| 1 | ###########################################################################
|
|---|
| 2 | #
|
|---|
| 3 | # Z3950Download.pm -- downloading class for z3950 protocol
|
|---|
| 4 | # A component of the Greenstone digital library software
|
|---|
| 5 | # from the New Zealand Digital Library Project at the
|
|---|
| 6 | # University of Waikato, New Zealand.
|
|---|
| 7 | #
|
|---|
| 8 | # Copyright (C) 1999 New Zealand Digital Library Project
|
|---|
| 9 | #
|
|---|
| 10 | # This program is free software; you can redistribute it and/or modify
|
|---|
| 11 | # it under the terms of the GNU General Public License as published by
|
|---|
| 12 | # the Free Software Foundation; either version 2 of the License, or
|
|---|
| 13 | # (at your option) any later version.
|
|---|
| 14 | #
|
|---|
| 15 | # This program is distributed in the hope that it will be useful,
|
|---|
| 16 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|---|
| 17 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|---|
| 18 | # GNU General Public License for more details.
|
|---|
| 19 | #
|
|---|
| 20 | # You should have received a copy of the GNU General Public License
|
|---|
| 21 | # along with this program; if not, write to the Free Software
|
|---|
| 22 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
|---|
| 23 | #
|
|---|
| 24 | ###########################################################################
|
|---|
| 25 |
|
|---|
| 26 | package Z3950Download;
|
|---|
| 27 |
|
|---|
| 28 | eval {require bytes};
|
|---|
| 29 |
|
|---|
| 30 | # suppress the annoying "subroutine redefined" warning that various
|
|---|
| 31 | # plugins cause under perl 5.6
|
|---|
| 32 | $SIG{__WARN__} = sub {warn($_[0]) unless ($_[0] =~ /Subroutine\s+\S+\sredefined/)};
|
|---|
| 33 |
|
|---|
| 34 | use strict;
|
|---|
| 35 |
|
|---|
| 36 | use BaseDownload;
|
|---|
| 37 | use IPC::Open2;
|
|---|
| 38 | use POSIX ":sys_wait_h"; # for waitpid, http://perldoc.perl.org/functions/waitpid.html
|
|---|
| 39 |
|
|---|
| 40 | sub BEGIN {
|
|---|
| 41 | @Z3950Download::ISA = ('BaseDownload');
|
|---|
| 42 | }
|
|---|
| 43 |
|
|---|
| 44 | my $arguments =
|
|---|
| 45 | [ { 'name' => "host",
|
|---|
| 46 | 'disp' => "{Z3950Download.host_disp}",
|
|---|
| 47 | 'desc' => "{Z3950Download.host}",
|
|---|
| 48 | 'type' => "string",
|
|---|
| 49 | 'reqd' => "yes"},
|
|---|
| 50 | { 'name' => "port",
|
|---|
| 51 | 'disp' => "{Z3950Download.port_disp}",
|
|---|
| 52 | 'desc' => "{Z3950Download.port}",
|
|---|
| 53 | 'type' => "string",
|
|---|
| 54 | 'reqd' => "yes"},
|
|---|
| 55 | { 'name' => "database",
|
|---|
| 56 | 'disp' => "{Z3950Download.database_disp}",
|
|---|
| 57 | 'desc' => "{Z3950Download.database}",
|
|---|
