1 | #!/usr/bin/perl -w
|
---|
2 |
|
---|
3 | ###########################################################################
|
---|
4 | #
|
---|
5 | # solr_passes.pl -- perl wrapper, akin to mgpp_passes, for Solr
|
---|
6 | # A component of the Greenstone digital library software
|
---|
7 | # from the New Zealand Digital Library Project at the
|
---|
8 | # University of Waikato, New Zealand.
|
---|
9 | #
|
---|
10 | # Copyright (C) 1999 New Zealand Digital Library Project
|
---|
11 | #
|
---|
12 | # This program is free software; you can redistribute it and/or modify
|
---|
13 | # it under the terms of the GNU General Public License as published by
|
---|
14 | # the Free Software Foundation; either version 2 of the License, or
|
---|
15 | # (at your option) any later version.
|
---|
16 | #
|
---|
17 | # This program is distributed in the hope that it will be useful,
|
---|
18 | # but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
19 | # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
20 | # GNU General Public License for more details.
|
---|
21 | #
|
---|
22 | # You should have received a copy of the GNU General Public License
|
---|
23 | # along with this program; if not, write to the Free Software
|
---|
24 | # Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
25 | #
|
---|
26 | ###########################################################################
|
---|
27 |
|
---|
28 | # Heavily based on lucene_passes.pl, but does not need a SolrWrapper.jar
|
---|
29 | # style solution as Solr has its own XML syntax:
|
---|
30 | #
|
---|
31 | # http://wiki.apache.org/solr/UpdateXmlMessages
|
---|
32 | #
|
---|
33 | # This syntax is rather similar to what we already use, so the
|
---|
34 | # main task of monitor_xml() is to translate the XML syntax Greenstone uses
|
---|
35 | # into that needed by the solr server.
|
---|
36 |
|
---|
37 | my $jetty_stop_key="greenstone-solr";
|
---|
38 |
|
---|
39 | BEGIN {
|
---|
40 | die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
|
---|
41 | die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
|
---|
42 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
|
---|
43 | die "GEXT_SOLR not set\n" unless defined $ENV{'GEXT_SOLR'};
|
---|
44 | }
|
---|
45 |
|
---|
46 | use strict;
|
---|
47 | use util;
|
---|
48 |
|
---|
49 |
|
---|
50 | # Not quite OO, but close enough for now
|
---|
51 | #
|
---|
52 | my $self = { 'full_server_jar' => undef,
|
---|
53 | ## 'full_post_jar' = undef,
|
---|
54 | ## 'full_index_propfile' = undef,
|
---|
55 | 'jetty_explicitly_started' => undef
|
---|
56 | };
|
---|
57 |
|
---|
58 |
|
---|
59 |
|
---|
60 | sub locate_file
|
---|
61 | {
|
---|
62 | my ($search_path,$suffix) = @_;
|
---|
63 |
|
---|
64 | foreach my $sp (@$search_path) {
|
---|
65 | my $full_path = &util::filename_cat($sp,$suffix);
|
---|
66 |
|
---|
67 | if (-f $full_path) {
|
---|
68 | return $full_path;
|
---|
69 | }
|
---|
70 | }
|
---|
71 |
|
---|
72 | # if get to here, then failed to find match
|
---|
73 |
|
---|
74 | print STDERR "Error: Failed to find '$suffix'\n";
|
---|
75 | print STDERR " Looked in: ", join(", ", @$search_path), "\n";
|
---|
76 | exit -1;
|
---|
77 | }
|
---|
78 |
|
---|
79 | sub start_solr_server
|
---|
80 | {
|
---|
81 | my ($search_path) = @_;
|
---|
82 |
|
