1 # Copyright (C) 2020-2021 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
4 # Combine any combination of PublicInbox::Search,
5 # PublicInbox::ExtSearch, and PublicInbox::LeiSearch objects
7 package PublicInbox::LeiXSearch;
10 use parent qw(PublicInbox::LeiSearch PublicInbox::IPC);
11 use PublicInbox::DS qw(dwaitpid now);
12 use PublicInbox::PktOp qw(pkt_do);
13 use PublicInbox::Import;
14 use File::Temp 0.19 (); # 0.19 for ->newdir
16 use PublicInbox::Search qw(xap_terms);
17 use PublicInbox::Spawn qw(popen_rd spawn which);
18 use PublicInbox::MID qw(mids);
19 use PublicInbox::Smsg;
20 use Fcntl qw(SEEK_SET F_SETFL O_APPEND O_RDWR);
24 PublicInbox::Search::load_xapian();
26 qp_flags => $PublicInbox::Search::QP_FLAGS |
27 PublicInbox::Search::FLAG_PURE_NOT(),
32 my ($self, $ibxish) = @_; # ibxish = ExtSearch or Inbox
33 my $desc = $ibxish->{inboxdir} // $ibxish->{topdir};
34 my $srch = $ibxish->search or
35 return warn("$desc not indexed for Xapian\n");
36 my @shards = $srch->xdb_shards_flat or
37 return warn("$desc has no Xapian shards\n");
39 if (delete $self->{xdb}) { # XXX: do we need this?
40 # clobber existing {xdb} if amending
41 my $expect = delete $self->{nshard};
42 my $shards = delete $self->{shards_flat};
43 scalar(@$shards) == $expect or die
44 "BUG: {nshard}$expect != shards=".scalar(@$shards);
47 for my $old_ibxish (@{$self->{shard2ibx}}) {
48 next if $prev == $old_ibxish;
50 my @shards = $old_ibxish->search->xdb_shards_flat;
51 push @{$self->{shards_flat}}, @shards;
53 my $nr = scalar(@{$self->{shards_flat}});
55 "BUG: reloaded $nr shards, expected $expect"
57 push @{$self->{shards_flat}}, @shards;
58 push(@{$self->{shard2ibx}}, $ibxish) for (@shards);
61 # returns a list of local inboxes (or count in scalar context)
62 sub locals { @{$_[0]->{locals} // []} }
64 sub remotes { @{$_[0]->{remotes} // []} }
66 # called by PublicInbox::Search::xdb
67 sub xdb_shards_flat { @{$_[0]->{shards_flat} // []} }
71 my ($self, $mitem) = @_;
72 # cf. https://trac.xapian.org/wiki/FAQ/MultiDatabaseDocumentID
73 my $nshard = $self->{nshard};
74 my $docid = $mitem->get_docid;
75 my $shard = ($docid - 1) % $nshard;
76 my $num = int(($docid - 1) / $nshard) + 1;
77 my $ibx = $self->{shard2ibx}->[$shard];
78 my $smsg = $ibx->over->get_art($num);
79 if (ref($ibx->can('msg_keywords'))) {
80 my $kw = xap_terms('K', $mitem->get_document);
81 $smsg->{kw} = [ sort keys %$kw ];
83 $smsg->{docid} = $docid;
88 my ($self, $qstr, $opt) = @_;
90 $opt->{relevance} //= -2;
91 $self->mset($qstr //= 'bytes:1..', $opt);
98 my $size = $mset->size;
99 $size >= $mo->{limit} && (($mo->{offset} += $size) < $mo->{limit});
102 # $startq will EOF when query_prepare is done augmenting and allow
103 # query_mset and query_thread_mset to proceed.