| 58 | 'type' => "string",
|
|---|
| 59 | 'reqd' => "yes"},
|
|---|
| 60 | { 'name' => "find",
|
|---|
| 61 | 'disp' => "{Z3950Download.find_disp}",
|
|---|
| 62 | 'desc' => "{Z3950Download.find}",
|
|---|
| 63 | 'type' => "string",
|
|---|
| 64 | 'deft' => "",
|
|---|
| 65 | 'reqd' => "yes"},
|
|---|
| 66 | { 'name' => "max_records",
|
|---|
| 67 | 'disp' => "{Z3950Download.max_records_disp}",
|
|---|
| 68 | 'desc' => "{Z3950Download.max_records}",
|
|---|
| 69 | 'type' => "int",
|
|---|
| 70 | 'deft' => "500",
|
|---|
| 71 | 'reqd' => "no"}];
|
|---|
| 72 |
|
|---|
| 73 | my $options = { 'name' => "Z3950Download",
|
|---|
| 74 | 'desc' => "{Z3950Download.desc}",
|
|---|
| 75 | 'abstract' => "no",
|
|---|
| 76 | 'inherits' => "yes",
|
|---|
| 77 | 'args' => $arguments };
|
|---|
| 78 |
|
|---|
| 79 |
|
|---|
| 80 | sub new
|
|---|
| 81 | {
|
|---|
| 82 | my ($class) = shift (@_);
|
|---|
| 83 | my ($getlist,$inputargs,$hashArgOptLists) = @_;
|
|---|
| 84 | push(@$getlist, $class);
|
|---|
| 85 |
|
|---|
| 86 | push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});
|
|---|
| 87 | push(@{$hashArgOptLists->{"OptList"}},$options);
|
|---|
| 88 |
|
|---|
| 89 | my $self = new BaseDownload($getlist,$inputargs,$hashArgOptLists);
|
|---|
| 90 |
|
|---|
| 91 | if ($self->{'info_only'}) {
|
|---|
| 92 | # don't worry about any options etc
|
|---|
| 93 | return bless $self, $class;
|
|---|
| 94 | }
|
|---|
| 95 |
|
|---|
| 96 | # Must set $self->{'url'}, since GLI use $self->{'url'} to calculate the log file name!
|
|---|
| 97 | $self->{'url'} = $self->{'host'}.":".$self->{'port'};
|
|---|
| 98 |
|
|---|
| 99 | $self->{'yaz'} = &FileUtils::filenameConcatenate($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "yaz-client");
|
|---|
| 100 |
|
|---|
| 101 | return bless $self, $class;
|
|---|
| 102 |
|
|---|
| 103 | }
|
|---|
| 104 |
|
|---|
| 105 | sub download
|
|---|
| 106 | {
|
|---|
| 107 | my ($self) = shift (@_);
|
|---|
| 108 | my ($hashGeneralOptions) = @_;
|
|---|
| 109 | my ($strOpen,$strBase,$strFind,$strResponse,$intAmount,$intMaxRecords,$strRecords);
|
|---|
| 110 |
|
|---|
| 111 | my $url = $self->{'url'};
|
|---|
| 112 |
|
|---|
| 113 | print STDERR "<<Defined Maximum>>\n";
|
|---|
| 114 |
|
|---|
| 115 | $strOpen = $self->start_yaz($url);
|
|---|
| 116 |
|
|---|
| 117 | print STDERR "Access database: \"$self->{'database'}\"\n";
|
|---|
| 118 | $self->run_command_without_output("base $self->{'database'}");
|
|---|
| 119 | print STDERR "Searching for keyword: \"$self->{'find'}\"\n";
|
|---|
| 120 | $intAmount = $self->findAmount($self->{'find'});
|
|---|
| 121 |
|
|---|
| 122 | if($intAmount <= 0)
|
|---|
| 123 | {
|
|---|
| 124 | ($intAmount == -1)?
