X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearch.pm;h=f7fdf85478cfd387a728b935ecf5d212df35be3b;hb=f91753f00b760c0e06eb0384792e5b18bd99bb0f;hp=bc2b69854a8765add7c4956f0f7a3c23d9edf0a0;hpb=de243560e2caa1d19bcbf518edfaf8b016161245;p=public-inbox.git diff --git a/lib/PublicInbox/Search.pm b/lib/PublicInbox/Search.pm index bc2b6985..f7fdf854 100644 --- a/lib/PublicInbox/Search.pm +++ b/lib/PublicInbox/Search.pm @@ -1,5 +1,5 @@ -# Copyright (C) 2015 all contributors -# License: AGPLv3 or later (https://www.gnu.org/licenses/agpl-3.0.txt) +# Copyright (C) 2015-2018 all contributors +# License: AGPL-3.0+ # based on notmuch, but with no concept of folders, files or flags # # Read-only search interface for use by the web and NNTP interfaces @@ -8,16 +8,14 @@ use strict; use warnings; # values for searching -use constant TS => 0; # timestamp -use constant NUM => 1; # NNTP article number -use constant BYTES => 2; # :bytes as defined in RFC 3977 -use constant LINES => 3; # :lines as defined in RFC 3977 -use constant YYYYMMDD => 4; # for searching in the WWW UI +use constant TS => 0; # Received: header in Unix time +use constant YYYYMMDD => 1; # for searching in the WWW UI use Search::Xapian qw/:standard/; use PublicInbox::SearchMsg; use PublicInbox::MIME; -use PublicInbox::MID qw/mid_clean id_compress/; +use PublicInbox::MID qw/id_compress/; +use PublicInbox::Over; # This is English-only, everything else is non-standard and may be confused as # a prefix common in patch emails @@ -41,44 +39,46 @@ use constant { # 12 - change YYYYMMDD value column to numeric # 13 - fix threading for empty References/In-Reply-To # (commit 83425ef12e4b65cdcecd11ddcb38175d4a91d5a0) - SCHEMA_VERSION => 13, + # 14 - fix ghost root vivification + SCHEMA_VERSION => 15, # n.b. FLAG_PURE_NOT is expensive not suitable for a public website # as it could become a denial-of-service vector QP_FLAGS => FLAG_PHRASE|FLAG_BOOLEAN|FLAG_LOVEHATE|FLAG_WILDCARD, }; -# setup prefixes -my %bool_pfx_internal = ( - type => 'T', # "mail" or "ghost" - thread => 'G', # newsGroup (or similar entity - e.g. a web forum name) -); - my %bool_pfx_external = ( - # do we still need these? probably not.. - path => 'XPATH', - mid => 'Q', # uniQue id (Message-ID) + mid => 'Q', # Message-ID (full/exact), this is mostly uniQue ); +my $non_quoted_body = 'XNQ XDFN XDFA XDFB XDFHH XDFCTX XDFPRE XDFPOST'; my %prob_prefix = ( # for mairix compatibility s => 'S', - m => 'XMID', # 'mid:' (bool) is exact, 'm:' (prob) can do partial + m => 'XM', # 'mid:' (bool) is exact, 'm:' (prob) can do partial f => 'A', t => 'XTO', tc => 'XTO XCC', c => 'XCC', tcf => 'XTO XCC A', a => 'XTO XCC A', - b => 'XNQ XQUOT', - bs => 'XNQ XQUOT S', + b => $non_quoted_body . ' XQUOT', + bs => $non_quoted_body . ' XQUOT S', n => 'XFN', q => 'XQUOT', - nq => 'XNQ', + nq => $non_quoted_body, + dfn => 'XDFN', + dfa => 'XDFA', + dfb => 'XDFB', + dfhh => 'XDFHH', + dfctx => 'XDFCTX', + dfpre => 'XDFPRE', + dfpost => 'XDFPOST', + dfblob => 'XDFPRE XDFPOST', # default: - '' => 'XMID S A XNQ XQUOT XFN', + '' => 'XM S A XQUOT XFN ' . $non_quoted_body, ); # not documenting m: and mid: for now, the using the URLs works w/o Xapian @@ -99,83 +99,113 @@ EOF 'a:' => 'match within the To, Cc, and From headers', 'tc:' => 'match within the To and Cc headers', 'bs:' => 'match within the Subject and body', + 'dfn:' => 'match filename from diff', + 'dfa:' => 'match diff removed (-) lines', + 'dfb:' => 'match diff added (+) lines', + 'dfhh:' => 'match diff hunk header context (usually a function name)', + 'dfctx:' => 'match diff context lines', + 'dfpre:' => 'match pre-image git blob ID', + 'dfpost:' => 'match post-image git blob ID', + 'dfblob:' => 'match either pre or post-image git blob ID', ); chomp @HELP; -# TODO: -# df (filenames from diff) -# da (diff a/ removed lines) -# db (diff b/ added lines) - -my %all_pfx = (%bool_pfx_internal, %bool_pfx_external, %prob_prefix); - -sub xpfx { $all_pfx{$_[0]} } - -my $mail_query = Search::Xapian::Query->new(xpfx('type') . 