1 # Copyright (C) 2015-2019 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 # There may be many of these objects loaded in memory at once
7 # for large threads in our WWW UI.
8 package PublicInbox::SearchMsg;
11 use base qw(Exporter);
12 our @EXPORT_OK = qw(subject_normalized);
13 use PublicInbox::MID qw/mid_clean mid_mime/;
14 use PublicInbox::Address;
15 use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
16 use Time::Local qw(timegm);
19 my ($class, $mime) = @_;
20 bless { mime => $mime }, $class;
24 my ($class, $mid) = @_;
25 bless { mid => $mid }, $class;
30 # sortable_unserialise is defined by PublicInbox::Search::load_xapian()
31 sortable_unserialise($doc->get_value($col));
35 my ($self, $oid, $mid0) = @_;
36 $oid = '' unless defined $oid;
50 sub load_from_data ($$) {
51 my ($self) = $_[0]; # data = $_[1]
57 # To: and Cc: are stored to optimize HDR/XHDR in NNTP since
58 # some NNTP clients will use that for message displays.
59 # NNTP only, and only stored in Over(view), not Xapian
69 ) = split(/\n/, $_[1]);
73 my ($self, $doc) = @_;
74 my $data = $doc->get_data or return;
75 $self->{ts} = get_val($doc, PublicInbox::Search::TS());
76 my $dt = get_val($doc, PublicInbox::Search::DT());
77 my ($yyyy, $mon, $dd, $hh, $mm, $ss) = unpack('A4A2A2A2A2A2', $dt);
78 $self->{ds} = timegm($ss, $mm, $hh, $dd, $mon - 1, $yyyy);
80 load_from_data($self, $data);
86 from_name($self); # fill in {from_name} so we can delete {from}
88 # drop NNTP-only fields which aren't relevant to PSGI results:
89 # saves ~80K on a 200 item search result:
90 delete @$self{qw(from ts to cc bytes lines)};
94 # Only called by PSGI interface, not NNTP
96 my ($mitem, $srch) = @_;
97 return $srch->retry_reopen(\&from_mitem, $mitem) if $srch;
98 my $self = bless {}, __PACKAGE__;
99 psgi_cull(load_expand($self, $mitem->get_document));
102 # :bytes and :lines metadata in RFC 3977
103 sub bytes ($) { $_[0]->{bytes} }
104 sub lines ($) { $_[0]->{lines} }
107 my ($self, $field) = @_;
108 my $val = $self->{$field};
109 return $val if defined $val;
111 my $mime = $self->{mime} or return;
112 my @raw = $mime->header($field);
113 $val = join(', ', @raw);
116 $self->{$field} = $val;
119 sub subject ($) { __hdr($_[0], 'subject') }
120 sub to ($) { __hdr($_[0], 'to') }
121 sub cc ($) { __hdr($_[0], 'cc') }
123 # no strftime, that is locale-dependent and not for RFC822
124 my @DoW = qw(Sun Mon Tue Wed Thu Fri Sat);
125 my @MoY = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
129 my $ds = $self->{ds};
130 return unless defined $ds;
131 my ($sec, $min, $hour, $mday, $mon, $year, $wday) = gmtime($ds);
132 "$DoW[$wday], " . sprintf("%02d $MoY[$mon] %04d %02d:%02d:%02d +0000",
133 $mday, $year+1900, $hour, $min, $sec);
139 my $from = __hdr($self, 'from');
140 if (defined $from && !defined $self->{from_name}) {
141 my @n = PublicInbox::Address::names($from);
142 $self->{from_name} = join(', ', @n);
149 my $from_name = $self->{from_name};
150 return $from_name if defined $from_name;
157 $self->{ts} ||= eval { msg_timestamp($self->{mime}->header_obj) } || 0;
162 $self->{ds} ||= eval { msg_datestamp($self->{mime}->header_obj); } || 0;
167 my $x = $self->{references};
168 defined $x ? $x : '';
172 my ($self, $mid) = @_;
176 } elsif (defined(my $rv = $self->{mid})) {
179 die "NO {mime} for mid\n" unless $self->{mime};
180 $self->_extract_mid; # v1 w/o Xapian
184 sub _extract_mid { mid_clean(mid_mime($_[0]->{mime})) }
186 our $REPLY_RE = qr/^re:\s+/i;
188 sub subject_normalized ($) {
190 $subj =~ s/\A\s+//s; # no leading space
191 $subj =~ s/\s+\z//s; # no trailing space
192 $subj =~ s/\s+/ /gs; # no redundant spaces
193 $subj =~ s/\.+\z//; # no trailing '.'
194 $subj =~ s/$REPLY_RE//igo; # remove reply prefix