---|
83 | my $solr_home = $ENV{'GEXT_SOLR'};
|
---|
84 | my $jetty_stop_port = $ENV{'JETTY_STOP_PORT'};
|
---|
85 |
|
---|
86 | chdir($solr_home);
|
---|
87 |
|
---|
88 | my $solr_etc = &util::filename_cat($solr_home,"etc");
|
---|
89 |
|
---|
90 | my $server_props = "-DSTOP.PORT=$jetty_stop_port";
|
---|
91 | $server_props .= " -DSTOP.KEY=$jetty_stop_key";
|
---|
92 | $server_props .= " -Dsolr.solr.home=$solr_etc";
|
---|
93 |
|
---|
94 | my $server_jar = &util::filename_cat("lib","java","solr-jetty-server.jar");
|
---|
95 | my $full_server_jar = locate_file($search_path,$server_jar);
|
---|
96 | $self->{'full_server_jar'} = $full_server_jar;
|
---|
97 |
|
---|
98 | my $server_java_cmd = "java $server_props -jar \"$full_server_jar\"";
|
---|
99 | # if ($ENV{'GSDLOS'} eq "windows") {
|
---|
100 | # $server_java_cmd = "start $server_java_cmd";
|
---|
101 | # }
|
---|
102 | # else {
|
---|
103 | # $server_java_cmd .= " &";
|
---|
104 | # }
|
---|
105 |
|
---|
106 |
|
---|
107 | if (open(SIN,"$server_java_cmd 2>&1 |")) {
|
---|
108 |
|
---|
109 | my $line;
|
---|
110 | while (defined($line=<SIN>)) {
|
---|
111 | # Scan through output until you see a line like:
|
---|
112 | # 2011-08-22 .. :INFO::Started [email protected]:8983
|
---|
113 | # which signifies that the server has started up and is
|
---|
114 | # "ready and listening"
|
---|
115 |
|
---|
116 | last if ($line =~ m/INFO::Started SocketConnector/);
|
---|
117 | }
|
---|
118 |
|
---|
119 | print STDERR "Jetty server ready and listening for connections\n";
|
---|
120 |
|
---|
121 | # now we know the server is ready to accept connections, fork a
|
---|
122 | # child process that continues to listen to the output and
|
---|
123 | # prints out any lines that are not INFO lines
|
---|
124 |
|
---|
125 | if (fork()==0) {
|
---|
126 | # child process
|
---|
127 |
|
---|
128 | my $line;
|
---|
129 | while (defined ($line = <SIN>)) {
|
---|
130 | next if ($line =~ m/^INFO:/);
|
---|
131 | next if ($line =~ m/^[0-9 :-]*INFO::/);
|
---|
132 | next if ($line =~ m/^\d{2}\/\d{2}\/\d{4}\s+/);
|
---|
133 | }
|
---|
134 | close(SIN);
|
---|
135 |
|
---|
136 | # And now stop nicely
|
---|
137 | exit 0;
|
---|
138 | }
|
---|
139 | }
|
---|
140 | else {
|
---|
141 | print STDERR "Error: failed to start solr-jetty-server\n";
|
---|
142 | print STDERR "!$\n";
|
---|
143 | exit -1;
|
---|
144 | }
|
---|
145 |
|
---|
146 | # If get to here then server started (and ready and listening)
|
---|
147 | # *and* we are the parent process of the fork()
|
---|
148 |
|
---|
149 | $self->{'jetty_explicitly_started'} = 1;
|
---|
150 |
|
---|
151 | }
|
---|
152 |
|
---|
153 |
|
---|
154 |
|
---|
155 | sub stop_solr_server
|
---|
156 | {
|
---|
157 | my $full_server_jar = $self->{'full_server_jar'};
|
---|
158 | my $jetty_stop_port = $ENV{'JETTY_STOP_PORT'};
|
---|
159 |
|
---|
160 | my $server_props = "-DSTOP.PORT=$jetty_stop_port -DSTOP.KEY=$jetty_stop_key";
|
---|
161 | my $server_java_cmd = "java $server_props -jar \"$full_server_jar\" --stop";
|
---|
162 |
|
---|
163 | my $server_status = system($server_java_cmd);
|
---|
164 |
|
---|
165 | if ($server_status!=0) {
|
---|
166 | print STDERR "Error: failed to stop solr-jetty-server\n";
|
---|
167 | print STDERR "!$\n";
|
---|
168 | exit -1;
|
---|
169 | }
|
