diff options
Diffstat (limited to 'lib/PublicInbox/SearchView.pm')
| -rw-r--r-- | lib/PublicInbox/SearchView.pm | 257 |
1 files changed, 123 insertions, 134 deletions
diff --git a/lib/PublicInbox/SearchView.pm b/lib/PublicInbox/SearchView.pm index ab0ff19a..50a2c01c 100644 --- a/lib/PublicInbox/SearchView.pm +++ b/lib/PublicInbox/SearchView.pm @@ -8,14 +8,18 @@ use warnings; use PublicInbox::SearchMsg; use PublicInbox::Hval qw/ascii_html/; use PublicInbox::View; -use PublicInbox::MID qw(mid2path mid_clean mid_mime); +use PublicInbox::WwwAtomStream; +use PublicInbox::MID qw(mid2path mid_mime mid_clean mid_escape); use Email::MIME; require PublicInbox::Git; +require PublicInbox::SearchThread; our $LIM = 50; +sub noop {} + sub sres_top_html { my ($ctx) = @_; - my $q = PublicInbox::SearchQuery->new($ctx->{cgi}); + my $q = PublicInbox::SearchQuery->new($ctx->{qp}); my $code = 200; # double the limit for expanded views: @@ -26,72 +30,92 @@ sub sres_top_html { relevance => $q->{r}, }; my ($mset, $total); - eval { - $mset = $ctx->{srch}->query($q->{q}, $opts); + $mset = $ctx->{srch}->query($q->{'q'}, $opts); $total = $mset->get_matches_estimated; }; my $err = $@; - my $res = html_start($q, $ctx) . '<pre>'; + ctx_prepare($q, $ctx); + my $cb; if ($err) { $code = 400; - $res .= err_txt($ctx, $err) . "</pre><hr /><pre>" . foot($ctx); + $ctx->{-html_tip} = '<pre>'.err_txt($ctx, $err).'</pre><hr>'; + $cb = *noop; } elsif ($total == 0) { $code = 404; - $res .= "\n\n[No results found]</pre><hr /><pre>".foot($ctx); + $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'; - $res .= search_nav_top($mset, $q) . "\n\n"; + $ctx->{-html_tip} = search_nav_top($mset, $q) . "\n\n"; if ($x eq 't') { - return sub { tdump($_[0], $res, $mset, $q, $ctx) }; + $cb = mset_thread($ctx, $mset, $q); + } else { + $cb = mset_summary($ctx, $mset, $q); } - dump_mset(\$res, $mset); - $res .= '</pre>' . search_nav_bot($mset, $q) . - "\n\n" . foot($ctx); } + PublicInbox::WwwStream->response($ctx, $code, $cb); +} + +# allow undef for individual doc loads... +sub load_doc_retry { + my ($srch, $mitem) = @_; - $res .= "</pre></body></html>"; - [$code, ['Content-Type'=>'text/html; charset=UTF-8'], [$res]]; + eval { + $srch->retry_reopen(sub { + PublicInbox::SearchMsg->load_doc($mitem->get_document) + }); + } } # display non-threaded search results similar to what users expect from # regular WWW search engines: -sub dump_mset { - my ($res, $mset) = @_; +sub mset_summary { + my ($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 $smsg = load_doc_retry($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)->as_href; + 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 .= search_nav_bot($mset, $q); + *noop; } sub err_txt { my ($ctx, $err) = @_; - my $u = '//xapian.org/docs/queryparser.html'; - $u = PublicInbox::Hval::prurl($ctx->{cgi}->{env}, $u); + my $u = $ctx->{-inbox}->base_url($ctx->{env}) . '_/text/help/'; $err =~ s/^\s*Exception:\s*//; # bad word to show users :P $err = ascii_html($err); - "\n\nBad query: <b>$err</b>\n" . - qq{See <a\nhref="$u">$u</a> for Xapian query syntax}; + "\nBad query: <b>$err</b>\n" . + qq{See <a\nhref="$u">$u</a> for help on using search}; } sub search_nav_top { my ($mset, $q) = @_; - my $rv = "Search results ordered by ["; + my $rv = "<pre>Search results ordered by ["; if ($q->{r}) { my $d = $q->qs_html(r => 0); $rv .= qq{<a\nhref="?$d">date</a>|<b>relevance</b>}; @@ -121,156 +145,121 @@ sub search_nav_bot { my $o = $q->{o}; my $end = $o + $nr; my $beg = $o + 1; - my $rv = "<hr /><pre>Results $beg-$end of $total"; + my $rv = "</pre><hr><pre>Results $beg-$end of $total"; my $n = $o + $LIM; if ($n < $total) { my $qs = $q->qs_html(o => $n); - $rv .= qq{, <a\nhref="?$qs">next</a>} + $rv .= qq{, <a\nhref="?$qs"\nrel=next>next</a>} } if ($o > 0) { $rv .= $n < $total ? '/' : ', '; my $p = $o - $LIM; my $qs = $q->qs_html(o => ($p > 0 ? $p : 0)); - $rv .= qq{<a\nhref="?$qs">prev</a>}; + $rv .= qq{<a\nhref="?$qs"\nrel=prev>prev</a>}; } - $rv; + $rv .= '</pre>'; } -sub tdump { - my ($cb, $res, $mset, $q, $ctx) = @_; - my $fh = $cb->([200, ['Content-Type'=>'text/html; charset=UTF-8']]); - $fh->write($res .