
| 531 |
|
---|
| 532 | sub extract_vocabulary {
|
---|
| 533 | my ($phindex_dir, $language, $verbosity) = @_;
|
---|
| 534 |
|
---|
| 535 | my ($w, $l, $line, $word);
|
---|
| 536 |
|
---|
| 537 | my ($first_delimiter, $last_delimiter,
|
---|
| 538 | $first_stopword, $last_stopword,
|
---|
| 539 | $first_extractword, $last_extractword,
|
---|
| 540 | $first_contentword, $last_contentword,
|
---|
| 541 | $phrasedelimiter);
|
---|
| 542 |
|
---|
| 543 | my ($use_thesaurus, %thesaurus, $first_thesaurusword, $last_thesaurusword);
|
---|
| 544 |
|
---|
| 545 |
|
---|
| 546 | my %symbol;
|
---|
| 547 | my (%freq);
|
---|
| 548 |
|
---|
| 549 | print "Calculating vocabulary\n" if ($verbosity > 1);
|
---|
| 550 |
|
---|
| 551 | # Read and store the stopwords
|
---|
| 552 | my $words = `find $ENV{'GSDLHOME'}/etc/phind/$language -name "*.sw" | xargs cat`;
|
---|
| 553 | my %stopwords;
|
---|
| 554 | foreach my $w (split(/\s+/, $words)) {
|
---|
| 555 | $l = lc($w);
|
---|
| 556 | $stopwords{$l} = $w;
|
---|
| 557 | }
|
---|
| 558 |
|
---|
| 559 | # Read and store the thesaurus terms
|
---|
| 560 | $use_thesaurus = 0;
|
---|
| 561 | my $lex_file = &util::filename_cat("$ENV{'GSDLHOME'}", "etc", "phind",
|
---|
| 562 | "$language", "agrovoc.lex");
|
---|
| 563 | if (-e "$lex_file") {
|
---|
| 564 | open(TH, "<$lex_file");
|
---|
| 565 | while(<TH>) {
|
---|
| 566 | s/^\d+ //;
|
---|
| 567 | s/\(.*\)//;
|
---|
| 568 | foreach my $w (split(/\s+/, $_)) {
|
---|
| 569 | $thesaurus{lc($w)} = $w;
|
---|
| 570 | }
|
---|
| 571 | }
|
---|
| 572 | close TH;
|
---|
| 573 | $use_thesaurus = 1;
|
---|
| 574 | }
|
---|
| 575 |
|
---|
| 576 | # Read words in the text and count occurences
|
---|
| 577 | open(TXT, "<$phindex_dir/clauses");
|
---|
| 578 | my @words;
|
---|
| 579 |
|
---|
| 580 | while(<TXT>) {
|
---|
| 581 | $line = $_;
|
---|
| 582 | next unless ($line =~ /./);
|
---|
| 583 |
|
---|
| 584 | @words = split(/\s+/, $line);
|
---|
| 585 | foreach $w (@words) {
|
---|
| 586 | $l = lc($w);
|
---|
| 587 | $w = $l if ((defined $stopwords{$l}) || (defined $thesaurus{$l}));
|
---|
| 588 | $freq{$w}++;
|
---|
| 589 | }
|
---|
| 590 | $freq{$senlimit}++;
|
---|
| 591 | }
|
---|
| 592 |
|
---|
| 593 | # Calculate the "best" form of each word
|
---|
| 594 | my (%bestform, %totalfreq, %bestfreq);
|
---|
| 595 |
|
---|
| 596 | foreach $w (sort (keys %freq)) {
|
---|
| 597 | $l = lc($w);
|
---|
| 598 |
|
---|
| 599 | # totalfreq is the number of times a term appears in any form
|
---|
| 600 | $totalfreq{$l} += $freq{$w};
|
---|
| 601 |
|
---|
| 602 | if (defined $stopwords{$l}) {
|
---|
| 603 | $bestform{$l} = $stopwords{$l};
|
---|
| 604 |
|
---|
| 605 | } elsif (defined $thesaurus{$l}) {
|
---|
| 606 | $bestform{$l} = $thesaurus{$l};
|
---|
| 607 |
|
---|
| 608 | } elsif (!$bestform{$l} || ($freq{$w} > $bestfreq{$l})) {
