[PATCH 0/3] enhance search results

[PATCH 0/3] enhance search results

[Eric Wong]

Allow users to search and expand threads at the same time.  This
allows easily reading entire threads which match a given subject
or phrase (such as a commit message).

[PATCH 1/3] view: more consistent prefix for ghost links

[Eric Wong]

This will be reused for search views.
---
 lib/PublicInbox/View.pm | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index 9f87546..7d85555 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -598,7 +598,7 @@ sub ghost_parent {
 	$mid = PublicInbox::Hval->new_msgid($mid);
 	my $href = $mid->as_href;
 	my $html = $mid->as_html;
-	qq{[parent not found: &lt;<a\nhref="$upfx../$href/">$html</a>&gt;]};
+	qq{[parent not found: &lt;<a\nhref="$upfx$href/">$html</a>&gt;]};
 }
 
 sub __thread_entry {
@@ -619,7 +619,7 @@ sub __thread_entry {
 		foreach my $g (@$ghost) {
 			$$cb->write("<table\nsummary=ghost><tr><td>" .
 				(INDENT x $g->[1]) . "</td><td>" .
-				PRE_WRAP . ghost_parent('../', $g->[0]) .
+				PRE_WRAP . ghost_parent('../../', $g->[0]) .
 				'</pre></td></table>');
 		}
 	}
@@ -751,7 +751,8 @@ sub inline_dump {
 	} else {
 		my $dot = $level == 0 ? '' : '` ';
 		my $pfx = (INDENT x $level) . $dot;
-		$$dst .= $pfx . ghost_parent($upfx, $node->messageid) . "\n";
+		$$dst .= $pfx;
+		$$dst .= ghost_parent("$upfx../", $node->messageid) . "\n";
 	}
 	inline_dump($dst, $state, $upfx, $node->child, $level+1);
 	inline_dump($dst, $state, $upfx, $node->next, $level);
-- 
EW

[PATCH 2/3] searchview: hoist out subroutines for clarity

[Eric Wong]

We'll be expanding the search view to handle expanded views.
---
 lib/PublicInbox/SearchView.pm | 93 +++++++++++++++++++++++++------------------
 1 file changed, 55 insertions(+), 38 deletions(-)

diff --git a/lib/PublicInbox/SearchView.pm b/lib/PublicInbox/SearchView.pm
index 13f4de2..82b97f3 100644
--- a/lib/PublicInbox/SearchView.pm
+++ b/lib/PublicInbox/SearchView.pm
@@ -38,52 +38,19 @@ sub sres_top_html {
 	my $foot = $ctx->{footer} || '';
 	$foot = qq{Back to <a\nhref=".">index</a>.};
 	if ($err) {
-		my $u = 'http://xapian.org/docs/queryparser.html';
 		$code = 400;
-		$err =~ s/^\s*Exception:\s*//; # bad word to show users :P
-		$err = PublicInbox::Hval->new_oneline($err)->as_html;
-		$res .= "\n\nBad query: <b>$err</b>\n";
-		$res .= qq{See <a\nhref="$u">$u</a> for Xapian query syntax};
-		$res .= "</pre><hr /><pre>$foot";
+		$res .= err_txt($err) . "</pre><hr /><pre>$foot";
 	} elsif ($total == 0) {
 		$code = 404;
 		$res .= "\n\n[No results found]</pre><hr /><pre>$foot";
 	} else {
 		$q = $query->as_href;
 		$q =~ s/%20/+/g; # improve URL readability
-		my $qp = "?q=$q";
-		$qp .= "&amp;o=$o" if $o;
+		$res .= search_nav_top($q, $o, $r);
+		$res .= "\n\n";
 
-		$res .= "Search results ordered by [";
-		if ($r) {
-			$res .= qq{<a\nhref="$qp">date</a>|<b>relevance</b>};
-		} else {
-			$qp .= '&amp;r';
-			$res .= qq{<b>date</b>|<a\nhref="$qp">relevance</a>};
-		}
-		$res .= "]\n\n";
-
-		dump_mset(\$res, $mset);
-		my $nr = scalar $mset->items;
-		my $end = $o + $nr;
-		my $beg = $o + 1;
-		$res .= "<hr /><pre>";
-		$res .= "Results $beg-$end of $total";
-
-		my $n = $o + $LIM;
-		if ($n < $total) {
-			$qp = "q=$q&amp;o=$n";
-			$qp .= "&amp;r" if $r;
-			$res .= qq{, <a\nhref="?$qp">next</a>}
-		}
-		if ($o > 0) {
-			$res .= $n < $total ? '/' : ',      ';
-			my $p = $o - $LIM;
-			$qp = "q=$q";
-			$qp .= "&amp;o=$p" if $p > 0;
-			$qp .= "&amp;r" if $r;
-			$res .= qq{<a\nhref="?$qp">prev</a>};
-		}
+		dump_mset(\$res, $mset, $o);
+		$res .= search_nav_bot($mset, $q, $o, $r);
 		$res .= "\n\n" . $foot;
 	}
 