= '</pre>'); +sub mset_thread { + my ($ctx, $mset, $q) = @_; my %pct; - my @m = map { + my $msgs = $ctx->{srch}->retry_reopen(sub { [ 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); + my $smsg = PublicInbox::SearchMsg->load_doc($i->get_document); + $pct{$smsg->mid} = $i->get_percent; + $smsg; + } ($mset->items) ]}); - require PublicInbox::Thread; - my $th = PublicInbox::Thread->new(@m); - { - no warnings 'once'; - $Mail::Thread::nosubject = 0; - } + my $th = PublicInbox::SearchThread->new($msgs); $th->thread; - if ($q->{r}) { + if ($q->{r}) { # order by relevance $th->order(sub { - sort { (eval { $pct{$b->topmost->messageid} } || 0) + [ sort { (eval { $pct{$b->topmost->{id}} } || 0) <=> - (eval { $pct{$a->topmost->messageid} } || 0) - } @_; + (eval { $pct{$a->topmost->{id}} } || 0) + } @{$_[0]} ]; }); - } else { - no warnings 'once'; - $th->order(*PublicInbox::View::rsort_ts); + } else { # order by time (default for threaded view) + $th->order(*PublicInbox::View::sort_ts); } - - my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir}); - my $state = { - ctx => $ctx, - anchor_idx => 0, - pct => \%pct, - cur_level => 0, - fh => $fh, + my $skel = search_nav_bot($mset, $q). "<pre>"; + 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($th, $ctx, + *PublicInbox::View::pre_thread); + + my $mime; + sub { + return unless $msgs; + while ($mime = shift @$msgs) { + $mime = $inbox->msg_by_smsg($mime) and last; + } + if ($mime) { + $mime = Email::MIME->new($mime); + return PublicInbox::View::index_entry($mime, $ctx, + scalar @$msgs); + } + $msgs = undef; + $skel .= "\n</pre>"; }; - $ctx->{searchview} = 1; - tdump_ent($git, $state, $_, 0) for $th->rootset; - PublicInbox::View::thread_adj_level($state, 0); - Email::Address->purge_cache; - - $fh->write(search_nav_bot($mset, $q). "\n\n" . - foot($ctx). '</pre></body></html>'); - - $fh->close; } -sub tdump_ent { - my ($git, $state, $node, $level) = @_; - return unless $node; - my $mime = $node->message; - - if ($mime) { - # lazy load the full message from mini_mime: - my $mid = mid_mime($mime); - $mime = eval { - my $path = mid2path(mid_clean($mid)); - Email::MIME->new($git->cat_file('HEAD:'.$path)); - }; - } - if ($mime) { - my $end = PublicInbox::View::thread_adj_level($state, $level); - PublicInbox::View::index_entry($mime, $level, $state); - $state->{fh}->write($end) if $end; - } else { - my $mid = $node->messageid; - PublicInbox::View::ghost_flush($state, '', $mid, $level); - } - tdump_ent($git, $state, $node->child, $level + 1); - tdump_ent($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 { +sub ctx_prepare { my ($q, $ctx) = @_; my $qh = ascii_html($q->{'q'}); - my $A = $q->qs_html(x => 'A', r => undef); - my $res = '<html><head>' . PublicInbox::Hval::STYLE . - "<title>$qh - search results</title>" . - qq{<link\nrel=alternate\ntitle="Atom feed"\n} . - qq!href="?$A"\ntype="application/atom+xml"/></head>! . - qq{<body><form\naction="">} . - qq{<input\nname=q\nvalue="$qh"\ntype=text />}; - - $res .= qq{<input\ntype=hidden\nname=r />} if $q->{r}; + $ctx->{-q_value_html} = $qh; + $ctx->{-atom} = '?'.$q->qs_html(x => 'A', r => undef); + $ctx->{-title_html} = "$qh - search results"; + my $extra = ''; + $extra .= qq{<input\ntype=hidden\nname=r />} if $q->{r}; if (my $x = $q->{x}) { $x = ascii_html($x); - $res .= qq{<input\ntype=hidden\nname=x\nvalue="$x" />}; + $extra .= qq{<input\ntype=hidden\nname=x\nvalue="$x" />}; } - - $res .= qq{<input\ntype=submit\nvalue=search /></form>}; + $ctx->{-extra_form_html} = $extra; } sub adump { my ($cb, $mset, $q, $ctx) = @_; - my $fh = $cb->([ 200, ['Content-Type' => 'application/atom+xml']]); - my $git = $ctx->{git} ||= PublicInbox::Git->new($ctx->{git_dir}); - 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); - PublicInbox::Feed::add_to_feed($feed_opts, $fh, $x, $git); - } - PublicInbox::Feed::end_feed($fh); + 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 Email::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, $cgi) = @_; - my $r = $cgi->param('r'); + my ($class, $qp) = @_; + + my $r = $qp->{r}; bless { - q => $cgi->param('q'), - x => $cgi->param('x') || '', - o => int($cgi->param('o') || 0) || 0, + q => $qp->{'q'}, + x => $qp->{x} || '', + o => (($qp->{o} || '0') =~ /(\d+)/), r => (defined $r && $r ne '0'), }, $class; } @@ -286,7 +275,7 @@ sub qs_html { $self = $tmp; } - my $q = PublicInbox::Hval->new($self->{'q'})->as_href; + my $q = mid_escape($self->{'q'}); $q =~ s/%20/+/g; # improve URL readability my $qs = "q=$q"; |