|
---|
| 609 | $bestfreq{$l} = $freq{$w};
|
---|
| 610 | $bestform{$l} = $w;
|
---|
| 611 | }
|
---|
| 612 | }
|
---|
| 613 |
|
---|
| 614 | undef %freq;
|
---|
| 615 | undef %bestfreq;
|
---|
| 616 |
|
---|
| 617 |
|
---|
| 618 | # Assign symbol numbers to tokens
|
---|
| 619 | my $nextsymbol = 1;
|
---|
| 620 | my (@vocab);
|
---|
| 621 |
|
---|
| 622 | # Delimiters
|
---|
| 623 | $first_delimiter = 1;
|
---|
| 624 |
|
---|
| 625 | foreach $word (@delimiters) {
|
---|
| 626 |
|
---|
| 627 | $word = lc($word);
|
---|
| 628 | $bestform{$word} = uc($word);
|
---|
| 629 | $vocab[$nextsymbol] = $word;
|
---|
| 630 | $symbol{$word} = $nextsymbol;
|
---|
| 631 | $nextsymbol++;
|
---|
| 632 | }
|
---|
| 633 | $last_delimiter = $nextsymbol - 1;
|
---|
| 634 |
|
---|
| 635 | # Stopwords
|
---|
| 636 | $first_stopword = $nextsymbol;
|
---|
| 637 |
|
---|
| 638 | foreach my $word (sort keys %stopwords) {
|
---|
| 639 |
|
---|
| 640 | # don't incluse stopword unless it occurs in the text
|
---|
| 641 | $word = lc($word);
|
---|
| 642 | next unless ($totalfreq{$word});
|
---|
| 643 | next if ($symbol{$word});
|
---|
| 644 |
|
---|
| 645 | $vocab[$nextsymbol] = $word;
|
---|
| 646 | $symbol{$word} = $nextsymbol;
|
---|
| 647 | $nextsymbol++;
|
---|
| 648 | }
|
---|
| 649 | $last_stopword = $nextsymbol - 1;
|
---|
| 650 | $first_contentword = $nextsymbol;
|
---|
| 651 |
|
---|
| 652 | # Thesaurus terms
|
---|
| 653 | if ($use_thesaurus) {
|
---|
| 654 | $first_thesaurusword = $nextsymbol;
|
---|
| 655 |
|
---|
| 656 | foreach my $word (sort keys %thesaurus) {
|
---|
| 657 |
|
---|
| 658 | $word = lc($word);
|
---|
| 659 | next if ($symbol{$word});
|
---|
| 660 | $bestform{$word} = $thesaurus{$word};
|
---|
| 661 |
|
---|
| 662 | $vocab[$nextsymbol] = $word;
|
---|
| 663 | $symbol{$word} = $nextsymbol;
|
---|
| 664 | $nextsymbol++;
|
---|
| 665 |
|
---|
| 666 | }
|
---|
| 667 | $last_thesaurusword = $nextsymbol - 1;
|
---|
| 668 | }
|
---|
| 669 |
|
---|
| 670 | # Other content words
|
---|
| 671 | $first_extractword = $nextsymbol;
|
---|
| 672 |
|
---|
| 673 | foreach my $word (sort (keys %bestform)) {
|
---|
| 674 |
|
---|
| 675 | next if ($symbol{$word});
|
---|
| 676 |
|
---|
| 677 | $vocab[$nextsymbol] = $word;
|
---|
| 678 | $symbol{$word} = $nextsymbol;
|
---|
| 679 | $nextsymbol++;
|
---|
| 680 | }
|
---|
| 681 | $last_extractword = $nextsymbol - 1;
|
---|
| 682 | $last_contentword = $nextsymbol - 1;
|
---|
| 683 |
|
---|
| 684 |
|
---|
| 685 | # Outut the words
|
---|
| 686 | print "Saving vocabulary in $phindex_dir/clauses.vocab\n" if ($verbosity > 1);
|
---|
| 687 | open(VOC, ">$phindex_dir/clauses.vocab");
|
---|
| 688 |
|
---|
| 689 | for (my $i = 1; $i < $nextsymbol; $i++) {
|
---|
| 690 | $w = $vocab[$i];
|
---|
| 691 |
|
---|
| 692 | print VOC "$bestform{$w}\n";
|
---|