104 sub wait_startq ($) {
107 read($startq, my $query_prepare_done, 1);
112 return unless $lei->{-progress};
113 if ($lei->{pkt_op}) { # called via pkt_op/pkt_do from workers
114 pkt_do($lei->{pkt_op}, 'mset_progress', @_);
115 } else { # single lei-daemon consumer
116 my ($desc, $mset_size, $mset_total_est) = @_;
117 $lei->{-mset_total} += $mset_size;
118 $lei->err("# $desc $mset_size/$mset_total_est");
122 sub query_thread_mset { # for --thread
123 my ($self, $lei, $ibxish) = @_;
124 local $0 = "$0 query_thread_mset";
125 $lei->atfork_child_wq($self);
126 my $startq = delete $lei->{startq};
128 my ($srch, $over) = ($ibxish->search, $ibxish->over);
129 my $desc = $ibxish->{inboxdir} // $ibxish->{topdir};
130 return warn("$desc not indexed by Xapian\n") unless ($srch && $over);
131 my $mo = { %{$lei->{mset_opt}} };
133 my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei, $ibxish);
135 $mset = $srch->mset($mo->{qstr}, $mo);
136 mset_progress($lei, $desc, $mset->size,
137 $mset->get_matches_estimated);
138 my $ids = $srch->mset_to_artnums($mset, $mo);
139 my $ctx = { ids => $ids };
141 my %n2item = map { ($ids->[$i++], $_) } $mset->items;
142 while ($over->expand_thread($ctx)) {
143 for my $n (@{$ctx->{xids}}) {
144 my $smsg = $over->get_art($n) or next;
145 wait_startq($startq) if $startq;
146 my $mitem = delete $n2item{$smsg->{num}};
147 $each_smsg->($smsg, $mitem);
149 @{$ctx->{xids}} = ();
151 } while (_mset_more($mset, $mo));
152 undef $each_smsg; # drops @io for l2m->{each_smsg_done}
153 $lei->{ovv}->ovv_atexit_child($lei);
156 sub query_mset { # non-parallel for non-"--thread" users
157 my ($self, $lei) = @_;
158 local $0 = "$0 query_mset";
159 $lei->atfork_child_wq($self);
160 my $startq = delete $lei->{startq};
161 my $mo = { %{$lei->{mset_opt}} };
163 for my $loc (locals($self)) {
164 attach_external($self, $loc);
166 my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei, $self);
168 $mset = $self->mset($mo->{qstr}, $mo);
169 mset_progress($lei, 'xsearch', $mset->size,
170 $mset->size, $mset->get_matches_estimated);
171 for my $mitem ($mset->items) {
172 my $smsg = smsg_for($self, $mitem) or next;
173 wait_startq($startq) if $startq;
174 $each_smsg->($smsg, $mitem);
176 } while (_mset_more($mset, $mo));
177 undef $each_smsg; # drops @io for l2m->{each_smsg_done}
178 $lei->{ovv}->ovv_atexit_child($lei);
181 sub each_eml { # callback for MboxReader->mboxrd
182 my ($eml, $self, $lei, $each_smsg) = @_;
183 my $smsg = bless {}, 'PublicInbox::Smsg';
184 $smsg->populate($eml);
185 $smsg->parse_references($eml, mids($eml));
186 $smsg->{$_} //= '' for qw(from to cc ds subject references mid);
187 delete @$smsg{qw(From Subject -ds -ts)};
188 if (my $startq = delete($lei->{startq})) { wait_startq($startq) }
189 if ($lei->{-progress}) {
190 ++$lei->{-nr_remote_eml};
192 my $next = $lei->{-next_progress} //= ($now + 1);
194 $lei->{-next_progress} = $now + 1;
195 my $nr = $lei->{-nr_remote_eml};
196 $lei->err("# $lei->{-current_url} $nr/?");
199 $each_smsg->($smsg, undef, $eml);
202 # PublicInbox::OnDestroy callback
205 kill('KILL', $pid); # spawn() blocks other signals
209 sub query_remote_mboxrd {
210 my ($self, $lei, $uris) = @_;
211 local $0 = "$0 query_remote_mboxrd";
212 $lei->atfork_child_wq($self);
213 local $SIG{TERM} = sub { exit(0) }; # for DESTROY (File::Temp, $reap)
214 my ($opt, $env) = @$lei{qw(opt env)};
215 my @qform = (q => $lei->{mset_opt}->{qstr}, x => 'm');