|
|---|
| 125 | print STDERR "Something wrong with the arguments,downloading can not be performed\n":
|
|---|
| 126 | print STDERR "No Record is found\n";
|
|---|
| 127 | print STDERR "<<Finished>>\n";
|
|---|
| 128 | return 0;
|
|---|
| 129 | }
|
|---|
| 130 | $intMaxRecords = ($self->{'max_records'} > $intAmount)? $intAmount : $self->{'max_records'};
|
|---|
| 131 | print STDERR "<<Total number of record(s):$intMaxRecords>>\n";
|
|---|
| 132 | $strRecords = "Records: $intMaxRecords\n".$self->getRecords($intMaxRecords);
|
|---|
| 133 |
|
|---|
| 134 | $self->saveRecords($strRecords,$hashGeneralOptions->{'cache_dir'},$intMaxRecords);
|
|---|
| 135 | print STDERR "Closing connection...\n";
|
|---|
| 136 |
|
|---|
| 137 | $self->quit_yaz();
|
|---|
| 138 | return 1;
|
|---|
| 139 | }
|
|---|
| 140 |
|
|---|
| 141 |
|
|---|
| 142 | sub start_yaz
|
|---|
| 143 | {
|
|---|
| 144 | my ($self, $url) = @_;
|
|---|
| 145 |
|
|---|
| 146 | my $yaz = $self->{'yaz'};
|
|---|
| 147 |
|
|---|
| 148 | my $childpid = open2(*YAZOUT, *YAZIN, $yaz)
|
|---|
| 149 | or (print STDERR "<<Finished>>\n" and die "can't open pipe to yaz-client: $!");
|
|---|
| 150 | $self->{'pid'} = $childpid;
|
|---|
| 151 | $self->{'YAZOUT'} = *YAZOUT;
|
|---|
| 152 | $self->{'YAZIN'} = *YAZIN;
|
|---|
| 153 |
|
|---|
| 154 | if (defined $url) {
|
|---|
| 155 | print STDERR "Opening connection to $url\n";
|
|---|
| 156 | my $strOpen = $self->open_connection("open $url");
|
|---|
| 157 |
|
|---|
| 158 | if (!$strOpen) {
|
|---|
| 159 | print STDERR "Cannot connect to $url\n";
|
|---|
| 160 | print STDERR "<<Finished>>\n";
|
|---|
| 161 | return 0;
|
|---|
| 162 | }
|
|---|
| 163 |
|
|---|
| 164 | return $strOpen;
|
|---|
| 165 | }
|
|---|
| 166 | return 1;
|
|---|
| 167 | }
|
|---|
| 168 |
|
|---|
| 169 | sub quit_yaz
|
|---|
| 170 | {
|
|---|
| 171 | my ($self) = shift (@_);
|
|---|
| 172 |
|
|---|
| 173 | # can't send a "close" cmd to close the database here, since the close command is only
|
|---|
| 174 | # recognised by Z3950, not by SRW. (This method is also used by the subclass for SRW.)
|
|---|
| 175 |
|
|---|
| 176 | print STDERR "<<Finished>>\n";
|
|---|
| 177 |
|
|---|
| 178 | # need to send the quit command, else yaz-client is still running in the background
|
|---|
| 179 | $self->run_command_without_output("quit");
|
|---|
| 180 | close($self->{'YAZIN'}); # close the input to yaz. It also flushes quit command to yaz.
|
|---|
| 181 |
|
|---|
| 182 | # make sure nothing is being output by yaz
|
|---|
| 183 | # flush the yaz-client process' outputstream, else we'll be stuck in an infinite
|
|---|
| 184 | # loop waiting for the process to quit.
|
|---|
| 185 | my $output = $self->{'YAZOUT'};
|
|---|
| 186 | my $line;
|
|---|
| 187 | while (defined ($line = <$output>)) {
|
|---|
| 188 | if($line !~ m/\w/s) { # print anything other than plain whitespace in case it is important
|
|---|
| 189 | print STDERR "***### $line";
|
|---|
| 190 | }
|
|---|
| 191 | }
|
|---|
| 192 |
|
|---|
| 193 | close($self->{'YAZOUT'});
|
|---|
| 194 |
|
|---|
| 195 | # Is the following necessary? The PerlDoc on open2 (http://perldoc.perl.org/IPC/Open2.html)
|
|---|
| 196 | # says that waitpid must be called to "reap the child process", or otherwise it will hang
|
|---|
| 197 | # around like a zombie process in the background. Adding it here makes the code work as
|
|---|
| 198 | # before, but it is certainly necessary to call waitpid on wget (see WgetDownload.pm).
