| 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 | package Phind;
|
|---|
| 33 |
|
|---|
| 34 | use BaseClassifier;
|
|---|
| 35 | use FileUtils;
|
|---|
| 36 | use util;
|
|---|
| 37 | use ghtml;
|
|---|
| 38 | use unicode;
|
|---|
| 39 |
|
|---|
| 40 | use strict;
|
|---|
| 41 | no strict 'refs'; # allow filehandles to be variables and viceversa
|
|---|
| 42 |
|
|---|
| 43 | my @removedirs = ();
|
|---|
| 44 |
|
|---|
| 45 | my %wanted_index_files = ('td'=>1,
|
|---|
| 46 | 't'=>1,
|
|---|
| 47 | 'ti'=>1,
|
|---|
| 48 | 'tl'=>1,
|
|---|
| 49 | 'tsd'=>1,
|
|---|
| 50 | 'idb'=>1,
|
|---|
| 51 | 'ib1'=>1,
|
|---|
| 52 | 'ib2'=>1,
|
|---|
| 53 | 'ib3'=>1,
|
|---|
| 54 | 'i'=>1,
|
|---|
| 55 | 'il'=>1,
|
|---|
| 56 | 'w'=>1,
|
|---|
| 57 | 'wa'=>1);
|
|---|
| 58 |
|
|---|
| 59 | sub BEGIN {
|
|---|
| 60 | @Phind::ISA = ('BaseClassifier');
|
|---|
| 61 | }
|
|---|
| 62 |
|
|---|
| 63 | sub END {
|
|---|
| 64 |
|
|---|
| 65 | # Tidy up stray files - we do this here as there's some weird problem
|
|---|
| 66 | # preventing us from doing it in the get_classify_info() function (on
|
|---|
| 67 | # windows at least) where the close() appears to fail on txthandle and
|
|---|
| 68 | # dochandle, thus preventing us from deleting those files
|
|---|
| 69 |
|
|---|
| 70 | foreach my $dir (@removedirs) {
|
|---|
| 71 | if (-d $dir && opendir (DIR, $dir)) {
|
|---|
| 72 | my @files = readdir DIR;
|
|---|
| 73 | closedir DIR;
|
|---|
| 74 |
|
|---|
| 75 | foreach my $file (@files) {
|
|---|
| 76 | next if $file =~ /^\.\.?$/;
|
|---|
| 77 | my ($suffix) = $file =~ /\.([^\.]+)$/;
|
|---|
| 78 | if (!defined $suffix || !defined $wanted_index_files{$suffix}) {
|
|---|
| 79 | # delete it!
|
|---|
| 80 | &FileUtils::removeFiles (&FileUtils::filenameConcatenate ($dir, $file));
|
|---|
| 81 | }
|
|---|
| 82 | }
|
|---|
| 83 | }
|
|---|
| 84 | }
|
|---|
| 85 | }
|
|---|
| 86 |
|
|---|
| 87 | my $arguments =
|
|---|
| 88 | [ { 'name' => "text",
|
|---|
| 89 | 'desc' => "{Phind.text}",
|
|---|
| 90 | 'type' => "string",
|
|---|
| 91 | 'deft' => "section:Title,section:text",
|
|---|
| 92 | 'reqd' => "no" },
|
|---|
| 93 | { 'name' => "title",
|
|---|
| 94 | 'desc' => "{Phind.title}",
|
|---|
| 95 | 'type' => "metadata",
|
|---|
| 96 | 'deft' => "Title",
|
|---|
| 97 | 'reqd' => "no" },
|
|---|
| 98 | { 'name' => "buttonname",
|
|---|
| 99 | 'desc' => "{BasClas.buttonname}",
|
|---|
| 100 | 'type' => "string",
|
|---|
| 101 | 'deft' => "Phrase",
|
|---|
| 102 | 'reqd' => "no" },
|
|---|
| 103 | { 'name' => "language",
|
|---|
| 104 | 'desc' => "{Phind.language}",
|
|---|
| 105 | 'type' => "string",
|
|---|
| 106 | 'deft' => "en",
|
|---|
| 107 | 'reqd' => "no" },
|
|---|
| 108 | { 'name' => "savephrases",
|
|---|
| 109 | 'desc' => "{Phind.savephrases}",
|
|---|
| 110 | 'type' => "string",
|
|---|
| 111 | 'deft' => "",
|
|---|
| 112 | 'reqd' => "no" },
|
|---|
| 113 | { 'name' => "suffixmode",
|
|---|
| 114 | 'desc' => "{Phind.suffixmode}",
|
|---|
| 115 | 'type' => "int",
|
|---|
| 116 | 'deft' => "1",
|
|---|
| 117 | 'range' => "0,1",
|
|---|
| 118 | 'reqd' => "no" },
|
|---|
| 119 | { 'name' => "min_occurs",
|
|---|
| 120 | 'desc' => "{Phind.min_occurs}",
|
|---|
| 121 | 'type' => "int",
|
|---|
| 122 | 'deft' => "2",
|
|---|
| 123 | 'range' => "1,",
|
|---|
| 124 | 'reqd' => "no" },
|
|---|
| 125 | { 'name' => "thesaurus",
|
|---|
| 126 | 'desc' => "{Phind.thesaurus}",
|
|---|
| 127 | 'type' => "string",
|
|---|
| 128 | 'deft' => "",
|
|---|
| 129 | 'reqd' => "no" },
|
|---|
| 130 | { 'name' => "untidy",
|
|---|
| 131 | 'desc' => "{Phind.untidy}",
|
|---|
| 132 | 'type' => "flag",
|
|---|
| 133 | 'reqd' => "no" } ];
|
|---|
| 134 |
|
|---|
| 135 | my $options = { 'name' => "Phind",
|
|---|
| 136 | 'desc' => "{Phind.desc}",
|
|---|
| 137 | 'abstract' => "no",
|
|---|
| 138 | 'inherits' => "yes",
|
|---|
| 139 | 'args' => $arguments };
|
|---|
| 140 |
|
|---|
| 141 |
|
|---|
| 142 | # Phrase delimiter symbols - these should be abstracted out someplace
|
|---|
| 143 |
|
|---|
| 144 | my $colstart = "COLLECTIONSTART";
|
|---|
| 145 | my $colend = "COLLECTIONEND";
|
|---|
| 146 | my $doclimit = "DOCUMENTLIMIT";
|
|---|
| 147 | my $senlimit = "SENTENCELIMIT";
|
|---|
| 148 | my @delimiters = ($colstart, $colend, $doclimit, $senlimit);
|
|---|
| 149 |
|
|---|
| 150 |
|
|---|
| 151 | # Create a new Phind browser based on collect.cfg
|
|---|
| 152 |
|
|---|
| 153 | sub new {
|
|---|
| 154 | my ($class) = shift (@_);
|
|---|
| 155 | my ($classifierslist,$inputargs,$hashArgOptLists) = @_;
|
|---|
| 156 | push(@$classifierslist, $class);
|
|---|
| 157 |
|
|---|
| 158 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 159 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 160 |
|
|---|
| 161 | my $self = new BaseClassifier($classifierslist, $inputargs, $hashArgOptLists);
|
|---|
| 162 |
|
|---|
| 163 | if ($self->{'info_only'}) {
|
|---|
| 164 | # don't worry about any options etc
|
|---|
| 165 | return bless $self, $class;
|
|---|
| 166 | }
|
|---|
| 167 |
|
|---|
| 168 | # Ensure the Phind generate scripts are in place
|
|---|
| 169 | my $file1 = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "suffix");
|
|---|
| 170 | $file1 .= ".exe" if $ENV{'GSDLOS'} =~ /^windows$/;
|
|---|
| 171 | my $src = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "src", "phind", "generate");
|
|---|
| 172 | if (!(-e $file1)) {
|
|---|
| 173 | print STDERR "Phind.pm: ERROR: The Phind \"suffix\" program is not installed.\n\n";
|
|---|
| 174 | exit(1);
|
|---|
| 175 | }
|
|---|
| 176 |
|
|---|
| 177 | # things that may have ex. in them that need to be stripped off
|
|---|
| 178 | $self->{'text'} = $self->strip_ex_from_metadata($self->{'text'});
|
|---|
| 179 | $self->{'title'} = $self->strip_ex_from_metadata($self->{'title'});
|
|---|
| 180 |
|
|---|
| 181 | # Transfer value from Auto Parsing to the variable name that used in previous GreenStone.
|
|---|
| 182 |
|
|---|
| 183 | $self->{"indexes"} = $self->{"text"};
|
|---|
| 184 |
|
|---|
| 185 | # Further setup
|
|---|
| 186 | $self->{'collection'} = $ENV{'GSDLCOLLECTION'}; # classifier information
|
|---|
| 187 | $self->{'collectiondir'} = $ENV{'GSDLCOLLECTDIR'}; # collection directories
|
|---|
| 188 | if (! defined $self->{'builddir'}) {
|
|---|
| 189 | $self->{'builddir'} = &FileUtils::filenameConcatenate($ENV{'GSDLCOLLECTDIR'}, "building");
|
|---|
| 190 | }
|
|---|
| 191 | $self->{'total'} = 0;
|
|---|
| 192 |
|
|---|
| 193 | # we set phind to be rtl if language is arabic
|
|---|
| 194 | if ($self->{'language'} eq "ar") {
|
|---|
| 195 | $self->{'textorientation'} = "rtl";
|
|---|
| 196 | }
|
|---|
| 197 | # Clean out the unused keys
|
|---|
| 198 | delete $self->{"text"};
|
|---|
| 199 |
|
|---|
| 200 | return bless $self, $class;
|
|---|
| 201 | }
|
|---|
| 202 |
|
|---|
| 203 |
|
|---|
| 204 | # Initialise the Phind classifier
|
|---|
| 205 |
|
|---|
| 206 | sub init {
|
|---|
| 207 | my $self = shift (@_);
|
|---|
| 208 |
|
|---|
| 209 | # ensure we have a build directory
|
|---|
| 210 | my $builddir = $self->{'builddir'};
|
|---|
| 211 | die unless (-e "$builddir");
|
|---|
| 212 |
|
|---|
| 213 | # create Phind directory
|
|---|
| 214 | my $phnumber = 1;
|
|---|
| 215 | my $phinddir = &FileUtils::filenameConcatenate($builddir, "phind1");
|
|---|
| 216 | while (-e "$phinddir") {
|
|---|
| 217 | $phnumber++;
|
|---|
| 218 | $phinddir = &FileUtils::filenameConcatenate($builddir, "phind$phnumber");
|
|---|
| 219 | }
|
|---|
| 220 | &util::mk_dir("$phinddir");
|
|---|
| 221 | $self->{'phinddir'} = $phinddir;
|
|---|
| 222 | $self->{'phindnumber'} = $phnumber;
|
|---|
| 223 |
|
|---|
| 224 | push(@removedirs, $phinddir) unless $self->{'untidy'};
|
|---|
| 225 |
|
|---|
| 226 | # open filehandles for documents and text
|
|---|
| 227 | my $clausefile = &FileUtils::filenameConcatenate("$phinddir", "clauses");
|
|---|
| 228 | &FileUtils::removeFiles($clausefile) if (-e $clausefile);
|
|---|
| 229 |
|
|---|
| 230 | my $txthandle = 'TEXT' . $phnumber;
|
|---|
| 231 | open($txthandle, ">$clausefile") || die "Cannot open $clausefile: $!";
|
|---|
| 232 | $self->{'txthandle'} = $txthandle;
|
|---|
| 233 |
|
|---|
| 234 | my $docfile = &FileUtils::filenameConcatenate("$phinddir", "docs.txt");
|
|---|
| 235 | &FileUtils::removeFiles($docfile) if (-e $docfile);
|
|---|
| 236 |
|
|---|
| 237 | my $dochandle = 'DOC' . $phnumber;
|
|---|
| 238 | open($dochandle, ">$docfile") || die "Cannot open $docfile: $!";
|
|---|
| 239 | $self->{'dochandle'} = $dochandle;
|
|---|
| 240 |
|
|---|
| 241 | }
|
|---|
| 242 |
|
|---|
| 243 |
|
|---|
| 244 | # Classify each document.
