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);
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
110 my ($req, $smsg) = @_;
111 my $self = $req->{self};
112 my $docid = $smsg->{num};
113 my $idx = $self->idx_shard($docid);
114 my $oid = $req->{oid};
115 my $xibx = $req->{ibx};
116 my $eml = $req->{eml};
117 if (my $new_smsg = $req->{new_smsg}) { # 'm' on cross-posted message
118 my $xnum = $req->{xnum};
119 $self->{oidx}->add_xref3($docid, $xnum, $oid, $xibx->eidx_key);
120 $idx->shard_add_eidx_info($docid, $oid, $xibx, $eml);
122 $idx->shard_remove_eidx_info($docid, $oid, $xibx, $eml);
126 # called by V2Writable::sync_prepare
127 sub artnum_max { $_[0]->{oidx}->eidx_max }
129 sub index_unseen ($) {
131 my $new_smsg = $req->{new_smsg} or die 'BUG: {new_smsg} unset';
132 my $eml = delete $req->{eml};
133 $new_smsg->populate($eml, $req);
134 my $self = $req->{self};
135 my $docid = $self->{oidx}->adj_counter('eidx_docid', '+');
136 $new_smsg->{num} = $docid;
137 my $idx = $self->idx_shard($docid);
138 $self->{oidx}->add_overview($eml, $new_smsg);
139 my $oid = $new_smsg->{blob};
140 my $ibx = delete $req->{ibx} or die 'BUG: {ibx} unset';
141 $self->{oidx}->add_xref3($docid, $req->{xnum}, $oid, $ibx->eidx_key);
142 $idx->index_raw(undef, $eml, $new_smsg, $ibx);
145 sub do_finalize ($) {
147 if (my $indexed = $req->{indexed}) {
148 do_xpost($req, $_) for @$indexed;
149 } elsif (exists $req->{new_smsg}) { # totally unseen messsage
152 warn "W: ignoring delete $req->{oid} (not found)\n";
156 sub do_step ($) { # main iterator for adding messages to the index
158 my $self = $req->{self};
160 if (my $next_arg = $req->{next_arg}) {
161 if (my $smsg = $self->{oidx}->next_by_mid(@$next_arg)) {
162 $req->{cur_smsg} = $smsg;
163 $self->git->cat_async($smsg->{blob},
164 \&ck_existing, $req);
165 return; # ck_existing calls do_step
167 delete $req->{cur_smsg};
168 delete $req->{next_arg};
170 my $mid = shift(@{$req->{mids}});
171 last unless defined $mid;
173 $req->{next_arg} = [ $mid, \$id, \$prev ];
179 sub ck_existing { # git->cat_async callback
180 my ($bref, $oid, $type, $size, $req) = @_;
181 my $smsg = $req->{cur_smsg} or die 'BUG: {cur_smsg} missing';
182 return if is_bad_blob($oid, $type, $size, $smsg->{blob});
183 my $cur = PublicInbox::Eml->new($bref);
184 if (content_hash($cur) eq $req->{chash}) {
185 push @{$req->{indexed}}, $smsg; # for do_xpost
186 } # else { index_unseen later }
190 # is the messages visible in the inbox currently being indexed?
