X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearchView.pm;h=7ef6b075bf60a5b18080f43dafc6380b407258c9;hb=6e6f7999361925e4c944f308df4bc32a1842cd69;hp=ebeb41f7a153ffbd69d137600261a85d5ea5d14e;hpb=712b8bb3a61cc261a4b8c1bfdb6e39f40cee3188;p=public-inbox.git

diff --git a/lib/PublicInbox/SearchView.pm b/lib/PublicInbox/SearchView.pm
index ebeb41f7..7ef6b075 100644
--- a/lib/PublicInbox/SearchView.pm
+++ b/lib/PublicInbox/SearchView.pm
@@ -1,54 +1,74 @@
-# Copyright (C) 2015 all contributors <meta@public-inbox.org>
-# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt)
+# Copyright (C) 2015-2018 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
 #
 # Displays search results for the web interface
 package PublicInbox::SearchView;
 use strict;
 use warnings;
+use URI::Escape qw(uri_unescape uri_escape);
 use PublicInbox::SearchMsg;
-use PublicInbox::Hval qw/ascii_html/;
+use PublicInbox::Hval qw/ascii_html obfuscate_addrs/;
 use PublicInbox::View;
-use PublicInbox::MID qw(mid2path mid_mime mid_clean mid_escape);
-use Email::MIME;
+use PublicInbox::WwwAtomStream;
+use PublicInbox::MID qw(MID_ESC);
+use PublicInbox::MIME;
 require PublicInbox::Git;
 require PublicInbox::SearchThread;
-our $LIM = 50;
+our $LIM = 200;
 
 sub noop {}
 
+sub mbox_results {
+	my ($ctx) = @_;
+	my $q = PublicInbox::SearchQuery->new($ctx->{qp});
+	my $x = $q->{x};
+	require PublicInbox::Mbox;
+	return PublicInbox::Mbox::mbox_all($ctx, $q->{'q'}) if $x eq 'm';
+	sres_top_html($ctx);
+}
+
 sub sres_top_html {
 	my ($ctx) = @_;
 	my $q = PublicInbox::SearchQuery->new($ctx->{qp});
-	my $code = 200;
+	my $x = $q->{x};
+	my $query = $q->{'q'};
 
+	my $code = 200;
 	# double the limit for expanded views:
 	my $opts = {
-		limit => $LIM,
+		limit => $q->{l},
 		offset => $q->{o},
 		mset => 1,
 		relevance => $q->{r},
 	};
-	my ($mset, $total);
+	my ($mset, $total, $err, $cb);
+retry:
 	eval {
-		$mset = $ctx->{srch}->query($q->{'q'}, $opts);
+		$mset = $ctx->{srch}->query($query, $opts);
 		$total = $mset->get_matches_estimated;
 	};
-	my $err = $@;
+	$err = $@;
 	ctx_prepare($q, $ctx);
-	my $cb;
 	if ($err) {
 		$code = 400;
 		$ctx->{-html_tip} = '<pre>'.err_txt($ctx, $err).'</pre><hr>';
 		$cb = *noop;
 	} elsif ($total == 0) {
+		if (defined($ctx->{-uxs_retried})) {
+			# undo retry damage:
+			$q->{'q'} = $ctx->{-uxs_retried};
+		} elsif (index($q->{'q'}, '%') >= 0) {
+			$ctx->{-uxs_retried} = $q->{'q'};
+			$q->{'q'} = uri_unescape($q->{'q'});
+			goto retry;
+		}
 		$code = 404;
 		$ctx->{-html_tip} = "<pre>\n[No results found]</pre><hr>";
 		$cb = *noop;
 	} else {
-		my $x = $q->{x};
-		return sub { adump($_[0], $mset, $q, $ctx) } if ($x eq 'A');
+		return adump($_[0], $mset, $q, $ctx) if $x eq 'A';
 
-		$ctx->{-html_tip} = search_nav_top($mset, $q) . "\n\n";
+		$ctx->{-html_tip} = search_nav_top($mset, $q, $ctx);
 		if ($x eq 't') {
 			$cb = mset_thread($ctx, $mset, $q);
 		} else {
@@ -58,7 +78,18 @@ sub sres_top_html {
 	PublicInbox::WwwStream->response($ctx, $code, $cb);
 }
 