|
|---|
| 245 | #
|
|---|
| 246 | # Each document is passed here in turn. The classifier extracts the
|
|---|
| 247 | # text of each and stores it in the clauses file. Document details are
|
|---|
| 248 | # stored in the docs.txt file.
|
|---|
| 249 |
|
|---|
| 250 | sub classify {
|
|---|
| 251 | my $self = shift (@_);
|
|---|
| 252 | my ($doc_obj) = @_;
|
|---|
| 253 |
|
|---|
| 254 | my $verbosity = $self->{'verbosity'};
|
|---|
| 255 | my $top_section = $doc_obj->get_top_section();
|
|---|
| 256 |
|
|---|
| 257 | my $titlefield = $self->{'title'};
|
|---|
| 258 |
|
|---|
| 259 | my $title = $doc_obj->get_metadata_element ($top_section, $titlefield);
|
|---|
| 260 | if (!defined($title)) {
|
|---|
| 261 | $title = "";
|
|---|
| 262 | print STDERR "Phind: document has no title\n";
|
|---|
| 263 | }
|
|---|
| 264 | print "process: $title\n" if ($verbosity > 2);
|
|---|
| 265 |
|
|---|
| 266 | # Only consider the file if it is in the correct language
|
|---|
| 267 | my $doclanguage = $doc_obj->get_metadata_element ($top_section, "Language");
|
|---|
| 268 | my $phrlanguage = $self->{'language'};
|
|---|
| 269 | return if ($doclanguage && ($doclanguage !~ /$phrlanguage/i));
|
|---|
| 270 |
|
|---|
| 271 | # record this file
|
|---|
| 272 | $self->{'total'} ++;
|
|---|
| 273 | # what is $file ???
|
|---|
| 274 | # print "file $self->{'total'}: $file\n" if ($self->{'$verbosity'});
|
|---|
| 275 |
|
|---|
| 276 | # Store document details
|
|---|
| 277 | my $OID = $doc_obj->get_OID();
|
|---|
| 278 | $OID = "NULL" unless defined $OID;
|
|---|
| 279 | my $dochandle = $self->{'dochandle'};
|
|---|
| 280 | print $dochandle "<Document>\t$OID\t$title\n";
|
|---|
| 281 |
|
|---|
| 282 | # Store the text occuring in this object
|
|---|
| 283 |
|
|---|
| 284 | # output the document delimiter
|
|---|
| 285 | my $txthandle = $self->{'txthandle'};
|
|---|
| 286 | print $txthandle "$doclimit\n";
|
|---|
| 287 |
|
|---|
| 288 | # iterate over the required indexes and store their text
|
|---|
| 289 | my $indexes = $self->{'indexes'};
|
|---|
| 290 | my $text = "";
|
|---|
| 291 | my ($part, $level, $field, $section, $data, $dataref);
|
|---|
| 292 |
|
|---|
| 293 | foreach $part (split(/,/, $indexes)) {
|
|---|
| 294 |
|
|---|
| 295 | # Each field has a level and a data element ((e.g. document:Title)
|
|---|
| 296 | ($level, $field) = split(/:/, $part);
|
|---|
| 297 | die unless ($level && $field);
|
|---|
| 298 |
|
|---|
| 299 | # Extract the text from every section
|
|---|
| 300 | # (In phind, document:text and section:text are equivalent)
|
|---|
| 301 | if ($field eq "text") {
|
|---|
| 302 | $data = "";
|
|---|
| 303 | $section = $doc_obj->get_top_section();
|
|---|
| 304 | while (defined($section)) {
|
|---|
| 305 | $data .= $doc_obj->get_text($section) . "\n";
|
|---|
| 306 | $section = $doc_obj->get_next_section($section);
|
|---|
| 307 | }
|
|---|
| 308 | $text .= convert_gml_to_tokens($phrlanguage, $data) . "\n";
|
|---|
| 309 | }
|
|---|
| 310 |
|
|---|
| 311 | # Extract a metadata field from a document
|
|---|
| 312 | # (If there is more than one element of the given type, get them all.)
|
|---|
| 313 | elsif ($level eq "document") {
|
|---|
| 314 | $dataref = $doc_obj->get_metadata($doc_obj->get_top_section(), $field);
|
|---|
| 315 | foreach $data (@$dataref) {
|
|---|
| 316 | $text .= convert_gml_to_tokens($phrlanguage, $data) . "\n";
|
|---|
| 317 | }
|
|---|
| 318 | }
|
|---|
| 319 |
|
|---|
| 320 | # Extract metadata from every section in a document
|
|---|
| 321 | elsif ($level eq "section") {
|
|---|
| 322 | $data = "";
|
|---|
| 323 | $section = $doc_obj->get_top_section();
|
|---|
| 324 | while (defined($section)) {
|
|---|
| 325 | $dataref = $doc_obj->get_metadata($section, $field);
|
|---|
| 326 | $data .= join("\n", @$dataref) . "\n";
|
|---|
| 327 | $section = $doc_obj->get_next_section($section);
|
|---|
| 328 | }
|
|---|
| 329 | $text .= convert_gml_to_tokens($phrlanguage, $data) . "\n";
|
|---|
| 330 | }
|
|---|
| 331 |
|
|---|
| 332 | # Some sort of specification which I don't understand
|
|---|
| 333 | else {
|
|---|
| 334 | die "Unknown level ($level) in Phind index ($part)\n";
|
|---|
| 335 | }
|
|---|
| 336 |
|
|---|
| 337 | }
|
|---|
| 338 |
|
|---|
| 339 | # output the text
|
|---|
| 340 | $text =~ tr/\n//s;
|
|---|
| 341 | print $txthandle "$text";
|
|---|
| 342 | }
|
|---|
| 343 |
|
|---|
| 344 |
|
|---|
| 345 | # Construct the classifier from the information already gathered
|
|---|
| 346 | #
|
|---|
| 347 | # When get_classify_info is called, the clauses and docs.txt files have
|
|---|
| 348 | # already been constructed in the Phind directory. This function will
|
|---|
| 349 | # translate them into compressed, indexed MGPP files that can be read by
|
|---|
| 350 | # the phindcgi script. It will also register our classifier so that it
|
|---|
| 351 | # shows up in the navigation bar.
|
|---|
| 352 |
|
|---|
| 353 | sub get_classify_info {
|
|---|
| 354 | my $self = shift (@_);
|
|---|
| 355 | my ($gli) = @_;
|
|---|
| 356 |
|
|---|
| 357 | close $self->{'dochandle'};
|
|---|
| 358 | close $self->{'txthandle'};
|
|---|
| 359 | my $verbosity = $self->{'verbosity'};
|
|---|
| 360 | my $out = $self->{'outhandle'};
|
|---|
| 361 | my $phinddir = $self->{'phinddir'};
|
|---|
| 362 |
|
|---|
| 363 | my $osextra = "";
|
|---|
| 364 | if ($ENV{'GSDLOS'} !~ /^windows$/i) {
|
|---|
| 365 | $osextra = " -d /";
|
|---|
| 366 | }
|
|---|
| 367 |
|
|---|
| 368 | print STDERR "</Stage>\n" if $gli;
|
|---|
| 369 |
|
|---|
| 370 | if ($verbosity) {
|
|---|
| 371 | print $out "\n*** Phind.pm generating indexes for ", $self->{'indexes'}, "\n";
|
|---|
| 372 | print $out "*** in ", $self->{'phinddir'}, "\n";
|
|---|
| 373 | }
|
|---|
| 374 |
|
|---|
| 375 | print STDERR "<Stage name='Phind'>\n" if $gli;
|
|---|
| 376 |
|
|---|
| 377 | # Construct phind indexes
|
|---|
| 378 | my $suffixmode = $self->{'suffixmode'};
|
|---|
| 379 | my $min_occurs = $self->{'min_occurs'};
|
|---|
| 380 | my ($command, $status);
|
|---|
| 381 |
|
|---|
| 382 | # Generate the vocabulary, symbol statistics, and numbers file
|
|---|
| 383 | # from the clauses file
|
|---|
| 384 | print $out "\nExtracting vocabulary and statistics\n" if $verbosity;
|
|---|
| 385 | print STDERR "<Phase name='ExtractingVocab'/>\n" if $gli;
|
|---|
| 386 | &extract_vocabulary($self);
|
|---|
| 387 |
|
|---|
| 388 | # Use the suffix program to generate the phind/phrases file
|
|---|
| 389 | print $out "\nExtracting phrases from processed text (with suffix)\n" if $verbosity;
|
|---|
| 390 | print STDERR "<Phase name='ExtractingPhrase'/>\n" if $gli;
|
|---|
| 391 | &execute("suffix \"$phinddir\" $suffixmode $min_occurs $verbosity", $verbosity, $out);
|
|---|
| 392 |
|
|---|
| 393 | # check that we generated some files. It's not necessarily an error if
|
|---|
| 394 | # we didn't (execute() would have quit on error), but we can't go on.
