X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearchMsg.pm;h=42384936b923364bb17fc515bd6b39c0d83bc35e;hb=95bdac7f09c69036efed537a4d03d5bdd2ae4eb6;hp=e3a04605b9b0925efdd97e390ba38496b94331ef;hpb=f76f265a851944b5dedcc3be5f3b5224b6ebda89;p=public-inbox.git

diff --git a/lib/PublicInbox/SearchMsg.pm b/lib/PublicInbox/SearchMsg.pm
index e3a04605..42384936 100644
--- a/lib/PublicInbox/SearchMsg.pm
+++ b/lib/PublicInbox/SearchMsg.pm
@@ -1,214 +1,198 @@
-# Copyright (C) 2015 all contributors <meta@public-inbox.org>
-# License: GPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
+# Copyright (C) 2015-2020 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 # based on notmuch, but with no concept of folders, files or flags
+#
+# Wraps a document inside our Xapian search index.
+# There may be many of these objects loaded in memory at once
+# for large threads in our WWW UI.
 package PublicInbox::SearchMsg;
 use strict;
 use warnings;
-use Search::Xapian;
-use Email::Address qw//;
-use Email::Simple qw//;
-use POSIX qw//;
-use Date::Parse qw/str2time/;
-use PublicInbox::MID qw/mid_clean/;
-use Encode qw/find_encoding/;
-my $enc_utf8 = find_encoding('UTF-8');
-our $PFX2TERM_RE = undef;
-use constant EPOCH_822 => 'Thu, 01 Jan 1970 00:00:00 +0000';
+use base qw(Exporter);
+our @EXPORT_OK = qw(subject_normalized);
+use PublicInbox::MID qw/mid_clean mid_mime/;
+use PublicInbox::Address;
+use PublicInbox::MsgTime qw(msg_timestamp msg_datestamp);
+use Time::Local qw(timegm);
 
 sub new {
 	my ($class, $mime) = @_;
-	my $doc = Search::Xapian::Document->new;
-	$doc->add_term(PublicInbox::Search::xpfx('type') . 'mail');
-
-	bless { type => 'mail', doc => $doc, mime => $mime }, $class;
+	bless { mime => $mime }, $class;
 }
 
 sub wrap {
-	my ($class, $doc, $mid) = @_;
-	bless { doc => $doc, mime => undef, mid => $mid }, $class;
-}
-
-sub load_doc {
-	my ($class, $doc) = @_;
-	my $data = $doc->get_data;
-	my $ts = eval {
-		no strict 'subs';
-		$doc->get_value(PublicInbox::Search::TS);
-	};
-	$ts = Search::Xapian::sortable_unserialise($ts);
-	$data = $enc_utf8->decode($data);
-	my ($subj, $from, $refs) = split(/\n/, $data);
-	bless {
-		doc => $doc,
-		subject => $subj,
-		ts => $ts,
-		from_name => $from,
-		references_sorted => $refs,
-	}, $class;
-}
-
-sub subject {
-	my ($self) = @_;
-	my $subj = $self->{subject};
-	return $subj if defined $subj;
-	$subj = $self->{mime}->header('Subject');
-	$subj = '' unless defined $subj;
-	$subj =~ tr/\n/ /;
-	$self->{subject} = $subj;
+	my ($class, $mid) = @_;
+	bless { mid => $mid }, $class;
 }
 
-sub from {
-	my ($self) = @_;
-	my $from = $self->mime->header('From') || '';
-	my @from;
-
-	if ($from) {
-		$from =~ tr/\n/ /;
-		@from = Email::Address->parse($from);
-		$self->{from} = $from[0];
-		$from = $from[0]->name;
-	}
-	$self->{from_name} = $from;
-	$self->{from};
+sub get_val ($$) {
+	my ($doc, $col) = @_;
+	# sortable_unserialise is defined by PublicInbox::Search::load_xapian()
+	sortable_unserialise($doc->get_value($col));
 }
 
-sub from_name {
-	my ($self) = @_;
-	my $from_name = $self->{from_name};
-	return $from_name if defined $from_name;
-	$self->from;
-	$self->{from_name};
+sub to_doc_data {
+	my ($self, $oid, $mid0) = @_;
+	$oid = '' unless defined $oid;
+	join("\n",
+		$self->subject,
+		$self->from,
+		$self->references,
+		$self->to,
+		$self->cc,
+		$oid,
+		$mid0,
+		$self->{bytes} || '',
+		$self->{lines} || ''
+	);
 }
 