191 # return the number if so
192 sub cur_ibx_xnum ($$) {
193 my ($req, $bref) = @_;
194 my $ibx = $req->{ibx} or die 'BUG: current {ibx} missing';
197 git_blob_digest($bref)->hexdigest eq $req->{oid} or die
198 "BUG: blob mismatch $req->{oid}";
200 $req->{eml} = PublicInbox::Eml->new($bref);
201 $req->{chash} = content_hash($req->{eml});
202 $req->{mids} = mids($req->{eml});
203 my @q = @{$req->{mids}}; # copy
204 while (defined(my $mid = shift @q)) {
206 while (my $x = $ibx->over->next_by_mid($mid, \$id, \$prev)) {
207 return $x->{num} if $x->{blob} eq $req->{oid};
213 sub index_oid { # git->cat_async callback for 'm'
214 my ($bref, $oid, $type, $size, $req) = @_;
215 return if is_bad_blob($oid, $type, $size, $req->{oid});
216 my $new_smsg = $req->{new_smsg} = bless {
218 }, 'PublicInbox::Smsg';
219 $new_smsg->{bytes} = $size + crlf_adjust($$bref);
220 defined($req->{xnum} = cur_ibx_xnum($req, $bref)) or return;
224 sub unindex_oid { # git->cat_async callback for 'd'
225 my ($bref, $oid, $type, $size, $req) = @_;
226 return if is_bad_blob($oid, $type, $size, $req->{oid});
227 return if defined(cur_ibx_xnum($req, $bref)); # was re-added
231 # overrides V2Writable::last_commits, called by sync_ranges via sync_prepare
233 my ($self, $sync) = @_;
235 my $ekey = $sync->{ibx}->eidx_key;
236 my $uv = $sync->{ibx}->uidvalidity;
237 for my $i (0..$sync->{epoch_max}) {
238 $heads->[$i] = $self->{oidx}->eidx_meta("lc-v2:$ekey//$uv;$i");
243 sub _sync_inbox ($$$) {
244 my ($self, $opt, $ibx) = @_;
246 need_checkpoint => \(my $bool = 0),
247 reindex => $opt->{reindex},
252 my $v = $ibx->version;
253 my $ekey = $ibx->eidx_key;
256 defined($ibx->git_dir_latest(\$epoch_max)) or return;
257 $sync->{epoch_max} = $epoch_max;
258 sync_prepare($self, $sync) or return; # fills $sync->{todo}
260 my $uv = $ibx->uidvalidity;
261 my $lc = $self->{oidx}->eidx_meta("lc-v1:$ekey//$uv");
262 my $stk = prepare_stack($sync, $lc ? "$lc..HEAD" : 'HEAD');
263 my $unit = { stack => $stk, git => $ibx->git };
264 push @{$sync->{todo}}, $unit;
266 warn "E: $ekey unsupported inbox version (v$v)\n";
269 index_todo($self, $sync, $_) for @{$sync->{todo}};
272 sub eidx_sync { # main entry point
273 my ($self, $opt) = @_;
274 $self->idx_init($opt); # acquire lock via V2Writable::_idx_init
275 $self->{oidx}->rethread_prepare($opt);
277 _sync_inbox($self, $opt, $_) for (@{$self->{ibx_list}});
279 $self->{oidx}->rethread_done($opt);
281 PublicInbox::V2Writable::done($self);
284 sub update_last_commit {
285 my ($self, $sync, $unit, $latest_cmt) = @_;
287 my $ALL = $self->git;
288 # while (scalar(@{$ALL->{inflight_c}}) || scalar(@{$ALL->{inflight}})) {
289 # $ALL->check_async_wait;
290 # $ALL->cat_async_wait;
295 sub idx_init { # similar to V2Writable
296 my ($self, $opt) = @_;
297 return if $self->{idx_shards};
301 my $ALL = $self->git->{git_dir}; # ALL.git
302 PublicInbox::Import::init_bare($ALL) unless -d $ALL;
303 my $info_dir = "$ALL/objects/info";
304 my $alt = "$info_dir/alternates";
306 my (%old, @old, %new, @new);
308 open(my $fh, '<', $alt) or die "open $alt: $!";
309 $mode = (stat($fh))[2] & 07777;
311 push @old, $_ if !$old{$_}++;
314 for my $ibx (@{$self->{ibx_list}}) {
315 my $line = $ibx->git->{git_dir} . "/objects\n";
321 PublicInbox::V2Writable::write_alternates($info_dir, $mode, \@old);
322 $self->parallel_init($self->{indexlevel});
323 $self->umask_prepare;
324 $self->with_umask(\&PublicInbox::V2Writable::_idx_init, $self, $opt);
325 $self->{oidx}->begin_lazy;
326 $self->{oidx}->eidx_prep;
330 *done = \&PublicInbox::V2Writable::done;
331 *umask_prepare = \&PublicInbox::InboxWritable::umask_prepare;
332 *with_umask = \&PublicInbox::InboxWritable::with_umask;
333 *parallel_init = \&PublicInbox::V2Writable::parallel_init;
334 *nproc_shards = \&PublicInbox::V2Writable::nproc_shards;
335 *sync_prepare = \&PublicInbox::V2Writable::sync_prepare;
336 *index_todo = \&PublicInbox::V2Writable::index_todo;
337 *count_shards = \&PublicInbox::V2Writable::count_shards;
338 *atfork_child = \&PublicInbox::V2Writable::atfork_child;
339 *idx_shard = \&PublicInbox::V2Writable::idx_shard;