-# display non-threaded search results similar to what users expect from
+# allow undef for individual doc loads...
+sub load_doc_retry {
+	my ($srch, $mitem) = @_;
+
+	eval {
+		$srch->retry_reopen(sub {
+			PublicInbox::SearchMsg->load_doc($mitem->get_document)
+		});
+	}
+}
+
+# display non-nested search results similar to what users expect from
 # regular WWW search engines:
 sub mset_summary {
 	my ($ctx, $mset, $q) = @_;
@@ -67,17 +98,31 @@ sub mset_summary {
 	my $pad = length("$total");
 	my $pfx = ' ' x $pad;
 	my $res = \($ctx->{-html_tip});
+	my $srch = $ctx->{srch};
+	my $ibx = $ctx->{-inbox};
+	my $obfs_ibx = $ibx->{obfuscate} ? $ibx : undef;
 	foreach my $m ($mset->items) {
 		my $rank = sprintf("%${pad}d", $m->get_rank + 1);
 		my $pct = $m->get_percent;
-		my $smsg = PublicInbox::SearchMsg->load_doc($m->get_document);
+		my $smsg = load_doc_retry($srch, $m);
+		unless ($smsg) {
+			eval {
+				$m = "$m ".$m->get_docid . " expired\n";
+				$ctx->{env}->{'psgi.errors'}->print($m);
+			};
+			next;
+		}
 		my $s = ascii_html($smsg->subject);
 		my $f = ascii_html($smsg->from_name);
-		my $ts = PublicInbox::View::fmt_ts($smsg->ts);
+		if ($obfs_ibx) {
+			obfuscate_addrs($obfs_ibx, $s);
+			obfuscate_addrs($obfs_ibx, $f);
+		}
+		my $date = PublicInbox::View::fmt_ts($smsg->ds);
 		my $mid = PublicInbox::Hval->new_msgid($smsg->mid)->{href};
 		$$res .= qq{$rank. <b><a\nhref="$mid/">}.
 			$s . "</a></b>\n";
-		$$res .= "$pfx  - by $f @ $ts UTC [$pct%]\n\n";
+		$$res .= "$pfx  - by $f @ $date UTC [$pct%]\n\n";
 	}
 	$$res .= search_nav_bot($mset, $q);
 	*noop;
@@ -93,9 +138,22 @@ sub err_txt {
 }
 
 sub search_nav_top {
-	my ($mset, $q) = @_;
+	my ($mset, $q, $ctx) = @_;
+	my $m = $q->qs_html(x => 'm', r => undef);
+	my $rv = qq{<form\naction="?$m"\nmethod="post"><pre>};
+	my $initial_q = $ctx->{-uxs_retried};
+	if (defined $initial_q) {
+		my $rewritten = $q->{'q'};
+		utf8::decode($initial_q);
+		utf8::decode($rewritten);
+		$initial_q = ascii_html($initial_q);
+		$rewritten = ascii_html($rewritten);
+		$rv .= " Warning: Initial query:\n <b>$initial_q</b>\n";
+		$rv .= " returned no results, used:\n";
+		$rv .= " <b>$rewritten</b>\n instead\n\n";
+	}
 
-	my $rv = "<pre>Search results ordered by [";
+	$rv .= 'Search results ordered by [';
 	if ($q->{r}) {
 		my $d = $q->qs_html(r => 0);
 		$rv .= qq{<a\nhref="?$d">date</a>|<b>relevance</b>};
@@ -109,62 +167,71 @@ sub search_nav_top {
 	my $x = $q->{x};
 	if ($x eq '') {
 		my $t = $q->qs_html(x => 't');
-		$rv .= qq{<b>summary</b>|<a\nhref="?$t">threaded</a>}
+		$rv .= qq{<b>summary</b>|<a\nhref="?$t">nested</a>}
 	} elsif ($q->{x} eq 't') {
 		my $s = $q->qs_html(x => '');
-		$rv .= qq{<a\nhref="?$s">summary</a>|<b>threaded</b>};
+		$rv .= qq{<a\nhref="?$s">summary</a>|<b>nested</b>};
 	}
 	my $A = $q->qs_html(x => 'A', r => undef);
 	$rv .= qq{|<a\nhref="?$A">Atom feed</a>]};
+	$rv .= qq{\n\t\t\t\t\t\tdownload: };
+	$rv .= qq{<input\ntype=submit\nvalue="mbox.gz"/></pre></form><pre>};
 }
 
