2 # ---------------------------------------------------------------
3 # Copyright © 2013,2014 Merrimack Valley Library Consortium
4 # Jason Stephenson <jstephenson@mvlc.org>
6 # This program is free software; you can redistribute it and/or modify
7 # it under the terms of the GNU General Public License as published by
8 # the Free Software Foundation; either version 2 of the License, or
9 # (at your option) any later version.
11 # This program is distributed in the hope that it will be useful,
12 # but WITHOUT ANY WARRANTY; without even the implied warranty of
13 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 # GNU General Public License for more details.
15 # ---------------------------------------------------------------
16 # TODO: Document with POD.
17 # This guy parallelizes a reingest.
23 # Globals for the command line options: --
25 # You will want to adjust the next two based on your database size,
26 # i.e. number of bib records as well as the number of cores on your
27 # database server. Using roughly number of cores/2 doesn't seem to
28 # have much impact in off peak times.
29 my $batch_size = 10000; # records processed per batch
30 my $max_child = 8; # max number of parallel worker processes
32 my $skip_browse; # Skip the browse reingest.
33 my $skip_attrs; # Skip the record attributes reingest.
34 my $skip_search; # Skip the search reingest.
35 my $skip_facets; # Skip the facets reingest.
36 my $start_id; # start processing at this bib ID.
37 my $end_id; # stop processing when this bib ID is reached.
38 my $max_duration; # max processing duration in seconds
39 my $help; # show help text
40 my $opt_pipe; # Read record ids from STDIN.
43 'batch-size=i' => \$batch_size,
44 'max-child=i' => \$max_child,
45 'skip-browse' => \$skip_browse,
46 'skip-attrs' => \$skip_attrs,
47 'skip-search' => \$skip_search,
48 'skip-facets' => \$skip_facets,
49 'start-id=i' => \$start_id,
50 'end-id=i' => \$end_id,
52 'max-duration=i' => \$max_duration,
59 $0 --batch-size $batch_size --max-child $max_child \
60 --start-id 1 --end-id 500000 --duration 14400
63 Number of records to process per batch
66 Max number of worker processes
72 Skip the selected reingest component
75 Start processing at this record ID.
78 Stop processing when this record ID is reached
81 Read record IDs to reingest from standard input.
82 This option conflicts with --start-id and/or --end-id.
85 Stop processing after this many total seconds have passed.
96 # Check for mutually exclusive options:
97 if ($opt_pipe && ($start_id || $end_id)) {
98 warn('Mutually exclusive options');
102 my $where = "WHERE deleted = 'f'";
103 if ($start_id && $end_id) {
104 $where .= " AND id BETWEEN $start_id AND $end_id";
105 } elsif ($start_id) {
106 $where .= " AND id >= $start_id";
108 $where .= " AND id <= $end_id";
111 # "Gimme the keys! I'll drive!"
114 FROM biblio.record_entry
119 # Stuffs needed for looping, tracking how many lists of records we
120 # have, storing the actual list of records, and the list of the lists
122 my ($count, $lists, $records) = (0,0,[]);
124 # To do the browse-only ingest:
127 my $start_epoch = time;
129 sub duration_expired {
130 return 1 if $max_duration && (time - $start_epoch) >= $max_duration;
134 # All of the DBI->connect() calls in this file assume that you have
135 # configured the PGHOST, PGPORT, PGDATABASE, PGUSER, and PGPASSWORD
136 # variables in your execution environment. If you have not, you have
141 # 2) edit the DBI->connect() calls in this program so that it can
142 # connect to your database.
144 # Get the input records from either standard input or the database.
148 # Want only numbers, one per line.
149 if ($_ =~ /([0-9]+)/) {
154 my $dbh = DBI->connect('DBI:Pg:');
155 @input = @{$dbh->selectcol_arrayref($q)};
159 foreach my $record (@input) {
160 push(@blist, $record); # separate list of browse-only ingest
161 push(@$records, $record);
162 if (++$count == $batch_size) {
163 $lol[$lists++] = $records;
168 $lol[$lists++] = $records if ($count); # Last batch is likely to be
171 # We're going to reuse $count to keep track of the total number of
175 # @running keeps track of the running child processes.
178 # We start the browse-only ingest before starting the other ingests.
179 browse_ingest(@blist) unless ($skip_browse);
181 # We loop until we have processed all of the batches stored in @lol
182 # or the maximum processing duration has been reached.
183 while ($count < $lists) {
184 my $duration_expired = duration_expired();
186 if (scalar(@lol) && scalar(@running) < $max_child && !$duration_expired) {
187 # Reuse $records for the lulz.
188 $records = shift(@lol);
189 if ($skip_search && $skip_facets && $skip_attrs) {
196 if (grep {$_ == $pid} @running) {
197 @running = grep {$_ != $pid} @running;
199 print "$count of $lists processed\n";
203 if ($duration_expired && scalar(@running) == 0) {
204 warn "Exiting on max_duration ($max_duration)\n";
209 # This subroutine forks a process to do the browse-only ingest on the
210 # @blist above. It cannot be parallelized, but can run in parrallel
211 # to the other ingests.
215 if (!defined($pid)) {
216 die "failed to spawn child";
218 # Add our browser to the list of running children.
219 push(@running, $pid);
220 # Increment the number of lists, because this list was not
221 # previously counted.
223 } elsif ($pid == 0) {
224 my $dbh = DBI->connect('DBI:Pg:');
225 my $sth = $dbh->prepare("SELECT metabib.reingest_metabib_field_entries(?, TRUE, FALSE, TRUE)");
227 if ($sth->execute($_)) {
228 my $crap = $sth->fetchall_arrayref();
230 warn ("Browse ingest failed for record $_");
232 if (duration_expired()) {
233 warn "browse_ingest() stopping on record $_ ".
234 "after max duration reached\n";
243 # Fork a child to do the other reingests:
248 if (!defined($pid)) {
249 die "Failed to spawn a child";
251 push(@running, $pid);
252 } elsif ($pid == 0) {
253 my $dbh = DBI->connect('DBI:Pg:');
254 reingest_attributes($dbh, $list) unless ($skip_attrs);
255 reingest_field_entries($dbh, $list)
256 unless ($skip_facets && $skip_search);
262 # Reingest metabib field entries on a list of records.
263 sub reingest_field_entries {
266 my $sth = $dbh->prepare("SELECT metabib.reingest_metabib_field_entries(?, ?, TRUE, ?)");
267 # Because reingest uses "skip" options we invert the logic of do variables.
268 $sth->bind_param(2, ($skip_facets) ? 1 : 0);
269 $sth->bind_param(3, ($skip_search) ? 1 : 0);
271 $sth->bind_param(1, $_);
272 if ($sth->execute()) {
273 my $crap = $sth->fetchall_arrayref();
275 warn ("metabib.reingest_metabib_field_entries failed for record $_");
280 # Reingest record attributes on a list of records.
281 sub reingest_attributes {
284 my $sth = $dbh->prepare(<<END_OF_INGEST
285 SELECT metabib.reingest_record_attributes(id, NULL::TEXT[], marc)
286 FROM biblio.record_entry
291 $sth->bind_param(1, $_);
292 if ($sth->execute()) {
293 my $crap = $sth->fetchall_arrayref();
295 warn ("metabib.reingest_record_attributes failed for record $_");