---|
170 | else {
|
---|
171 | wait(); # let the child process finish
|
---|
172 | }
|
---|
173 | }
|
---|
174 |
|
---|
175 |
|
---|
176 | sub open_java_solr
|
---|
177 | {
|
---|
178 | my ($collect, $doc_tag_level,$full_builddir,$indexdir,$removeold) = @_;
|
---|
179 |
|
---|
180 |
|
---|
181 | # if removeold set, then delete the curring $full_builddir
|
---|
182 | if ($removeold) {
|
---|
183 | my $full_indexdir = &util::filename_cat($full_builddir,$indexdir);
|
---|
184 | &util::rm_r($full_indexdir);
|
---|
185 | }
|
---|
186 |
|
---|
187 | # No longer used, as solr now run as two cores per collection (-Doc and -Sec)
|
---|
188 | #
|
---|
189 | # # Set up an 'index.properties' file in $full_builddir, with the line
|
---|
190 | # # index=$indexdir
|
---|
191 | # $full_index_propfile = &util::filename_cat($full_builddir,"index.properties");
|
---|
192 |
|
---|
193 | # if (open(IPOUT,">$full_index_propfile")) {
|
---|
194 | # print IPOUT "index=$indexdir\n";
|
---|
195 | # close(IPOUT);
|
---|
196 | # }
|
---|
197 | # else {
|
---|
198 | # print STDERR "Failed to create $full_index_propfile\n";
|
---|
199 | # print STDERR "!$\n";
|
---|
200 | # exit -2;
|
---|
201 | # }
|
---|
202 |
|
---|
203 |
|
---|
204 |
|
---|
205 | my $search_path = [];
|
---|
206 |
|
---|
207 | push(@$search_path,$ENV{'GSDLCOLLECTDIR'}) if defined $ENV{'GSDLCOLLECTDIR'};
|
---|
208 | push(@$search_path,$ENV{'GSDLHOME'}) if defined $ENV{'GSDLHOME'};
|
---|
209 | push(@$search_path,$ENV{'GEXT_SOLR'}) if defined $ENV{'GEXT_SOLR'};
|
---|
210 |
|
---|
211 |
|
---|
212 | # The following returns once Jetty has generated its
|
---|
213 | # "reading and listening" line
|
---|
214 | #
|
---|
215 | start_solr_server($search_path);
|
---|
216 |
|
---|
217 |
|
---|
218 | # Now run the solr-post command
|
---|
219 |
|
---|
220 | chdir($ENV{'GEXT_SOLR'});
|
---|
221 |
|
---|
222 |
|
---|
223 | my $post_jar = &util::filename_cat("lib","java","solr-post.jar");
|
---|
224 | my $full_post_jar = locate_file($search_path,$post_jar);
|
---|
225 | ## $self->{'full_post_jar'} = $full_post_jar;
|
---|
226 |
|
---|
227 | my $jetty_server_port = $ENV{'SOLR_JETTY_PORT'};
|
---|
228 |
|
---|
229 | # Now run solr-post command
|
---|
230 | my $post_props = "-Durl=http://localhost:$jetty_server_port/solr/$collect-$doc_tag_level/update";
|
---|
231 | $post_props .= " -Ddata=stdin";
|
---|
232 | $post_props .= " -Dcommit=yes";
|
---|
233 |
|
---|
234 | my $post_java_cmd = "java $post_props -jar \"$full_post_jar\"";
|
---|
235 |
|
---|
236 | print STDERR "**** post cmd = $post_java_cmd\n";
|
---|
237 |
|
---|
238 | open (PIPEOUT, "| $post_java_cmd")
|
---|
239 | || die "Error in solr_passes.pl: Failed to run $post_java_cmd\n!$\n";
|
---|
240 | }
|
---|
241 |
|
---|
242 |
|
---|
243 |
|
---|
244 | sub close_java_solr
|
---|
245 | {
|
---|
246 | # closing the pipe has the effect of shutting down solr-post.jar
|
---|
247 | close(PIPEOUT);
|
---|
248 |
|
---|
249 | if ($self->{'jetty_explicitly_started'}) {
|
---|
250 | stop_solr_server();
|
---|
251 | }
|
---|
252 |
|
---|
253 | # No longer used
|
---|
254 | # # $full_index_propfile is set up as a global variable so it can be shared
|
---|
255 | # # between open_java_solr() and here
|
---|
256 | # &util::rm($full_index_propfile);
|
---|
257 | }