@@ -112,4 +79,54 @@ sub dump_mset {
 	}
 }
 
+sub err_txt {
+	my ($err) = @_;
+	my $u = 'http://xapian.org/docs/queryparser.html';
+	$err =~ s/^\s*Exception:\s*//; # bad word to show users :P
+	$err = PublicInbox::Hval->new_oneline($err)->as_html;
+	"\n\nBad query: <b>$err</b>\n" .
+		qq{See <a\nhref="$u">$u</a> for Xapian query syntax};
+}
+
+sub search_nav_top {
+	my ($q, $o, $r) = @_;
+	my $qs = "q=$q";
+	$qs .= "&amp;o=$o" if $o;
+
+	my $rv = "Search results ordered by [";
+	if ($r) {
+		$rv .= qq{<a\nhref="?$qs">date</a>|<b>relevance</b>};
+	} else {
+		$qs .= '&amp;r';
+		$rv .= qq{<b>date</b>|<a\nhref="?$qs">relevance</a>};
+	}
+	$rv .= ']';
+}
+
+sub search_nav_bot {
+	my ($mset, $q, $o, $r) = @_;
+	my $total = $mset->get_matches_estimated;
+	my $nr = scalar $mset->items;
+	my $end = $o + $nr;
+	my $beg = $o + 1;
+
+	my $rv = "<hr /><pre>Results $beg-$end of $total";
+
+	my $n = $o + $LIM;
+	if ($n < $total) {
+		my $qs = "q=$q&amp;o=$n";
+		$qs .= "&amp;r" if $r;
+		$rv .= qq{, <a\nhref="?$qs">next</a>}
+	}
+	if ($o > 0) {
+		$rv .= $n < $total ? '/' : ',      ';
+		my $p = $o - $LIM;
+		my $qs = "q=$q";
+		$qs .= "&amp;o=$p" if $p > 0;
+		$qs .= "&amp;r" if $r;
+		$rv .= qq{<a\nhref="?$qs">prev</a>};
+	}
+	$rv;
+}
+
 1;
-- 
EW

[PATCH 3/3] searchview: support displaying entire threads

[Eric Wong]

This hopefully makes it easy to perform queries to display
an entire thread.  Raise the limit in the threaded view to
display more results and hopefully improve the output of
thread display.
---
 lib/PublicInbox/SearchView.pm | 239 ++++++++++++++++++++++++++++++++++--------
 lib/PublicInbox/View.pm       |  19 ++--
 lib/PublicInbox/WWW.pm        |   5 +-
 3 files changed, 208 insertions(+), 55 deletions(-)

diff --git a/lib/PublicInbox/SearchView.pm b/lib/PublicInbox/SearchView.pm
index 82b97f3..c15a046 100644
--- a/lib/PublicInbox/SearchView.pm
+++ b/lib/PublicInbox/SearchView.pm
@@ -6,52 +6,49 @@ use warnings;
 use PublicInbox::SearchMsg;
 use PublicInbox::Hval;
 use PublicInbox::View;
+use PublicInbox::MID qw(mid2path mid_clean);
+use Email::MIME;
 use POSIX qw/strftime/;
-our $LIM = 25;
+our $LIM = 50;
 
