1 | ###########################################################################
|
---|
2 | #
|
---|
3 | # WebDownload.pm -- base class for all the import plugins
|
---|
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 BasDownload;
|
---|
37 | use IPC::Open2;
|
---|
38 |
|
---|
39 | sub BEGIN {
|
---|
40 | @Z3950Download::ISA = ('BasDownload');
|
---|
41 | }
|
---|
42 |
|
---|
43 | my $arguments =
|
---|
44 | [ { 'name' => "host",
|
---|
45 | 'disp' => "{Z3950Download.host_disp}",
|
---|
46 | 'desc' => "{Z3950Download.host}",
|
---|
47 | 'type' => "string",
|
---|
48 | 'reqd' => "yes"},
|
---|
49 | { 'name' => "port",
|
---|
50 | 'disp' => "{Z3950Download.port_disp}",
|
---|
51 | 'desc' => "{Z3950Download.port}",
|
---|
52 | 'type' => "string",
|
---|
53 | 'reqd' => "yes"},
|
---|
54 | { 'name' => "database",
|
---|
55 | 'disp' => "{Z3950Download.database_disp}",
|
---|
56 | 'desc' => "{Z3950Download.database}",
|
---|
57 | 'type' => "string",
|
---|
58 | 'reqd' => "yes"},
|
---|
59 | { 'name' => "find",
|
---|
60 | 'disp' => "{Z3950Download.find_disp}",
|
---|
61 | 'desc' => "{Z3950Download.find}",
|
---|
62 | 'type' => "string",
|
---|
63 | 'deft' => "",
|
---|
64 | 'reqd' => "yes"},
|
---|
65 | { 'name' => "max_records",
|
---|
66 | 'disp' => "{Z3950Download.max_records_disp}",
|
---|
67 | 'desc' => "{Z3950Download.max_records}",
|
---|
68 | 'type' => "int",
|
---|
69 | 'deft' => "500",
|
---|
70 | 'reqd' => "no"}];
|
---|
71 |
|
---|
72 | my $options = { 'name' => "Z3950Download",
|
---|
73 | 'desc' => "{Z3950Download.desc}",
|
---|
74 | 'abstract' => "no",
|
---|
75 | 'inherits' => "yes",
|
---|
76 | 'args' => $arguments };
|
---|
77 |
|
---|
78 |
|
---|
79 | sub new
|
---|
80 | {
|
---|
81 | my ($class) = shift (@_);
|
---|
82 | my ($getlist,$inputargs,$hashArgOptLists) = @_;
|
---|
83 | push(@$getlist, $class);
|
---|
84 |
|
---|
85 | if(defined $arguments){ push(@{$hashArgOptLists->{"ArgList"}},@{$arguments});}
|
---|
86 | if(defined $options) { push(@{$hashArgOptLists->{"OptList"}},$options)};
|
---|
87 |
|
---|
88 | my $self = (defined $hashArgOptLists)? new BasDownload($getlist,$inputargs,$hashArgOptLists): new BasDownload($getlist,$inputargs);
|
---|
89 |
|
---|
90 | if ($self->{'info_only'}) {
|
---|
91 | # don't worry about any options etc
|
---|
92 | return bless $self, $class;
|
---|
93 | }
|
---|
94 |
|
---|
95 | # Must set $self->{'url'}, since GLI use $self->{'url'} to calculate the log file name!