-sub ts {
-	my ($self) = @_;
-	$self->{ts} ||= eval { str2time($self->mime->header('Date')) } || 0;
-}
+sub load_from_data ($$) {
+	my ($self) = $_[0]; # data = $_[1]
+	(
+		$self->{subject},
+		$self->{from},
+		$self->{references},
 
-sub to_doc_data {
-	my ($self) = @_;
-	PublicInbox::Search::subject_summary($self->subject) . "\n" .
-	$self->from_name . "\n".
-	$self->references_sorted;
+		# To: and Cc: are stored to optimize HDR/XHDR in NNTP since
+		# some NNTP clients will use that for message displays.
+		# NNTP only, and only stored in Over(view), not Xapian
+		$self->{to},
+		$self->{cc},
+
+		$self->{blob},
+		$self->{mid},
+
+		# NNTP only
+		$self->{bytes},
+		$self->{lines}
+	) = split(/\n/, $_[1]);
 }
 
-sub references_sorted {
-	my ($self) = @_;
-	my $x = $self->{references_sorted};
-	defined $x ? $x : '';
+sub load_expand {
+	my ($self, $doc) = @_;
+	my $data = $doc->get_data or return;
+	$self->{ts} = get_val($doc, PublicInbox::Search::TS());
+	my $dt = get_val($doc, PublicInbox::Search::DT());
+	my ($yyyy, $mon, $dd, $hh, $mm, $ss) = unpack('A4A2A2A2A2A2', $dt);
+	$self->{ds} = timegm($ss, $mm, $hh, $dd, $mon - 1, $yyyy);
+	utf8::decode($data);
+	load_from_data($self, $data);
+	$self;
 }
 
-sub ensure_metadata {
+sub psgi_cull ($) {
 	my ($self) = @_;
-	my $doc = $self->{doc};
-	my $end = $doc->termlist_end;
+	from_name($self); # fill in {from_name} so we can delete {from}
 
-	unless (defined $PFX2TERM_RE) {
-		my $or = join('|', keys %PublicInbox::Search::PFX2TERM_RMAP);
-		$PFX2TERM_RE = qr/\A($or)/;
-	}
+	# drop NNTP-only fields which aren't relevant to PSGI results:
+	# saves ~80K on a 200 item search result:
+	delete @$self{qw(from ts to cc bytes lines)};
+	$self;
+}
 
-	while (my ($pfx, $field) = each %PublicInbox::Search::PFX2TERM_RMAP) {
-		# ideally we'd move this out of the loop:
-		my $i = $doc->termlist_begin;
+# Only called by PSGI interface, not NNTP
+sub from_mitem {
+	my ($mitem, $srch) = @_;
+	return $srch->retry_reopen(\&from_mitem, $mitem) if $srch;
+	my $self = bless {}, __PACKAGE__;
+	psgi_cull(load_expand($self, $mitem->get_document));
+}
 
-		$i->skip_to($pfx);
-		if ($i != $end) {
-			my $val = $i->get_termname;
+# :bytes and :lines metadata in RFC 3977
+sub bytes ($) { $_[0]->{bytes} }
+sub lines ($) { $_[0]->{lines} }
 
-			if ($val =~ s/$PFX2TERM_RE//o) {
-				$self->{$field} = $val;
-			}
-		}
-	}
+sub __hdr ($$) {
+	my ($self, $field) = @_;
+	my $val = $self->{$field};
+	return $val if defined $val;
+
+	my $mime = $self->{mime} or return;
+	my @raw = $mime->header($field);
+	$val = join(', ', @raw);
+	$val =~ tr/\t\n/  /;
+	$val =~ tr/\r//d;
+	$self->{$field} = $val;
 }
 
-# for threading only
-sub mini_mime {
-	my ($self) = @_;
-	$self->ensure_metadata;
-	my @hs = (
-		'Subject' => $self->subject,
-		'X-PI-From' => $self->from_name,
-	);
+sub subject ($) { __hdr($_[0], 'subject') }
+sub to ($) { __hdr($_[0], 'to') }
+sub cc ($) { __hdr($_[0], 'cc') }
 