|
|---|
| 395 | my $phrasesfile=&FileUtils::filenameConcatenate($self->{'phinddir'}, 'phrases');
|
|---|
| 396 | if (! -r $phrasesfile) {
|
|---|
| 397 | print STDERR "<Warning name='NoPhrasesFound'/>\n" if $gli;
|
|---|
| 398 | print $out "\nNo phrases found for Phind classifier!\n";
|
|---|
| 399 | return;
|
|---|
| 400 | }
|
|---|
| 401 |
|
|---|
| 402 | # Create the phrase file and put phrase numbers in phind/phrases
|
|---|
| 403 | print $out "\nSorting and renumbering phrases for input to mgpp\n" if $verbosity;
|
|---|
| 404 | print STDERR "<Phase name='SortAndRenumber'/>\n" if $gli;
|
|---|
| 405 | &renumber_phrases($self);
|
|---|
| 406 |
|
|---|
| 407 | print $out "\nCreating phrase databases\n";
|
|---|
| 408 | print STDERR "<Phase name='PhraseDatabases'/>\n" if $gli;
|
|---|
| 409 | my $mg_input = &FileUtils::filenameConcatenate($phinddir, "pdata.txt");
|
|---|
| 410 | my $mg_stem = &FileUtils::filenameConcatenate($phinddir, "pdata");
|
|---|
| 411 |
|
|---|
| 412 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T1 \"$mg_input\"", $verbosity, $out);
|
|---|
| 413 | &execute("mgpp_compression_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
|---|
| 414 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T2 \"$mg_input\"", $verbosity, $out);
|
|---|
| 415 |
|
|---|
| 416 | # create the mg index of words
|
|---|
| 417 | print $out "\nCreating word-level search indexes\n";
|
|---|
| 418 | print STDERR "<Phase name='WordLevelIndexes'/>\n" if $gli;
|
|---|
| 419 | $mg_input = &FileUtils::filenameConcatenate($phinddir, "pword.txt");
|
|---|
| 420 | $mg_stem = &FileUtils::filenameConcatenate($phinddir, "pword");
|
|---|
| 421 |
|
|---|
| 422 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T1 -I1 \"$mg_input\"", $verbosity, $out);
|
|---|
| 423 | &execute("mgpp_compression_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
|---|
| 424 | &execute("mgpp_perf_hash_build $osextra -f \"$mg_stem\"", $verbosity, $out);
|
|---|
| 425 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T2 -I2 \"$mg_input\"", $verbosity, $out);
|
|---|
| 426 | &execute("mgpp_weights_build $osextra -f \"$mg_stem\"", $verbosity, $out);
|
|---|
| 427 | &execute("mgpp_invf_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
|---|
| 428 |
|
|---|
| 429 | &execute("mgpp_stem_idx $osextra -f \"$mg_stem\" -s 1", $verbosity, $out);
|
|---|
| 430 | &execute("mgpp_stem_idx $osextra -f \"$mg_stem\" -s 2", $verbosity, $out);
|
|---|
| 431 | &execute("mgpp_stem_idx $osextra -f \"$mg_stem\" -s 3", $verbosity, $out);
|
|---|
| 432 |
|
|---|
| 433 | # create the mg document information database
|
|---|
| 434 | print $out "\nCreating document information databases\n";
|
|---|
| 435 | print STDERR "<Phase name='DocInfoDatabases'/>\n" if $gli;
|
|---|
| 436 | $mg_input = &FileUtils::filenameConcatenate($phinddir, "docs.txt");
|
|---|
| 437 | $mg_stem = &FileUtils::filenameConcatenate($phinddir, "docs");
|
|---|
| 438 |
|
|---|
| 439 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T1 \"$mg_input\"", $verbosity, $out);
|
|---|
| 440 | &execute("mgpp_compression_dict $osextra -f \"$mg_stem\"", $verbosity, $out);
|
|---|
| 441 | &execute("mgpp_passes $osextra -f \"$mg_stem\" -T2 \"$mg_input\"", $verbosity, $out);
|
|---|
| 442 |
|
|---|
| 443 | my $parameters = "phindnumber=$self->{'phindnumber'}";
|
|---|
| 444 | if (defined ($self->{'textorientation'})) {
|
|---|
| 445 | $parameters .= ";textorientation=$self->{'textorientation'}";
|
|---|
| 446 | }
|
|---|
| 447 | # Return the information about the classifier that we'll later want to
|
|---|
| 448 | # use to create macros when the Phind classifier document is displayed.
|
|---|
| 449 | my %classifyinfo = ('thistype'=>'Invisible',
|
|---|
| 450 | 'childtype'=>'Phind',
|
|---|
| 451 | 'Title'=>$self->{'buttonname'},
|
|---|
| 452 | 'parameters'=>$parameters,
|
|---|
| 453 | 'contains'=>[]);
|
|---|
| 454 |
|
|---|
| 455 | my $collection = $self->{'collection'};
|
|---|
| 456 | my $url = "library?a=p&p=phind&c=$collection";
|
|---|
| 457 | push (@{$classifyinfo{'contains'}}, {'OID'=>$url});
|
|---|
| 458 |
|
|---|
| 459 | return \%classifyinfo;
|
|---|
| 460 | }
|
|---|
| 461 |
|
|---|
| 462 |
|
|---|
| 463 |
|
|---|
| 464 | sub convert_gml_to_tokens {
|
|---|
| 465 |
|
|---|
| 466 | my ($language_exp, $text) = @_;
|
|---|
| 467 |
|
|---|
| 468 | # escape any magic words... - jrm21
|
|---|
| 469 | foreach my $delim (@delimiters) {
|
|---|
| 470 | my $replacement=lc($delim);
|
|---|
| 471 | my $num= $text=~ s/$delim/$replacement/g;
|
|---|
| 472 | if (!$num) {$num=0;}
|
|---|
| 473 | }
|
|---|
| 474 |
|
|---|
| 475 | if ($language_exp =~ /^en$/) {
|
|---|
| 476 | return &convert_gml_to_tokens_EN($text);
|
|---|
| 477 | }
|
|---|
| 478 |
|
|---|
| 479 | if ($language_exp =~ /zh/) {
|
|---|
| 480 | return &convert_gml_to_tokens_ZH($text);
|
|---|
| 481 | }
|
|---|
| 482 |
|
|---|
| 483 | $_ = $text;
|
|---|
| 484 |
|
|---|
| 485 | # 1. remove GML tags
|
|---|
| 486 |
|
|---|
| 487 | # Remove everything that is in a tag
|
|---|
| 488 | s/\s*<p>\s*/ PARAGRAPHBREAK /isgo;
|
|---|
| 489 | s/\s*<br>\s*/ LINEBREAK /isgo;
|
|---|
| 490 | s/<[^>]*>/ /sgo;
|
|---|
| 491 |
|
|---|
| 492 | # Now we have the text, but it may contain HTML
|
|---|
| 493 | # elements coded as > etc. Remove these tags.
|
|---|
| 494 | s/&/&/sgo;
|
|---|
| 495 | s/</</sgo;
|
|---|
| 496 | s/>/>/sgo;
|
|---|
| 497 | s/\s*<p>\s*/ PARAGRAPHBREAK /isgo;
|
|---|
| 498 | s/\s*<br>\s*/ LINEBREAK /isgo;
|
|---|
| 499 | s/<[^>]*>/ /sgo;
|
|---|
| 500 |
|
|---|
| 501 | # replace<p> and <br> placeholders with clause break symbol (\n)
|
|---|
| 502 | s/\s+/ /gso;
|
|---|
| 503 | s/PARAGRAPHBREAK/\n/sgo;
|
|---|
| 504 | s/LINEBREAK/\n/sgo;
|
|---|
| 505 |
|
|---|
| 506 |
|
|---|
| 507 | # 2. Split the remaining text into space-delimited tokens
|
|---|
| 508 |
|
|---|
| 509 | # Convert any HTML special characters (like ") to their UTF8 equivalent
|
|---|
| 510 | s/&([^;]+);/&unicode::ascii2utf8(\&ghtml::getcharequiv($1,1))/gse;
|
|---|
| 511 |
|
|---|
| 512 | # Split text at word boundaries
|
|---|
| 513 | s/\b/ /go;
|
|---|
| 514 |
|
|---|
| 515 | # 3. Convert the remaining text to "clause format"
|
|---|
| 516 |
|
|---|
| 517 | # Insert newline if the end of a sentence is detected
|
|---|
| 518 | # (delimter is: "[\.\?\!]\s")
|
|---|
| 519 | # s/\s*[\.\?\!]\s+/\n/go;
|
|---|
| 520 |
|
|---|
| 521 | # remove unnecessary punctuation and replace with clause break symbol (\n)
|
|---|
| 522 | # the following very nicely removes all non alphanumeric characters. too bad if you are not using english...
|
|---|
| 523 | #s/[^\w ]/\n/go;
|
|---|
| 524 | # replace punct with new lines - is this what we want??
|
|---|
| 525 | s/\s*[\?\;\:\!\,\.\"\[\]\{\}\(\)]\s*/\n/go; #"
|
|---|
| 526 | # then remove other punct with space
|
|---|
| 527 | s/[\'\`\\\_]/ /go;
|
|---|
| 528 |
|
|---|
| 529 | # remove extraneous whitespace
|
|---|
| 530 | s/ +/ /sgo;
|
|---|
| 531 | s/^\s+//mgo;
|
|---|
| 532 | s/\s*$/\n/mgo;
|
|---|
| 533 |
|
|---|
| 534 | # remove lines that contain one word or less
|
|---|
| 535 | s/^\S*$//mgo;
|
|---|
| 536 | s/^\s*$//mgo;
|
|---|
| 537 | tr/\n//s;
|
|---|
| 538 |
|
|---|
| 539 | return $_;
|
|---|
| 540 | }
|
|---|
| 541 |
|
|---|
| 542 | # a chinese version
|
|---|
| 543 | sub convert_gml_to_tokens_ZH {
|
|---|
| 544 |
|
|---|
| 545 | $_ = shift @_;
|
|---|
| 546 |
|
|---|
| 547 | # Replace all whitespace with a simple space
|
|---|
| 548 | s/\s+/ /gs;
|
|---|
| 549 | # Remove everything that is in a tag
|
|---|
| 550 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
|---|
| 551 | s/\s*<br>\s*/ LINEBREAK /isg;
|
|---|
| 552 | s/<[^>]*>/ /sg;
|
|---|
| 553 |
|
|---|
| 554 | # Now we have the text, but it may contain HTML
|
|---|
| 555 | # elements coded as > etc. Remove these tags.