 sub search_nav_bot {
 	my ($mset, $q) = @_;
 	my $total = $mset->get_matches_estimated;
-	my $nr = scalar $mset->items;
 	my $o = $q->{o};
-	my $end = $o + $nr;
+	my $l = $q->{l};
+	my $end = $o + $mset->size;
 	my $beg = $o + 1;
-	my $rv = "</pre><hr><pre>Results $beg-$end of $total";
-	my $n = $o + $LIM;
+	my $rv = '</pre><hr><pre id=t>';
+	if ($beg <= $end) {
+		$rv .= "Results $beg-$end of $total";
+		$rv .= ' (estimated)' if $end != $total;
+	} else {
+		$rv .= "No more results, only $total";
+	}
+	my $n = $o + $l;
 
 	if ($n < $total) {
-		my $qs = $q->qs_html(o => $n);
-		$rv .= qq{, <a\nhref="?$qs"\nrel=next>next</a>}
+		my $qs = $q->qs_html(o => $n, l => $l);
+		$rv .= qq{  <a\nhref="?$qs"\nrel=next>next</a>}
 	}
 	if ($o > 0) {
-		$rv .= $n < $total ? '/' : ',      ';
-		my $p = $o - $LIM;
+		$rv .= $n < $total ? '/' : '       ';
+		my $p = $o - $l;
 		my $qs = $q->qs_html(o => ($p > 0 ? $p : 0));
 		$rv .= qq{<a\nhref="?$qs"\nrel=prev>prev</a>};
 	}
 	$rv .= '</pre>';
 }
 
+sub sort_relevance {
+	my ($pct) = @_;
+	sub {
+		[ sort { (eval { $pct->{$b->topmost->{id}} } || 0)
+				<=>
+			(eval { $pct->{$a->topmost->{id}} } || 0)
+	} @{$_[0]} ] };
+}
+
 sub mset_thread {
 	my ($ctx, $mset, $q) = @_;
 	my %pct;
-	my @m = map {
+	my $srch = $ctx->{srch};
+	my $msgs = $srch->retry_reopen(sub { [ map {
 		my $i = $_;
-		my $m = PublicInbox::SearchMsg->load_doc($i->get_document);
-		$pct{$m->mid} = $i->get_percent;
-		$m;
-	} ($mset->items);
-
-	my $th = PublicInbox::SearchThread->new(\@m);
-	$th->thread;
-	if ($q->{r}) { # order by relevance
-		$th->order(sub {
-			[ sort { (eval { $pct{$b->topmost->{id}} } || 0)
-					<=>
-				(eval { $pct{$a->topmost->{id}} } || 0)
-			} @{$_[0]} ];
-		});
-	} else { # order by time (default for threaded view)
-		$th->order(*PublicInbox::View::sort_ts);
-	}
+		my $smsg = PublicInbox::SearchMsg->load_doc($i->get_document);
+		$pct{$smsg->mid} = $i->get_percent;
+		$smsg;
+	} ($mset->items) ]});
+	my $r = $q->{r};
+	my $rootset = PublicInbox::SearchThread::thread($msgs,
+		$r ? sort_relevance(\%pct) : *PublicInbox::View::sort_ds,
+		$ctx);
 	my $skel = search_nav_bot($mset, $q). "<pre>";
-	my $inbox = $ctx->{-inbox};
+	my $ibx = $ctx->{-inbox};
 	$ctx->{-upfx} = '';
 	$ctx->{anchor_idx} = 1;
 	$ctx->{cur_level} = 0;
@@ -173,22 +240,22 @@ sub mset_thread {
 	$ctx->{pct} = \%pct;
 	$ctx->{prev_attr} = '';
 	$ctx->{prev_level} = 0;
-	$ctx->{seen} = {};
-	$ctx->{s_nr} = scalar(@m).'+ results';
+	$ctx->{s_nr} = scalar(@$msgs).'+ results';
 
-	PublicInbox::View::walk_thread($th, $ctx,
+	# reduce hash lookups in skel_dump
+	$ctx->{-obfuscate} = $ctx->{-inbox}->{obfuscate};
+	PublicInbox::View::walk_thread($rootset, $ctx,
 		*PublicInbox::View::pre_thread);
 