 sub sres_top_html {
-	my ($ctx, $q) = @_;
-	my $cgi = $ctx->{cgi};
+	my ($ctx) = @_;
+	my $q = PublicInbox::SearchQuery->new($ctx->{cgi});
 	my $code = 200;
-	# $q ||= $cgi->param('q');
-	my $o = int($cgi->param('o') || 0);
-	my $r = $cgi->param('r');
-	$r = (defined $r && $r ne '0');
-	my $opts = { limit => $LIM, offset => $o, mset => 1, relevance => $r };
+
+	# double the limit for expanded views:
+	my $opts = {
+		limit => $LIM,
+		offset => $q->{o},
+		mset => 1,
+		relevance => $q->{r},
+	};
 	my ($mset, $total);
+
 	eval {
-		$mset = $ctx->{srch}->query($q, $opts);
+		$mset = $ctx->{srch}->query($q->{q}, $opts);
 		$total = $mset->get_matches_estimated;
 	};
 	my $err = $@;
-	my $query = PublicInbox::Hval->new_oneline($q);
-	my $qh = $query->as_html;
-	my $res = "<html><head><title>$qh - search results</title></head>" .
-		  qq{<body><form\naction="">} .
-		  qq{<input\nname=q\nvalue="$qh"\ntype=text />};
-
-	$res .= qq{<input\ntype=hidden\nname=r />} if $r;
-
-	$res .= qq{<input\ntype=submit\nvalue=search /></form>} .
-		  PublicInbox::View::PRE_WRAP;
-
-	my $foot = $ctx->{footer} || '';
-	$foot = qq{Back to <a\nhref=".">index</a>.};
+	my $res = html_start($q) . PublicInbox::View::PRE_WRAP;
 	if ($err) {
 		$code = 400;
-		$res .= err_txt($err) . "</pre><hr /><pre>$foot";
+		$res .= err_txt($err) . "</pre><hr /><pre>" . foot($ctx);
 	} elsif ($total == 0) {
 		$code = 404;
-		$res .= "\n\n[No results found]</pre><hr /><pre>$foot";
+		$res .= "\n\n[No results found]</pre><hr /><pre>".foot($ctx);
 	} else {
-		$q = $query->as_href;
-		$q =~ s/%20/+/g; # improve URL readability
-		$res .= search_nav_top($q, $o, $r);
-		$res .= "\n\n";
+		my $x = $q->{x};
+		# TODO
+		#return sub { adump($_[0], $mset, $q, $ctx) } if ($x eq 'A');
 
-		dump_mset(\$res, $mset, $o);
-		$res .= search_nav_bot($mset, $q, $o, $r);
-		$res .= "\n\n" . $foot;
+		$res .= search_nav_top($mset, $q);
+		if ($x eq 't') {
+			return sub { tdump($_[0], $res, $mset, $q, $ctx) };
+		}
+		$res .= "\n\n";
+		dump_mset(\$res, $mset);
+		$res .= search_nav_bot($mset, $q) . "\n\n" . foot($ctx);
 	}
 
 	$res .= "</pre></body></html>";
@@ -89,44 +86,194 @@ sub err_txt {
 }
 
 sub search_nav_top {
-	my ($q, $o, $r) = @_;
-	my $qs = "q=$q";
-	$qs .= "&amp;o=$o" if $o;
+	my ($mset, $q) = @_;
 
 	my $rv = "Search results ordered by [";
-	if ($r) {
-		$rv .= qq{<a\nhref="?$qs">date</a>|<b>relevance</b>};
+	if ($q->{r}) {
+		my $d = $q->qs_html(r => 0);
+		$rv .= qq{<a\nhref="?$d">date</a>|<b>relevance</b>};
 	} else {
-		$qs .= '&amp;r';
-		$rv .= qq{<b>date</b>|<a\nhref="?$qs">relevance</a>};
+		my $d = $q->qs_html(r => 1);
+		$rv .= qq{<b>date</b>|<a\nhref="?$d">relevance</a>};
+	}
+
+	$rv .= ']  view[';
+
+	my $x = $q->{x};
+	if ($x eq '') {
+		my $t = $q->qs_html(x => 't');
+		$rv .= qq{<b>summary</b>|};
+		$rv .= qq{<a\nhref="?$t">threaded</a>}
+	} elsif ($q->{x} eq 't') {
+		my $s = $q->qs_html(x => '');
+		$rv .= qq{<a\nhref="?$s">summary</a>|};
+		$rv .= qq{<b>threaded</b>};
 	}
+	# my $A = $q->qs_html(x => 'a');
+	# $rv .= qq{|<a\nhref="?$A">Atom</a>}; # TODO
 	$rv .= ']';
 }
 
