X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;ds=sidebyside;f=lib%2FPublicInbox%2FSearchView.pm;h=cec87c6ad1c84c891fe41d4eac62ec556e6b3a97;hb=4ba430497bfb4763691b0ad8af573eb6ff420b96;hp=82b97f32ef6aecdf54c18f8f436cab2437d233db;hpb=3fcca93c1391c570e71a81f9467b95d7fcd1910e;p=public-inbox.git diff --git a/lib/PublicInbox/SearchView.pm b/lib/PublicInbox/SearchView.pm index 82b97f32..cec87c6a 100644 --- a/lib/PublicInbox/SearchView.pm +++ b/lib/PublicInbox/SearchView.pm @@ -1,132 +1,299 @@ # Copyright (C) 2015 all contributors # License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) +# +# Displays search results for the web interface package PublicInbox::SearchView; use strict; use warnings; use PublicInbox::SearchMsg; -use PublicInbox::Hval; +use PublicInbox::Hval qw/ascii_html/; use PublicInbox::View; -use POSIX qw/strftime/; -our $LIM = 25; +use PublicInbox::WwwAtomStream; +use PublicInbox::MID qw(mid2path mid_mime mid_clean mid_escape); +use PublicInbox::MIME; +require PublicInbox::Git; +require PublicInbox::SearchThread; +our $LIM = 200; + +sub noop {} sub sres_top_html { - my ($ctx, $q) = @_; - my $cgi = $ctx->{cgi}; + my ($ctx) = @_; + my $q = PublicInbox::SearchQuery->new($ctx->{qp}); 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 = "
$foot"; + $ctx->{-html_tip} = ''.err_txt($ctx, $err).'
'; + $cb = *noop; } elsif ($total == 0) { $code = 404; - $res .= "\n\n[No results found]
$foot"; + $ctx->{-html_tip} = ""; - [$code, ['Content-Type'=>'text/html; charset=UTF-8'], [$res]]; + eval { + $srch->retry_reopen(sub { + PublicInbox::SearchMsg->load_doc($mitem->get_document) + }); + } } -sub dump_mset { - my ($res, $mset) = @_; +# display non-threaded search results similar to what users expect from +# regular WWW search engines: +sub mset_summary { + my ($ctx, $mset, $q) = @_; my $total = $mset->get_matches_estimated; my $pad = length("$total"); my $pfx = ' ' x $pad; + my $res = \($ctx->{-html_tip}); + my $srch = $ctx->{srch}; 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 $s = PublicInbox::Hval->new_oneline($smsg->subject); - my $f = $smsg->from_name; - $f = PublicInbox::Hval->new_oneline($f)->as_html; - my $d = strftime('%Y-%m-%d %H:%M', gmtime($smsg->ts)); - my $mid = PublicInbox::Hval->new_msgid($smsg->mid)->as_href; + 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); + my $mid = PublicInbox::Hval->new_msgid($smsg->mid)->{href}; $$res .= qq{$rank. }. - $s->as_html . "\n"; - $$res .= "$pfx - by $f @ $d UTC [$pct%]\n\n"; + $s . "\n"; + $$res .= "$pfx - by $f @ $ts UTC [$pct%]\n\n"; } + $$res .= search_nav_bot($mset, $q); + *noop; } sub err_txt { - my ($err) = @_; - my $u = 'http://xapian.org/docs/queryparser.html'; + my ($ctx, $err) = @_; + my $u = $ctx->{-inbox}->base_url($ctx->{env}) . '_/text/help/'; $err =~ s/^\s*Exception:\s*//; # bad word to show users :P - $err = PublicInbox::Hval->new_oneline($err)->as_html; - "\n\nBad query: $err\n" . - qq{See $u for Xapian query syntax}; + $err = ascii_html($err); + "\nBad query: $err\n" . + qq{See $u for help on using search}; } sub search_nav_top { - my ($q, $o, $r) = @_; - my $qs = "q=$q"; - $qs .= "&o=$o" if $o; + my ($mset, $q) = @_; - my $rv = "Search results ordered by ["; - if ($r) { - $rv .= qq{date|relevance}; + my $rv = "\n[No results found]
"; + $cb = *noop; } else { - $q = $query->as_href; - $q =~ s/%20/+/g; # improve URL readability - $res .= search_nav_top($q, $o, $r); - $res .= "\n\n"; - - dump_mset(\$res, $mset, $o); - $res .= search_nav_bot($mset, $q, $o, $r); - $res .= "\n\n" . $foot; + my $x = $q->{x}; + return adump($_[0], $mset, $q, $ctx) if $x eq 'A'; + + $ctx->{-html_tip} = search_nav_top($mset, $q) . "\n\n"; + if ($x eq 't') { + $cb = mset_thread($ctx, $mset, $q); + } else { + $cb = mset_summary($ctx, $mset, $q); + } } + PublicInbox::WwwStream->response($ctx, $code, $cb); +} + +# allow undef for individual doc loads... +sub load_doc_retry { + my ($srch, $mitem) = @_; - $res .= "