216 push(@qform, t => 1) if $opt->{thread};
217 my @cmd = ($self->{curl}, qw(-sSf -d), '');
218 my $verbose = $opt->{verbose};
220 my $cerr = File::Temp->new(TEMPLATE => 'curl.err-XXXX', TMPDIR => 1);
221 fcntl($cerr, F_SETFL, O_APPEND|O_RDWR) or warn "set O_APPEND: $!";
222 my $rdr = { 2 => $cerr };
225 # spawn a process to force line-buffering, otherwise curl
226 # will write 1 character at-a-time and parallel outputs
227 # mmmaaayyy llloookkk llliiikkkeee ttthhhiiisss
229 my $o = { 1 => $lei->{2}, 2 => $lei->{2} };
230 my $pid = spawn(['tail', '-f', $cerr->filename], undef, $o);
231 $reap = PublicInbox::OnDestroy->new(\&kill_reap, $pid);
233 for my $o ($lei->curl_opt) {
234 $o =~ s/\|[a-z0-9]\b//i; # remove single char short option
235 if ($o =~ s/=[is]@\z//) {
236 my $ary = $opt->{$o} or next;
237 push @cmd, map { ("--$o", $_) } @$ary;
238 } elsif ($o =~ s/=[is]\z//) {
239 my $val = $opt->{$o} // next;
240 push @cmd, "--$o", $val;
241 } elsif ($opt->{$o}) {
245 $opt->{torsocks} = 'false' if $opt->{'no-torsocks'};
246 my $tor = $opt->{torsocks} //= 'auto';
247 my $each_smsg = $lei->{ovv}->ovv_each_smsg_cb($lei);
248 for my $uri (@$uris) {
249 $lei->{-current_url} = $uri->as_string;
250 $lei->{-nr_remote_eml} = 0;
251 $uri->query_form(@qform);
252 my $cmd = [ @cmd, $uri->as_string ];
253 if ($tor eq 'auto' && substr($uri->host, -6) eq '.onion' &&
254 (($env->{LD_PRELOAD}//'') !~ /torsocks/)) {
255 unshift @$cmd, which('torsocks');
256 } elsif (PublicInbox::Config::git_bool($tor)) {
257 unshift @$cmd, which('torsocks');
260 # continue anyways if torsocks is missing; a proxy may be
261 # specified via CLI, curlrc, environment variable, or even
263 shift(@$cmd) if !$cmd->[0];
265 $lei->err("# @$cmd") if $verbose;
266 my ($fh, $pid) = popen_rd($cmd, $env, $rdr);
267 $fh = IO::Uncompress::Gunzip->new($fh);
268 PublicInbox::MboxReader->mboxrd($fh, \&each_eml, $self,
270 waitpid($pid, 0) == $pid or die "BUG: waitpid (curl): $!";
272 my $nr = $lei->{-nr_remote_eml};
273 mset_progress($lei, $lei->{-current_url}, $nr, $nr);
276 seek($cerr, $coff, SEEK_SET) or warn "seek(curl stderr): $!\n";
277 my $e = do { local $/; <$cerr> } //
278 die "read(curl stderr): $!\n";
281 next if (($? >> 8) == 22 && $e =~ /\b404\b/);
282 $lei->child_error($?);
283 $uri->query_form(q => $lei->{mset_opt}->{qstr});
284 # --verbose already showed the error via tail(1)
285 $lei->err("E: $uri \$?=$?\n", $verbose ? () : $e);
288 $lei->{ovv}->ovv_atexit_child($lei);
294 my $tmp = File::Temp->newdir('lei_xsrch_git-XXXXXXXX', TMPDIR => 1);
295 for my $ibx (@{$self->{shard2ibx} // []}) {
296 my $d = File::Spec->canonpath($ibx->git->{git_dir});
297 $seen{$d} //= push @dirs, "$d/objects\n"
299 my $git_dir = $tmp->dirname;
300 PublicInbox::Import::init_bare($git_dir);
301 my $f = "$git_dir/objects/info/alternates";
302 open my $alt, '>', $f or die "open($f): $!";
303 print $alt @dirs or die "print $f: $!";
304 close $alt or die "close $f: $!";
305 my $git = PublicInbox::Git->new($git_dir);
310 sub query_done { # EOF callback
312 my $has_l2m = exists $lei->{l2m};
313 for my $f (qw(lxs l2m)) {
314 my $wq = delete $lei->{$f} or next;
315 $wq->wq_wait_old($lei);
317 $lei->{ovv}->ovv_end($lei);
318 if ($has_l2m) { # close() calls LeiToMail reap_compress
319 if (my $out = delete $lei->{old_1}) {
320 if (my $mbout = $lei->{1}) {
321 close($mbout) or return $lei->fail(<<"");
