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