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 |
|
---|
38 | BEGIN {
|
---|
39 | die "GSDLHOME not set\n" unless defined $ENV{'GSDLHOME'};
|
---|
40 | die "GSDLOS not set\n" unless defined $ENV{'GSDLOS'};
|
---|
41 | unshift (@INC, "$ENV{'GSDLHOME'}/perllib");
|
---|
42 | die "GEXT_SOLR not set\n" unless defined $ENV{'GEXT_SOLR'};
|
---|
43 |
|
---|
44 | my $solr_ext = $ENV{'GEXT_SOLR'};
|
---|
45 | unshift (@INC, "$solr_ext/perllib");
|
---|
46 | }
|
---|
47 |
|
---|
48 | use strict;
|
---|
49 | use util;
|
---|
50 | use solrutil;
|
---|
51 | use solrserver;
|
---|
52 |
|
---|
53 | my $DOC_BATCH_SIZE = 20;
|
---|
54 |
|
---|
55 | # Not quite OO, but close enough for now
|
---|
56 | #
|
---|
57 | my $self = { 'solr_server' => undef };
|
---|
58 |
|
---|
59 | sub open_java_solr
|
---|
60 | {
|
---|
61 | my ($core,$full_builddir,$indexdir) = @_;
|
---|
62 |
|
---|
63 | # If the Solr/Jetty server is not already running, the following starts
|
---|
64 | # it up, and only returns when the server is "reading and listening"
|
---|
65 |
|
---|
66 | my $solr_server = new solrserver($full_builddir);
|
---|
67 | $solr_server->start();
|
---|
68 | $self->{'solr_server'} = $solr_server;
|
---|
69 |
|
---|
70 | # Now start up the solr-post command and store the open cmd that is returned
|
---|
71 | $self->{'post_java_cmd'} = &solrutil::open_post_pipe($core, $solr_server->get_solr_base_url());
|
---|
72 | }
|
---|
73 |
|
---|
74 | # To commit any stream of data to solr that's amassed so far on the pipe to SimplePostTool,
|
---|
75 | # close the pipe. Then reopen it to continue streaming data to it.
|
---|
76 | sub flush_and_reopen_java_solr
|
---|
77 | {
|
---|
78 | &solrutil::close_post_pipe();
|
---|
79 | &solrutil::reopen_post_pipe($self->{'post_java_cmd'});
|
---|
80 | }
|
---|
81 |
|
---|
82 | sub close_java_solr
|
---|
83 | {
|
---|
84 | &solrutil::close_post_pipe();
|
---|
85 |
|
---|
86 | my $solr_server = $self->{'solr_server'};
|
---|
87 | if ($solr_server->explicitly_started()) {
|
---|
88 | $solr_server->stop();
|
---|
89 | }
|
---|
90 | }
|
---|
91 |
|
---|
92 | #----
|
---|
93 |
|
---|
94 | sub save_xml_doc
|
---|
95 | {
|
---|
96 | # This is identical to the one in lucene_passes.pl, and should be
|
---|
97 | # moved in to a package and shared ####
|
---|
98 |
|
---|
99 | my ($full_textdir,$output_filename,$doc_xml) = @_;
|
---|
100 |
|
---|
101 | my $dir_sep = &util::get_os_dirsep();
|
---|
102 |
|
---|
103 | my $full_output_filename = &util::filename_cat($full_textdir,$output_filename);
|
---|
104 | my ($full_output_dir) = ($full_output_filename =~ m/^(.*$dir_sep)/x);
|
---|
105 | &FileUtils::makeAllDirectories($full_output_dir);
|
---|
106 |
|
---|
107 | open(DOCOUT,">$full_output_filename")
|
---|
108 | || die "Unable to open $full_output_filename";
|
---|
109 |
|
---|
110 | print DOCOUT $doc_xml;
|
---|
111 | close(DOCOUT);
|
---|
112 |
|
---|
113 | # What this the purpose of the following? ####
|
---|
114 | my @secs = ($doc_xml =~ m/<Sec\s+gs2:id="\d+"\s*>.*?<\/Sec>/sg);
|
---|
115 | }
|
---|
116 |
|
---|
117 |
|
---|
118 | sub compress_xml_doc
