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 $help = <<EOF; # the following should fit w/o scrolling in 80x24 term:
12 usage: public-inbox-index [options] INBOX_DIR
14 Create and update per-inbox search indices
18 --no-fsync speed up indexing, risk corruption on power outage
19 -L LEVEL `basic', `medium', or `full' (default: full)
20 -E EXTINDEX update extindex (default: `all')
21 --all index all configured inboxes
22 --compact | -c run public-inbox-compact(1) after indexing
23 --sequential-shard index Xapian shards sequentially for slow storage
24 --jobs=NUM set or disable parallelization (NUM=0)
25 --batch-size=BYTES flush changes to OS after a given number of bytes
26 --max-size=BYTES do not index messages larger than the given size
27 --reindex index previously indexed data (if upgrading)
28 --since=DATE limit --reindex to changes after DATE
29 --until=DATE limit --reindex to changes before DATE
30 --rethread regenerate thread IDs (if upgrading, use sparingly)
31 --prune prune git storage on discontiguous history
32 --verbose | -v increase verbosity (may be repeated)
34 BYTES may use `k', `m', and `g' suffixes (e.g. `10m' for 10 megabytes)
35 See public-inbox-index(1) man page for full documentation.
38 quiet => -1, compact => 0, max_size => undef, fsync => 1,
39 'update-extindex' => [], # ":s@" optional arg sets '' if no arg given
41 GetOptions($opt, qw(verbose|v+ reindex rethread compact|c+ jobs|j=i prune
42 fsync|sync! xapian_only|xapian-only dangerous
43 indexlevel|index-level|L=s max_size|max-size=s
44 batch_size|batch-size=s
45 since|after=s until|before=s
46 sequential-shard|seq-shard
47 no-update-extindex update-extindex|E=s@
48 fast-noop|F skip-docdata all C=s@ help|h))
50 if ($opt->{help}) { print $help; exit 0 };
51 die "--jobs must be >= 0\n" if defined $opt->{jobs} && $opt->{jobs} < 0;
52 if ($opt->{xapian_only} && !$opt->{reindex}) {
53 die "--xapian-only requires --reindex\n";
55 if ($opt->{reindex} && delete($opt->{'fast-noop'})) {
56 warn "--fast-noop ignored with --reindex\n";
59 # require lazily to speed up --help
60 require PublicInbox::Admin;
61 PublicInbox::Admin::require_or_die('-index');
62 PublicInbox::Admin::do_chdir(delete $opt->{C});
64 my $cfg = PublicInbox::Config->new; # Config is loaded by Admin
66 my @ibxs = PublicInbox::Admin::resolve_inboxes(\@ARGV, $opt, $cfg);
67 PublicInbox::Admin::require_or_die('-index');
68 unless (@ibxs) { print STDERR $help; exit 1 }
70 my (@eidx, %eidx_seen);
71 my $update_extindex = $opt->{'update-extindex'};
72 if (!scalar(@$update_extindex) && (my $ALL = $cfg->ALL)) {
73 # extindex and normal inboxes may have different owners
74 push(@$update_extindex, 'all') if -w $ALL->{topdir};
76 @$update_extindex = () if $opt->{'no-update-extindex'};
77 if (scalar @$update_extindex) {
78 PublicInbox::Admin::require_or_die('-search');
79 require PublicInbox::ExtSearchIdx;
81 for my $ei_name (@$update_extindex) {
82 my $es = $cfg->lookup_ei($ei_name);
84 if (!$es && -d $ei_name) { # allow dirname or config section name
87 $topdir = $es->{topdir};
89 die "extindex `$ei_name' not configured or found\n";
92 delete $o->{indexlevel} if ($o->{indexlevel}//'') eq 'basic';
93 $eidx_seen{$topdir} //=
94 push(@eidx, PublicInbox::ExtSearchIdx->new($topdir, $o));
97 my @eidx_unconfigured;
98 foreach my $ibx (@ibxs) {
99 # detect_indexlevel may also set $ibx->{-skip_docdata}
100 my $detected = PublicInbox::Admin::detect_indexlevel($ibx);
101 # XXX: users can shoot themselves in the foot, with opt->{indexlevel}
102 $ibx->{indexlevel} //= $opt->{indexlevel} // ($opt->{xapian_only} ?
104 PublicInbox::Admin::scan_ibx_modules($mods, $ibx);
105 if (@eidx && $ibx->{-unconfigured}) {
106 push @eidx_unconfigured, " $ibx->{inboxdir}\n";
109 warn <<EOF if @eidx_unconfigured;
110 The following inboxes are unconfigured and will not be updated in
111 @$update_extindex:\n@eidx_unconfigured
114 # "Search::Xapian" includes SWIG "Xapian", too:
115 $opt->{compact} = 0 if !$mods->{'Search::Xapian'};
117 PublicInbox::Admin::require_or_die(keys %$mods);
118 my $env = PublicInbox::Admin::index_prepare($opt, $cfg);
119 local %ENV = (%ENV, %$env) if $env;
120 require PublicInbox::InboxWritable;
121 PublicInbox::Xapcmd::check_compact() if $opt->{compact};
122 PublicInbox::Admin::progress_prepare($opt);
123 for my $ibx (@ibxs) {
124 $ibx = PublicInbox::InboxWritable->new($ibx);
125 if ($opt->{compact} >= 2) {
126 PublicInbox::Xapcmd::run($ibx, 'compact', $opt->{compact_opt});
128 $ibx->{-no_fsync} = 1 if !$opt->{fsync};
129 $ibx->{-dangerous} = 1 if $opt->{dangerous};
130 $ibx->{-skip_docdata} //= $opt->{'skip-docdata'};
133 if (defined(my $s = $ibx->{lc('indexSequentialShard')})) {
134 defined(my $v = $cfg->git_bool($s)) or die <<EOL;
135 publicInbox.$ibx->{name}.indexSequentialShard not boolean
137 $ibx_opt = { %$opt, 'sequential-shard' => $v };
139 my $nidx = PublicInbox::Admin::index_inbox($ibx, undef, $ibx_opt);
140 last if $ibx_opt->{quit};
141 if (my $copt = $opt->{compact_opt}) {
142 local $copt->{jobs} = 0 if $ibx_opt->{'sequential-shard'};
143 PublicInbox::Xapcmd::run($ibx, 'compact', $copt);
145 last if $ibx_opt->{quit};
146 next if $ibx->{-unconfigured} || !$nidx;
147 for my $eidx (@eidx) {
148 $eidx->attach_inbox($ibx);
151 my $pr = $opt->{-progress};
152 for my $eidx (@eidx) {
153 $pr->("indexing $eidx->{topdir} ...\n") if $pr;
154 $eidx->eidx_sync($opt);
155 last if $opt->{quit};