|
|---|
| 556 | s/</</sg;
|
|---|
| 557 | s/>/>/sg;
|
|---|
| 558 |
|
|---|
| 559 | s/\s+/ /sg;
|
|---|
| 560 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
|---|
| 561 | s/\s*<br>\s*/ LINEBREAK /isg;
|
|---|
| 562 | s/<[^>]*>/ /sg;
|
|---|
| 563 |
|
|---|
| 564 | # remove & and other miscellaneous markup tags
|
|---|
| 565 | s/&/&/sg;
|
|---|
| 566 | s/</</sg;
|
|---|
| 567 | s/>/>/sg;
|
|---|
| 568 | s/&/&/sg;
|
|---|
| 569 |
|
|---|
| 570 | # replace<p> and <br> placeholders with carriage returns
|
|---|
| 571 | s/PARAGRAPHBREAK/\n/sg;
|
|---|
| 572 | s/LINEBREAK/\n/sg;
|
|---|
| 573 |
|
|---|
| 574 |
|
|---|
| 575 | # print STDERR "text:$_\n";
|
|---|
| 576 | return $_;
|
|---|
| 577 | }
|
|---|
| 578 |
|
|---|
| 579 | # A version of convert_gml_to_tokens that is fine-tuned to the English language.
|
|---|
| 580 |
|
|---|
| 581 | sub convert_gml_to_tokens_EN {
|
|---|
| 582 | $_ = shift @_;
|
|---|
| 583 |
|
|---|
| 584 | # FIRST, remove GML tags
|
|---|
| 585 |
|
|---|
| 586 | # Replace all whitespace with a simple space
|
|---|
| 587 | s/\s+/ /gs;
|
|---|
| 588 |
|
|---|
| 589 | # Remove everything that is in a tag
|
|---|
| 590 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
|---|
| 591 | s/\s*<br>\s*/ LINEBREAK /isg;
|
|---|
| 592 | s/<[^>]*>/ /sg;
|
|---|
| 593 |
|
|---|
| 594 | # Now we have the text, but it may contain HTML
|
|---|
| 595 | # elements coded as > etc. Remove these tags.
|
|---|
| 596 | s/</</sg;
|
|---|
| 597 | s/>/>/sg;
|
|---|
| 598 |
|
|---|
| 599 | s/\s+/ /sg;
|
|---|
| 600 | s/\s*<p>\s*/ PARAGRAPHBREAK /isg;
|
|---|
| 601 | s/\s*<br>\s*/ LINEBREAK /isg;
|
|---|
| 602 | s/<[^>]*>/ /sg;
|
|---|
| 603 |
|
|---|
| 604 | # remove & and other miscellaneous markup tags
|
|---|
| 605 | s/&/&/sg;
|
|---|
| 606 | s/</</sg;
|
|---|
| 607 | s/>/>/sg;
|
|---|
| 608 | s/&/&/sg;
|
|---|
| 609 |
|
|---|
| 610 | # replace<p> and <br> placeholders with carriage returns
|
|---|
| 611 | s/PARAGRAPHBREAK/\n/sg;
|
|---|
| 612 | s/LINEBREAK/\n/sg;
|
|---|
| 613 |
|
|---|
| 614 |
|
|---|
| 615 | # Exceptional punctuation
|
|---|
| 616 | #
|
|---|
| 617 | # We make special cases of some punctuation
|
|---|
| 618 |
|
|---|
| 619 | # remove any apostrophe that indicates omitted letters
|
|---|
| 620 | s/(\w+)\'(\w*\s)/ $1$2 /g;
|
|---|
| 621 |
|
|---|
| 622 | # remove period that appears in a person's initals
|
|---|
| 623 | s/\s([A-Z])\./ $1 /g;
|
|---|
| 624 |
|
|---|
| 625 | # replace hyphens in hypheanted words and names with a space
|
|---|
| 626 | s/([A-Za-z])-\s*([A-Za-z])/$1 $2/g;
|
|---|
| 627 |
|
|---|
| 628 | # Convert the remaining text to "clause format",
|
|---|
| 629 | # This means removing all excess punctuation and garbage text,
|
|---|
| 630 | # normalising valid punctuation to fullstops and commas,
|
|---|
| 631 | # then putting one cluse on each line.
|
|---|
| 632 |
|
|---|
| 633 | # Insert newline when the end of a sentence is detected
|
|---|
| 634 | # (delimter is: "[\.\?\!]\s")
|
|---|
| 635 | s/\s*[\.\?\!]\s+/\n/g;
|
|---|
| 636 |
|
|---|
| 637 | # split numbers after four digits
|
|---|
| 638 | s/(\d\d\d\d)/$1 /g;
|
|---|
| 639 |
|
|---|
| 640 | # split words after 32 characters
|
|---|
| 641 |
|
|---|
| 642 | # squash repeated punctuation
|
|---|
| 643 | tr/A-Za-z0-9 //cs;
|
|---|
| 644 |
|
|---|
| 645 | # save email addresses
|
|---|
| 646 | # s/\w+@\w+\.[\w\.]+/EMAIL/g;
|
|---|
| 647 |
|
|---|
| 648 | # normalise clause breaks (mostly punctuation symbols) to commas
|
|---|
| 649 | s/[^A-Za-z0-9 \n]+/ , /g;
|
|---|
| 650 |
|
|---|
| 651 | # Remove repeated commas, and replace with newline
|
|---|
| 652 | s/\s*,[, ]+/\n/g;
|
|---|
| 653 |
|
|---|
| 654 | # remove extra whitespace
|
|---|
| 655 | s/ +/ /sg;
|
|---|
| 656 | s/^\s+//mg;
|
|---|
| 657 | s/\s*$/\n/mg;
|
|---|
| 658 |
|
|---|
| 659 | # remove lines that contain one word or less
|
|---|
| 660 | s/^\w*$//mg;
|
|---|
| 661 | s/^\s*$//mg;
|
|---|
| 662 | tr/\n//s;
|
|---|
| 663 |
|
|---|
| 664 | return $_;
|
|---|
| 665 |
|
|---|
| 666 | }
|
|---|
| 667 |
|
|---|
| 668 |
|
|---|
| 669 |
|
|---|
| 670 | # Execute a system command
|
|---|
| 671 |
|
|---|
| 672 | sub execute {
|
|---|
| 673 | my ($command, $verbosity, $outhandle) = @_;
|
|---|
| 674 | print $outhandle "Executing: $command\n" if ($verbosity > 2);
|
|---|
| 675 | $! = 0;
|
|---|
| 676 | my $status = system($command);
|
|---|
| 677 | if ($status != 0) {
|
|---|
| 678 | print STDERR "Phind - Error executing '$command': $!\n";
|
|---|
| 679 | exit($status); # this causes the build to fail...
|
|---|
| 680 | }
|
|---|
| 681 | }
|
|---|
| 682 |
|
|---|
| 683 |
|
|---|
| 684 | # Generate the vocabulary, symbol statistics, and numbers file from the
|
|---|
| 685 | # clauses file. This is legacy code, so is a bit messy and probably wont
|
|---|
| 686 | # run under windows.
|
|---|
| 687 |
|
|---|
| 688 | sub extract_vocabulary {
|
|---|
| 689 | my ($self) = @_;
|
|---|
| 690 |
|
|---|
| 691 | my $verbosity = $self->{'verbosity'};
|
|---|
| 692 | my $out = $self->{'outhandle'};
|
|---|
| 693 |
|
|---|
| 694 | my $collectiondir = $self->{'collectiondir'};
|
|---|
| 695 | my $phinddir = $self->{'phinddir'};
|
|---|
| 696 |
|
|---|
| 697 | my $language_exp = $self->{'language'};
|
|---|
| 698 |
|
|---|
| 699 | my ($w, $l, $line, $word);
|
|---|
| 700 |
|
|---|
| 701 | my ($first_delimiter, $last_delimiter,
|
|---|
| 702 | $first_stopword, $last_stopword,
|
|---|
| 703 | $first_extractword, $last_extractword,
|
|---|
| 704 | $first_contentword, $last_contentword,
|
|---|
| 705 | $phrasedelimiter);
|
|---|
| 706 |
|
|---|
| 707 | my $thesaurus = $self->{'thesaurus'};
|
|---|
| 708 | my ($thesaurus_links, $thesaurus_terms,
|
|---|
| 709 | %thesaurus, $first_thesaurusword, $last_thesaurusword);
|
|---|
| 710 |
|
|---|
| 711 | my %symbol;
|
|---|
| 712 | my (%freq);
|
|---|
| 713 |
|
|---|
| 714 | print $out "Calculating vocabulary\n" if ($verbosity > 1);
|
|---|
| 715 |
|
|---|
| 716 | # Read and store the stopwords
|
|---|
| 717 | my $stopdir = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "etc", "packages", "phind", "stopword");
|
|---|
| 718 | my $stopword_files = ();
|
|---|
| 719 | my ($language, $language_dir, $file, $file_name);
|
|---|
| 720 | my %stopwords;
|
|---|
| 721 |
|
|---|
| 722 | # Examine each directory in the stopword directory
|
|---|
| 723 | opendir(STOPDIR, $stopdir);
|
|---|
| 724 | foreach $language (readdir STOPDIR) {
|
|---|
| 725 |
|
|---|
| 726 | # Ignore entries that do not match the classifier's language
|
|---|
| 727 | next unless ($language =~ /$language_exp/);
|
|---|
| 728 | $language_dir = &FileUtils::filenameConcatenate($stopdir, $language);
|
|---|
| 729 | next unless (-d "$language_dir");
|
|---|
| 730 |
|
|---|
| 731 | opendir(LANGDIR, $language_dir);
|
|---|
| 732 | foreach $file (readdir LANGDIR) {
|
|---|
| 733 |
|
|---|
| 734 | # Ignore entries that are not stopword files
|
|---|
| 735 | next unless ($file =~ /sw$/);
|
|---|
| 736 | $file_name = &FileUtils::filenameConcatenate($language_dir, $file);
|
|---|
| 737 | next unless (-f "$file_name");
|
|---|
| 738 |
|
|---|
| 739 | # Read the stopwords
|
|---|
| 740 | open(STOPFILE, "<$file_name");
|
|---|
| 741 | while (<STOPFILE>) {
|
|---|
| 742 | s/^\s+//;
|
|---|
| 743 | s/\s.*//;
|
|---|
| 744 | $word = $_;
|
|---|
| 745 | $l = lc($word);
|
|---|
| 746 | $stopwords{$l} = $word;
|
|---|
| 747 | }
|
|---|
| 748 | close STOPFILE;
|
|---|
| 749 |
|
|---|
| 750 | }
|
|---|
| 751 | closedir LANGDIR;
|
|---|
| 752 | }
|
|---|
| 753 | closedir STOPDIR;
|
|---|
| 754 |
|
|---|
| 755 | # Read thesaurus information
|
|---|
