1 # Copyright (C) 2020 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # Detached/external index cross inbox search indexing support
5 # read-write counterpart to PublicInbox::ExtSearch
7 # It's based on the same ideas as public-inbox-v2-format(5) using
8 # over.sqlite3 for dedupe and sharded Xapian. msgmap.sqlite3 is
9 # missing, so there is no Message-ID conflict resolution, meaning
10 # no NNTP support for now.
12 # v2 has a 1:1 mapping of index:inbox or msgmap for NNTP support.
13 # This is intended to be an M:N index:inbox mapping, but it'll likely
14 # be 1:N in common practice (M==1)
16 package PublicInbox::ExtSearchIdx;
19 use parent qw(PublicInbox::ExtSearch PublicInbox::Lock);
20 use Carp qw(croak carp);
21 use PublicInbox::Search;
22 use PublicInbox::SearchIdx qw(crlf_adjust prepare_stack is_ancestor);
23 use PublicInbox::OverIdx;
24 use PublicInbox::MID qw(mids);
25 use PublicInbox::V2Writable;
26 use PublicInbox::InboxWritable;
27 use PublicInbox::ContentHash qw(content_hash);
32 my (undef, $dir, $opt) = @_;
33 my $l = $opt->{indexlevel} // 'full';
34 $l !~ $PublicInbox::SearchIdx::INDEXLEVELS and
35 die "invalid indexlevel=$l\n";
36 $l eq 'basic' and die "E: indexlevel=basic not yet supported\n";
38 xpfx => "$dir/ei".PublicInbox::Search::SCHEMA_VERSION,
40 creat => $opt->{creat},
41 ibx_map => {}, # (newsgroup//inboxdir) => $ibx
48 lock_path => "$dir/ei.lock",
50 $self->{shards} = $self->count_shards || nproc_shards($opt->{creat});
51 my $oidx = PublicInbox::OverIdx->new("$self->{xpfx}/over.sqlite3");
52 $oidx->{-no_fsync} = 1 if $opt->{-no_fsync};
53 $self->{oidx} = $oidx;
58 my ($self, $ibx) = @_;
59 $ibx = PublicInbox::InboxWritable->new($ibx);
60 my $key = $ibx->eidx_key;
61 if (!$ibx->over || !$ibx->mm) {
62 warn "W: skipping $key (unindexed)\n";
65 if (!defined($ibx->uidvalidity)) {
66 warn "W: skipping $key (no UIDVALIDITY)\n";
69 my $ibxdir = File::Spec->canonpath($ibx->{inboxdir});
70 if ($ibxdir ne $ibx->{inboxdir}) {
71 warn "W: `$ibx->{inboxdir}' canonicalized to `$ibxdir'\n";
72 $ibx->{inboxdir} = $ibxdir;
74 $ibx = PublicInbox::InboxWritable->new($ibx);
75 $self->{ibx_map}->{$key} //= do {
76 push @{$self->{ibx_list}}, $ibx;
81 sub _ibx_attach { # each_inbox callback
82 my ($ibx, $self) = @_;
83 attach_inbox($self, $ibx);
87 my ($self, $cfg) = @_;
88 $cfg->each_inbox(\&_ibx_attach, $self);
91 sub git_blob_digest ($) {
93 my $dig = Digest::SHA->new(1); # XXX SHA256 later
94 $dig->add('blob '.length($$bref)."\0");
99 sub is_bad_blob ($$$$) {
100 my ($oid, $type, $size, $expect_oid) = @_;
101 if ($type ne 'blob') {
102 carp "W: $expect_oid is not a blob (type=$type)";
105 croak "BUG: $oid != $expect_oid" if $oid ne $expect_oid;
106 $size == 0 ? 1 : 0; # size == 0 means purged
109 sub check_batch_limit ($) {
111 my $self = $req->{self};
112 my $new_smsg = $req->{new_smsg};
114 # {raw_bytes} may be unset, so just use {bytes}
115 my $n = $self->{transact_bytes} += $new_smsg->{bytes};
117 # set flag for PublicInbox::V2Writable::index_todo:
118 ${$req->{need_checkpoint}} = 1 if $n >= $self->{batch_bytes};
122 my ($req, $smsg) = @_;
123 my $self = $req->{self};
124 my $docid = $smsg->{num};
125 my $idx = $self->idx_shard($docid);
126 my $oid = $req->{oid};
127 my $xibx = $req->{ibx};
128 my $eml = $req->{eml};
129 if (my $new_smsg = $req->{new_smsg}) { # 'm' on cross-posted message
130 my $xnum = $req->{xnum};
131 $self->{oidx}->add_xref3($docid, $xnum, $oid, $xibx->eidx_key);
132 $idx->shard_add_eidx_info($docid, $oid, $xibx, $eml);
133 check_batch_limit($req);
135 $self->{oidx}->remove_xref3($docid, $oid, $xibx->eidx_key);
136 $idx->shard_remove_eidx_info($docid, $oid, $xibx, $eml);
140 # called by V2Writable::sync_prepare
141 sub artnum_max { $_[0]->{oidx}->eidx_max }
143 sub index_unseen ($) {
145 my $new_smsg = $req->{new_smsg} or die 'BUG: {new_smsg} unset';
146 my $eml = delete $req->{eml};
147 $new_smsg->populate($eml, $req);
148 my $self = $req->{self};
149 my $docid = $self->{oidx}->adj_counter('eidx_docid', '+');
150 $new_smsg->{num} = $docid;
151 my $idx = $self->idx_shard($docid);
152 $self->{oidx}->add_overview($eml, $new_smsg);
153 my $oid = $new_smsg->{blob};
154 my $ibx = delete $req->{ibx} or die 'BUG: {ibx} unset';
155 $self->{oidx}->add_xref3($docid, $req->{xnum}, $oid, $ibx->eidx_key);
156 $idx->index_raw(undef, $eml, $new_smsg, $ibx);
157 check_batch_limit($req);
160 sub do_finalize ($) {
162 if (my $indexed = $req->{indexed}) {
163 do_xpost($req, $_) for @$indexed;
164 } elsif (exists $req->{new_smsg}) { # totally unseen messsage
167 # `d' message was already unindexed in the v1/v2 inboxes,
168 # so it's too noisy to warn, here.
