X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearch.pm;h=cb669e8733e7cd8596d6483ed77f4bb33dd16d66;hb=8c3016372fb40ad3f19bc69c90e49cbd5a5f081f;hp=d861cf471932aacddff0915f6ede3e8ec103fd6e;hpb=792c643ba76e8b06ae17d3653402d381f59857c2;p=public-inbox.git diff --git a/lib/PublicInbox/Search.pm b/lib/PublicInbox/Search.pm index d861cf47..cb669e87 100644 --- a/lib/PublicInbox/Search.pm +++ b/lib/PublicInbox/Search.pm @@ -1,4 +1,4 @@ -# Copyright (C) 2015-2018 all contributors +# Copyright (C) 2015-2020 all contributors # License: AGPL-3.0+ # based on notmuch, but with no concept of folders, files or flags # @@ -12,25 +12,45 @@ use constant TS => 0; # Received: header in Unix time use constant YYYYMMDD => 1; # Date: header for searching in the WWW UI use constant DT => 2; # Date: YYYYMMDDHHMMSS -use PublicInbox::SearchMsg; -use PublicInbox::MIME; -use PublicInbox::MID qw/id_compress/; +use PublicInbox::Smsg; use PublicInbox::Over; my $QP_FLAGS; +our %X = map { $_ => 0 } qw(BoolWeight Database Enquire + NumberValueRangeProcessor QueryParser Stem); +our $Xap; # 'Search::Xapian' or 'Xapian' +my $ENQ_ASCENDING; + sub load_xapian () { - $QP_FLAGS ||= eval { - require Search::Xapian; - Search::Xapian->import(qw(:standard)); + return 1 if defined $Xap; + for my $x (qw(Search::Xapian Xapian)) { + eval "require $x"; + next if $@; + + $x->import(qw(:standard)); + $Xap = $x; + $X{$_} = $Xap.'::'.$_ for (keys %X); + + # ENQ_ASCENDING doesn't seem exported by SWIG Xapian.pm, + # so lets hope this part of the ABI is stable because it's + # just an integer: + $ENQ_ASCENDING = $x eq 'Xapian' ? + 1 : Search::Xapian::ENQ_ASCENDING(); + # for Smsg: + *PublicInbox::Smsg::sortable_unserialise = + $Xap.'::sortable_unserialise'; # n.b. FLAG_PURE_NOT is expensive not suitable for a public # website as it could become a denial-of-service vector # FLAG_PHRASE also seems to cause performance problems chert # (and probably earlier Xapian DBs). glass seems fine... # TODO: make this an option, maybe? # or make indexlevel=medium as default - FLAG_PHRASE()|FLAG_BOOLEAN()|FLAG_LOVEHATE()|FLAG_WILDCARD(); - }; -}; + $QP_FLAGS = FLAG_PHRASE() | FLAG_BOOLEAN() | FLAG_LOVEHATE() | + FLAG_WILDCARD(); + return 1; + } + undef; +} # This is English-only, everything else is non-standard and may be confused as # a prefix common in patch emails @@ -54,11 +74,20 @@ use constant { # 13 - fix threading for empty References/In-Reply-To # (commit 83425ef12e4b65cdcecd11ddcb38175d4a91d5a0) # 14 - fix ghost root vivification + # 15 - see public-inbox-v2-format(5) + # further bumps likely unnecessary, we'll suggest in-place + # "--reindex" use for further fixes and tweaks + # + # public-inbox v1.5.0 adds (still SCHEMA_VERSION=15): + # * "lid:" and "l:" for List-Id searches SCHEMA_VERSION => 15, }; +# note: the non-X term prefix allocations are shared with +# Xapian omega, see xapian-applications/omega/docs/termprefixes.rst my %bool_pfx_external = ( mid => 'Q', # Message-ID (full/exact), this is mostly uniQue + lid => 'G', # newsGroup (or similar entity), just inside <> dfpre => 'XDFPRE', dfpost => 'XDFPOST', dfblob => 'XDFPRE