-	my @h = (
-		# prevent Email::Simple::Creator from running,
-		# this header is useless for threading as we use X-PI-TS
-		# for sorting and display:
-		'Date' => EPOCH_822,
-		'Message-ID' => "<$self->{mid}>",
-		'X-PI-TS' => $self->ts,
-	);
-	if (my $refs = $self->{references_sorted}) {
-		push @h, References => $refs;
-	}
-	my $mime = Email::MIME->create(header_str => \@hs, header => \@h);
-	my $h = $mime->header_obj;
+# no strftime, that is locale-dependent and not for RFC822
+my @DoW = qw(Sun Mon Tue Wed Thu Fri Sat);
+my @MoY = qw(Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec);
 
-	# set these headers manually since Encode::encode('MIME-Q', ...)
-	# will add spaces to long values when using header_str above.
+sub date ($) {
+	my ($self) = @_;
+	my $ds = $self->{ds};
+	return unless defined $ds;
+	my ($sec, $min, $hour, $mday, $mon, $year, $wday) = gmtime($ds);
+	"$DoW[$wday], " . sprintf("%02d $MoY[$mon] %04d %02d:%02d:%02d +0000",
+				$mday, $year+1900, $hour, $min, $sec);
 
-	# drop useless headers Email::MIME set for us
-	$h->header_set('Date');
-	$h->header_set('MIME-Version');
-	$mime;
 }
 
-sub mid {
-	my ($self, $mid) = @_;
-
-	if (defined $mid) {
-		$self->{mid} = $mid;
-	} elsif (my $rv = $self->{mid}) {
-		$rv;
-	} else {
-		$self->ensure_metadata; # needed for ghosts
-		$self->{mid} ||= $self->_extract_mid;
+sub from ($) {
+	my ($self) = @_;
+	my $from = __hdr($self, 'from');
+	if (defined $from && !defined $self->{from_name}) {
+		my @n = PublicInbox::Address::names($from);
+		$self->{from_name} = join(', ', @n);
 	}
+	$from;
 }
 
-sub _extract_mid {
+sub from_name {
 	my ($self) = @_;
+	my $from_name = $self->{from_name};
+	return $from_name if defined $from_name;
+	$self->from;
+	$self->{from_name};
+}
 
-	my $mid = $self->mime->header('Message-ID');
-	defined $mid ? mid_clean($mid) : $mid;
+sub ts {
+	my ($self) = @_;
+	$self->{ts} ||= eval { msg_timestamp($self->{mime}->header_obj) } || 0;
 }
 
-sub mime {
-	my ($self, $mime) = @_;
-	if (defined $mime) {
-		$self->{mime} = $mime;
-	} else {
-		# TODO load from git
-		$self->{mime};
-	}
+sub ds {
+	my ($self) = @_;
+	$self->{ds} ||= eval { msg_datestamp($self->{mime}->header_obj); } || 0;
+}
+
+sub references {
+	my ($self) = @_;
+	my $x = $self->{references};
+	defined $x ? $x : '';
 }
 
-sub doc_id {
-	my ($self, $doc_id) = @_;
-	if (defined $doc_id) {
-		$self->{doc_id} = $doc_id;
+sub mid ($;$) {
+	my ($self, $mid) = @_;
+
+	if (defined $mid) {
+		$self->{mid} = $mid;
+	} elsif (defined(my $rv = $self->{mid})) {
+		$rv;
 	} else {
-		# TODO load from xapian
-		$self->{doc_id};
+		die "NO {mime} for mid\n" unless $self->{mime};
+		$self->_extract_mid; # v1 w/o Xapian
 	}
 }
 
-sub thread_id {
-	my ($self) = @_;
-	my $tid = $self->{thread};
-	return $tid if defined $tid;
-	$self->ensure_metadata;
-	$self->{thread};
-}
+sub _extract_mid { mid_clean(mid_mime($_[0]->{mime})) }
 
-sub path {
-	my ($self) = @_;
-	my $path = $self->{path};
-	return $path if defined $path;
-	$self->ensure_metadata;
-	$self->{path};
+our $REPLY_RE = qr/^re:\s+/i;
+
+sub subject_normalized ($) {
+	my ($subj) = @_;
+	$subj =~ s/\A\s+//s; # no leading space
+	$subj =~ s/\s+\z//s; # no trailing space
+	$subj =~ s/\s+/ /gs; # no redundant spaces
+	$subj =~ s/\.+\z//; # no trailing '.'
+	$subj =~ s/$REPLY_RE//igo; # remove reply prefix
+	$subj;
 }
 
 1;