| 756 | if ($thesaurus) {
|
|---|
| 757 |
|
|---|
| 758 | # link file exists
|
|---|
| 759 | $thesaurus_links = &FileUtils::filenameConcatenate($collectiondir, "etc", "$thesaurus.lnk");
|
|---|
| 760 | die "Cannot find thesaurus link file" unless (-e "$thesaurus_links");
|
|---|
| 761 |
|
|---|
| 762 | # ensure term file exists in the correct language
|
|---|
| 763 | if ($language_exp =~ /^([a-z][a-z])/) {
|
|---|
| 764 | $language = $1;
|
|---|
| 765 | } else {
|
|---|
| 766 | $language = 'en';
|
|---|
| 767 | }
|
|---|
| 768 | $thesaurus_terms = &FileUtils::filenameConcatenate($collectiondir, "etc", "$thesaurus.$language");
|
|---|
| 769 | die "Cannot find thesaurus term file" unless (-e "$thesaurus_terms");
|
|---|
| 770 |
|
|---|
| 771 |
|
|---|
| 772 | # Read the thesaurus terms
|
|---|
| 773 | open(TH, "<$thesaurus_terms");
|
|---|
| 774 | while(<TH>) {
|
|---|
| 775 | s/^\d+ //;
|
|---|
| 776 | s/\(.*\)//;
|
|---|
| 777 | foreach $w (split(/\s+/, $_)) {
|
|---|
| 778 | $thesaurus{lc($w)} = $w;
|
|---|
| 779 | }
|
|---|
| 780 | }
|
|---|
| 781 | close TH;
|
|---|
| 782 | }
|
|---|
| 783 |
|
|---|
| 784 | # Read words in the text and count occurences
|
|---|
| 785 | open(TXT, "<$phinddir/clauses");
|
|---|
| 786 |
|
|---|
| 787 | my @words;
|
|---|
| 788 | while(<TXT>) {
|
|---|
| 789 | $line = $_;
|
|---|
| 790 | next unless ($line =~ /./);
|
|---|
| 791 |
|
|---|
| 792 | @words = split(/\s+/, $line);
|
|---|
| 793 | foreach $w (@words) {
|
|---|
| 794 | $l = lc($w);
|
|---|
| 795 | $w = $l if ((defined $stopwords{$l}) || (defined $thesaurus{$l}));
|
|---|
| 796 | $freq{$w}++;
|
|---|
| 797 | }
|
|---|
| 798 | $freq{$senlimit}++;
|
|---|
| 799 | }
|
|---|
| 800 |
|
|---|
| 801 | close TXT;
|
|---|
| 802 |
|
|---|
| 803 | # Calculate the "best" form of each word
|
|---|
| 804 | my (%bestform, %totalfreq, %bestfreq);
|
|---|
| 805 |
|
|---|
| 806 | foreach $w (sort (keys %freq)) {
|
|---|
| 807 | $l = lc($w);
|
|---|
| 808 |
|
|---|
| 809 | # totalfreq is the number of times a term appears in any form
|
|---|
| 810 | $totalfreq{$l} += $freq{$w};
|
|---|
| 811 |
|
|---|
| 812 | if (defined $stopwords{$l}) {
|
|---|
| 813 | $bestform{$l} = $stopwords{$l};
|
|---|
| 814 |
|
|---|
| 815 | } elsif (defined $thesaurus{$l}) {
|
|---|
| 816 | $bestform{$l} = $thesaurus{$l};
|
|---|
| 817 |
|
|---|
| 818 | } elsif (!$bestform{$l} || ($freq{$w} > $bestfreq{$l})) {
|
|---|
| 819 | $bestfreq{$l} = $freq{$w};
|
|---|
| 820 | $bestform{$l} = $w;
|
|---|
| 821 | }
|
|---|
| 822 | }
|
|---|
| 823 | undef %freq;
|
|---|
| 824 | undef %bestfreq;
|
|---|
| 825 |
|
|---|
| 826 |
|
|---|
| 827 | # Assign symbol numbers to tokens
|
|---|
| 828 | my $nextsymbol = 1;
|
|---|
| 829 | my (@vocab);
|
|---|
| 830 |
|
|---|
| 831 | # Delimiters
|
|---|
| 832 | $first_delimiter = 1;
|
|---|
| 833 |
|
|---|
| 834 | foreach $word (@delimiters) {
|
|---|
| 835 |
|
|---|
| 836 | # $word = lc($word); # jrm21
|
|---|
| 837 | $word = uc($word);
|
|---|
| 838 | $bestform{$word} = $word;
|
|---|
| 839 | $vocab[$nextsymbol] = $word;
|
|---|
| 840 | $symbol{$word} = $nextsymbol;
|
|---|
| 841 | $nextsymbol++;
|
|---|
| 842 | }
|
|---|
| 843 | $last_delimiter = $nextsymbol - 1;
|
|---|
| 844 | # Stopwords
|
|---|
| 845 | $first_stopword = $nextsymbol;
|
|---|
| 846 |
|
|---|
| 847 | foreach my $word (sort keys %stopwords) {
|
|---|
| 848 | # don't include stopword unless it occurs in the text
|
|---|
| 849 | $word = lc($word);
|
|---|
| 850 | next unless ($totalfreq{$word});
|
|---|
| 851 | next if ($symbol{$word});
|
|---|
| 852 |
|
|---|
| 853 | $vocab[$nextsymbol] = $word;
|
|---|
| 854 | $symbol{$word} = $nextsymbol;
|
|---|
| 855 | $nextsymbol++;
|
|---|
| 856 | }
|
|---|
| 857 | $last_stopword = $nextsymbol - 1;
|
|---|
| 858 | $first_contentword = $nextsymbol;
|
|---|
| 859 |
|
|---|
| 860 | # Thesaurus terms
|
|---|
| 861 | if ($thesaurus) {
|
|---|
| 862 | $first_thesaurusword = $nextsymbol;
|
|---|
| 863 |
|
|---|
| 864 | foreach my $word (sort keys %thesaurus) {
|
|---|
| 865 |
|
|---|
| 866 | $word = lc($word);
|
|---|
| 867 | next if ($symbol{$word});
|
|---|
| 868 | $bestform{$word} = $thesaurus{$word};
|
|---|
| 869 |
|
|---|
| 870 | $vocab[$nextsymbol] = $word;
|
|---|
| 871 | $symbol{$word} = $nextsymbol;
|
|---|
| 872 | $nextsymbol++;
|
|---|
| 873 |
|
|---|
| 874 | }
|
|---|
| 875 | $last_thesaurusword = $nextsymbol - 1;
|
|---|
| 876 | }
|
|---|
| 877 |
|
|---|
| 878 | # Other content words
|
|---|
| 879 | $first_extractword = $nextsymbol;
|
|---|
| 880 |
|
|---|
| 881 | foreach my $word (sort (keys %bestform)) {
|
|---|
| 882 |
|
|---|
| 883 | next if ($symbol{$word});
|
|---|
| 884 |
|
|---|
| 885 | $vocab[$nextsymbol] = $word;
|
|---|
| 886 | $symbol{$word} = $nextsymbol;
|
|---|
| 887 | $nextsymbol++;
|
|---|
| 888 | }
|
|---|
| 889 | $last_extractword = $nextsymbol - 1;
|
|---|
| 890 | $last_contentword = $nextsymbol - 1;
|
|---|
| 891 |
|
|---|
| 892 | # Outut the words
|
|---|
| 893 | print $out "Saving vocabulary in $phinddir/clauses.vocab\n" if ($verbosity > 1);
|
|---|
| 894 | open(VOC, ">$phinddir/clauses.vocab");
|
|---|
| 895 |
|
|---|
| 896 | for (my $i = 1; $i < $nextsymbol; $i++) {
|
|---|
| 897 | $w = $vocab[$i];
|
|---|
| 898 |
|
|---|
| 899 | print VOC "$bestform{$w}\n";
|
|---|
| 900 | $totalfreq{$w} = 0 unless ($totalfreq{$w});
|
|---|
| 901 | }
|
|---|
| 902 | close VOC;
|
|---|
| 903 |
|
|---|
| 904 |
|
|---|
| 905 | # Create statistics file
|
|---|
| 906 | # Output statistics about the vocablary
|
|---|
| 907 | print $out "Saving statistics in $phinddir/clauses.stats\n" if ($verbosity > 1);
|
|---|
| 908 | &FileUtils::removeFiles("$phinddir/clauses.stats") if (-e "$phinddir/clauses.stats");
|
|---|
| 909 |
|
|---|
| 910 | open(STAT, ">$phinddir/clauses.stats")
|
|---|
| 911 | || die "Cannot open $phinddir/clauses.stats: $!";
|
|---|
| 912 |
|
|---|
| 913 | print STAT "first_delimiter $first_delimiter\n";
|
|---|
| 914 | print STAT "last_delimiter $last_delimiter\n";
|
|---|
| 915 | print STAT "first_stopword $first_stopword\n";
|
|---|
| 916 | print STAT "last_stopword $last_stopword\n";
|
|---|
| 917 | if ($thesaurus) {
|
|---|
| 918 | print STAT "first_thesaurusword $first_thesaurusword\n";
|
|---|
| 919 | print STAT "last_thesaurusword $last_thesaurusword\n";
|
|---|
| 920 | }
|
|---|
| 921 | print STAT "first_extractword $first_extractword\n";
|
|---|
| 922 | print STAT "last_extractword $last_extractword\n";
|
|---|
| 923 | print STAT "first_contentword $first_contentword\n";
|
|---|
| 924 | print STAT "last_contentword $last_contentword\n";
|
|---|
| 925 | print STAT "first_symbol $first_delimiter\n";
|
|---|
| 926 | print STAT "last_symbol $last_contentword\n";
|
|---|
| 927 | print STAT "first_word $first_stopword\n";
|
|---|
| 928 | print STAT "last_word $last_contentword\n";
|
|---|
| 929 | close STAT;
|
|---|
| 930 |
|
|---|
| 931 | undef @vocab;
|
|---|
| 932 |
|
|---|
| 933 |
|
|---|
| 934 | # Create numbers file
|
|---|
| 935 | # Save text as symbol numbers
|
|---|
| 936 | print $out "Saving text as numbers in $phinddir/clauses.numbers\n" if ($verbosity > 1);
|
|---|
| 937 |
|
|---|
| 938 | open(TXT, "<$phinddir/clauses");
|
|---|
| 939 | open(NUM, ">$phinddir/clauses.numbers");
|
|---|
| 940 |
|
|---|
| 941 | ## $phrasedelimiter = $symbol{lc($senlimit)}; # jrm21
|
|---|
| 942 | ## print NUM "$symbol{lc($colstart)}\n"; # jrm21
|
|---|
| 943 | $phrasedelimiter = $symbol{$senlimit};
|
|---|
| 944 | print NUM "$symbol{$colstart}\n";
|
|---|
| 945 |
|
|---|
| 946 | # set up the special symbols that delimit documents and sentences
|
|---|
| 947 | while(<TXT>) {
