2 # Copyright (C) 2008 Equinox Software, Inc.
3 # Author: Bill Erickson <erickson@esilibrary.com>
5 # This program is free software; you can redistribute it and/or
6 # modify it under the terms of the GNU General Public License
7 # as published by the Free Software Foundation; either version 2
8 # of the License, or (at your option) any later version.
10 # This program is distributed in the hope that it will be useful,
11 # but WITHOUT ANY WARRANTY; without even the implied warranty of
12 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 # GNU General Public License for more details.
16 use strict; use warnings;
17 use Net::Server::PreFork;
18 use base qw/Net::Server::PreFork/;
22 use MARC::File::USMARC;
25 use File::Basename qw/fileparse/;
27 use Getopt::Long qw(:DEFAULT GetOptionsFromArray);
30 use OpenSRF::Utils::Logger qw/$logger/;
31 use OpenILS::Utils::Cronscript;
32 require 'oils_header.pl';
33 use vars qw/$apputils/;
41 # 'osrf-config=s' => '/openils/conf/opensrf_core.xml',
51 $OpenILS::Utils::Cronscript::debug=1 if $debug;
52 $Getopt::Long::debug=1 if $debug > 1;
53 my $o = OpenILS::Utils::Cronscript->new(\%defaults);
57 if (grep {$_ eq '--'} @ARGV) {
58 print "Splitting options into groups\n" if $debug;
61 $_ eq '--' and last; # stop at the first --
62 push @script_args, $_;
69 print "Calling MyGetOptions ",
70 (@script_args ? "with options: " . join(' ', @script_args) : 'without options from command line'),
73 my $real_opts = $o->MyGetOptions(\@script_args);
75 # GetOptionsFromArray(\@script_args, \%defaults, %defaults); # similar to
77 $real_opts->{tempdir} ||= tempdir_setting(); # This doesn't go in defaults because it reads config, must come after bootstrap
79 my $bufsize = $real_opts->{buffsize};
80 my $bib_source = $real_opts->{source};
81 my $osrf_config = $real_opts->{'osrf-config'};
82 my $oils_username = $real_opts->{user};
83 my $oils_password = $real_opts->{password};
84 my $help = $real_opts->{help};
85 my $merge_profile = $real_opts->{merge_profile};
86 my $queue_id = $real_opts->{queue};
87 my $tempdir = $real_opts->{tempdir};
88 $debug += $real_opts->{debug};
90 foreach (keys %$real_opts) {
91 print("real_opt->{$_} = ", $real_opts->{$_}, "\n") if $real_opts->{debug} or $debug;
93 my $wait_time = $real_opts->{wait};
96 # DEFAULTS for Net::Server
97 my $filename = fileparse($0, '.pl');
98 my $conf_file = (-r "$filename.conf") ? "$filename.conf" : undef;
99 # $conf_file is the Net::Server config for THIS script (not EG), if it exists and is readable
104 pod2usage(1) if $help;
105 unless ($oils_password) {
106 print STDERR "\nERROR: password option required for session login\n\n";
110 print Dumper($o) if $debug;
113 foreach my $ref (qw/bufsize bib_source osrf_config oils_username oils_password help conf_file debug/) {
115 printf "%16s => %s\n", $ref, (eval("\$$ref") || '');
120 print Dumper($real_opts);
124 sub tempdir_setting {
125 my $ret = $apputils->simplereq( qw# opensrf.settings opensrf.settings.xpath.get
126 /opensrf/default/apps/open-ils.vandelay/app_settings/databases/importer # );
127 return $ret->[0] || '/tmp';
133 WARNING: This script provides no security layer. Any client that has
134 access to the server+port can inject MARC records into the system.
140 return $apputils->simplereq(
142 'open-ils.cat.biblio.record.xml.import',
147 sub old_process_batch_data {
148 my $data = shift or $logger->error("process_batch_data called without any data");
152 open $handle, '<', \$data;
153 my $batch = MARC::Batch->new('USMARC', $handle);
161 eval { $rec = $batch->next; };
164 $logger->error("Failed parsing MARC record $index");
167 last unless $rec; # The only way out
169 my $resp = xml_import($authtoken, $rec->as_xml_record, $bib_source);
171 # has the session timed out?