|
---|
96 | $self->{'url'} = $self->{'host'}.":".$self->{'port'};
|
---|
97 |
|
---|
98 | $self->{'yaz'} = &util::filename_cat($ENV{'GSDLHOME'}, "bin", $ENV{'GSDLOS'}, "yaz-client");
|
---|
99 |
|
---|
100 | return bless $self, $class;
|
---|
101 |
|
---|
102 | }
|
---|
103 |
|
---|
104 | sub download
|
---|
105 | {
|
---|
106 | my ($self) = shift (@_);
|
---|
107 | my ($hashGeneralOptions) = @_;
|
---|
108 | my ($strOpen,$strBase,$strFind,$strResponse,$intAmount,$intMaxRecords,$strRecords);
|
---|
109 |
|
---|
110 | my $url = $self->{'url'};
|
---|
111 |
|
---|
112 | print STDERR "<<Defined Maximum>>\n";
|
---|
113 |
|
---|
114 | print STDERR "Opening connection to $url\n";
|
---|
115 |
|
---|
116 | my $yaz = $self->{'yaz'};
|
---|
117 |
|
---|
118 | my $childpid = open2(*YAZOUT, *YAZIN, $yaz)
|
---|
119 | or (print STDERR "<<Finished>>\n" and die "can't open pipe to yaz-client: $!");
|
---|
120 | $self->{'YAZOUT'} = *YAZOUT;
|
---|
121 | $self->{'YAZIN'} = *YAZIN;
|
---|
122 |
|
---|
123 | $strOpen = $self->open_connection("open $url");
|
---|
124 |
|
---|
125 | if (!$strOpen) {
|
---|
126 | print STDERR "Cannot connect to $url\n";
|
---|
127 | print STDERR "<<Finished>>\n";
|
---|
128 | return 0;
|
---|
129 | }
|
---|
130 |
|
---|
131 | print STDERR "Access database: \"$self->{'database'}\"\n";
|
---|
132 | $self->run_command_without_output("base $self->{'database'}");
|
---|
133 | print STDERR "Searching for keyword: \"$self->{'find'}\"\n";
|
---|
134 | $intAmount = $self->findAmount($self->{'find'});
|
---|
135 |
|
---|
136 | if($intAmount <= 0)
|
---|
137 | {
|
---|
138 | ($intAmount == -1)?
|
---|
139 | print STDERR "Something wrong with the arguments,downloading can not be performed\n":
|
---|
140 | print STDERR "No Record is found\n";
|
---|
141 | print STDERR "<<Finished>>\n";
|
---|
142 | return 0;
|
---|
143 | }
|
---|
144 | $intMaxRecords = ($self->{'max_records'} > $intAmount)? $intAmount : $self->{'max_records'};
|
---|
145 | print STDERR "<<Total number of record(s):$intMaxRecords>>\n";
|
---|
146 | $strRecords = "Records: $intMaxRecords\n".$self->getRecords($intMaxRecords);
|
---|
147 |
|
---|
148 | $self->saveRecords($strRecords,$hashGeneralOptions->{'cache_dir'},$intMaxRecords);
|
---|
149 | print STDERR "Closing connection...\n";
|
---|
150 | print STDERR "<<Finished>>\n";
|
---|
151 |
|
---|
152 | close(YAZOUT);
|
---|
153 | close(YAZIN);
|
---|
154 | return 1;
|
---|
155 | }
|
---|
156 |
|
---|
157 | sub open_connection{
|
---|
158 | my ($self,$strCommand) = (@_);
|
---|
159 |
|
---|
160 | $self->run_command($strCommand);
|
---|
161 |
|
---|
162 | my $out = $self->{'YAZOUT'};
|
---|
163 |
|
---|
164 | $_ = <$out>;
|
---|
165 |
|
---|
166 | return (/Connecting...OK/i)? 1: 0;
|
---|
167 |
|
---|
168 | }
|
---|
169 |
|
---|
170 | sub findAmount
|
---|
171 | {
|
---|
172 | my ($self) = shift (@_);
|
---|
173 | my($strFindTarget) = @_;
|
---|
174 | my $strResponse = $self->run_command_with_output("find $strFindTarget","^Number of hits:");
|
---|
175 | return ($strResponse =~ m/^Number of hits: (\d+)/m)? $1:-1;
|
---|
176 | }
|
---|
177 |
|
---|
178 | sub getRecords
|
---|
179 | {
|
---|
180 | my ($self) = shift (@_);
|
---|
