1 # Copyright (C) 2015-2018 all contributors <meta@public-inbox.org>
2 # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
3 # based on notmuch, but with no concept of folders, files or flags
5 # Wraps a document inside our Xapian search index.
6 package PublicInbox::SearchMsg;
9 use PublicInbox::MID qw/mid_clean mid_mime/;
10 use PublicInbox::Address;
11 use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
12 use Time::Local qw(timegm);
15 my ($class, $mime) = @_;
16 bless { mime => $mime }, $class;
20 my ($class, $mid) = @_;
21 bless { mid => $mid }, $class;
25 my ($class, $head, $db, $mid) = @_;
26 my $doc_id = $head->get_docid;
27 load_expand(wrap($class, $mid), $db->get_document($doc_id));
32 Search::Xapian::sortable_unserialise($doc->get_value($col));
36 my ($self, $oid, $mid0) = @_;
37 $oid = '' unless defined $oid;
51 sub load_from_data ($$) {
52 my ($self) = $_[0]; # data = $_[1]
58 # To: and Cc: are stored to optimize HDR/XHDR in NNTP since
59 # some NNTP clients will use that for message displays.
60 # NNTP only, and only stored in Over(view), not Xapian
70 ) = split(/\n/, $_[1]);
74 my ($self, $doc) = @_;
75 my $data = $doc->get_data or return;
76 $self->{ts} = get_val($doc, PublicInbox::Search::TS());
77 my $dt = get_val($doc, PublicInbox::Search::DT());
78 my ($yyyy, $mon, $dd, $hh, $mm, $ss) = unpack('A4A2A2A2A2A2', $dt);
79 $self->{ds} = timegm($ss, $mm, $hh, $dd, $mon - 1, $yyyy);
81 load_from_data($self, $data);
85 # Only called by PSGI interface, not NNTP
87 my ($class, $doc) = @_;
88 my $self = bless {}, $class;
89 my $smsg = load_expand($self, $doc);
91 from_name($smsg); # fill in {from_name} so we can delete {from}
93 # drop NNTP-only fields which aren't relevant to PSGI results:
94 # saves ~80K on a 200 item search result:
95 delete @$smsg{qw(from ts to cc bytes lines)};
99 # :bytes and :lines metadata in RFC 3977
100 sub bytes ($) { $_[0]->{bytes} }
101 sub lines ($) { $_[0]->{lines} }
104 my ($self, $field) = @_;
105 my $val = $self->{$field};
106 return $val if defined $val;
108 my $mime = $self->{mime} or return;
109 $val = $mime->header($field);
110 $val = '' unless defined $val;
113 $self->{$field} = $val;
116 sub subject ($) { __hdr($_[0], 'subject') }
117 sub to ($) { __hdr($_[0], 'to') }
118 sub cc ($) { __hdr($_[0], 'cc') }
120 # no strftime, that is locale-dependent and not for RFC822
121 my @DoW = qw(Sun Mon Tue Wed Thu Fri Sat);
122 my @MoY = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
126 my $ds = $self->{ds};
127 return unless defined $ds;
128 my ($sec, $min, $hour, $mday, $mon, $year, $wday) = gmtime($ds);
129 "$DoW[$wday], " . sprintf("%02d $MoY[$mon] %04d %02d:%02d:%02d +0000",
130 $mday, $year+1900, $hour, $min, $sec);
136 my $from = __hdr($self, 'from');
137 if (defined $from && !defined $self->{from_name}) {
138 my @n = PublicInbox::Address::names($from);
139 $self->{from_name} = join(', ', @n);
146 my $from_name = $self->{from_name};
147 return $from_name if defined $from_name;
154 $self->{ts} ||= eval { msg_timestamp($self->{mime}->header_obj) } || 0;
159 $self->{ds} ||= eval { msg_datestamp($self->{mime}->header_obj); } || 0;
164 my $x = $self->{references};
165 defined $x ? $x : '';
169 my ($self, $mid) = @_;
173 } elsif (defined(my $rv = $self->{mid})) {
176 die "NO {mime} for mid\n" unless $self->{mime};
177 $self->_extract_mid; # v1 w/o Xapian
181 sub _extract_mid { mid_clean(mid_mime($_[0]->{mime})) }