| 693 | $totalfreq{$w} = 0 unless ($totalfreq{$w});
|
---|
| 694 | }
|
---|
| 695 | close VOC;
|
---|
| 696 |
|
---|
| 697 |
|
---|
| 698 | # Output statistics about the vocablary
|
---|
| 699 | print "Saving statistics in $phindex_dir/clauses.stats\n" if ($verbosity > 1);
|
---|
| 700 | &util::rm("$phindex_dir/clauses.stats") if (-e "$phindex_dir/clauses.stats");
|
---|
| 701 | open(STAT, ">$phindex_dir/clauses.stats")
|
---|
| 702 | || die "Cannot open $phindex_dir/clauses.stats: $!";
|
---|
| 703 |
|
---|
| 704 | print STAT "first_delimiter $first_delimiter\n";
|
---|
| 705 | print STAT "last_delimiter $last_delimiter\n";
|
---|
| 706 | print STAT "first_stopword $first_stopword\n";
|
---|
| 707 | print STAT "last_stopword $last_stopword\n";
|
---|
| 708 | if ($use_thesaurus) {
|
---|
| 709 | print STAT "first_thesaurusword $first_thesaurusword\n";
|
---|
| 710 | print STAT "last_thesaurusword $last_thesaurusword\n";
|
---|
| 711 | }
|
---|
| 712 | print STAT "first_extractword $first_extractword\n";
|
---|
| 713 | print STAT "last_extractword $last_extractword\n";
|
---|
| 714 | print STAT "first_contentword $first_contentword\n";
|
---|
| 715 | print STAT "last_contentword $last_contentword\n";
|
---|
| 716 | print STAT "first_symbol $first_delimiter\n";
|
---|
| 717 | print STAT "last_symbol $last_contentword\n";
|
---|
| 718 | print STAT "first_word $first_stopword\n";
|
---|
| 719 | print STAT "last_word $last_contentword\n";
|
---|
| 720 | close STAT;
|
---|
| 721 |
|
---|
| 722 | undef @vocab;
|
---|
| 723 |
|
---|
| 724 |
|
---|
| 725 | # Save text as symbol numbers
|
---|
| 726 | print "Saving text as numbers in $phindex_dir/clauses.numbers\n" if ($verbosity > 1);
|
---|
| 727 |
|
---|
| 728 | open(TXT, "<$phindex_dir/clauses");
|
---|
| 729 | open(NUM, ">$phindex_dir/clauses.numbers");
|
---|
| 730 |
|
---|
| 731 | $phrasedelimiter = $symbol{lc($senlimit)};
|
---|
| 732 | print NUM "$symbol{lc($colstart)}\n";
|
---|
| 733 |
|
---|
| 734 | # set up the special symbols that delimit documents and sentences
|
---|
| 735 | while(<TXT>) {
|
---|
| 736 |
|
---|
| 737 | # split sentence into a list of tokens
|
---|
| 738 | $line = $_;
|
---|
| 739 | next unless ($line =~ /./);
|
---|
| 740 | @words = split(/\s+/, $line);
|
---|
| 741 |
|
---|
| 742 | # output one token at a time
|
---|
| 743 | foreach $word (@words) {
|
---|
| 744 | $word = lc($word);
|
---|
| 745 | print NUM "$symbol{$word}\n";
|
---|
| 746 | }
|
---|
| 747 |
|
---|
| 748 | # output phrase delimiter
|
---|
| 749 | print NUM "$phrasedelimiter\n";
|
---|
| 750 | }
|
---|
| 751 |
|
---|
| 752 | print NUM "$symbol{lc($colend)}\n";
|
---|
| 753 |
|
---|
| 754 | }
|
---|
| 755 |
|
---|
| 756 |
|
---|
| 757 | # Prepare the phrases file to be input to mgpp.
|
---|
| 758 | # This means renumbering the phrases in order of decreasing frequency.