172 sub do_step ($) { # main iterator for adding messages to the index
174 my $self = $req->{self};
176 if (my $next_arg = $req->{next_arg}) {
177 if (my $smsg = $self->{oidx}->next_by_mid(@$next_arg)) {
178 $req->{cur_smsg} = $smsg;
179 $self->git->cat_async($smsg->{blob},
180 \&ck_existing, $req);
181 return; # ck_existing calls do_step
183 delete $req->{cur_smsg};
184 delete $req->{next_arg};
186 my $mid = shift(@{$req->{mids}});
187 last unless defined $mid;
189 $req->{next_arg} = [ $mid, \$id, \$prev ];
195 sub _blob_missing ($) { # called when req->{cur_smsg}->{blob} is bad
197 my $smsg = $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
198 my $self = $req->{self};
199 my $xref3 = $self->{oidx}->get_xref3($smsg->{num});
200 my @keep = grep(!/:$smsg->{blob}\z/, @$xref3);
202 $keep[0] =~ /:([a-f0-9]{40,}+)\z/ or
203 die "BUG: xref $keep[0] has no OID";
205 $self->{oidx}->remove_xref3($smsg->{num}, $smsg->{blob});
206 my $upd = $self->{oidx}->update_blob($smsg, $oidhex);
207 my $saved = $self->{oidx}->get_art($smsg->{num});
209 $self->{oidx}->delete_by_num($smsg->{num});
213 sub ck_existing { # git->cat_async callback
214 my ($bref, $oid, $type, $size, $req) = @_;
215 my $smsg = $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
216 if ($type eq 'missing') {
218 } elsif (!is_bad_blob($oid, $type, $size, $smsg->{blob})) {
219 my $cur = PublicInbox::Eml->new($bref);
220 if (content_hash($cur) eq $req->{chash}) {
221 push @{$req->{indexed}}, $smsg; # for do_xpost
222 } # else { index_unseen later }
227 # is the messages visible in the inbox currently being indexed?
228 # return the number if so
229 sub cur_ibx_xnum ($$) {
230 my ($req, $bref) = @_;
231 my $ibx = $req->{ibx} or die 'BUG: current {ibx} missing';
234 git_blob_digest($bref)->hexdigest eq $req->{oid} or die
235 "BUG: blob mismatch $req->{oid}";
237 $req->{eml} = PublicInbox::Eml->new($bref);
238 $req->{chash} = content_hash($req->{eml});
239 $req->{mids} = mids($req->{eml});
240 my @q = @{$req->{mids}}; # copy
241 while (defined(my $mid = shift @q)) {
243 while (my $x = $ibx->over->next_by_mid($mid, \$id, \$prev)) {
244 return $x->{num} if $x->{blob} eq $req->{oid};
250 sub index_oid { # git->cat_async callback for 'm'
251 my ($bref, $oid, $type, $size, $req) = @_;
252 return if is_bad_blob($oid, $type, $size, $req->{oid});
253 my $new_smsg = $req->{new_smsg} = bless {
255 }, 'PublicInbox::Smsg';
256 $new_smsg->{bytes} = $size + crlf_adjust($$bref);
257 defined($req->{xnum} = cur_ibx_xnum($req, $bref)) or return;
261 sub unindex_oid { # git->cat_async callback for 'd'
262 my ($bref, $oid, $type, $size, $req) = @_;
263 return if is_bad_blob($oid, $type, $size, $req->{oid});
264 return if defined(cur_ibx_xnum($req, $bref)); # was re-added
268 # overrides V2Writable::last_commits, called by sync_ranges via sync_prepare
270 my ($self, $sync) = @_;
272 my $ekey = $sync->{ibx}->eidx_key;
273 my $uv = $sync->{ibx}->uidvalidity;
274 for my $i (0..$sync->{epoch_max}) {
275 $heads->[$i] = $self->{oidx}->eidx_meta("lc-v2:$ekey//$uv;$i");
280 sub _sync_inbox ($$$) {
281 my ($self, $opt, $ibx) = @_;
283 need_checkpoint => \(my $bool = 0),
284 reindex => $opt->{reindex},
289 my $v = $ibx->version;
290 my $ekey = $ibx->eidx_key;
293 defined($ibx->git_dir_latest(\$epoch_max)) or return;