 sub search_nav_bot {
-	my ($mset, $q, $o, $r) = @_;
+	my ($mset, $q) = @_;
 	my $total = $mset->get_matches_estimated;
 	my $nr = scalar $mset->items;
+	my $o = $q->{o};
 	my $end = $o + $nr;
 	my $beg = $o + 1;
-
 	my $rv = "<hr /><pre>Results $beg-$end of $total";
-
 	my $n = $o + $LIM;
+
 	if ($n < $total) {
-		my $qs = "q=$q&amp;o=$n";
-		$qs .= "&amp;r" if $r;
+		my $qs = $q->qs_html(o => $n);
 		$rv .= qq{, <a\nhref="?$qs">next</a>}
 	}
 	if ($o > 0) {
 		$rv .= $n < $total ? '/' : ',      ';
 		my $p = $o - $LIM;
-		my $qs = "q=$q";
-		$qs .= "&amp;o=$p" if $p > 0;
-		$qs .= "&amp;r" if $r;
+		my $qs = $q->qs_html(o => ($p > 0 ? $p : 0));
 		$rv .= qq{<a\nhref="?$qs">prev</a>};
 	}
 	$rv;
 }
 
+sub tdump {
+	my ($cb, $res, $mset, $q, $ctx) = @_;
+	my $fh = $cb->([200, ['Content-Type'=>'text/html; charset=UTF-8']]);
+	$fh->write($res);
+	my %pct;
+	my @m = map {
+		my $i = $_;
+		my $m = PublicInbox::SearchMsg->load_doc($i->get_document);
+		$pct{$m->mid} = $i->get_percent;
+		$m = $m->mini_mime;
+		$m;
+	} ($mset->items);
+
+	require PublicInbox::Thread;
+	my $th = PublicInbox::Thread->new(@m);
+	{
+		no warnings 'once';
+		$Mail::Thread::nosubject = 0;
+	}
+	$th->thread;
+	if ($q->{r}) {
+		$th->order(sub {
+			sort { (eval { $pct{$b->topmost->messageid} } || 0)
+					<=>
+				(eval { $pct{$a->topmost->messageid} } || 0)
+			} @_;
+		});
+	} else {
+		no warnings 'once';
+		$th->order(*PublicInbox::View::rsort_ts);
+	}
+
+	require PublicInbox::GitCatFile;
+	my $git = PublicInbox::GitCatFile->new($ctx->{git_dir});
+	my $state = { ctx => $ctx, anchor_idx => 0, pct => \%pct };
+	$ctx->{searchview} = 1;
+	tdump_ent($fh, $git, $state, $_, 0) for $th->rootset;
+	$git = undef;
+	Email::Address->purge_cache;
+
+	$fh->write(search_nav_bot($mset, $q). "\n\n" .
+			foot($ctx). '</pre></body></html>');
+
+	$fh->close;
+}
+
+sub tdump_ent {
+	my ($fh, $git, $state, $node, $level) = @_;
+	return unless $node;
+	my $mime = $node->message;
+
+	if ($mime) {
+		# lazy load the full message from mini_mime:
+		my $mid = $mime->header('Message-ID');
+		$mime = eval {
+			my $path = mid2path(mid_clean($mid));
+			Email::MIME->new($git->cat_file('HEAD:'.$path));
+		};
+	}
+	if ($mime) {
+		PublicInbox::View::index_entry($fh, $mime, $level, $state);
+	} else {
+		my $mid = $node->messageid;
+		$fh->write(PublicInbox::View::ghost_table('', $mid, $level));
+	}
+	tdump_ent($fh, $git, $state, $node->child, $level + 1);
+	tdump_ent($fh, $git, $state, $node->next, $level);
+}
+
+sub foot {
+	my ($ctx) = @_;
+	my $foot = $ctx->{footer} || '';
+	qq{Back to <a\nhref=".">index</a>.\n$foot};
+}
+
+sub html_start {
+	my ($q) = @_;
+	my $query = PublicInbox::Hval->new_oneline($q->{q});
+
+	my $qh = $query->as_html;
+	my $res = "<html><head><title>$qh - search results</title></head>" .
+		  qq{<body><form\naction="">} .
+		  qq{<input\nname=q\nvalue="$qh"\ntype=text />};
+
+	$res .= qq{<input\ntype=hidden\nname=r />} if $q->{r};
+	if (my $x = $q->{x}) {
+		my $xh = PublicInbox::Hval->new_oneline($x)->as_html;
+		$res .= qq{<input\ntype=hidden\nname=x\nvalue="$xh" />};
+	}
+
+	$res .= qq{<input\ntype=submit\nvalue=search /></form>};
+}
+
+package PublicInbox::SearchQuery;
+use strict;
+use warnings;
+use fields qw(q o t x r);
+use PublicInbox::Hval;
+
+sub new {
+	my ($class, $cgi) = @_;
+	my $self = fields::new($class);
+	$self->{q} = $cgi->param('q');
+	$self->{x} = $cgi->param('x') || '';
+	$self->{o} = int($cgi->param('o') || 0) || 0;
+	my $r = $cgi->param('r');
+	$self->{r} = (defined $r && $r ne '0');
+
+	$self;
+}
+
+sub qs_html {
+	my ($self, %over) = @_;
+
+	if (keys %over) {
+		my $tmp = fields::new(ref($self));
+		%$tmp = %$self;
+		foreach my $k (keys %over) {
+			$tmp->{$k} = $over{$k};
+		}
+		$self = $tmp;
+	}
+
+	my $q = PublicInbox::Hval->new($self->{q})->as_href;
+	$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 $r = $self->{r}) {
+		$qs .= "&amp;r";
+	}
+	if (my $x = $self->{x}) {
+		$qs .= "&amp;x=$x" if ($x eq 't' || $x eq 'A');
+	}
+	$qs;
+}
+
 1;
diff --git a/lib/PublicInbox/View.pm b/lib/PublicInbox/View.pm
index 7d85555..de2d667 100644
--- a/lib/PublicInbox/View.pm
+++ b/lib/PublicInbox/View.pm
@@ -72,7 +72,7 @@ sub index_entry {
 	my $subj = $mime->header('Subject');
 	my $header_obj = $mime->header_obj;
 