|
|---|
| 199 | # http://perldoc.perl.org/functions/waitpid.html
|
|---|
| 200 | my $kidpid;
|
|---|
| 201 | do {
|
|---|
| 202 | $kidpid = waitpid($self->{'pid'}, WNOHANG);
|
|---|
| 203 | } while $kidpid > 0; # waiting for pid to become -1
|
|---|
| 204 | }
|
|---|
| 205 |
|
|---|
| 206 | sub open_connection{
|
|---|
| 207 | my ($self,$strCommand) = (@_);
|
|---|
| 208 |
|
|---|
| 209 | $self->run_command($strCommand);
|
|---|
| 210 |
|
|---|
| 211 | my $out = $self->{'YAZOUT'};
|
|---|
| 212 |
|
|---|
| 213 | my $opening_line = <$out>;
|
|---|
| 214 |
|
|---|
| 215 | return ($opening_line =~ m/Connecting...OK/i)? 1: 0;
|
|---|
| 216 |
|
|---|
| 217 | }
|
|---|
| 218 |
|
|---|
| 219 | sub findAmount
|
|---|
| 220 | {
|
|---|
| 221 | my ($self) = shift (@_);
|
|---|
| 222 | my($strFindTarget) = @_;
|
|---|
| 223 | my $strResponse = $self->run_command_with_output("find $strFindTarget","^Number of hits:");
|
|---|
| 224 | return ($strResponse =~ m/^Number of hits: (\d+)/m)? $1:-1;
|
|---|
| 225 | }
|
|---|
| 226 |
|
|---|
| 227 | sub getRecords
|
|---|
| 228 | {
|
|---|
| 229 | my ($self) = shift (@_);
|
|---|
| 230 | my ($intMaxRecords) = @_;
|
|---|
| 231 | my ($strShow,$intStartNumber,$numRecords,$strResponse,$strRecords,$intRecordsLeft);
|
|---|
| 232 |
|
|---|
| 233 | $intStartNumber = 1;
|
|---|
| 234 | $intRecordsLeft = $intMaxRecords;
|
|---|
| 235 | $numRecords = 0;
|
|---|
| 236 | $strResponse ="";
|
|---|
| 237 |
|
|---|
| 238 | while ($intRecordsLeft > 0)
|
|---|
| 239 | {
|
|---|
| 240 | if($intRecordsLeft > 50)
|
|---|
| 241 | {
|
|---|
| 242 |
|
|---|
| 243 | print STDERR "Yaz is Gathering records: $intStartNumber - ".($intStartNumber+49)."\n";
|
|---|
| 244 | $numRecords = 50;
|
|---|
| 245 | $strShow = "show $intStartNumber+50";
|
|---|
| 246 | $intStartNumber = $intStartNumber + 50;
|
|---|
| 247 | $intRecordsLeft = $intRecordsLeft - 50;
|
|---|
| 248 |
|
|---|
| 249 | }
|
|---|
| 250 | else
|
|---|
| 251 | {
|
|---|
| 252 | $numRecords = $intRecordsLeft;
|
|---|
| 253 | print STDERR "Yaz is Gathering records: $intStartNumber - ".($intStartNumber+$intRecordsLeft-1)."\n";
|
|---|
| 254 | $strShow = "show $intStartNumber+$intRecordsLeft";
|
|---|
| 255 | $intRecordsLeft = 0;
|
|---|
| 256 |
|
|---|
| 257 | }
|
|---|
| 258 |
|
|---|
| 259 | $strResponse .= $self->get($strShow,$numRecords);
|
|---|
| 260 |
|
|---|
| 261 | if ($strResponse eq ""){
|
|---|
| 262 | print STDERR "<<ERROR: failed to get $numRecords records>>\n";
|
|---|
| 263 | }
|
|---|
| 264 | else{
|
|---|
