2 # Copyright (C) 2015-2020 all contributors <meta@public-inbox.org>
3 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # Basic tool to create a Xapian search index for a public-inbox.
5 # Usage with libeatmydata <https://www.flamingspork.com/projects/libeatmydata/>
6 # highly recommended: eatmydata public-inbox-index INBOX_DIR
10 use Getopt::Long qw(:config gnu_getopt no_ignore_case auto_abbrev);
11 my $usage = "public-inbox-index INBOX_DIR";
12 use PublicInbox::Admin;
13 PublicInbox::Admin::require_or_die('-index');
14 use PublicInbox::Xapcmd;
17 my $opt = { quiet => -1, compact => 0, maxsize => undef, sync => 1 };
18 GetOptions($opt, qw(verbose|v+ reindex rethread compact|c+ jobs|j=i prune sync!
19 indexlevel|L=s maxsize|max-size=s batchsize|batch-size=s
20 sequentialshard|seq-shard|sequential-shard))
21 or die "bad command-line args\n$usage";
22 die "--jobs must be >= 0\n" if defined $opt->{jobs} && $opt->{jobs} < 0;
24 if ($opt->{compact}) {
25 require PublicInbox::Xapcmd;
26 PublicInbox::Xapcmd::check_compact();
27 $compact_opt = { -coarse_lock => 1, compact => 1 };
28 if (defined(my $jobs = $opt->{jobs})) {
29 $compact_opt->{jobs} = $jobs;
33 my $cfg = PublicInbox::Config->new;
34 my @ibxs = PublicInbox::Admin::resolve_inboxes(\@ARGV, undef, $cfg);
35 PublicInbox::Admin::require_or_die('-index');
36 unless (@ibxs) { print STDERR "Usage: $usage\n"; exit 1 }
38 my $max_size = $opt->{maxsize} // $cfg->{lc('publicInbox.indexMaxSize')};
39 if (defined $max_size) {
40 PublicInbox::Admin::parse_unsigned(\$max_size) or
41 die "`publicInbox.indexMaxSize=$max_size' not parsed\n";
44 if (my $bs = $opt->{batchsize} // $cfg->{lc('publicInbox.indexBatchSize')}) {
45 PublicInbox::Admin::parse_unsigned(\$bs) or
46 die "`publicInbox.indexBatchSize=$bs' not parsed\n";
47 $PublicInbox::SearchIdx::BATCH_BYTES = $bs;
50 my $s = $opt->{sequentialshard} //
51 $cfg->{lc('publicInbox.indexSequentialShard')};
53 my $v = $cfg->git_bool($s);
55 die "`publicInbox.indexSequentialShard=$s' not boolean\n";
56 $opt->{sequentialshard} = $v;
60 foreach my $ibx (@ibxs) {
61 # XXX: users can shoot themselves in the foot, with opt->{indexlevel}
62 $ibx->{indexlevel} //= $opt->{indexlevel} //
63 PublicInbox::Admin::detect_indexlevel($ibx);
64 $ibx->{index_max_size} = $max_size;
65 PublicInbox::Admin::scan_ibx_modules($mods, $ibx);
68 PublicInbox::Admin::require_or_die(keys %$mods);
69 PublicInbox::Admin::progress_prepare($opt);
71 $ibx = PublicInbox::InboxWritable->new($ibx);
72 if ($opt->{compact} >= 2) {
73 PublicInbox::Xapcmd::run($ibx, 'compact', $compact_opt);
75 $ibx->{-no_sync} = 1 if !$opt->{sync};
78 if (defined(my $s = $ibx->{indexsequentialshard})) {
79 defined(my $v = $cfg->git_bool($s)) or die <<EOL;
80 publicInbox.$ibx->{name}.indexSequentialShard not boolean
82 $ibx_opt = { %$opt, sequentialshard => $v };
84 PublicInbox::Admin::index_inbox($ibx, undef, $ibx_opt);
85 PublicInbox::Xapcmd::run($ibx, 'compact', $compact_opt) if $compact_opt;