|
---|
258 |
|
---|
259 |
|
---|
260 |
|
---|
261 |
|
---|
262 |
|
---|
263 | sub save_xml_doc
|
---|
264 | {
|
---|
265 | # This is identical to the one in lucene_passes.pl, and should be
|
---|
266 | # moved in to a package and shared ####
|
---|
267 |
|
---|
268 | my ($full_textdir,$output_filename,$doc_xml) = @_;
|
---|
269 |
|
---|
270 | my $dir_sep = &util::get_os_dirsep();
|
---|
271 |
|
---|
272 | my $full_output_filename = &util::filename_cat($full_textdir,$output_filename);
|
---|
273 | my ($full_output_dir) = ($full_output_filename =~ m/^(.*$dir_sep)/x);
|
---|
274 | &util::mk_all_dir($full_output_dir);
|
---|
275 |
|
---|
276 | open(DOCOUT,">$full_output_filename")
|
---|
277 | || die "Unable to open $full_output_filename";
|
---|
278 |
|
---|
279 | print DOCOUT $doc_xml;
|
---|
280 | close(DOCOUT);
|
---|
281 |
|
---|
282 | # What this the purpose of the following? ####
|
---|
283 | my @secs = ($doc_xml =~ m/<Sec\s+gs2:id="\d+"\s*>.*?<\/Sec>/sg);
|
---|
284 | }
|
---|
285 |
|
---|
286 |
|
---|
287 | sub compress_xml_doc
|
---|
288 | {
|
---|
289 | # This is identical to the one in lucene_passes.pl, and should be
|
---|
290 | # moved in to a package and shared ####
|
---|
291 |
|
---|
292 | my ($full_textdir,$output_filename) = @_;
|
---|
293 |
|
---|
294 | my $full_output_filename
|
---|
295 | = &util::filename_cat($full_textdir,$output_filename);
|
---|
296 |
|
---|
297 | # Greenstone ships with gzip for Windows
|
---|
298 | `gzip $full_output_filename`;
|
---|
299 | }
|
---|
300 |
|
---|
301 |
|
---|
302 | sub monitor_xml_stream
|
---|
303 | {
|
---|
304 | # based on lucene's monitor_xml_stream, but simplified
|
---|
305 | # as only now used when in "text" mode
|
---|
306 |
|
---|
307 | my ($full_textdir) = @_;
|
---|
308 |
|
---|
309 | my $doc_xml = "";
|
---|
310 | my $output_filename = "";
|
---|
311 |
|
---|
312 | my $line;
|
---|
313 | while (defined ($line = <STDIN>)) {
|
---|
314 |
|
---|
315 | $doc_xml .= $line;
|
---|
316 |
|
---|
317 | if ($line =~ m/^<Doc.+file=\"(.*?)\".*>$/) {
|
---|
318 | $output_filename = $1;
|
---|
319 | }
|
---|
320 |
|
---|
321 | if ($line =~ m/^<\/Doc>$/) {
|
---|
322 | save_xml_doc($full_textdir,$output_filename,$doc_xml);
|
---|
323 |
|
---|
324 | # Compress file
|
---|
325 | #
|
---|
326 | # The compress option was taken out for efficiency
|
---|
327 | # reasons. Consider putting it back in but making it a
|
---|
328 | # switch so a collection builder can decide for themselves on a
|
---|
329 | # case by case basis if they want to save on diskspace, but have
|
---|
330 | # the overhead of uncompressing at runtime
|
---|
331 |
|
---|
332 | ### compress_xml_doc($full_textdir,$output_filename);
|
---|
333 |
|
---|
334 | $doc_xml = "";
|
---|
335 | $output_filename = "";
|
---|
336 | }
|
---|
337 | }
|
---|
338 | }
|
---|
339 |
|
---|
340 |
|
---|
341 | sub pass_on_xml_stream
|
---|
342 | {
|
---|
343 | my $line;
|
---|
344 | while (defined ($line = <STDIN>)) {
|
---|
345 | print PIPEOUT $line;
|
---|
346 | }
|
---|
347 | }
|
---|
348 |
|
---|
349 |
|
---|
350 |
|
---|
351 |
|
---|
352 | # /** This checks the arguments on the command line, filters the
|
---|
353 | # * unknown command line arguments and then calls the open_java_solr
|
---|
354 | # * function to begin processing.