| 265 | print STDERR "<<Done:$numRecords>>\n";
|
|---|
| 266 | }
|
|---|
| 267 | }
|
|---|
| 268 |
|
|---|
| 269 | return "$strResponse\n";
|
|---|
| 270 |
|
|---|
| 271 | }
|
|---|
| 272 |
|
|---|
| 273 | sub saveRecords
|
|---|
| 274 | {
|
|---|
| 275 | my ($self,$strRecords,$strOutputDir,$intMaxRecords) = @_;
|
|---|
| 276 |
|
|---|
| 277 | # setup directories
|
|---|
| 278 | # Currently only gather the MARC format
|
|---|
| 279 | my $strFileName = $self->generateFileName($intMaxRecords);
|
|---|
| 280 |
|
|---|
| 281 | $strOutputDir =~ s/"//g; #"
|
|---|
| 282 |
|
|---|
| 283 | my $strOutputFile = &FileUtils::filenameConcatenate($strOutputDir,$self->{'host'},"$strFileName.marc");
|
|---|
| 284 | # prepare subdirectory for record (if needed)
|
|---|
| 285 | my ($strSubDirPath,$unused) = $self->dirFileSplit($strOutputFile);
|
|---|
| 286 |
|
|---|
| 287 | &FileUtils::makeAllDirectories($strSubDirPath);
|
|---|
| 288 |
|
|---|
| 289 | print STDERR "Saving records to \"$strOutputFile\"\n";
|
|---|
| 290 |
|
|---|
| 291 | # save record
|
|---|
| 292 | open (ZOUT,">$strOutputFile")
|
|---|
| 293 | || die "Unable to save Z3950 record: $!\n";
|
|---|
| 294 | print ZOUT $strRecords;
|
|---|
| 295 | close(ZOUT);
|
|---|
| 296 | }
|
|---|
| 297 |
|
|---|
| 298 |
|
|---|
| 299 | sub run_command_with_output
|
|---|
| 300 | {
|
|---|
| 301 | my ($self,$strCMD,$strStopRE) =@_;
|
|---|
| 302 |
|
|---|
| 303 | $self->run_command($strCMD);
|
|---|
| 304 |
|
|---|
| 305 | return $self->get_output($strStopRE);
|
|---|
| 306 |
|
|---|
| 307 | }
|
|---|
| 308 |
|
|---|
| 309 | sub get{
|
|---|
| 310 | my ($self,$strShow,$numRecord) = @_;
|
|---|
| 311 |
|
|---|
| 312 | $self->run_command($strShow);
|
|---|
| 313 |
|
|---|
| 314 | my $strFullOutput="";
|
|---|
| 315 | my $count=0;
|
|---|
| 316 | my $readRecord = 0;
|
|---|
| 317 |
|
|---|
| 318 | my $output = $self->{'YAZOUT'};
|
|---|
| 319 | while (my $strLine = <$output>)
|
|---|
| 320 | {
|
|---|
| 321 |
|
|---|
| 322 | if ($strLine =~ m/Records: ([\d]*)/i ){
|
|---|
| 323 | $readRecord = 1;
|
|---|
| 324 | next;
|
|---|
| 325 | }
|
|---|
| 326 |
|
|---|
| 327 | return $strFullOutput if ($strLine =~ m/nextResultSetPosition|Not connected/i);
|
|---|
| 328 |
|
|---|
| 329 | next if(!$readRecord);
|
|---|
| 330 |
|
|---|
| 331 | $strFullOutput .= $strLine;
|
|---|
| 332 | }
|
|---|
| 333 |
|
|---|
| 334 | }
|
|---|
| 335 |
|
|---|
| 336 | sub run_command_without_output
|
|---|
| 337 | {
|
|---|
| 338 | my ($self) = shift (@_);
|
|---|
| 339 | my ($strCMD) = @_;
|
|---|
| 340 |
|
|---|