|
---|
| 759 |
|
---|
| 760 |
|
---|
| 761 | # This is legacy code, and a little ugly, and may be unix-specific
|
---|
| 762 | # (particularly the sort command).
|
---|
| 763 |
|
---|
| 764 | sub renumber_phrases {
|
---|
| 765 | my ($phindex_dir, $verbosity) = @_;
|
---|
| 766 |
|
---|
| 767 | # Sort the phrases into order of increasing frequency
|
---|
| 768 | # This means the expansions will be sorted correctly later on.
|
---|
| 769 | print "Sorting phrases into freq order\n" if ($verbosity);
|
---|
| 770 | system("sort -rnt ':' +2 -o $phindex_dir/phrases $phindex_dir/phrases");
|
---|
| 771 |
|
---|
| 772 | my @symbol;
|
---|
| 773 |
|
---|
| 774 | # Read the vocabulary
|
---|
| 775 | print "Reading the vocabulary\n" if ($verbosity);
|
---|
| 776 | open(V, "<$phindex_dir/clauses.vocab")
|
---|
| 777 | || die "Cannot open $phindex_dir/clauses.vocab: $!";
|
---|
| 778 |
|
---|
| 779 | my $i = 1;
|
---|
| 780 | while(<V>) {
|
---|
| 781 | chomp;
|
---|
| 782 | $symbol[$i++] = $_;
|
---|
| 783 | }
|
---|
| 784 |
|
---|
| 785 | # Create file for phrase data
|
---|
| 786 | #
|
---|
| 787 | # The phrases file looks something like this
|
---|
| 788 | # 159396-1:s5175:4:1:116149-2:3:d2240,2;d2253;d2254
|
---|
| 789 | # 159409-1:s5263:6:1:159410-2:6:d2122;d2128;d2129;d2130;d2215;d2380
|
---|
| 790 | # 159415-1:s5267:9:1:159418-2:8:d3,2;d632;d633;d668;d1934;d2010;d2281;d2374
|
---|
| 791 | # 159426-1:s5273:5:2:159429-2,115168-17:5:d252;d815;d938;d939;d2361
|
---|
| 792 |
|
---|
| 793 | # The first field on each line is a unique phrase identifier.
|
---|
| 794 | # We need to calculate phrase numbers for each phrase
|
---|
| 795 | print "Calculate phrase numbers\n" if ($verbosity);
|
---|
| 796 |
|
---|
| 797 | my %phrasenumber;
|
---|
| 798 | my $nextphrase = 1;
|
---|
| 799 | my ($line);
|
---|
| 800 |
|
---|
| 801 | open(IN, "<$phindex_dir/phrases");
|
---|
| 802 | while(<IN>) {
|
---|
| 803 |
|
---|
| 804 | # read the line
|
---|
| 805 | chomp;
|
---|
| 806 | $line = $_;
|
---|
| 807 |
|
---|
| 808 | # we're only interested in the first field
|
---|
| 809 | $line =~ s/:.*//;
|
---|
| 810 |
|
---|
| 811 | # get a phrase number for this line
|
---|
| 812 | $phrasenumber{$line} = $nextphrase;
|
---|
| 813 | $nextphrase++;
|
---|
| 814 | }
|
---|
| 815 |
|
---|
| 816 |
|
---|
| 817 | # Now we create a new phrase file using phrase numbers, not the old IDs.