294 $sync->{epoch_max} = $epoch_max;
295 sync_prepare($self, $sync) or return; # fills $sync->{todo}
297 my $uv = $ibx->uidvalidity;
298 my $lc = $self->{oidx}->eidx_meta("lc-v1:$ekey//$uv");
299 my $stk = prepare_stack($sync, $lc ? "$lc..HEAD" : 'HEAD');
300 my $unit = { stack => $stk, git => $ibx->git };
301 push @{$sync->{todo}}, $unit;
303 warn "E: $ekey unsupported inbox version (v$v)\n";
306 index_todo($self, $sync, $_) for @{$sync->{todo}};
309 sub eidx_sync { # main entry point
310 my ($self, $opt) = @_;
311 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
312 $self->{oidx}->rethread_prepare($opt);
314 my $warn_cb = $SIG{__WARN__} || sub { print STDERR @_ };
315 local $self->{current_info} = '';
316 local $SIG{__WARN__} = sub {
317 $warn_cb->($self->{current_info}, ': ', @_);
320 # don't use $_ here, it'll get clobbered by reindex_checkpoint
321 for my $ibx (@{$self->{ibx_list}}) {
322 _sync_inbox($self, $opt, $ibx);
325 $self->{oidx}->rethread_done($opt);
327 PublicInbox::V2Writable::done($self);
330 sub update_last_commit { # overrides V2Writable
331 my ($self, $sync, $unit, $latest_cmt) = @_;
332 return unless defined $latest_cmt;
334 $self->git->async_wait_all;
335 my $ibx = $sync->{ibx} or die 'BUG: {ibx} missing';
336 my $ekey = $ibx->eidx_key;
337 my $uv = $ibx->uidvalidity;
338 my $epoch = $unit->{epoch};
340 my $v = $ibx->version;
342 die 'No {epoch} for v2 unit' unless defined $epoch;
343 $meta_key = "lc-v2:$ekey//$uv;$epoch";
345 die 'Unexpected {epoch} for v1 unit' if defined $epoch;
346 $meta_key = "lc-v1:$ekey//$uv";
348 die "Unsupported inbox version: $v";
350 my $last = $self->{oidx}->eidx_meta($meta_key);
351 if (defined $last && is_ancestor($unit->{git}, $last, $latest_cmt)) {
352 my @cmd = (qw(rev-list --count), "$last..$latest_cmt");
353 chomp(my $n = $unit->{git}->qx(@cmd));
354 return if $n ne '' && $n == 0;
356 $self->{oidx}->eidx_meta($meta_key, $latest_cmt);
359 sub idx_init { # similar to V2Writable
360 my ($self, $opt) = @_;
361 return if $self->{idx_shards};
365 my $ALL = $self->git->{git_dir}; # ALL.git
366 PublicInbox::Import::init_bare($ALL) unless -d $ALL;
367 my $info_dir = "$ALL/objects/info";
368 my $alt = "$info_dir/alternates";
370 my (%old, @old, %new, @new);
372 open(my $fh, '<', $alt) or die "open $alt: $!";
373 $mode = (stat($fh))[2] & 07777;
375 push @old, $_ if !$old{$_}++;
378 for my $ibx (@{$self->{ibx_list}}) {
379 my $line = $ibx->git->{git_dir} . "/objects\n";
387 PublicInbox::V2Writable::write_alternates($info_dir, $mode, $o);
389 $self->parallel_init($self->{indexlevel});
390 $self->umask_prepare;
391 $self->with_umask(\&PublicInbox::V2Writable::_idx_init, $self, $opt);
392 $self->{oidx}->begin_lazy;
393 $self->{oidx}->eidx_prep;
397 *done = \&PublicInbox::V2Writable::done;
398 *umask_prepare = \&PublicInbox::InboxWritable::umask_prepare;
399 *with_umask = \&PublicInbox::InboxWritable::with_umask;
400 *parallel_init = \&PublicInbox::V2Writable::parallel_init;
401 *nproc_shards = \&PublicInbox::V2Writable::nproc_shards;
402 *sync_prepare = \&PublicInbox::V2Writable::sync_prepare;
403 *index_todo = \&PublicInbox::V2Writable::index_todo;
404 *count_shards = \&PublicInbox::V2Writable::count_shards;
405 *atfork_child = \&PublicInbox::V2Writable::atfork_child;
406 *idx_shard = \&PublicInbox::V2Writable::idx_shard;