-	my $mid_raw = $header_obj->header('Message-ID');
+	my $mid_raw = mid_clean($header_obj->header('Message-ID'));
 	my $id = anchor_for($mid_raw);
 	my $seen = $state->{seen};
 	$seen->{$id} = "#$id"; # save the anchor for children, later
@@ -139,7 +139,9 @@ sub index_entry {
 		}
 		$rv .= " <a\nhref=\"$parent_anchor\">parent</a>";
 	}
-	if ($srch) {
+	if (my $pct = $state->{pct}) {
+		$rv .= " [$pct->{$mid_raw}%]";
+	} elsif ($srch) {
 		if ($ctx->{flat}) {
 			$rv .= " [<a\nhref=\"${path}$href/t/#u\">threaded</a>" .
 				"|<b>flat</b>]";
@@ -601,6 +603,14 @@ sub ghost_parent {
 	qq{[parent not found: &lt;<a\nhref="$upfx$href/">$html</a>&gt;]};
 }
 
+sub ghost_table {
+	my ($upfx, $mid, $level) = @_;
+	"<table\nsummary=ghost><tr><td>" .
+		(INDENT x $level) . "</td><td>" .
+		PRE_WRAP . ghost_parent($upfx, $mid) .
+		'</pre></td></table>';
+}
+
 sub __thread_entry {
 	my ($cb, $git, $state, $mime, $level) = @_;
 
@@ -617,10 +627,7 @@ sub __thread_entry {
 	if (my $ghost = delete $state->{ghost}) {
 		# n.b. ghost messages may only be parents, not children
 		foreach my $g (@$ghost) {
-			$$cb->write("<table\nsummary=ghost><tr><td>" .
-				(INDENT x $g->[1]) . "</td><td>" .
-				PRE_WRAP . ghost_parent('../../', $g->[0]) .
-				'</pre></td></table>');
+			$$cb->write(ghost_table('../../', @$g));
 		}
 	}
 	index_entry($$cb, $mime, $level, $state);
diff --git a/lib/PublicInbox/WWW.pm b/lib/PublicInbox/WWW.pm
index 508abf7..58a4256 100644
--- a/lib/PublicInbox/WWW.pm
+++ b/lib/PublicInbox/WWW.pm
@@ -130,11 +130,10 @@ sub get_index {
 	my ($ctx) = @_;
 	require PublicInbox::Feed;
 	my $srch = searcher($ctx);
-	my $q = $ctx->{cgi}->param('q');
 	footer($ctx);
-	if (defined $q) {
+	if (defined $ctx->{cgi}->param('q')) {
 		require PublicInbox::SearchView;
-		PublicInbox::SearchView::sres_top_html($ctx, $q);
+		PublicInbox::SearchView::sres_top_html($ctx);
 	} else {
 		PublicInbox::Feed::generate_html_index($ctx);
 	}
-- 
EW