|
---|
| 818 | print "Format phrase data for MGPP\n" if ($verbosity);
|
---|
| 819 |
|
---|
| 820 | open(IN, "<$phindex_dir/phrases");
|
---|
| 821 | open(DATA, ">$phindex_dir/pdata.txt");
|
---|
| 822 | open(IDX, ">$phindex_dir/pword.txt");
|
---|
| 823 |
|
---|
| 824 | my ($key, $tf, $num, $countexp, $expansions, $countdocs, $documents, $text, $word);
|
---|
| 825 | my @fields;
|
---|
| 826 | my @documents;
|
---|
| 827 | my (@newexp, $k, $n);
|
---|
| 828 |
|
---|
| 829 | my $linenumber = 0;
|
---|
| 830 |
|
---|
| 831 | while(<IN>) {
|
---|
| 832 |
|
---|
| 833 | # read the line
|
---|
| 834 | chomp;
|
---|
| 835 | $line = $_;
|
---|
| 836 | @fields = split(/:/, $line);
|
---|
| 837 |
|
---|
| 838 | # get a phrase number for this line
|
---|
| 839 | $key = shift @fields;
|
---|
| 840 | die unless (defined($phrasenumber{$key}));
|
---|
| 841 | $num = $phrasenumber{$key};
|
---|
| 842 |
|
---|
| 843 | # get the text of the phrase
|
---|
| 844 | $text = shift @fields;
|
---|
| 845 | $text =~ s/s(\d+)/$symbol[$1]/g;
|
---|
| 846 | if ($text =~ / /) {
|
---|
| 847 | $word = "";
|
---|
| 848 | } else {
|
---|
| 849 | $word = $text;
|
---|
| 850 | }
|
---|
| 851 |
|
---|
| 852 | $linenumber++;
|
---|
| 853 | if ($linenumber % 1000 == 0) {
|
---|
| 854 | print "line $linenumber:\t$num\t$key\t($text)\n" if ($verbosity > 2);
|
---|
| 855 | }
|
---|
| 856 | print "$num: $key\t($text)\n" if ($verbosity > 3);
|
---|
| 857 |
|
---|
| 858 | # get the phrase frequency
|
---|
| 859 | $tf = shift @fields;
|
---|
| 860 |
|
---|
| 861 | # get the number of expansions
|
---|
| 862 | $countexp = shift @fields;
|
---|
| 863 |
|
---|
| 864 | # get the expansions and convert them into phrase numbers
|
---|
| 865 | $expansions = shift @fields;
|
---|
| 866 | @newexp = ();
|
---|
| 867 | foreach $k (split(/,/, $expansions)) {
|
---|
| 868 | die "ERROR - no phrase number for: $k" unless (defined($phrasenumber{$k}));
|
---|
| 869 | $n = $phrasenumber{$k};
|
---|
| 870 | push @newexp, $n;
|
---|
| 871 | }
|
---|
| 872 | @newexp = sort numerically @newexp;
|
---|
| 873 |
|
---|
| 874 | # get the number of documents
|
---|
| 875 | $countdocs = shift @fields;
|
---|
| 876 |
|
---|
| 877 | # get the documents
|
---|
| 878 | $documents = shift @fields;
|
---|
| 879 | $documents =~ s/d//g;
|
---|
| 880 | @documents = split(/;/, $documents);
|
---|
| 881 | @documents = sort by_frequency @documents;
|
---|
| 882 |
|
---|
| 883 | # output the phrase data
|
---|
| 884 | print DATA "<Document>";
|
---|
| 885 | print DATA "$num:$text:$tf:$countexp:$countdocs:";
|
---|
| 886 | print DATA join(",", @newexp), ":", join(";", @documents), "\n";
|
---|
| 887 |
|
---|
| 888 | # output the word index search data
|
---|
| 889 | print IDX "<Document>$word\n";
|
---|
| 890 |
|
---|
| 891 |
|
---|
| 892 | }
|
---|
| 893 | }
|
---|
| 894 |
|
---|
| 895 | # sort routines used to renumber phrases
|
---|
| 896 |
|
---|
| 897 | sub numerically { $a <=> $b }
|
---|
| 898 |
|
---|
| 899 | sub by_frequency {
|
---|
| 900 | my $fa = 1;
|
---|
| 901 | if ($a =~ /,/) {
|
---|
| 902 | $fa = $a;
|
---|
| 903 | $fa =~ s/\d+,//;
|
---|
| 904 | }
|
---|
| 905 | my $fb = 1;
|
---|
| 906 | if ($b =~ /,/) {
|
---|
| 907 | $fb = $b;
|
---|
| 908 | $fb =~ s/\d+,//;
|
---|
| 909 | }
|
---|
| 910 |
|
---|
| 911 | return ($fb <=> $fa);
|
---|
| 912 | }
|
---|
| 913 |
|
---|
| 914 |
|
---|
| 915 | 1;
|
---|