181 | my ($intMaxRecords) = @_;
|
---|
182 | my ($strShow,$intStartNumber,$numRecords,$strResponse,$strRecords,$intRecordsLeft);
|
---|
183 |
|
---|
184 | $intStartNumber = 1;
|
---|
185 | $intRecordsLeft = $intMaxRecords;
|
---|
186 | $numRecords = 0;
|
---|
187 | $strResponse ="";
|
---|
188 |
|
---|
189 | while ($intRecordsLeft > 0)
|
---|
190 | {
|
---|
191 | if($intRecordsLeft > 50)
|
---|
192 | {
|
---|
193 |
|
---|
194 | print STDERR "Yaz is Gathering records: $intStartNumber - ".($intStartNumber+49)."\n";
|
---|
195 | $numRecords = 50;
|
---|
196 | $strShow = "show $intStartNumber+50";
|
---|
197 | $intStartNumber = $intStartNumber + 50;
|
---|
198 | $intRecordsLeft = $intRecordsLeft - 50;
|
---|
199 |
|
---|
200 | }
|
---|
201 | else
|
---|
202 | {
|
---|
203 | $numRecords = $intRecordsLeft;
|
---|
204 | print STDERR "Yaz is Gathering records: $intStartNumber - ".($intStartNumber+$intRecordsLeft-1)."\n";
|
---|
205 | $strShow = "show $intStartNumber+$intRecordsLeft";
|
---|
206 | $intRecordsLeft = 0;
|
---|
207 |
|
---|
208 | }
|
---|
209 |
|
---|
210 | $strResponse .= $self->get($strShow,$numRecords);
|
---|
211 |
|
---|
212 | if ($strResponse eq ""){
|
---|
213 | print STDERR "<<ERROR: failed to get $numRecords records>>\n";
|
---|
214 | }
|
---|
215 | else{
|
---|
216 | print STDERR "<<Done:$numRecords>>\n";
|
---|
217 | }
|
---|
218 | }
|
---|
219 |
|
---|
220 | return "$strResponse\n";
|
---|
221 |
|
---|
222 | }
|
---|
223 |
|
---|
224 | sub saveRecords
|
---|
225 | {
|
---|
226 | my ($self,$strRecords,$strOutputDir,$intMaxRecords) = @_;
|
---|
227 |
|
---|
228 | # setup directories
|
---|
229 | # Currently only gather the MARC format
|
---|
230 | my $strFileName = $self->generateFileName($intMaxRecords);
|
---|
231 |
|
---|
232 | $strOutputDir =~ s/"//g; #"
|
---|
233 |
|
---|
234 | my $strOutputFile = &util::filename_cat($strOutputDir,$self->{'host'},"$strFileName.marc");
|
---|
235 | # prepare subdirectory for record (if needed)
|
---|
236 | my ($strSubDirPath,$unused) = $self->dirFileSplit($strOutputFile);
|
---|
237 |
|
---|
238 | &util::mk_all_dir($strSubDirPath);
|
---|
239 |
|
---|
240 | print STDERR "Saving records to \"$strOutputFile\"\n";
|
---|
241 |
|
---|
242 | # save record
|
---|
243 | open (ZOUT,">$strOutputFile")
|
---|
244 | || die "Unable to save Z3950 record: $!\n";
|
---|
245 | print ZOUT $strRecords;
|
---|
246 | close(ZOUT);
|
---|
247 | }
|
---|
248 |
|
---|
249 |
|
---|
250 | sub run_command_with_output
|
---|
251 | {
|
---|
252 | my ($self,$strCMD,$strStopRE) =@_;
|
---|
253 |
|
---|
254 | $self->run_command($strCMD);
|
---|
255 |
|
---|
256 | return $self->get_output($strStopRE);
|
---|
257 |
|
---|
258 | }
|
---|
259 |
|
---|
260 | sub get{
|
---|
261 | my ($self,$strShow,$numRecord) = @_;
|
---|
262 |
|
---|
263 | $self->run_command($strShow);
|
---|
264 |
|
---|
265 | my $strFullOutput="";
|
---|
266 | my $count=0;
|
---|
267 | my $readRecord = 0;
|
---|
268 |
|
---|
269 | while (my $strLine = <YAZOUT>)
|
---|
270 | {
|
---|
271 |
|
---|
272 | if ($strLine =~ m/Records: ([\d]*)/i ){
|
---|
273 | $readRecord = 1;
|
---|
274 | next;
|
---|
275 | }
|
---|
276 |
|
---|