Search results ordered by ["; + if ($q->{r}) { + my $d = $q->qs_html(r => 0); + $rv .= qq{date|relevance}; } else { - $qs .= '&r'; - $rv .= qq{date|relevance}; + my $d = $q->qs_html(r => 1); + $rv .= qq{date|relevance}; } - $rv .= ']'; + + $rv .= '] view['; + + my $x = $q->{x}; + if ($x eq '') { + my $t = $q->qs_html(x => 't'); + $rv .= qq{summary|threaded} + } elsif ($q->{x} eq 't') { + my $s = $q->qs_html(x => ''); + $rv .= qq{summary|threaded}; + } + my $A = $q->qs_html(x => 'A', r => undef); + $rv .= qq{|Atom feed]}; } sub search_nav_bot { - my ($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 = "Results $beg-$end of $total"; - + my $rv = ''; + if ($beg <= $end) { + $rv .= "Results $beg-$end of $total"; + $rv .= ' (estimated)' if $end != $total; + } else { + $rv .= "No more results, only $total"; + } my $n = $o + $LIM; + if ($n < $total) { - my $qs = "q=$q&o=$n"; - $qs .= "&r" if $r; - $rv .= qq{, next} + my $qs = $q->qs_html(o => $n); + $rv .= qq{ next} } if ($o > 0) { - $rv .= $n < $total ? '/' : ', '; + $rv .= $n < $total ? '/' : ' '; my $p = $o - $LIM; - my $qs = "q=$q"; - $qs .= "&o=$p" if $p > 0; - $qs .= "&r" if $r; - $rv .= qq{prev}; + my $qs = $q->qs_html(o => ($p > 0 ? $p : 0)); + $rv .= qq{prev}; + } + $rv .= ''; +} + +sub sort_relevance { + my ($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 $msgs = $ctx->{srch}->retry_reopen(sub { [ map { + my $i = $_; + 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_ts); + my $skel = search_nav_bot($mset, $q). ""; + my $inbox = $ctx->{-inbox}; + $ctx->{-upfx} = ''; + $ctx->{anchor_idx} = 1; + $ctx->{cur_level} = 0; + $ctx->{dst} = \$skel; + $ctx->{mapping} = {}; + $ctx->{pct} = \%pct; + $ctx->{prev_attr} = ''; + $ctx->{prev_level} = 0; + $ctx->{seen} = {}; + $ctx->{s_nr} = scalar(@$msgs).'+ results'; + + PublicInbox::View::walk_thread($rootset, $ctx, + *PublicInbox::View::pre_thread); + @$msgs = reverse @$msgs if $r; + my $mime; + sub { + return unless $msgs; + while ($mime = pop @$msgs) { + $mime = $inbox->msg_by_smsg($mime) and last; + } + if ($mime) { + $mime = PublicInbox::MIME->new($mime); + return PublicInbox::View::index_entry($mime, $ctx, + scalar @$msgs); + } + $msgs = undef; + $skel .= "\n"; + }; +} + +sub ctx_prepare { + my ($q, $ctx) = @_; + my $qh = ascii_html($q->{'q'}); + $ctx->{-q_value_html} = $qh; + $ctx->{-atom} = '?'.$q->qs_html(x => 'A', r => undef); + $ctx->{-title_html} = "$qh - search results"; + my $extra = ''; + $extra .= qq{} if $q->{r}; + if (my $x = $q->{x}) { + $x = ascii_html($x); + $extra .= qq{}; + } + $ctx->{-extra_form_html} = $extra; +} + +sub adump { + my ($cb, $mset, $q, $ctx) = @_; + my $ibx = $ctx->{-inbox}; + 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 = $ibx->msg_by_smsg($x) and + return PublicInbox::MIME->new($x); + } + return undef; + }); +} + +package PublicInbox::SearchQuery; +use strict; +use warnings; +use PublicInbox::Hval; +use PublicInbox::MID qw(mid_escape); + +sub new { + my ($class, $qp) = @_; + + my $r = $qp->{r}; + bless { + q => $qp->{'q'}, + x => $qp->{x} || '', + o => (($qp->{o} || '0') =~ /(\d+)/), + r => (defined $r && $r ne '0'), + }, $class; +} + +sub qs_html { + my ($self, %over) = @_; + + if (keys %over) { + my $tmp = bless { %$self }, ref($self); + foreach my $k (keys %over) { + $tmp->{$k} = $over{$k}; + } + $self = $tmp; + } + + my $q = mid_escape($self->{'q'}); + $q =~ s/%20/+/g; # improve URL readability + my $qs = "q=$q"; + + if (my $o = $self->{o}) { # ignore o == 0 + $qs .= "&o=$o"; + } + if (my $r = $self->{r}) { + $qs .= "&r"; + } + if (my $x = $self->{x}) { + $qs .= "&x=$x" if ($x eq 't' || $x eq 'A'); } - $rv; + $qs; } 1;