extmsg: use Xapian only for partial matches

[public-inbox.git] / lib / PublicInbox / Search.pm

diff --git a/lib/PublicInbox/Search.pm b/lib/PublicInbox/Search.pm
index f7fdf85478cfd387a728b935ecf5d212df35be3b..5aabda02f4a6b3c808b941da19132c4231589055 100644 (file)
--- a/lib/PublicInbox/Search.pm
+++ b/lib/PublicInbox/Search.pm
@@ -9,7 +9,8 @@ use warnings;
 
 # values for searching
 use constant TS => 0;  # Received: header in Unix time
-use constant YYYYMMDD => 1; # for searching in the WWW UI
+use constant YYYYMMDD => 1; # Date: header for searching in the WWW UI
+use constant DT => 2; # Date: YYYYMMDDHHMMSS
 
 use Search::Xapian qw/:standard/;
 use PublicInbox::SearchMsg;
@@ -88,6 +89,9 @@ our @HELP = (
 date range as YYYYMMDD  e.g. d:19931002..20101002
 Open-ended ranges such as d:19931002.. and d:..20101002
 are also supported
+EOF
+       'dt:' => <<EOF,
+date-time range as YYYYMMDDhhmmss (e.g. dt:19931002011000..19931002011200)
 EOF
        'b:' => 'match within message body, including text attachments',
        'nq:' => 'match non-quoted text within message body',
@@ -179,8 +183,8 @@ sub query {
 }
 
 sub get_thread {
-       my ($self, $mid, $opts) = @_;
-       $self->{over_ro}->get_thread($mid, $opts);
+       my ($self, $mid, $prev) = @_;
+       $self->{over_ro}->get_thread($mid, $prev);
 }
 
 sub retry_reopen {
@@ -216,11 +220,14 @@ sub _do_enquire {
 sub _enquire_once {
        my ($self, $query, $opts) = @_;
        my $enquire = enquire($self);
-       $query = Search::Xapian::Query->new(OP_AND,$query);
        $enquire->set_query($query);
        $opts ||= {};
         my $desc = !$opts->{asc};
-       if ($opts->{relevance}) {
+       if (($opts->{mset} || 0) == 2) {
+               $enquire->set_docid_order(Search::Xapian::ENQ_ASCENDING());
+               $enquire->set_weighting_scheme(Search::Xapian::BoolWeight->new);
+               delete $self->{enquire};
+       } elsif ($opts->{relevance}) {
                $enquire->set_sort_by_relevance_then_value(TS, $desc);
        } else {
                $enquire->set_sort_by_value_then_relevance(TS, $desc);
@@ -253,8 +260,11 @@ sub qp {
        $qp->set_database($self->{xdb});
        $qp->set_stemmer($self->stemmer);
        $qp->set_stemming_strategy(STEM_SOME);
+       $qp->set_max_wildcard_expansion(100);
        $qp->add_valuerangeprocessor(
                Search::Xapian::NumberValueRangeProcessor->new(YYYYMMDD, 'd:'));
+       $qp->add_valuerangeprocessor(
+               Search::Xapian::NumberValueRangeProcessor->new(DT, 'dt:'));
 
        while (my ($name, $prefix) = each %bool_pfx_external) {
                $qp->add_boolean_prefix($name, $prefix);
@@ -295,58 +305,14 @@ sub query_ts {
        $self->{over_ro}->query_ts($ts, $prev);
 }
 
-sub first_smsg_by_mid {
-       my ($self, $mid) = @_;
-       my $smsg;
-       retry_reopen($self, sub {
-               each_smsg_by_mid($self, $mid, sub { $smsg = $_[0]; undef });
-       });
-       $smsg;
-}
-
 sub lookup_article {
        my ($self, $num) = @_;
-       my $term = 'XNUM'.$num;
-       my $db = $self->{xdb};
-       retry_reopen($self, sub {
-               my $head = $db->postlist_begin($term);
-               my $tail = $db->postlist_end($term);
-               return if $head->equal($tail);
-               my $doc_id = $head->get_docid;
-               return unless defined $doc_id;
-               $head->inc;
-               if ($head->nequal($tail)) {
-                       warn "article #$num is not unique\n";
-               }
-               # raises on error:
-               my $doc = $db->get_document($doc_id);
-               my $smsg = PublicInbox::SearchMsg->wrap($doc);
-               $smsg->{doc_id} = $doc_id;
-               $smsg->load_expand;
-       });
+       $self->{over_ro}->get_art($num);
 }
 
-sub each_smsg_by_mid {
-       my ($self, $mid, $cb) = @_;
-       # XXX retry_reopen isn't necessary for V2Writable, but the PSGI
-       # interface will need it...
-       my $db = $self->{xdb};
-       my $term = 'Q' . $mid;
-       my $head = $db->postlist_begin($term);
-       my $tail = $db->postlist_end($term);
-       if ($head == $tail) {
-               $db->reopen;
-               $head = $db->postlist_begin($term);
-               $tail = $db->postlist_end($term);
-       }
-       return ($head, $tail, $db) if wantarray;
-       for (; $head->nequal($tail); $head->inc) {
-               my $doc_id = $head->get_docid;
-               my $doc = $db->get_document($doc_id);
-               my $smsg = PublicInbox::SearchMsg->wrap($doc, $mid);
-               $smsg->{doc_id} = $doc_id;
-               $cb->($smsg) or return;
-       }
+sub next_by_mid {
+       my ($self, $mid, $id, $prev) = @_;
+       $self->{over_ro}->next_by_mid($mid, $id, $prev);
 }
 
 # normalize subjects so they are suitable as pathnames for URLs