|
---|
355 | # */
|
---|
356 | sub main
|
---|
357 | {
|
---|
358 | my (@argv) = @_;
|
---|
359 | my $argc = scalar(@argv);
|
---|
360 |
|
---|
361 | my $removeold = 0;
|
---|
362 | my @filtered_argv = ();
|
---|
363 |
|
---|
364 | my $i = 0;
|
---|
365 | while ($i<$argc) {
|
---|
366 | if ($argv[$i] =~ m/^\-(.*)$/) {
|
---|
367 |
|
---|
368 | my $option = $1;
|
---|
369 |
|
---|
370 | # -removeold causes the existing index to be overwritten
|
---|
371 | if ($option eq "removeold") {
|
---|
372 | print STDERR "\n-removeold set (new index will be created)\n";
|
---|
373 | $removeold = 1;
|
---|
374 | }
|
---|
375 | # -verbosity <num>
|
---|
376 | elsif ($option eq "verbosity") {
|
---|
377 | $i++;
|
---|
378 | if ($i<$argc)
|
---|
379 | {
|
---|
380 | # solr indexing has no support for verbosity
|
---|
381 | # => parse to be compatible with calling program, but supress it
|
---|
382 | # for solr-post.jar
|
---|
383 | }
|
---|
384 | }
|
---|
385 | else {
|
---|
386 | print STDERR "Unrecognised minus option: -$option\n";
|
---|
387 | }
|
---|
388 | }
|
---|
389 | else {
|
---|
390 | push(@filtered_argv,$argv[$i]);
|
---|
391 | }
|
---|
392 | $i++;
|
---|
393 | }
|
---|
394 |
|
---|
395 | my $filtered_argc = scalar(@filtered_argv);
|
---|
396 |
|
---|
397 | if ($filtered_argc < 5) {
|
---|
398 | print STDERR "Usage: solr_passes.pl [-removeold|-verbosity num] collect \"text\"|\"index\" doc-tag-level build-dir index-name\n";
|
---|
399 | exit 1;
|
---|
400 | }
|
---|
401 |
|
---|
402 | my $collect = $filtered_argv[0];
|
---|
403 | my $mode = $filtered_argv[1];
|
---|
404 | my $doc_tag_level = $filtered_argv[2];
|
---|
405 | my $full_builddir = $filtered_argv[3];
|
---|
406 | my $indexdir = $filtered_argv[4];
|
---|
407 |
|
---|
408 | # We only need the Solr handle opened if we are indexing the
|
---|
409 | # documents, not if we are just storing the text
|
---|
410 | if ($mode eq "index") {
|
---|
411 | open_java_solr($collect, $doc_tag_level, $full_builddir, $indexdir, $removeold);
|
---|
412 | }
|
---|
413 |
|
---|
414 | if ($mode eq "text") {
|
---|
415 | print STDERR "Monitoring for input!\n";
|
---|
416 | my $full_textdir = &util::filename_cat($full_builddir,"text");
|
---|
417 | monitor_xml_stream($full_textdir);
|
---|
418 | }
|
---|
419 | else {
|
---|
420 | print STDERR "Streaming input onto solr server!\n";
|
---|
421 | pass_on_xml_stream();
|
---|
422 | }
|
---|
423 |
|
---|
424 |
|
---|
425 | if ($mode eq "index") {
|
---|
426 | close_java_solr();
|
---|
427 | }
|
---|
428 | }
|
---|
429 |
|
---|
430 |
|
---|
431 | &main(@ARGV);
|
---|