277 | return $strFullOutput if ($strLine =~ m/nextResultSetPosition|Not connected/i);
|
---|
278 |
|
---|
279 | next if(!$readRecord);
|
---|
280 |
|
---|
281 | $strFullOutput .= $strLine;
|
---|
282 | }
|
---|
283 |
|
---|
284 | }
|
---|
285 |
|
---|
286 | sub run_command_without_output
|
---|
287 | {
|
---|
288 | my ($self) = shift (@_);
|
---|
289 | my ($strCMD) = @_;
|
---|
290 |
|
---|
291 | $self->run_command($strCMD);
|
---|
292 | }
|
---|
293 |
|
---|
294 | sub run_command
|
---|
295 | {
|
---|
296 | my ($self,$strCMD) = @_;
|
---|
297 |
|
---|
298 | my $input = $self->{'YAZIN'};
|
---|
299 |
|
---|
300 | print $input "$strCMD\n";
|
---|
301 | }
|
---|
302 |
|
---|
303 | sub get_output{
|
---|
304 | my ($self,$strStopRE) = @_;
|
---|
305 |
|
---|
306 | if (!defined $strStopRE){return "";}
|
---|
307 | else
|
---|
308 | {
|
---|
309 | my $strFullOutput;
|
---|
310 | my $output = $self->{'YAZOUT'};
|
---|
311 | while (my $strLine = <$output>)
|
---|
312 | {
|
---|
313 | $strFullOutput .= $strLine;
|
---|
314 | if($strLine =~ m/^$strStopRE|Not connected/i){return $strFullOutput;}
|
---|
315 | }
|
---|
316 | }
|
---|
317 | }
|
---|
318 |
|
---|
319 | sub generateFileName
|
---|
320 | {
|
---|
321 | my ($self,$intMaxRecords) = @_;
|
---|
322 | my $strFileName = ($self->{'database'})."_".($self->{'find'})."_".($intMaxRecords);
|
---|
323 |
|
---|
324 | }
|
---|
325 |
|
---|
326 | sub dirFileSplit
|
---|
327 | {
|
---|
328 | my ($self,$strFile) = @_;
|
---|
329 |
|
---|
330 | my @aryDirs = split(/[\/\\]/,$strFile);
|
---|
331 |
|
---|
332 | my $strLocalFile = pop(@aryDirs);
|
---|
333 | my $strSubDirs = join("/",@aryDirs);
|
---|
334 |
|
---|
335 | return ($strSubDirs,$strLocalFile);
|
---|
336 | }
|
---|
337 |
|
---|
338 | sub url_information
|
---|
339 | {
|
---|
340 | my ($self,$url) = @_;
|
---|
341 |
|
---|
342 | $url = $self->{'url'} unless defined $url;
|
---|
343 |
|
---|
344 | my $yaz = $self->{'yaz'};
|
---|
345 |
|
---|
346 | my $childpid = open2(*YAZOUT, *YAZIN, $yaz)
|
---|
347 | or die "can't open pipe to yaz-client: $!";
|
---|
348 |
|
---|
349 | $self->{'YAZOUT'} = *YAZOUT;
|
---|
350 | $self->{'YAZIN'} = *YAZIN;
|
---|
351 |
|
---|
352 | my $strOpen = $self->open_connection("open $url");
|
---|
353 |
|
---|
354 | if (!$strOpen) {
|
---|
355 | print STDERR "Cannot connect to $url\n";
|
---|
356 | print STDERR "<<Finished>>\n";
|
---|
357 | return 0;
|
---|
358 | }
|
---|
359 |
|
---|
360 |
|
---|
361 | $strOpen = $self->run_command_with_output("open $url","^Options");
|
---|
362 |
|
---|
363 |
|
---|
364 | $strOpen =~ s/Z> //g;
|
---|
365 | $strOpen =~ s/Elapsed:.*//g;
|
---|
366 |
|
---|
367 | print STDERR $strOpen;
|
---|
368 |
|
---|
369 | print STDERR "<<Finished>>\n";
|
---|
370 |
|
---|
371 | close(YAZOUT);
|
---|
372 | close(YAZIN);
|
---|
373 |
|
---|
374 | return 0;
|
---|
375 |
|
---|
376 | }
|
---|
377 |
|
---|
378 | sub error
|
---|
379 | {
|
---|
380 | my ($self,$strFunctionName,$strError) = @_;
|
---|
381 | {
|
---|
382 | print STDERR "Error occoured in Z3950Download.pm\n".
|
---|
383 | "In Function:".$strFunctionName."\n".
|
---|
384 | "Error Message:".$strError."\n";
|
---|
385 | exit(-1);
|
---|
386 | }
|
---|
387 | }
|
---|
388 |
|
---|
389 | 1;
|
---|