XDFPOST', @@ -69,6 +98,7 @@ my %prob_prefix = ( # for mairix compatibility s => 'S', m => 'XM', # 'mid:' (bool) is exact, 'm:' (prob) can do partial + l => 'XL', # 'lid:' (bool) is exact, 'l:' (prob) can do partial f => 'A', t => 'XTO', tc => 'XTO XCC', @@ -92,6 +122,9 @@ my %prob_prefix = ( ); # not documenting m: and mid: for now, the using the URLs works w/o Xapian +# not documenting lid: for now, either, it is probably redundant with l:, +# especially since we don't offer boolean searches for To/Cc/From +# headers, either our @HELP = ( 's:' => 'match within Subject e.g. s:"a quick brown fox"', 'd:' => < 'match within the From header', 'a:' => 'match within the To, Cc, and From headers', 'tc:' => 'match within the To and Cc headers', + 'l:' => 'match contents of the List-Id header', 'bs:' => 'match within the Subject and body', 'dfn:' => 'match filename from diff', 'dfa:' => 'match diff removed (-) lines', @@ -125,15 +159,15 @@ chomp @HELP; sub xdir ($;$) { my ($self, $rdonly) = @_; - if ($self->{version} == 1) { - "$self->{mainrepo}/public-inbox/xapian" . SCHEMA_VERSION; + if ($self->{ibx_ver} == 1) { + "$self->{inboxdir}/public-inbox/xapian" . SCHEMA_VERSION; } else { - my $dir = "$self->{mainrepo}/xap" . SCHEMA_VERSION; + my $dir = "$self->{inboxdir}/xap" . SCHEMA_VERSION; return $dir if $rdonly; - my $part = $self->{partition}; - defined $part or die "partition not given"; - $dir .= "/$part"; + my $shard = $self->{shard}; + defined $shard or die "shard not given"; + $dir .= "/$shard"; } } @@ -142,20 +176,20 @@ sub _xdb ($) { my $dir = xdir($self, 1); my ($xdb, $slow_phrase); my $qpf = \($self->{qp_flags} ||= $QP_FLAGS); - if ($self->{version} >= 2) { - foreach my $part (<$dir/*>) { - -d $part && $part =~ m!/\d+\z! or next; - my $sub = Search::Xapian::Database->new($part); + if ($self->{ibx_ver} >= 2) { + foreach my $shard (<$dir/*>) { + -d $shard && $shard =~ m!/[0-9]+\z! or next; + my $sub = $X{Database}->new($shard); if ($xdb) { $xdb->add_database($sub); } else { $xdb = $sub; } - $slow_phrase ||= -f "$part/iamchert"; + $slow_phrase ||= -f "$shard/iamchert"; } } else { $slow_phrase = -f "$dir/iamchert"; - $xdb = Search::Xapian::Database->new($dir); + $xdb = $X{Database}->new($dir); } $$qpf |= FLAG_PHRASE() unless $slow_phrase; $xdb; @@ -170,17 +204,12 @@ sub xdb ($) { } sub new { - my ($class, $mainrepo, $altid) = @_; - my $version = 1; - my $ibx = $mainrepo; - if (ref $ibx) { - $version = $ibx->{version} || 1; - $mainrepo = $ibx->{mainrepo}; - } + my ($class, $ibx) = @_; + ref $ibx or die "BUG: expected PublicInbox::Inbox object: $ibx"; my $self = bless { - mainrepo => $mainrepo, - altid => $altid, - version => $version, + inboxdir => $ibx->{inboxdir}, + altid => $ibx->{altid}, + ibx_ver => $ibx->version, }, $class; my $dir = xdir($self, 1); $self->{over_ro} = PublicInbox::Over->new("$dir/over.sqlite3"); @@ -211,24 +240,26 @@ sub query { } sub retry_reopen { - my ($self, $cb) = @_; + my ($self, $cb, $arg) = @_; for my $i (1..10) { if (wantarray) { my @ret; - eval { @ret = $cb->() }; + eval { @ret = $cb->($arg) }; return @ret unless $@; } else { my $ret; - eval { $ret = $cb->() }; + eval { $ret = $cb->($arg) }; return $ret unless $@; } # Exception: The revision being read has been discarded - # you should call Xapian::Database::reopen() - if (ref($@) eq 'Search::Xapian::DatabaseModifiedError') { + if (ref($@) =~ /\bDatabaseModifiedError\b/) { warn "reopen try #$i on $@\n"; reopen($self); } else { - warn "ref: ", ref($@), "\n"; + # let caller decide how to spew, because ExtMsg queries + # get wonky and trigger: + # "something terrible happened at .../Xapian/Enquire.pm" die; } } @@ -237,19 +268,19 @@ sub retry_reopen { sub _do_enquire { my ($self, $query, $opts) = @_; - retry_reopen($self, sub { _enquire_once($self, $query, $opts) }); + retry_reopen($self, \&_enquire_once, [ $self, $query, $opts ]); } -sub _enquire_once { - my ($self, $query, $opts) = @_; +sub _enquire_once { # retry_reopen callback + my ($self, $query, $opts) = @{$_[0]}; my $xdb = xdb($self); - my $enquire = Search::Xapian::Enquire->new($xdb); + my $enquire = $X{Enquire}->new($xdb); $enquire->set_query($query); $opts ||= {}; my $desc = !$opts->{asc}; if (($opts->{mset} || 0) == 2) { - $enquire->set_docid_order(Search::Xapian::ENQ_ASCENDING()); - $enquire->set_weighting_scheme(Search::Xapian::BoolWeight->new); + $enquire->set_docid_order($ENQ_ASCENDING); + $enquire->set_weighting_scheme($X{BoolWeight}->new); } elsif ($opts->{relevance}) { $enquire->set_sort_by_relevance_then_value(TS, $desc); } else { @@ -259,16 +290,14 @@ sub _enquire_once { my $limit = $opts->{limit} || 50; my $mset = $enquire->get_mset($offset, $limit); return $mset if $opts->{mset}; - my @msgs = map { - PublicInbox::SearchMsg->load_doc($_->get_document); - } $mset->items; + my @msgs = map { PublicInbox::Smsg::from_mitem($_) } $mset->items; return \@msgs unless wantarray; ($mset->get_matches_estimated, \@msgs) } # read-write -sub stemmer { Search::Xapian::Stem->new($LANG) } +sub stemmer { $X{Stem}->new($LANG) } # read-only sub qp { @@ -278,16 +307,15 @@ sub qp { return $qp if $qp; my $xdb = xdb($self); # new parser - $qp = Search::Xapian::QueryParser->new; + $qp = $X{QueryParser}->new; $qp->set_default_op(OP_AND()); $qp->set_database($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:')); + my $nvrp = $X{NumberValueRangeProcessor}; + $qp->add_valuerangeprocessor($nvrp->new(YYYYMMDD, 'd:')); + $qp->add_valuerangeprocessor($nvrp->new(DT, 'dt:')); while (my ($name, $prefix) = each %bool_pfx_external) { $qp->add_boolean_prefix($name, $_) foreach split(/ /, $prefix); @@ -296,9 +324,11 @@ sub qp { # we do not actually create AltId objects, # just parse the spec to avoid the extra DB handles for now. if (my $altid = $self->{altid}) { - my $user_pfx = $self->{-user_pfx} ||= []; + my $user_pfx = $self->{-user_pfx} = []; for (@$altid) { # $_ = 'serial:gmane:/path/to/gmane.msgmap.sqlite3' + # note: Xapian supports multibyte UTF-8, /^[0-9]+$/, + # and '_' with prefixes matching \w+ /\Aserial:(\w+):/ or next; my $pfx = $1; push @$user_pfx, "$pfx:", <{query_parser} = $qp; } -sub lookup_article { - my ($self, $num) = @_; - $self->{over_ro}->get_art($num); -} - sub help { my ($self) = @_; $self->qp; # parse altids