172 if (oils_event_equals($resp, 'NO_SESSION')) {
174 $resp = xml_import($authtoken, $rec->as_xml_record, $bib_source); # try again w/ new token
176 oils_event_die($resp);
181 sub process_spool { # filename
182 $apputils->simplereq('open-ils.vandelay', 'open-ils.vandelay.bib.process_spool', $authtoken, undef,
183 $queue_id, 'import', shift, $bib_source );
185 sub bib_queue_import {
186 my $extra = {auto_overlay_exact => 1};
187 $extra->{merge_profile} = $merge_profile if $merge_profile;
188 $apputils->simplereq('open-ils.vandelay', 'open-ils.vandelay.bib_queue.import', $authtoken,
192 sub process_batch_data {
193 my $data = shift or $logger->error("process_batch_data called without any data");
196 my ($handle, $tempfile) = File::Temp(DIR => $tempdir) or die "Cannot write tempfile in $tempdir";
200 my $resp = process_spool($tempfile);
202 if (oils_event_equals($resp, 'NO_SESSION')) { # has the session timed out?
204 $resp = process_spool($tempfile); # try again w/ new token
207 $resp = bib_queue_import();
209 if (oils_event_equals($resp, 'NO_SESSION')) { # has the session timed out?
211 $resp = bib_queue_import(); # try again w/ new token
213 oils_event_die($resp);
216 sub process_request { # The core Net::Server method
218 my $socket = $self->{server}->{client};
222 # Reading <STDIN> blocks until the client is closed. Instead of waiting
223 # for that, give each inbound record $wait_time seconds to fully arrive
224 # and pull the data directly from the socket
226 local $SIG{ALRM} = sub { die "alarm\n" };
229 last unless $socket->sysread($buf, $bufsize);
234 if ($real_opts->{noqueue}) {
235 old_process_batch_data($data);
237 process_batch_data($data);
242 # the authtoken will timeout after the configured inactivity period.
243 # When that happens, get a new one.
245 $authtoken = oils_login($oils_username, $oils_password, 'staff')
246 or die "Unable to login to Evergreen as user $oils_username";
252 osrf_connect($osrf_config);
254 print "Calling Net::Server run ", (@ARGV ? "with command-line options: " . join(' ', @ARGV) : ''), "\n";
255 __PACKAGE__->run(conf_file => $conf_file);
261 marc_stream_importer.pl - Import MARC records via bare socket connection.
265 ./marc_stream_importer.pl [common opts ...] [script opts ...] -- [Net::Server opts ...] &
267 This script uses the EG common options from B<Cronscript>. See --help output for those.
269 Run C<perldoc marc_stream_importer.pl> for full documentation.
271 Note the extra C<--> to separate options for the script wrapper from options for the
272 underlying L<Net::Server> options.
274 Note: this script has to be run in the same directory as B<oils_header.pl>.
276 Typical execution will include a trailing C<&> to run in the background.
280 This script is a L<Net::Server::PreFork> instance for shoving records into Evergreen from a remote system.
284 The only required option is --password
286 --password =<eg_password>
287 --user =<eg_username> default: admin
288 --source =<bib_source> default: 1 Integer
289 --merge =<i> default: 0
290 --tempdir =</temp/dir/> default: from L<opensrf.conf> <open-ils.vandelay/app_settings/databases/importer>
291 --source =<i> default: 1
293 =head2 Old style: --noqueue and associated options
295 To bypass vandelay queue processing and push directly into the database (as the old style)
297 --noqueue default: OFF
298 --buffsize =<i> default: 4096 Buffer size. Only used by --noqueue
299 --wait =<i> default: 5 Seconds to read socket before processing. Only used by --noqueue
301 =head2 Net::Server Options
303 By default, the script will use the Net::Server configuration file B<marc_stream_importer.conf>. You can
304 override this by passing a filepath with the --conf_file option.
306 Other Net::Server options include: --port=<port> --min_servers=<X> --max_servers=<Y> and --log_file=[path/to/file]
308 See L<Net::Server> for a complete list.
312 =head3 OCLC Connexion
314 To use this script with OCLC Connexion, configure the client as follows:
316 Under Tools -> Options -> Export (tab)
317 Create -> Choose Connection -> OK -> Leave translation at "None"
318 -> Create -> Create -> choose TCP/IP (internet)
319 -> Enter hostname and Port, leave 'Use Telnet Protocol' checked
320 -> Create/OK your way out of the dialogs
321 Record Characteristics (button) -> Choose 'UTF-8 Unicode' for the Character Set
324 OCLC and Connexion are trademark/service marks of OCLC Online Computer Library Center, Inc.
328 WARNING: This script provides no inherent security layer. Any client that has
329 access to the server+port can inject MARC records into the system.
330 Use the available options (like allow/deny) in the Net::Server config file
331 or via the command line to restrict access as necessary.
335 ./marc_stream_importer.pl \
336 admin open-ils connexion --port 5555 --min_servers 2 \
337 --max_servers=20 --log_file=/openils/var/log/marc_net_importer.log &
341 L<Net::Server::PreFork>, L<marc_stream_importer.conf>
345 Bill Erickson <erickson@esilibrary.com>
346 Joe Atzberger <jatzberger@esilibrary.com>