-	my $msgs = \@m;
-	my $mime;
+	@$msgs = reverse @$msgs if $r;
 	sub {
 		return unless $msgs;
-		while ($mime = shift @$msgs) {
-			$mime = $inbox->msg_by_smsg($mime) and last;
+		my $smsg;
+		while (my $m = pop @$msgs) {
+			$smsg = $ibx->smsg_mime($m) and last;
 		}
-		if ($mime) {
-			$mime = Email::MIME->new($mime);
-			return PublicInbox::View::index_entry($mime, $ctx,
+		if ($smsg) {
+			return PublicInbox::View::index_entry($smsg, $ctx,
 				scalar @$msgs);
 		}
 		$msgs = undef;
@@ -198,7 +265,9 @@ sub mset_thread {
 
 sub ctx_prepare {
 	my ($q, $ctx) = @_;
-	my $qh = ascii_html($q->{'q'});
+	my $qh = $q->{'q'};
+	utf8::decode($qh);
+	$qh = ascii_html($qh);
 	$ctx->{-q_value_html} = $qh;
 	$ctx->{-atom} = '?'.$q->qs_html(x => 'A', r => undef);
 	$ctx->{-title_html} = "$qh - search results";
@@ -213,39 +282,37 @@ sub ctx_prepare {
 
 sub adump {
 	my ($cb, $mset, $q, $ctx) = @_;
-	my $fh = $cb->([ 200, ['Content-Type' => 'application/atom+xml']]);
 	my $ibx = $ctx->{-inbox};
-	my $feed_opts = PublicInbox::Feed::get_feedopts($ctx);
-	my $x = ascii_html($q->{'q'});
-	$x = qq{$x - search results};
-	$feed_opts->{atomurl} = $feed_opts->{url} . '?'. $q->qs_html;
-	$feed_opts->{url} .= '?'. $q->qs_html(x => undef);
-	$x = PublicInbox::Feed::atom_header($feed_opts, $x);
-	$fh->write($x. PublicInbox::Feed::feed_updated());
-
-	for ($mset->items) {
-		$x = PublicInbox::SearchMsg->load_doc($_->get_document)->mid;
-		$x = mid2path($x);
-		my $s = PublicInbox::Feed::feed_entry($feed_opts, $x, $ibx);
-		$fh->write($s) if defined $s;
-	}
-	PublicInbox::Feed::end_feed($fh);
+	my @items = $mset->items;
+	$ctx->{search_query} = $q;
+	my $srch = $ctx->{srch};
+	PublicInbox::WwwAtomStream->response($ctx, 200, sub {
+		while (my $x = shift @items) {
+			$x = load_doc_retry($srch, $x);
+			$x = $ibx->smsg_mime($x) and return $x;
+		}
+		return undef;
+	});
 }
 
 package PublicInbox::SearchQuery;
 use strict;
 use warnings;
+use URI::Escape qw(uri_escape);
 use PublicInbox::Hval;
-use PublicInbox::MID qw(mid_escape);
+use PublicInbox::MID qw(MID_ESC);
 
 sub new {
 	my ($class, $qp) = @_;
 
 	my $r = $qp->{r};
+	my ($l) = (($qp->{l} || '') =~ /(\d+)/);
+	$l = $LIM if !$l || $l > $LIM;
 	bless {
 		q => $qp->{'q'},
 		x => $qp->{x} || '',
 		o => (($qp->{o} || '0') =~ /(\d+)/),
+		l => $l,
 		r => (defined $r && $r ne '0'),
 	}, $class;
 }
@@ -261,18 +328,21 @@ sub qs_html {
 		$self = $tmp;
 	}
 
-	my $q = mid_escape($self->{'q'});
+	my $q = uri_escape($self->{'q'}, MID_ESC);
 	$q =~ s/%20/+/g; # improve URL readability
 	my $qs = "q=$q";
 
 	if (my $o = $self->{o}) { # ignore o == 0
 		$qs .= "&amp;o=$o";
 	}
+	if (my $l = $self->{l}) {
+		$qs .= "&amp;l=$l" unless $l == $LIM;
+	}
 	if (my $r = $self->{r}) {
 		$qs .= "&amp;r";
 	}
 	if (my $x = $self->{x}) {
-		$qs .= "&amp;x=$x" if ($x eq 't' || $x eq 'A');
+		$qs .= "&amp;x=$x" if ($x eq 't' || $x eq 'A' || $x eq 'm');
 	}
 	$qs;
 }