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 # read-only counterpart for PublicInbox::LeiStore
5 package PublicInbox::LeiSearch;
8 use parent qw(PublicInbox::ExtSearch);
9 use PublicInbox::Search qw(xap_terms);
10 use PublicInbox::ContentHash qw(content_digest content_hash);
11 use PublicInbox::MID qw(mids mids_for_index);
13 # get combined docid from over.num:
14 # (not generic Xapian, only works with our sharding scheme)
16 my ($self, $num) = @_;
17 my $nshard = $self->{nshard};
18 ($num - 1) * $nshard + $num % $nshard + 1;
22 my ($self, $num) = @_; # num_or_mitem
23 my $xdb = $self->xdb; # set {nshard};
24 my $docid = ref($num) ? $num->get_docid : num2docid($self, $num);
25 my $kw = xap_terms('K', $xdb, $docid);
26 warn "E: #$docid ($num): $@\n" if $@;
27 wantarray ? sort(keys(%$kw)) : $kw;
30 # returns undef if blob is unknown
32 my ($self, $oidhex) = @_;
33 my @num = $self->over->blob_exists($oidhex) or return;
34 my $xdb = $self->xdb; # set {nshard};
36 for my $num (@num) { # there should only be one...
37 my $doc = $xdb->get_document(num2docid($self, $num));
38 my $x = xap_terms('K', $doc);
44 # lookup keywords+labels for external messages
46 my ($self, $smsg, $want_label) = @_;
47 return if $smsg->{kw};
48 my $xdb = $self->xdb; # set {nshard};
49 my (%kw, %L, $doc, $x);
50 $kw{flagged} = 1 if delete($smsg->{lei_q_tt_flagged});
51 my @num = $self->over->blob_exists($smsg->{blob});
52 for my $num (@num) { # there should only be one...
53 $doc = $xdb->get_document(num2docid($self, $num));
54 $x = xap_terms('K', $doc);
56 if ($want_label) { # JSON/JMAP only
57 $x = xap_terms('L', $doc);
61 $smsg->{kw} = [ sort keys %kw ] if scalar(keys(%kw));
62 $smsg->{L} = [ sort keys %L ] if scalar(keys(%L));
65 # when a message has no Message-IDs at all, this is needed for
66 # unsent Draft messages, at least
69 my $dig = content_digest($eml);
70 my $chash = $dig->clone->digest;
71 my $mids = mids_for_index($eml);
73 $eml->{-lei_fake_mid} = $mids->[0] =
74 PublicInbox::Import::digest2mid($dig, $eml, 0);
79 sub _cmp_1st { # git->cat_async callback
80 my ($bref, $oid, $type, $size, $cmp) = @_;
81 # cmp: [chash, xoids, smsg, lms]
82 $bref //= $cmp->[3] ? $cmp->[3]->local_blob($oid, 1) : undef;
83 if ($bref && content_hash(PublicInbox::Eml->new($bref)) eq $cmp->[0]) {
84 $cmp->[1]->{$oid} = $cmp->[2]->{num};
88 # returns { OID => num } mapping for $eml matches
89 # The `num' hash value only makes sense from LeiSearch itself
90 # and is nonsense from the PublicInbox::LeiALE subclass
92 my ($self, $eml, $min) = @_;
93 my ($chash, $mids) = content_key($eml);
94 my @overs = ($self->over // $self->overs_all);
97 # no lms when used via {ale}:
98 my $lms = $self->{-lms_ro} //= lms($self) if defined($self->{topdir});
99 for my $mid (@$mids) {
102 while (my $cur = $o->next_by_mid($mid, \$id, \$prev)) {
103 next if $cur->{bytes} == 0 ||
104 $xoids->{$cur->{blob}};
105 $git->cat_async($cur->{blob}, \&_cmp_1st,
106 [$chash, $xoids, $cur, $lms]);
107 if ($min && scalar(keys %$xoids) >= $min) {
108 $git->cat_async_wait;
114 $git->cat_async_wait;
115 scalar(keys %$xoids) ? $xoids : undef;
118 # returns true if $eml is indexed by lei/store and keywords don't match
120 my ($self, $eml, $new_kw_sorted, $docids) = @_;
121 my $xoids = xoids_for($self, $eml) // return;
123 @$docids = sort { $a <=> $b } values %$xoids;
124 my $cur_kw = msg_keywords($self, $docids->[0]);
126 # RFC 5550 sec 5.9 on the $Forwarded keyword states:
127 # "Once set, the flag SHOULD NOT be cleared"
128 if (exists($cur_kw->{forwarded}) &&
129 !grep(/\Aforwarded\z/, @$new_kw_sorted)) {
130 delete $cur_kw->{forwarded};
132 $cur_kw = join("\0", sort keys %$cur_kw);
133 join("\0", @$new_kw_sorted) eq $cur_kw ? 0 : 1;
137 my ($self, $pfx) = @_;
138 my $xdb = $self->xdb;
139 my $cur = $xdb->allterms_begin($pfx);
140 my $end = $xdb->allterms_end($pfx);
142 for (; $cur != $end; $cur++) {
143 my $tn = $cur->get_termname;
144 index($tn, $pfx) == 0 and
145 $ret{substr($tn, length($pfx))} = undef;
147 wantarray ? (sort keys %ret) : \%ret;
152 my $qp = $self->SUPER::qparse_new; # PublicInbox::Search
153 $qp->add_boolean_prefix('kw', 'K');
154 $qp->add_boolean_prefix('L', 'L');
160 require PublicInbox::LeiMailSync;
161 my $f = "$self->{topdir}/mail_sync.sqlite3";
162 -f $f ? PublicInbox::LeiMailSync->new($f) : undef;
165 # allow SolverGit->resolve_patch to work with "lei index"
167 my ($self, $smsg) = @_;
168 PublicInbox::Inbox::smsg_eml($self, $smsg) // do {
169 my $lms = lms($self);
170 my $bref = $lms ? $lms->local_blob($smsg->{blob}, 1) : undef;
171 $bref ? PublicInbox::Eml->new($bref) : undef;