|
|---|
| 948 |
|
|---|
| 949 | # split sentence into a list of tokens
|
|---|
| 950 | $line = $_;
|
|---|
| 951 | next unless ($line =~ /./);
|
|---|
| 952 | @words = split(/\s+/, $line);
|
|---|
| 953 |
|
|---|
| 954 | # output one token at a time
|
|---|
| 955 | foreach $word (@words) {
|
|---|
| 956 | # don't lower-case special delimiters - jrm21
|
|---|
| 957 | if (!map {if ($word eq $_) {1} else {()}} @delimiters) {
|
|---|
| 958 | $word = lc($word);
|
|---|
| 959 | }
|
|---|
| 960 | print NUM "$symbol{$word}\n";
|
|---|
| 961 | }
|
|---|
| 962 |
|
|---|
| 963 | # output phrase delimiter
|
|---|
| 964 | print NUM "$phrasedelimiter\n";
|
|---|
| 965 | }
|
|---|
| 966 |
|
|---|
| 967 | close TXT;
|
|---|
| 968 | # print NUM "$symbol{lc($colend)}\n";# jrm21
|
|---|
| 969 | print NUM "$symbol{$colend}\n";
|
|---|
| 970 | close NUM;
|
|---|
| 971 |
|
|---|
| 972 | # Save thesaurus data in one convienient file
|
|---|
| 973 | if ($thesaurus) {
|
|---|
| 974 |
|
|---|
| 975 | my $thesaurusfile = &FileUtils::filenameConcatenate($phinddir, "$thesaurus.numbers");
|
|---|
| 976 |
|
|---|
| 977 |
|
|---|
| 978 | print $out "Saving thesaurus as numbers in $thesaurusfile\n"
|
|---|
| 979 | if ($verbosity > 1);
|
|---|
| 980 |
|
|---|
| 981 | # Read the thesaurus terms
|
|---|
| 982 | my ($num, $text, %thes_symbols);
|
|---|
| 983 |
|
|---|
| 984 | open(TH, "<$thesaurus_terms");
|
|---|
| 985 | while(<TH>) {
|
|---|
| 986 | chomp;
|
|---|
| 987 | @words = split(/\s+/, $_);
|
|---|
| 988 | $num = shift @words;
|
|---|
| 989 | $text = "";
|
|---|
| 990 |
|
|---|
| 991 | # translate words into symbol numbers
|
|---|
| 992 | foreach $word (@words) {
|
|---|
| 993 | $word = lc($word);
|
|---|
| 994 | if ($symbol{$word}) {
|
|---|
| 995 | $text .= "s$symbol{$word} ";
|
|---|
| 996 | } elsif ($verbosity) {
|
|---|
| 997 | print $out "Phind: No thesaurus symbol, ignoring \"$word\"\n";
|
|---|
| 998 | }
|
|---|
| 999 | }
|
|---|
| 1000 | $text =~ s/ $//;
|
|---|
| 1001 | $thes_symbols{$num} = $text;
|
|---|
| 1002 | }
|
|---|
| 1003 | close TH;
|
|---|
| 1004 |
|
|---|
| 1005 | # Read the thesaurus links and write the corresponding data
|
|---|
| 1006 | open(TH, "<$thesaurus_links");
|
|---|
| 1007 | open(THOUT, ">$thesaurusfile");
|
|---|
| 1008 |
|
|---|
| 1009 | while(<TH>) {
|
|---|
| 1010 | chomp;
|
|---|
| 1011 | ($num, $text) = split(/:/, $_);
|
|---|
| 1012 |
|
|---|
| 1013 | if (defined($thes_symbols{$num})) {
|
|---|
| 1014 | print THOUT "$num:$thes_symbols{$num}:$text\n";
|
|---|
| 1015 | } else {
|
|---|
| 1016 | print THOUT "$num:untranslated:$text\n";
|
|---|
| 1017 | }
|
|---|
| 1018 | }
|
|---|
| 1019 | close TH;
|
|---|
| 1020 | close THOUT;
|
|---|
| 1021 | }
|
|---|
| 1022 |
|
|---|
| 1023 |
|
|---|
| 1024 |
|
|---|
| 1025 |
|
|---|
| 1026 | }
|
|---|
| 1027 |
|
|---|
| 1028 |
|
|---|
| 1029 | # renumber_phrases
|
|---|
| 1030 | #
|
|---|
| 1031 | # Prepare the phrases file to be input to mgpp. The biggest problem is
|
|---|
| 1032 | # reconciling the phrase identifiers used by the suffix program (which
|
|---|
| 1033 | # we'll call suffix-id numbers) with the numbers used in the thesaurus
|
|---|
| 1034 | # (theesaurus-id) to create a ciommon set of phind id numbers (phind-id).
|
|---|
| 1035 | # Phind-id numbers must be sorted by frequency of occurance.
|
|---|
| 1036 | #
|
|---|
| 1037 | # Start creating a set of phind-id numbers from the sorted suffix-id
|
|---|
| 1038 | # numbers and (if required) the thesaurus-id numbers. Then add any other
|
|---|
| 1039 | # phrases occuring in the thesaurus.
|
|---|
| 1040 | #
|
|---|
| 1041 | # The last thing we have to do is restore the vocabulary information to the
|
|---|
| 1042 | # phrase file so that the phrases are stored as words, not as symbol
|
|---|
| 1043 | # numbers.
|
|---|
| 1044 |
|
|---|
| 1045 | # The original phrases file looks something like this:
|
|---|
| 1046 | # 159396-1:s5175:4:1:116149-2:3:d2240,2;d2253;d2254
|
|---|
| 1047 | # 159409-1:s5263:6:1:159410-2:6:d2122;d2128;d2129;d2130;d2215;d2380
|
|---|
| 1048 | # 159415-1:s5267:9:1:159418-2:8:d3,2;d632;d633;d668;d1934;d2010;d2281;d2374
|
|---|
| 1049 | # 159426-1:s5273:5:2:159429-2,115168-17:5:d252;d815;d938;d939;d2361
|
|---|
| 1050 |
|
|---|
| 1051 |
|
|---|
| 1052 | sub renumber_phrases {
|
|---|
| 1053 | my ($self) = @_;
|
|---|
| 1054 |
|
|---|
| 1055 | renumber_suffix_data($self);
|
|---|
| 1056 | renumber_thesaurus_data($self);
|
|---|
| 1057 | restore_vocabulary_data($self);
|
|---|
| 1058 |
|
|---|
| 1059 | }
|
|---|
| 1060 |
|
|---|
| 1061 |
|
|---|
| 1062 |
|
|---|
| 1063 | # renumber_suffix_data
|
|---|
| 1064 | #
|
|---|
| 1065 | # Translate phrases file to phrases.2 using phind keys instead
|
|---|
| 1066 | # of suffix keys and sorting the expansion data.
|
|---|
| 1067 |
|
|---|
| 1068 | sub renumber_suffix_data {
|
|---|
| 1069 | my ($self) = @_;
|
|---|
| 1070 |
|
|---|
| 1071 | my $verbosity = $self->{'verbosity'};
|
|---|
| 1072 | my $out = $self->{'outhandle'};
|
|---|
| 1073 | print $out "Translate phrases: suffix-ids become phind-id's\n"
|
|---|
| 1074 | if ($verbosity);
|
|---|
| 1075 |
|
|---|
| 1076 | my $phinddir = $self->{'phinddir'};
|
|---|
| 1077 | my $infile = &FileUtils::filenameConcatenate($phinddir, 'phrases');
|
|---|
| 1078 | my $outfile = &FileUtils::filenameConcatenate($phinddir, 'phrases.2');
|
|---|
| 1079 |
|
|---|
| 1080 | # Read the phrase file. Calculate initial set of phind-id
|
|---|
| 1081 | # numbers and store (suffixid -> frequency) relation.
|
|---|
| 1082 |
|
|---|
| 1083 | my %suffixtophind;
|
|---|
| 1084 | my @totalfrequency;
|
|---|
| 1085 | my (@fields, $suffixid);
|
|---|
| 1086 | my $nextphind = 1;
|
|---|
| 1087 |
|
|---|
| 1088 | open(IN, "<$infile");
|
|---|
| 1089 | while(<IN>) {
|
|---|
| 1090 |
|
|---|
| 1091 | chomp;
|
|---|
| 1092 | @fields = split(/:/, $_);
|
|---|
| 1093 |
|
|---|
| 1094 | # get next suffixid and phindid
|
|---|
| 1095 | $suffixid = shift @fields;
|
|---|
| 1096 | $suffixtophind{$suffixid} = $nextphind;
|
|---|
| 1097 |
|
|---|
| 1098 | # store total frequency
|
|---|
| 1099 | shift @fields;
|
|---|
| 1100 | $totalfrequency[$nextphind] = shift @fields;
|
|---|
| 1101 |
|
|---|
| 1102 | $nextphind++;
|
|---|
| 1103 | }
|
|---|
| 1104 | close IN;
|
|---|
| 1105 |
|
|---|
| 1106 |
|
|---|
| 1107 | # Translate phrases file to phrases.2. Use phind keys (not suffix
|
|---|
| 1108 | # keys), sort expansion and document occurance data in order of
|
|---|
| 1109 | # descending frequency..
|
|---|
| 1110 | open(IN, "<$infile");
|
|---|
| 1111 | open(OUT, ">$outfile");
|
|---|
| 1112 |
|
|---|
| 1113 | my ($phindid, $text, $tf, $countexp, $expansions, $countdocs, $documents);
|
|---|
| 1114 | my (@documwents, @newexp, $k, $n);
|
|---|
| 1115 | my $linenumber = 0;
|
|---|
| 1116 |
|
|---|
| 1117 | while(<IN>) {
|
|---|
| 1118 |
|
|---|
| 1119 | # read the line
|
|---|
| 1120 | chomp;
|
|---|
| 1121 | @fields = split(/:/, $_);
|
|---|
| 1122 |
|
|---|
| 1123 | # get a phrase number for this line
|
|---|
| 1124 | $suffixid = shift @fields;
|
|---|
| 1125 | die unless (defined($suffixtophind{$suffixid}));
|
|---|
| 1126 | $phindid = $suffixtophind{$suffixid};
|
|---|
| 1127 |
|
|---|
| 1128 | # get the symbols in the phrase
|
|---|
| 1129 | $text = shift @fields;
|
|---|
| 1130 |
|
|---|
| 1131 | # output status information
|
|---|
| 1132 | $linenumber++;
|
|---|
| 1133 | if ($verbosity > 2) {
|
|---|
| 1134 | if ($linenumber % 1000 == 0) {
|
|---|
| 1135 | print $out "line $linenumber:\t$phindid\t$suffixid\t($text)\n";
|
|---|
| 1136 | }
|
|---|
| 1137 | # what are $num and $key??
