2 # Copyright (C) 2015-2021 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 $help = <<EOF; # the following should fit w/o scrolling in 80x24 term:
12 usage: public-inbox-index [options] INBOX_DIR
14 Create and update per-inbox search indices
18 --no-fsync speed up indexing, risk corruption on power outage
19 -L LEVEL `basic', `medium', or `full' (default: full)
20 -E EXTINDEX update extindex (default: `all')
21 --all index all configured inboxes
22 --compact | -c run public-inbox-compact(1) after indexing
23 --sequential-shard index Xapian shards sequentially for slow storage
24 --jobs=NUM set or disable parallelization (NUM=0)
25 --batch-size=BYTES flush changes to OS after a given number of bytes
26 --max-size=BYTES do not index messages larger than the given size
27 --reindex index previously indexed data (if upgrading)
28 --rethread regenerate thread IDs (if upgrading, use sparingly)
29 --prune prune git storage on discontiguous history
30 --verbose | -v increase verbosity (may be repeated)
32 BYTES may use `k', `m', and `g' suffixes (e.g. `10m' for 10 megabytes)
33 See public-inbox-index(1) man page for full documentation.
36 quiet => -1, compact => 0, max_size => undef, fsync => 1,
37 'update-extindex' => [], # ":s@" optional arg sets '' if no arg given
39 GetOptions($opt, qw(verbose|v+ reindex rethread compact|c+ jobs|j=i prune
40 fsync|sync! xapian_only|xapian-only
41 indexlevel|index-level|L=s max_size|max-size=s
42 batch_size|batch-size=s
43 sequential-shard|seq-shard
44 no-update-extindex update-extindex|E=s@
45 fast-noop|F skip-docdata all C=s@ help|h))
47 if ($opt->{help}) { print $help; exit 0 };
48 die "--jobs must be >= 0\n" if defined $opt->{jobs} && $opt->{jobs} < 0;
49 if ($opt->{xapian_only} && !$opt->{reindex}) {
50 die "--xapian-only requires --reindex\n";
52 if ($opt->{reindex} && delete($opt->{'fast-noop'})) {
53 warn "--fast-noop ignored with --reindex\n";
56 # require lazily to speed up --help
57 require PublicInbox::Admin;
58 PublicInbox::Admin::require_or_die('-index');
59 PublicInbox::Admin::do_chdir(delete $opt->{C});
61 my $cfg = PublicInbox::Config->new; # Config is loaded by Admin
63 my @ibxs = PublicInbox::Admin::resolve_inboxes(\@ARGV, $opt, $cfg);
64 PublicInbox::Admin::require_or_die('-index');
65 unless (@ibxs) { print STDERR $help; exit 1 }
67 my (@eidx, %eidx_seen);
68 my $update_extindex = $opt->{'update-extindex'};
69 if (!scalar(@$update_extindex) && (my $ALL = $cfg->ALL)) {
70 # extindex and normal inboxes may have different owners
71 push(@$update_extindex, 'all') if -w $ALL->{topdir};
73 @$update_extindex = () if $opt->{'no-update-extindex'};
74 if (scalar @$update_extindex) {
75 PublicInbox::Admin::require_or_die('-search');
76 require PublicInbox::ExtSearchIdx;
78 for my $ei_name (@$update_extindex) {
79 my $es = $cfg->lookup_ei($ei_name);
81 if (!$es && -d $ei_name) { # allow dirname or config section name
84 $topdir = $es->{topdir};
86 die "extindex `$ei_name' not configured or found\n";
89 delete $o->{indexlevel} if ($o->{indexlevel}//'') eq 'basic';
90 $eidx_seen{$topdir} //=
91 push(@eidx, PublicInbox::ExtSearchIdx->new($topdir, $o));
94 my @eidx_unconfigured;
95 foreach my $ibx (@ibxs) {
96 # detect_indexlevel may also set $ibx->{-skip_docdata}
97 my $detected = PublicInbox::Admin::detect_indexlevel($ibx);
98 # XXX: users can shoot themselves in the foot, with opt->{indexlevel}
99 $ibx->{indexlevel} //= $opt->{indexlevel} // ($opt->{xapian_only} ?
101 PublicInbox::Admin::scan_ibx_modules($mods, $ibx);
102 if (@eidx && $ibx->{-unconfigured}) {
103 push @eidx_unconfigured, " $ibx->{inboxdir}\n";
106 warn <<EOF if @eidx_unconfigured;
107 The following inboxes are unconfigured and will not be updated in
108 @$update_extindex:\n@eidx_unconfigured
111 # "Search::Xapian" includes SWIG "Xapian", too:
112 $opt->{compact} = 0 if !$mods->{'Search::Xapian'};
114 PublicInbox::Admin::require_or_die(keys %$mods);
115 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
116 local %ENV = (%ENV, %$env) if $env;
117 require PublicInbox::InboxWritable;
118 PublicInbox::Xapcmd::check_compact() if $opt->{compact};
119 PublicInbox::Admin::progress_prepare($opt);
120 for my $ibx (@ibxs) {
121 $ibx = PublicInbox::InboxWritable->new($ibx);
122 if ($opt->{compact} >= 2) {
123 PublicInbox::Xapcmd::run($ibx, 'compact', $opt->{compact_opt});
125 $ibx->{-no_fsync} = 1 if !$opt->{fsync};
126 $ibx->{-skip_docdata} //= $opt->{'skip-docdata'};
129 if (defined(my $s = $ibx->{lc('indexSequentialShard')})) {
130 defined(my $v = $cfg->git_bool($s)) or die <<EOL;
131 publicInbox.$ibx->{name}.indexSequentialShard not boolean
133 $ibx_opt = { %$opt, 'sequential-shard' => $v };
135 my $nidx = PublicInbox::Admin::index_inbox($ibx, undef, $ibx_opt);
136 last if $ibx_opt->{quit};
137 if (my $copt = $opt->{compact_opt}) {
138 local $copt->{jobs} = 0 if $ibx_opt->{'sequential-shard'};
139 PublicInbox::Xapcmd::run($ibx, 'compact', $copt);
141 last if $ibx_opt->{quit};
142 next if $ibx->{-unconfigured} || !$nidx;
143 for my $eidx (@eidx) {
144 $eidx->attach_inbox($ibx);
147 my $pr = $opt->{-progress};
148 for my $eidx (@eidx) {
149 $pr->("indexing $eidx->{topdir} ...\n") if $pr;
150 $eidx->eidx_sync($opt);
151 last if $opt->{quit};