| 341 | $self->run_command($strCMD);
|
|---|
| 342 | }
|
|---|
| 343 |
|
|---|
| 344 | sub run_command
|
|---|
| 345 | {
|
|---|
| 346 | my ($self,$strCMD) = @_;
|
|---|
| 347 |
|
|---|
| 348 | my $input = $self->{'YAZIN'};
|
|---|
| 349 |
|
|---|
| 350 | print $input "$strCMD\n";
|
|---|
| 351 | }
|
|---|
| 352 |
|
|---|
| 353 | sub get_output{
|
|---|
| 354 | my ($self,$strStopRE) = @_;
|
|---|
| 355 |
|
|---|
| 356 | if (!defined $strStopRE){return "";}
|
|---|
| 357 | else
|
|---|
| 358 | {
|
|---|
| 359 | my $strFullOutput;
|
|---|
| 360 | my $output = $self->{'YAZOUT'};
|
|---|
| 361 | while (my $strLine = <$output>)
|
|---|
| 362 | {
|
|---|
| 363 | $strFullOutput .= $strLine;
|
|---|
| 364 | if($strLine =~ m/^$strStopRE|Not connected/i){return $strFullOutput;}
|
|---|
| 365 | }
|
|---|
| 366 | }
|
|---|
| 367 | }
|
|---|
| 368 |
|
|---|
| 369 | sub generateFileName
|
|---|
| 370 | {
|
|---|
| 371 | my ($self,$intMaxRecords) = @_;
|
|---|
| 372 | my $strFileName = ($self->{'database'})."_".($self->{'find'})."_".($intMaxRecords);
|
|---|
| 373 |
|
|---|
| 374 | }
|
|---|
| 375 |
|
|---|
| 376 | sub dirFileSplit
|
|---|
| 377 | {
|
|---|
| 378 | my ($self,$strFile) = @_;
|
|---|
| 379 |
|
|---|
| 380 | my @aryDirs = split(/[\/\\]/,$strFile);
|
|---|
| 381 |
|
|---|
| 382 | my $strLocalFile = pop(@aryDirs);
|
|---|
| 383 | my $strSubDirs = join("/",@aryDirs);
|
|---|
| 384 |
|
|---|
| 385 | return ($strSubDirs,$strLocalFile);
|
|---|
| 386 | }
|
|---|
| 387 |
|
|---|
| 388 | sub url_information
|
|---|
| 389 | {
|
|---|
| 390 | my ($self,$url) = @_;
|
|---|
| 391 |
|
|---|
| 392 | $url = $self->{'url'} unless defined $url;
|
|---|
| 393 |
|
|---|
| 394 | #my $strOpen = $self->start_yaz();
|
|---|
| 395 | $self->start_yaz();
|
|---|
| 396 |
|
|---|
| 397 | my $strOpen = $self->run_command_with_output("open $url","^Options");
|
|---|
| 398 |
|
|---|
| 399 | $strOpen =~ s/Z> //g;
|
|---|
| 400 | $strOpen =~ s/Elapsed:.*//g;
|
|---|
| 401 |
|
|---|
| 402 | print STDERR $strOpen;
|
|---|
| 403 |
|
|---|
| 404 | $self->quit_yaz();
|
|---|
| 405 |
|
|---|
| 406 | return 0;
|
|---|
| 407 |
|
|---|
| 408 | }
|
|---|
| 409 |
|
|---|
| 410 | sub error
|
|---|
| 411 | {
|
|---|
| 412 | my ($self,$strFunctionName,$strError) = @_;
|
|---|
| 413 | {
|
|---|
| 414 | print STDERR "Error occoured in Z3950Download.pm\n".
|
|---|
| 415 | "In Function:".$strFunctionName."\n".
|
|---|
| 416 | "Error Message:".$strError."\n";
|
|---|
| 417 | exit(-1);
|
|---|
| 418 | }
|
|---|
| 419 | }
|
|---|
| 420 |
|
|---|
| 421 | 1;
|
|---|