|
|---|
| 1138 | #print $out "$num: $key\t($text)\n" if ($verbosity > 3);
|
|---|
| 1139 | }
|
|---|
| 1140 |
|
|---|
| 1141 | # get the phrase frequency
|
|---|
| 1142 | $tf = shift @fields;
|
|---|
| 1143 |
|
|---|
| 1144 | # get the number of expansions
|
|---|
| 1145 | $countexp = shift @fields;
|
|---|
| 1146 |
|
|---|
| 1147 | # get the expansions, convert them into phind-id numbers, and sort them
|
|---|
| 1148 | $expansions = shift @fields;
|
|---|
| 1149 | @newexp = ();
|
|---|
| 1150 | foreach $k (split(/,/, $expansions)) {
|
|---|
| 1151 | die "ERROR - no phindid for: $k" unless (defined($suffixtophind{$k}));
|
|---|
| 1152 | $n = $suffixtophind{$k};
|
|---|
| 1153 | push @newexp, $n;
|
|---|
| 1154 | }
|
|---|
| 1155 | @newexp = sort {$totalfrequency[$b] <=> $totalfrequency[$a]} @newexp;
|
|---|
| 1156 |
|
|---|
| 1157 | # get the number of documents
|
|---|
| 1158 | $countdocs = shift @fields;
|
|---|
| 1159 |
|
|---|
| 1160 | # get the documents and sort them
|
|---|
| 1161 | $documents = shift @fields;
|
|---|
| 1162 | $documents =~ s/d//g;
|
|---|
| 1163 | my @documents = split(/;/, $documents);
|
|---|
| 1164 | @documents = sort by_doc_frequency @documents;
|
|---|
| 1165 |
|
|---|
| 1166 | # output the phrase data
|
|---|
| 1167 | print OUT "$phindid:$text:$tf:$countexp:$countdocs:";
|
|---|
| 1168 | print OUT join(",", @newexp), ",:", join(";", @documents), ";\n";
|
|---|
| 1169 |
|
|---|
| 1170 | }
|
|---|
| 1171 |
|
|---|
| 1172 | close IN;
|
|---|
| 1173 | close OUT;
|
|---|
| 1174 | }
|
|---|
| 1175 |
|
|---|
| 1176 |
|
|---|
| 1177 | # renumber_thesaurus_data
|
|---|
| 1178 | #
|
|---|
| 1179 | # Translate phrases.2 to phrases.3, adding thesaurus data if available.
|
|---|
| 1180 |
|
|---|
| 1181 | sub renumber_thesaurus_data {
|
|---|
| 1182 | my ($self) = @_;
|
|---|
| 1183 |
|
|---|
| 1184 | my $out = $self->{'outhandle'};
|
|---|
| 1185 | my $verbosity = $self->{'verbosity'};
|
|---|
| 1186 | my $thesaurus = $self->{'thesaurus'};
|
|---|
| 1187 |
|
|---|
| 1188 | my $phinddir = $self->{'phinddir'};
|
|---|
| 1189 | my $infile = &FileUtils::filenameConcatenate($phinddir, "phrases.2");
|
|---|
| 1190 | my $outfile = &FileUtils::filenameConcatenate($phinddir, "phrases.3");
|
|---|
| 1191 |
|
|---|
| 1192 |
|
|---|
| 1193 | # If no thesaurus is defined, simply move the phrases file.
|
|---|
| 1194 | if (!$thesaurus) {
|
|---|
| 1195 | print $out "Translate phrases.2: no thesaurus data\n"
|
|---|
| 1196 | if ($verbosity);
|
|---|
| 1197 | &FileUtils::moveFiles($infile, $outfile);
|
|---|
| 1198 | return;
|
|---|
| 1199 | }
|
|---|
| 1200 |
|
|---|
| 1201 | print $out "Translate phrases.2: add thesaurus data\n"
|
|---|
| 1202 | if ($verbosity);
|
|---|
| 1203 |
|
|---|
| 1204 | # 1.
|
|---|
| 1205 | # Read thesaurus file and store (symbols->thesaurusid) mapping
|
|---|
| 1206 | my $thesaurusfile = &FileUtils::filenameConcatenate($phinddir, "$thesaurus.numbers");
|
|---|
| 1207 | my %symbolstothesid;
|
|---|
| 1208 | my (@fields, $thesid, $symbols);
|
|---|
| 1209 |
|
|---|
| 1210 | open(TH, "<$thesaurusfile");
|
|---|
| 1211 |
|
|---|
| 1212 | while (<TH>) {
|
|---|
| 1213 |
|
|---|
| 1214 | chomp;
|
|---|
| 1215 | @fields = split(/:/, $_);
|
|---|
| 1216 |
|
|---|
| 1217 | # get id and text
|
|---|
| 1218 | $thesid = shift @fields;
|
|---|
| 1219 | $symbols = shift @fields;
|
|---|
| 1220 | $symbolstothesid{$symbols} = $thesid;
|
|---|
| 1221 | }
|
|---|
| 1222 | close TH;
|
|---|
| 1223 |
|
|---|
| 1224 | # 2.
|
|---|
| 1225 | # Read phrases file to find thesaurus entries that already
|
|---|
| 1226 | # have a phindid. Store their phind-ids for later translation,
|
|---|
| 1227 | # and store their frequency for later sorting.
|
|---|
| 1228 | my %thesaurustophindid;
|
|---|
| 1229 | my %phindidtofrequency;
|
|---|
| 1230 | my ($phindid, $freq);
|
|---|
| 1231 |
|
|---|
| 1232 | open(IN, "<$infile");
|
|---|
| 1233 |
|
|---|
| 1234 | while(<IN>) {
|
|---|
| 1235 |
|
|---|
| 1236 | chomp;
|
|---|
| 1237 | @fields = split(/:/, $_);
|
|---|
| 1238 |
|
|---|
| 1239 | # phindid and symbols for this line
|
|---|
| 1240 | $phindid = shift @fields;
|
|---|
| 1241 | $symbols = shift @fields;
|
|---|
| 1242 | $freq = shift @fields;
|
|---|
| 1243 |
|
|---|
| 1244 | # do we have a thesaurus id corresponding to this phrase?
|
|---|
| 1245 | if (defined($symbolstothesid{$symbols})) {
|
|---|
| 1246 | $thesid = $symbolstothesid{$symbols};
|
|---|
| 1247 | $thesaurustophindid{$thesid} = $phindid;
|
|---|
| 1248 | $phindidtofrequency{$phindid} = $freq;
|
|---|
| 1249 | }
|
|---|
| 1250 | }
|
|---|
| 1251 | close IN;
|
|---|
| 1252 |
|
|---|
| 1253 | undef %symbolstothesid;
|
|---|
| 1254 |
|
|---|
| 1255 | # 3.
|
|---|
| 1256 | # Create phind-id numbers for remaining thesaurus entries,
|
|---|
| 1257 | # and note that their frequency is 0 for later sorting.
|
|---|
| 1258 | my $nextphindid = $phindid + 1;
|
|---|
| 1259 |
|
|---|
| 1260 | open(TH, "<$thesaurusfile");
|
|---|
| 1261 | while(<TH>) {
|
|---|
| 1262 |
|
|---|
| 1263 | chomp;
|
|---|
| 1264 | @fields = split(/:/, $_);
|
|---|
| 1265 |
|
|---|
| 1266 | # read thesaurus-id and ensure it has a corresponding phind-id
|
|---|
| 1267 | $thesid = shift @fields;
|
|---|
| 1268 | if (!defined($thesaurustophindid{$thesid})) {
|
|---|
| 1269 | $thesaurustophindid{$thesid} = $nextphindid;
|
|---|
| 1270 | $phindidtofrequency{$nextphindid} = 0;
|
|---|
| 1271 | $nextphindid++;
|
|---|
| 1272 | }
|
|---|
| 1273 | }
|
|---|
| 1274 | close TH;
|
|---|
| 1275 |
|
|---|
| 1276 | # 4.
|
|---|
| 1277 | # Translate thesaurus file, replacing thesaurus-id numbers with
|
|---|
| 1278 | # phind-id numbers.
|
|---|
| 1279 | my $newthesaurusfile = &FileUtils::filenameConcatenate($phinddir, "$thesaurus.phindid");
|
|---|
| 1280 | my ($relations, $linkcounter, $linktext, $linktype, @linkdata);
|
|---|
| 1281 | my (@links, $linkid, %linkidtotype, $newrelation);
|
|---|
| 1282 |
|
|---|
| 1283 | open(TH, "<$thesaurusfile");
|
|---|
| 1284 | open(TO, ">$newthesaurusfile");
|
|---|
| 1285 | while(<TH>) {
|
|---|
| 1286 |
|
|---|
| 1287 | chomp;
|
|---|
| 1288 | @fields = split(/:/, $_);
|
|---|
| 1289 |
|
|---|
| 1290 | # phindid and symbols for this line
|
|---|
| 1291 | ($thesid, $symbols, $relations) = @fields;
|
|---|
| 1292 |
|
|---|
| 1293 | die unless ($thesid && $symbols);
|
|---|
| 1294 | die unless $thesaurustophindid{$thesid};
|
|---|
| 1295 | $phindid = $thesaurustophindid{$thesid};
|
|---|
| 1296 |
|
|---|
| 1297 | # convert each part of the relation string to use phind-id numbers
|
|---|
| 1298 | # at the same time, we want to sort the list by frequency.
|
|---|
| 1299 | undef %linkidtotype;
|
|---|
| 1300 |
|
|---|
| 1301 | foreach $linktext (split(/;/, $relations)) {
|
|---|
| 1302 | @linkdata = split(/,/, $linktext);
|
|---|
| 1303 |
|
|---|
| 1304 | # remember the linktype (e.g. BT, NT)
|
|---|
| 1305 | $linktype = shift @linkdata;
|
|---|
| 1306 |
|
|---|
| 1307 | # store the type of each link
|
|---|
| 1308 | foreach $thesid (@linkdata) {
|
|---|
| 1309 | die unless (defined($thesaurustophindid{$thesid}));
|
|---|
| 1310 | $linkidtotype{$thesaurustophindid{$thesid}} = $linktype;
|
|---|
| 1311 | }
|
|---|
| 1312 | }
|
|---|
| 1313 |
|
|---|
| 1314 | # sort the list of links, first by frequency, then by type.
