# 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;
# This is English-only, everything else is non-standard and may be confused as
# a prefix common in patch emails
-our $REPLY_RE = qr/^re:\s+/i;
our $LANG = 'english';
use constant {
my %bool_pfx_external = (
mid => 'Q', # Message-ID (full/exact), this is mostly uniQue
+ dfpre => 'XDFPRE',
+ dfpost => 'XDFPOST',
+ dfblob => 'XDFPRE XDFPOST',
);
my $non_quoted_body = 'XNQ XDFN XDFA XDFB XDFHH XDFCTX XDFPRE XDFPOST';
dfb => 'XDFB',
dfhh => 'XDFHH',
dfctx => 'XDFCTX',
- dfpre => 'XDFPRE',
- dfpost => 'XDFPOST',
- dfblob => 'XDFPRE XDFPOST',
# default:
'' => 'XM S A XQUOT XFN ' . $non_quoted_body,
date range as YYYYMMDD e.g. d:19931002..20101002
Open-ended ranges such as d:19931002.. and d:..20101002
are also supported
+EOF
+ 'dt:' => <<EOF,
+date-time range as YYYYMMDDhhmmss (e.g. dt:19931002011000..19931002011200)
EOF
'b:' => 'match within message body, including text attachments',
'nq:' => 'match non-quoted text within message body',
}
}
-sub get_thread {
- my ($self, $mid, $opts) = @_;
- $self->{over_ro}->get_thread($mid, $opts);
-}
-
sub retry_reopen {
my ($self, $cb) = @_;
for my $i (1..10) {
sub _enquire_once {
my ($self, $query, $opts) = @_;
- my $enquire = enquire($self);
- $query = Search::Xapian::Query->new(OP_AND,$query);
+ my $enquire = Search::Xapian::Enquire->new($self->{xdb});
$enquire->set_query($query);
$opts ||= {};
my $desc = !$opts->{asc};
- if ($opts->{relevance}) {
+ if (($opts->{mset} || 0) == 2) {
+ $enquire->set_docid_order(Search::Xapian::ENQ_ASCENDING());
+ $enquire->set_weighting_scheme(Search::Xapian::BoolWeight->new);
+ } elsif ($opts->{relevance}) {
$enquire->set_sort_by_relevance_then_value(TS, $desc);
} else {
$enquire->set_sort_by_value_then_relevance(TS, $desc);
$qp->set_database($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);
+ $qp->add_boolean_prefix($name, $_) foreach split(/ /, $prefix);
}
# we do not actually create AltId objects,
$self->{query_parser} = $qp;
}
-# only used for NNTP server
-sub query_xover {
- my ($self, $beg, $end, $offset) = @_;
- $self->{over_ro}->query_xover($beg, $end, $offset);
-}
-
-sub query_ts {
- my ($self, $ts, $offset) = @_;
- $self->{over_ro}->query_ts($ts, $offset);
-}
-
-sub first_smsg_by_mid {
- my ($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;
- $smsg->load_expand;
- });
-}
-
-sub each_smsg_by_mid {
- my ($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;
- }
-}
-
-# normalize subjects so they are suitable as pathnames for URLs
-# XXX: consider for removal
-sub subject_path {
- my $subj = pop;
- $subj = subject_normalized($subj);
- $subj =~ s![^a-zA-Z0-9_\.~/\-]+!_!g;
- lc($subj);
-}
-
-sub subject_normalized {
- my $subj = pop;
- $subj =~ s/\A\s+//s; # no leading space
- $subj =~ s/\s+\z//s; # no trailing space
- $subj =~ s/\s+/ /gs; # no redundant spaces
- $subj =~ s/\.+\z//; # no trailing '.'
- $subj =~ s/$REPLY_RE//igo; # remove reply prefix
- $subj;
-}
-
-sub enquire {
- my ($self) = @_;
- $self->{enquire} ||= Search::Xapian::Enquire->new($self->{xdb});
+ $self->{over_ro}->get_art($num);
}
sub help {