322 Error closing $lei->{ovv}->{dst}: $!
329 $lei->{-progress} and
330 $lei->err('# ', $lei->{-mset_total} // 0, " matches");
334 sub do_post_augment {
335 my ($lei, $zpipe, $au_done) = @_;
336 my $l2m = $lei->{l2m} or die 'BUG: no {l2m}';
337 eval { $l2m->post_augment($lei, $zpipe) };
339 if (my $lxs = delete $lei->{lxs}) {
345 close $au_done; # triggers wait_startq
348 my $MAX_PER_HOST = 4;
351 my ($self, $opt) = @_;
352 my $nl = $opt->{thread} ? locals($self) : 1;
353 my $nr = remotes($self);
354 $nr = $MAX_PER_HOST if $nr > $MAX_PER_HOST;
358 sub start_query { # always runs in main (lei-daemon) process
359 my ($self, $io, $lei) = @_;
360 if ($lei->{opt}->{thread}) {
361 for my $ibxish (locals($self)) {
362 $self->wq_do('query_thread_mset', $io, $lei, $ibxish);
364 } elsif (locals($self)) {
365 $self->wq_do('query_mset', $io, $lei);
369 for my $uri (remotes($self)) {
370 push @{$q->[$i++ % $MAX_PER_HOST]}, $uri;
373 $self->wq_do('query_remote_mboxrd', $io, $lei, $uris);
378 sub query_prepare { # called by wq_do
379 my ($self, $lei) = @_;
380 local $0 = "$0 query_prepare";
381 $lei->atfork_child_wq($self);
382 delete $lei->{l2m}->{-wq_s1};
383 eval { $lei->{l2m}->do_augment($lei) };
384 $lei->fail($@) if $@;
385 pkt_do($lei->{pkt_op}, '.') == 1 or die "do_post_augment trigger: $!"
388 sub fail_handler ($;$$) {
389 my ($lei, $code, $io) = @_;
390 for my $f (qw(lxs l2m)) {
391 my $wq = delete $lei->{$f} or next;
392 $wq->wq_wait_old($lei) if $wq->wq_kill_old; # lei-daemon
394 close($io) if $io; # needed to avoid warnings on SIGPIPE
395 $lei->x_it($code // (1 >> 8));
398 sub sigpipe_handler { # handles SIGPIPE from l2m/lxs workers
399 fail_handler($_[0], 13, delete $_[0]->{1});
403 my ($self, $lei) = @_;
404 $lei->{1}->autoflush(1);
405 my ($au_done, $zpipe);
406 my $l2m = $lei->{l2m};
408 pipe($lei->{startq}, $au_done) or die "pipe: $!";
410 fcntl($lei->{startq}, 1031, 4096) if $^O eq 'linux';
411 $zpipe = $l2m->pre_augment($lei);
414 '|' => [ \&sigpipe_handler, $lei ],
415 '!' => [ \&fail_handler, $lei ],
416 '.' => [ \&do_post_augment, $lei, $zpipe, $au_done ],
417 '' => [ \&query_done, $lei ],
418 'mset_progress' => [ \&mset_progress, $lei ],
419 'x_it' => [ $lei->can('x_it'), $lei ],
420 'child_error' => [ $lei->can('child_error'), $lei ],
422 (my $op, $lei->{pkt_op}) = PublicInbox::PktOp->pair($ops);
423 my ($lei_ipc, @io) = $lei->atfork_parent_wq($self);
424 delete($lei->{pkt_op});
426 $lei->event_step_init; # wait for shutdowns
428 $self->wq_do('query_prepare', \@io, $lei_ipc);
429 $io[1] = $zpipe->[1] if $zpipe;
431 start_query($self, \@io, $lei_ipc);
433 if ($lei->{oneshot}) {
434 # for the $lei_ipc->atfork_child_wq PIPE handler:
435 while ($op->{sock}) { $op->event_step }
440 my ($self, $uri) = @_;
441 if (my $curl = $self->{curl} //= which('curl') // 0) {
442 require PublicInbox::MboxReader;
443 require IO::Uncompress::Gunzip;
444 push @{$self->{remotes}}, $uri;
446 warn "curl missing, ignoring $uri\n";
450 sub prepare_external {
451 my ($self, $loc, $boost) = @_; # n.b. already ordered by boost
452 if (ref $loc) { # already a URI, or PublicInbox::Inbox-like object
453 return add_uri($self, $loc) if $loc->can('scheme');
454 } elsif ($loc =~ m!\Ahttps?://!) {
456 return add_uri($self, URI->new($loc));
457 } elsif (-f "$loc/ei.lock") {
458 require PublicInbox::ExtSearch;
459 $loc = PublicInbox::ExtSearch->new($loc);
460 } elsif (-f "$loc/inbox.lock" || -d "$loc/public-inbox") {
461 require PublicInbox::Inbox; # v2, v1
462 $loc = bless { inboxdir => $loc }, 'PublicInbox::Inbox';
464 warn "W: ignoring $loc, unable to determine type\n";
467 push @{$self->{locals}}, $loc;