|
|---|
| 1315 | @links = sort { ($phindidtofrequency{$b} <=> $phindidtofrequency{$a})
|
|---|
| 1316 | or ($linkidtotype{$a} cmp $linkidtotype{$b}) } (keys %linkidtotype);
|
|---|
| 1317 | $linkcounter = (scalar @links);
|
|---|
| 1318 |
|
|---|
| 1319 | # create a string describing the link information
|
|---|
| 1320 | $linktype = $linkidtotype{$links[0]};
|
|---|
| 1321 | $newrelation = $linktype;
|
|---|
| 1322 | foreach $linkid (@links) {
|
|---|
| 1323 | if ($linkidtotype{$linkid} ne $linktype) {
|
|---|
| 1324 | $linktype = $linkidtotype{$linkid};
|
|---|
| 1325 | $newrelation .= ";" . $linktype;
|
|---|
| 1326 | }
|
|---|
| 1327 | $newrelation .= "," . $linkid;
|
|---|
| 1328 | }
|
|---|
| 1329 | $newrelation .= ";";
|
|---|
| 1330 |
|
|---|
| 1331 |
|
|---|
| 1332 | # output the new line
|
|---|
| 1333 | print TO "$phindid:$symbols:$linkcounter:$newrelation:\n";
|
|---|
| 1334 | }
|
|---|
| 1335 | close TH;
|
|---|
| 1336 | close TO;
|
|---|
| 1337 |
|
|---|
| 1338 | undef %thesaurustophindid;
|
|---|
| 1339 | undef %linkidtotype;
|
|---|
| 1340 | undef %phindidtofrequency;
|
|---|
| 1341 |
|
|---|
| 1342 | # 5.
|
|---|
| 1343 | # Read thesaurus data (in phind-id format) into memory
|
|---|
| 1344 | my %thesaurusdata;
|
|---|
| 1345 |
|
|---|
| 1346 | open(TH, "<$newthesaurusfile");
|
|---|
| 1347 | while(<TH>) {
|
|---|
| 1348 | chomp;
|
|---|
| 1349 | ($phindid, $symbols, $linkcounter, $relations) = split(/:/, $_);
|
|---|
| 1350 | die unless ($phindid && $symbols);
|
|---|
| 1351 | $thesaurusdata{$phindid} = "$symbols:$linkcounter:$relations";
|
|---|
| 1352 | }
|
|---|
| 1353 | close TH;
|
|---|
| 1354 |
|
|---|
| 1355 | # 6.
|
|---|
| 1356 | # Add thesaurus data to phrases file
|
|---|
| 1357 | my ($text, $tf, $countexp, $expansions, $countdocs, $documents);
|
|---|
| 1358 | my (@documwents, @newexp, $k, $n);
|
|---|
| 1359 | my $linenumber = 0;
|
|---|
| 1360 |
|
|---|
| 1361 | open(IN, "<$infile");
|
|---|
| 1362 | open(OUT, ">$outfile");
|
|---|
| 1363 |
|
|---|
| 1364 | # Update existing phrases
|
|---|
| 1365 | while(<IN>) {
|
|---|
| 1366 |
|
|---|
| 1367 | chomp;
|
|---|
| 1368 | @fields = split(/:/, $_);
|
|---|
| 1369 |
|
|---|
| 1370 | # get data for this line
|
|---|
| 1371 | $phindid = shift @fields;
|
|---|
| 1372 |
|
|---|
| 1373 | # output the phrase data, with thesaurus information
|
|---|
| 1374 | print OUT "$phindid:", join(":", @fields);
|
|---|
| 1375 |
|
|---|
| 1376 | # add thesaurus data
|
|---|
| 1377 | if (defined($thesaurusdata{$phindid})) {
|
|---|
| 1378 | @fields = split(/:/, $thesaurusdata{$phindid});
|
|---|
| 1379 | shift @fields;
|
|---|
| 1380 | $linkcounter = shift @fields;
|
|---|
| 1381 | $relations = shift @fields;
|
|---|
| 1382 |
|
|---|
| 1383 | print OUT ":$linkcounter:$relations";
|
|---|
| 1384 | $thesaurusdata{$phindid} = "";
|
|---|
| 1385 | }
|
|---|
| 1386 | print OUT "\n";
|
|---|
| 1387 | }
|
|---|
| 1388 | close IN;
|
|---|
| 1389 |
|
|---|
| 1390 | # Add phrases that aren't already in the file
|
|---|
| 1391 | foreach $phindid (sort numerically keys %thesaurusdata) {
|
|---|
| 1392 | next unless ($thesaurusdata{$phindid});
|
|---|
| 1393 |
|
|---|
| 1394 | @fields = split(/:/, $thesaurusdata{$phindid});
|
|---|
| 1395 | $symbols = shift @fields;
|
|---|
| 1396 | $linkcounter = shift @fields;
|
|---|
| 1397 | $relations = shift @fields;
|
|---|
| 1398 |
|
|---|
| 1399 | print OUT "$phindid:$symbols:0:0:0:::$linkcounter:$relations\n";
|
|---|
| 1400 | }
|
|---|
| 1401 | close OUT;
|
|---|
| 1402 |
|
|---|
| 1403 | }
|
|---|
| 1404 |
|
|---|
| 1405 | # restore_vocabulary_data
|
|---|
| 1406 | #
|
|---|
| 1407 | # Read phrases.3 and restore vocabulary information. Then write
|
|---|
| 1408 | # this data to the MGPP input files (pword.txt and pdata.txt) and
|
|---|
| 1409 | # (if requested) to the saved phrases file.
|
|---|
| 1410 |
|
|---|
| 1411 | sub restore_vocabulary_data {
|
|---|
| 1412 | my ($self) = @_;
|
|---|
| 1413 |
|
|---|
| 1414 | my $out = $self->{'outhandle'};
|
|---|
| 1415 | my $verbosity = $self->{'verbosity'};
|
|---|
| 1416 | print $out "Translate phrases.3: restore vocabulary\n" if ($verbosity);
|
|---|
| 1417 |
|
|---|
| 1418 | my $phinddir = $self->{'phinddir'};
|
|---|
| 1419 | my $infile = &FileUtils::filenameConcatenate($phinddir, 'phrases.3');
|
|---|
| 1420 | my $vocabfile = &FileUtils::filenameConcatenate($phinddir, 'clauses.vocab');
|
|---|
| 1421 | my $datafile = &FileUtils::filenameConcatenate($phinddir, 'pdata.txt');
|
|---|
| 1422 | my $wordfile = &FileUtils::filenameConcatenate($phinddir, 'pword.txt');
|
|---|
| 1423 |
|
|---|
| 1424 | my $savephrases = $self->{'savephrases'};
|
|---|
| 1425 |
|
|---|
| 1426 | # 1.
|
|---|
| 1427 | # Read the vocabulary file
|
|---|
| 1428 | open(V, "<$vocabfile")
|
|---|
| 1429 | || die "Cannot open $vocabfile: $!";
|
|---|
| 1430 | my @symbol;
|
|---|
| 1431 | my $i = 1;
|
|---|
| 1432 | while(<V>) {
|
|---|
| 1433 | chomp;
|
|---|
| 1434 | $symbol[$i++] = $_;
|
|---|
| 1435 | }
|
|---|
| 1436 | close V;
|
|---|
| 1437 |
|
|---|
| 1438 | # 2.
|
|---|
| 1439 | # Translate phrases.3 to MGPP input files
|
|---|
| 1440 | my ($key, $text, $word, $isThesaurus, $line);
|
|---|
| 1441 | my @fields;
|
|---|
| 1442 | my $linenumber = 0;
|
|---|
| 1443 |
|
|---|
| 1444 | open(IN, "<$infile");
|
|---|
| 1445 | open(DATA, ">$datafile");
|
|---|
| 1446 | open(WORD, ">$wordfile");
|
|---|
| 1447 |
|
|---|
| 1448 | # Save the phrases in a separate text file
|
|---|
| 1449 | if ($savephrases) {
|
|---|
| 1450 | print $out "Saving phrases in $savephrases\n" if ($verbosity);
|
|---|
| 1451 | open(SAVE, ">$savephrases");
|
|---|
| 1452 | }
|
|---|
| 1453 |
|
|---|
| 1454 | while(<IN>) {
|
|---|
| 1455 |
|
|---|
| 1456 | # read the line
|
|---|
| 1457 | chomp;
|
|---|
| 1458 | $line = $_;
|
|---|
| 1459 | @fields = split(/:/, $line);
|
|---|
| 1460 |
|
|---|
| 1461 | # get a phrase number for this line
|
|---|
| 1462 | $key = shift @fields;
|
|---|
| 1463 |
|
|---|
| 1464 | # restore the text of the phrase
|
|---|
| 1465 | $text = shift @fields;
|
|---|
| 1466 | $text =~ s/s(\d+)/$symbol[$1]/g;
|
|---|
| 1467 | if ($text =~ / /) {
|
|---|
| 1468 | $word = "";
|
|---|
| 1469 | } elsif ($text ne 'untranslated') {
|
|---|
| 1470 | $word = $text;
|
|---|
| 1471 | }
|
|---|
| 1472 |
|
|---|
| 1473 | # output the phrase data
|
|---|
| 1474 | print DATA "<Document>";
|
|---|
| 1475 | print DATA "$key:$text:", join(":", @fields), ":\n";
|
|---|
| 1476 |
|
|---|
| 1477 | # output the word index search data
|
|---|
| 1478 | print WORD "<Document>$word\n";
|
|---|
| 1479 |
|
|---|
| 1480 | # output the phrases to a text file
|
|---|
| 1481 | if ($savephrases) {
|
|---|
| 1482 | if ((scalar @fields) == 7) {
|
|---|
| 1483 | $isThesaurus = 1;
|
|---|
| 1484 | } else {
|
|---|
| 1485 | $isThesaurus = 0;
|
|---|
| 1486 | }
|
|---|
| 1487 | print SAVE $fields[0], "\t", $fields[2], "\t$isThesaurus\t$text\n";
|
|---|
| 1488 | }
|
|---|
| 1489 | }
|
|---|
| 1490 | close IN;
|
|---|
| 1491 | close WORD;
|
|---|
| 1492 | close DATA;
|
|---|
| 1493 | close SAVE if ($savephrases);
|
|---|
| 1494 |
|
|---|
| 1495 | }
|
|---|
| 1496 |
|
|---|
| 1497 |
|
|---|
| 1498 |
|
|---|
| 1499 | # sort routines used to renumber phrases
|
|---|
| 1500 |
|
|---|
| 1501 | sub numerically { $a <=> $b }
|
|---|
| 1502 |
|
|---|
| 1503 | sub by_doc_frequency {
|
|---|
| 1504 | my $fa = 1;
|
|---|
| 1505 | if ($a =~ /,/) {
|
|---|
| 1506 | $fa = $a;
|
|---|
| 1507 | $fa =~ s/\d+,//;
|
|---|
| 1508 | }
|
|---|
| 1509 | my $fb = 1;
|
|---|
| 1510 | if ($b =~ /,/) {
|
|---|
| 1511 | $fb = $b;
|
|---|
| 1512 | $fb =~ s/\d+,//;
|
|---|
| 1513 | }
|
|---|
| 1514 |
|
|---|
| 1515 | return ($fb <=> $fa);
|
|---|
| 1516 | }
|
|---|
| 1517 |
|
|---|
| 1518 | 1;
|
|---|