|
---|
119 | {
|
---|
120 | # This is identical to the one in lucene_passes.pl, and should be
|
---|
121 | # moved in to a package and shared ####
|
---|
122 |
|
---|
123 | my ($full_textdir,$output_filename) = @_;
|
---|
124 |
|
---|
125 | my $full_output_filename
|
---|
126 | = &util::filename_cat($full_textdir,$output_filename);
|
---|
127 |
|
---|
128 | # Greenstone ships with gzip for Windows
|
---|
129 | `gzip $full_output_filename`;
|
---|
130 | }
|
---|
131 |
|
---|
132 |
|
---|
133 | sub monitor_xml_stream
|
---|
134 | {
|
---|
135 | # based on lucene's monitor_xml_stream, but simplified
|
---|
136 | # as only now used when in "text" mode
|
---|
137 |
|
---|
138 | my ($full_textdir) = @_;
|
---|
139 |
|
---|
140 | my $doc_xml = "";
|
---|
141 | my $output_filename = "";
|
---|
142 |
|
---|
143 | my $line;
|
---|
144 | while (defined ($line = <STDIN>)) {
|
---|
145 |
|
---|
146 | $doc_xml .= $line;
|
---|
147 |
|
---|
148 | if ($line =~ m/^<Doc.+file=\"(.*?)\".*>$/) {
|
---|
149 | $output_filename = $1;
|
---|
150 | }
|
---|
151 |
|
---|
152 | if ($line =~ m/^<\/Doc>$/) {
|
---|
153 | save_xml_doc($full_textdir,$output_filename,$doc_xml);
|
---|
154 |
|
---|
155 | # Compress file
|
---|
156 | #
|
---|
157 | # The compress option was taken out for efficiency
|
---|
158 | # reasons. Consider putting it back in but making it a
|
---|
159 | # switch so a collection builder can decide for themselves on a
|
---|
160 | # case by case basis if they want to save on diskspace, but have
|
---|
161 | # the overhead of uncompressing at runtime
|
---|
162 |
|
---|
163 | ### compress_xml_doc($full_textdir,$output_filename);
|
---|
164 |
|
---|
165 | $doc_xml = "";
|
---|
166 | $output_filename = "";
|
---|
167 | }
|
---|
168 | }
|
---|
169 | }
|
---|
170 |
|
---|
171 |
|
---|
172 | # Called when mode = index,
|
---|
173 | # This is the function that passes the contents of the docs for ingesting into solr
|
---|
174 | # as one long stream.
|
---|
175 | # Since if we have many docs in a collection, there will be one long stream of bytes
|
---|
176 | # sent to the SimplePostTool/(solr-)post.jar, which remain uncommitted until
|
---|
177 | # the pipe is closed. And for a long bytestream, this will result in an out of heap memory
|
---|
178 | # https://stackoverflow.com/questions/2082057/outputstream-outofmemoryerror-when-sending-http
|
---|
179 | # So we need to close and then reopen solr post pipe to force commit after some $DOC_BATCH_SIZE
|
---|
180 | # We could still have the same issue if any one document is very large (long stream of bytes),
|
---|
181 | # but then we'd need to take care of the problem at the root, see the StackOverFlow page
|
---|
182 | # and SimplePostTool.java, as the problem lies in HttpURLConnection.
|
---|
183 | sub pass_on_xml_stream
|
---|
184 | {
|
---|
185 | # the xml_stream sent to solr looks like:
|
---|
186 | # <update>
|
---|
187 | # <add>
|
---|
188 | # <doc>
|
---|
189 | # ....
|
---|
190 | # </doc>
|
---|
191 | # </add>
|
---|
192 | # <doc>
|
---|
193 | # ....
|
---|
194 | # </doc>
|
---|
195 | # </add>
|
---|
196 | # ...
