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 [options] INBOX_DIR';
12 my $help = <<EOF; # the following should fit w/o scrolling in 80x24 term:
15 Create and update search indices
19 --no-fsync speed up indexing, risk corruption on power outage
20 -L LEVEL `basic', `medium', or `full' (default: full)
21 --compact | -c run public-inbox-compact(1) after indexing
22 --sequential-shard index Xapian shards sequentially for slow storage
23 --jobs=NUM set or disable parallelization (NUM=0)
24 --batch-size=BYTES flush changes to OS after a given number of bytes
25 --max-size=BYTES do not index messages larger than the given size
26 --reindex index previously indexed data (if upgrading)
27 --rethread regenerate thread IDs (if upgrading, use sparingly)
28 --prune prune git storage on discontiguous history
29 --verbose | -v increase verbosity (may be repeated)
30 --help | -? show this help
32 BYTES may use `k', `m', and `g' suffixes (e.g. `10m' for 10 megabytes)
33 See public-inbox-index(1) man page for full documentation.
36 my $opt = { quiet => -1, compact => 0, max_size => undef, fsync => 1 };
37 GetOptions($opt, qw(verbose|v+ reindex rethread compact|c+ jobs|j=i prune
38 fsync|sync! xapian_only|xapian-only
39 indexlevel|index-level|L=s max_size|max-size=s
40 batch_size|batch-size=s
41 sequential_shard|seq-shard|sequential-shard
43 or die "bad command-line args\n$usage";
44 if ($opt->{help}) { print $help; exit 0 };
45 die "--jobs must be >= 0\n" if defined $opt->{jobs} && $opt->{jobs} < 0;
46 if ($opt->{xapian_only} && !$opt->{reindex}) {
47 die "--xapian-only requires --reindex\n";
50 # require lazily to speed up --help
51 require PublicInbox::Admin;
52 PublicInbox::Admin::require_or_die('-index');
54 if ($opt->{compact}) {
55 require PublicInbox::Xapcmd;
56 PublicInbox::Xapcmd::check_compact();
57 $compact_opt = { -coarse_lock => 1, compact => 1 };
58 if (defined(my $jobs = $opt->{jobs})) {
59 $compact_opt->{jobs} = $jobs;
63 my $cfg = PublicInbox::Config->new; # Config is loaded by Admin
64 my @ibxs = PublicInbox::Admin::resolve_inboxes(\@ARGV, undef, $cfg);
65 PublicInbox::Admin::require_or_die('-index');
66 unless (@ibxs) { print STDERR "Usage: $usage\n"; exit 1 }
68 my $max_size = $opt->{max_size} // $cfg->{lc('publicInbox.indexMaxSize')};
69 if (defined $max_size) {
70 PublicInbox::Admin::parse_unsigned(\$max_size) or
71 die "`publicInbox.indexMaxSize=$max_size' not parsed\n";
72 $opt->{max_size} = $max_size;
75 my $bs = $opt->{batch_size} // $cfg->{lc('publicInbox.indexBatchSize')};
77 PublicInbox::Admin::parse_unsigned(\$bs) or
78 die "`publicInbox.indexBatchSize=$bs' not parsed\n";
79 $opt->{batch_size} = $bs;
82 # out-of-the-box builds of Xapian 1.4.x are still limited to 32-bit
83 # https://getting-started-with-xapian.readthedocs.io/en/latest/concepts/indexing/limitations.html
84 local $ENV{XAPIAN_FLUSH_THRESHOLD} ||= '4294967295' if defined($bs);
86 my $s = $opt->{sequential_shard} //
87 $cfg->{lc('publicInbox.indexSequentialShard')};
89 my $v = $cfg->git_bool($s);
91 die "`publicInbox.indexSequentialShard=$s' not boolean\n";
92 $opt->{sequential_shard} = $v;
96 foreach my $ibx (@ibxs) {
97 # XXX: users can shoot themselves in the foot, with opt->{indexlevel}
98 $ibx->{indexlevel} //= $opt->{indexlevel} // ($opt->{xapian_only} ?
99 'full' : PublicInbox::Admin::detect_indexlevel($ibx));
100 PublicInbox::Admin::scan_ibx_modules($mods, $ibx);
103 PublicInbox::Admin::require_or_die(keys %$mods);
104 require PublicInbox::InboxWritable;
105 PublicInbox::Admin::progress_prepare($opt);
106 for my $ibx (@ibxs) {
107 $ibx = PublicInbox::InboxWritable->new($ibx);
108 if ($opt->{compact} >= 2) {
109 PublicInbox::Xapcmd::run($ibx, 'compact', $compact_opt);
111 $ibx->{-no_fsync} = 1 if !$opt->{fsync};
114 if (defined(my $s = $ibx->{lc('indexSequentialShard')})) {
115 defined(my $v = $cfg->git_bool($s)) or die <<EOL;
116 publicInbox.$ibx->{name}.indexSequentialShard not boolean
118 $ibx_opt = { %$opt, sequential_shard => $v };
120 PublicInbox::Admin::index_inbox($ibx, undef, $ibx_opt);
122 local $compact_opt->{jobs} = 0 if $ibx_opt->{sequential_shard};
123 PublicInbox::Xapcmd::run($ibx, 'compact', $compact_opt);