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_in);
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)) : $kw;
30 # lookup keywords+labels for external messages
32 my ($self, $smsg, $want_label) = @_;
33 return if $smsg->{kw};
34 my $xdb = $self->xdb; # set {nshard};
35 my (%kw, %L, $doc, $x);
36 $kw{flagged} = 1 if delete($smsg->{lei_q_tt_flagged});
37 my @num = $self->over->blob_exists($smsg->{blob});
38 for my $num (@num) { # there should only be one...
39 $doc = $xdb->get_document(num2docid($self, $num));
40 $x = xap_terms('K', $doc);
42 if ($want_label) { # JSON/JMAP only
43 $x = xap_terms('L', $doc);
47 $smsg->{kw} = [ sort keys %kw ] if scalar(keys(%kw));
48 $smsg->{L} = [ sort keys %L ] if scalar(keys(%L));
51 # when a message has no Message-IDs at all, this is needed for
52 # unsent Draft messages, at least
55 my $dig = content_digest($eml);
56 my $chash = $dig->clone->digest;
57 my $mids = mids_in($eml,
58 qw(Message-ID X-Alt-Message-ID Resent-Message-ID));
60 $eml->{-lei_fake_mid} = $mids->[0] =
61 PublicInbox::Import::digest2mid($dig, $eml);
66 sub _cmp_1st { # git->cat_async callback
67 my ($bref, $oid, $type, $size, $cmp) = @_; # cmp: [chash, xoids, smsg]
68 if ($bref && content_hash(PublicInbox::Eml->new($bref)) eq $cmp->[0]) {
69 $cmp->[1]->{$oid} = $cmp->[2]->{num};
73 # returns { OID => num } mapping for $eml matches
74 # The `num' hash value only makes sense from LeiSearch itself
75 # and is nonsense from the PublicInbox::LeiALE subclass
77 my ($self, $eml, $min) = @_;
78 my ($chash, $mids) = content_key($eml);
79 my @overs = ($self->over // $self->overs_all);
82 for my $mid (@$mids) {
85 while (my $cur = $o->next_by_mid($mid, \$id, \$prev)) {
86 next if $cur->{bytes} == 0 ||
87 $xoids->{$cur->{blob}};
88 $git->cat_async($cur->{blob}, \&_cmp_1st,
89 [ $chash, $xoids, $cur ]);
90 if ($min && scalar(keys %$xoids) >= $min) {
98 scalar(keys %$xoids) ? $xoids : undef;
101 # returns true if $eml is indexed by lei/store and keywords don't match
103 my ($self, $eml, $new_kw_sorted) = @_;
104 my $xoids = xoids_for($self, $eml, 1) // return;
105 my ($num) = values %$xoids;
106 my @cur_kw = msg_keywords($self, $num);
107 join("\0", @$new_kw_sorted) eq join("\0", @cur_kw) ? 0 : 1;
111 my ($self, $pfx) = @_;
112 my $xdb = $self->xdb;
113 my $cur = $xdb->allterms_begin($pfx);
114 my $end = $xdb->allterms_end($pfx);
116 for (; $cur != $end; $cur++) {
117 my $tn = $cur->get_termname;
118 index($tn, $pfx) == 0 and
119 $ret{substr($tn, length($pfx))} = undef;
121 wantarray ? (sort keys %ret) : \%ret;
126 my $qp = $self->SUPER::qparse_new; # PublicInbox::Search
127 $qp->add_boolean_prefix('kw', 'K');
128 $qp->add_boolean_prefix('L', 'L');