|
---|
197 | # </update>
|
---|
198 |
|
---|
199 | my $doc_count = 0;
|
---|
200 |
|
---|
201 | my $line;
|
---|
202 | while (defined ($line = <STDIN>)) {
|
---|
203 | # monitor for end of each document
|
---|
204 | if ($line =~ m/^\s*<\/add>$/) {
|
---|
205 | $doc_count++;
|
---|
206 | }
|
---|
207 | else {
|
---|
208 | if ($doc_count == $DOC_BATCH_SIZE) {
|
---|
209 | # evidence that there is still more documents to process
|
---|
210 | # => but reached batch size, so flush and reopen
|
---|
211 | # to force a commit of past docs to solr
|
---|
212 | # before sending this current line on its way
|
---|
213 |
|
---|
214 | if ($line =~ m/^\s*<add>$/) {
|
---|
215 | &solrutil::print_to_post_pipe("</update>\n");
|
---|
216 | flush_and_reopen_java_solr();
|
---|
217 |
|
---|
218 | # start next doc
|
---|
219 | &solrutil::print_to_post_pipe("<update>\n");
|
---|
220 | $doc_count = 0;
|
---|
221 | }
|
---|
222 | }
|
---|
223 | }
|
---|
224 |
|
---|
225 | &solrutil::print_to_post_pipe($line);
|
---|
226 |
|
---|
227 | }
|
---|
228 | }
|
---|
229 |
|
---|
230 |
|
---|
231 | # /** This checks the arguments on the command line, filters the
|
---|
232 | # * unknown command line arguments and then calls the open_java_solr
|
---|
233 | # * function to begin processing.
|
---|
234 | # */
|
---|
235 | sub main
|
---|
236 | {
|
---|
237 | my (@argv) = @_;
|
---|
238 | my $argc = scalar(@argv);
|
---|
239 |
|
---|
240 | my @filtered_argv = ();
|
---|
241 |
|
---|
242 | my $i = 0;
|
---|
243 | while ($i<$argc) {
|
---|
244 | if ($argv[$i] =~ m/^\-(.*)$/) {
|
---|
245 |
|
---|
246 | my $option = $1;
|
---|
247 |
|
---|
248 | # -verbosity <num>
|
---|
249 | if ($option eq "verbosity") {
|
---|
250 | $i++;
|
---|
251 | if ($i<$argc)
|
---|
252 | {
|
---|
253 | # solr indexing has no support for verbosity
|
---|
254 | # => parse to be compatible with calling program, but supress it
|
---|
255 | # for solr-post.jar
|
---|
256 | }
|
---|
257 | }
|
---|
258 | else {
|
---|
259 | print STDERR "Unrecognised minus option: -$option\n";
|
---|
260 | }
|
---|
261 | }
|
---|
262 | else {
|
---|
263 | push(@filtered_argv,$argv[$i]);
|
---|
264 | }
|
---|
265 | $i++;
|
---|
266 | }
|
---|
267 |
|
---|
268 | my $filtered_argc = scalar(@filtered_argv);
|
---|
269 |
|
---|
270 | if ($filtered_argc < 4) {
|
---|
271 | print STDERR "Usage: solr_passes.pl [-verbosity num] core \"text\"|\"index\" build-dir index-name\n";
|
---|
272 | exit 1;
|
---|
273 | }
|
---|
274 |
|
---|
275 | my $core = $filtered_argv[0];
|
---|
276 | my $mode = $filtered_argv[1];
|
---|
277 | my $full_builddir = $filtered_argv[2];
|
---|
278 | my $indexdir = $filtered_argv[3];
|
---|
279 |
|
---|
280 | # We only need the Solr handle opened if we are indexing the
|
---|
281 | # documents, not if we are just storing the text
|
---|
282 | if ($mode eq "index") {
|
---|
283 | open_java_solr($core, $full_builddir, $indexdir);
|
---|
284 | }
|
---|
285 |
|
---|
286 | if ($mode eq "text") {
|
---|
287 | print STDERR "Monitoring for input!\n";
|
---|
288 | my $full_textdir = &util::filename_cat($full_builddir,"text");
|
---|
289 | monitor_xml_stream($full_textdir);
|
---|
290 | }
|
---|
291 | else {
|
---|
292 | print STDERR "Streaming document input onto Solr server!\n";
|
---|
293 | pass_on_xml_stream();
|
---|
294 | }
|
---|
295 |
|
---|
296 |
|
---|
297 | if ($mode eq "index") {
|
---|
298 | close_java_solr();
|
---|
299 | }
|
---|
300 | }
|
---|
301 |
|
---|
302 |
|
---|
303 | &main(@ARGV);
|
---|