'mail'); sub xdir { - my (undef, $git_dir) = @_; - "$git_dir/public-inbox/xapian" . SCHEMA_VERSION; + my ($self) = @_; + if ($self->{version} == 1) { + "$self->{mainrepo}/public-inbox/xapian" . SCHEMA_VERSION; + } else { + my $dir = "$self->{mainrepo}/xap" . SCHEMA_VERSION; + my $part = $self->{partition}; + defined $part or die "partition not given"; + $dir .= "/$part"; + } } sub new { - my ($class, $git_dir, $altid) = @_; - my $dir = $class->xdir($git_dir); - my $db = Search::Xapian::Database->new($dir); - bless { xdb => $db, git_dir => $git_dir, altid => $altid }, $class; + my ($class, $mainrepo, $altid) = @_; + my $version = 1; + my $ibx = $mainrepo; + if (ref $ibx) { + $version = $ibx->{version} || 1; + $mainrepo = $ibx->{mainrepo}; + } + my $self = bless { + mainrepo => $mainrepo, + altid => $altid, + version => $version, + }, $class; + my $dir; + if ($version >= 2) { + $dir = "$self->{mainrepo}/xap" . SCHEMA_VERSION; + my $xdb; + my $parts = 0; + foreach my $part (<$dir/*>) { + -d $part && $part =~ m!/\d+\z! or next; + $parts++; + my $sub = Search::Xapian::Database->new($part); + if ($xdb) { + $xdb->add_database($sub); + } else { + $xdb = $sub; + } + } + $self->{xdb} = $xdb; + } else { + $dir = $self->xdir; + $self->{xdb} = Search::Xapian::Database->new($dir); + } + $self->{over_ro} = PublicInbox::Over->new("$dir/over.sqlite3"); + $self; } -sub reopen { $_[0]->{xdb}->reopen } +sub reopen { + my ($self) = @_; + $self->{xdb}->reopen; + $self; # make chaining easier +} # read-only sub query { my ($self, $query_string, $opts) = @_; - my $query; - $opts ||= {}; - unless ($query_string eq '') { - $query = $self->qp->parse_query($query_string, QP_FLAGS); + if ($query_string eq '' && !$opts->{mset}) { + $self->{over_ro}->recent($opts); + } else { + my $query = $self->qp->parse_query($query_string, QP_FLAGS); $opts->{relevance} = 1 unless exists $opts->{relevance}; + _do_enquire($self, $query, $opts); } - - _do_enquire($self, $query, $opts); } sub get_thread { my ($self, $mid, $opts) = @_; - my $smsg = eval { $self->lookup_message($mid) }; - - return { total => 0, msgs => [] } unless $smsg; - my $qtid = Search::Xapian::Query->new(xpfx('thread').$smsg->thread_id); - my $path = $smsg->path; - if (defined $path && $path ne '') { - my $path = id_compress($smsg->path); - my $qsub = Search::Xapian::Query->new(xpfx('path').$path); - $qtid = Search::Xapian::Query->new(OP_OR, $qtid, $qsub); - } - $opts ||= {}; - $opts->{limit} ||= 1000; - - # always sort threads by timestamp, this makes life easier - # for the threading algorithm (in SearchThread.pm) - $opts->{asc} = 1; - - _do_enquire($self, $qtid, $opts); + $self->{over_ro}->get_thread($mid, $opts); } sub retry_reopen { my ($self, $cb) = @_; - my $ret; - for (1..10) { - eval { $ret = $cb->() }; - return $ret unless $@; + for my $i (1..10) { + if (wantarray) { + my @ret; + eval { @ret = $cb->() }; + return @ret unless $@; + } else { + my $ret; + eval { $ret = $cb->() }; + return $ret unless $@; + } # Exception: The revision being read has been discarded - # you should call Xapian::Database::reopen() if (ref($@) eq 'Search::Xapian::DatabaseModifiedError') { + warn "reopen try #$i on $@\n"; reopen($self); } else { + warn "ref: ", ref($@), "\n"; die; } } + die "Too many Xapian database modifications in progress\n"; } sub _do_enquire { @@ -185,19 +215,13 @@ sub _do_enquire { sub _enquire_once { my ($self, $query, $opts) = @_; - my $enquire = $self->enquire; - if (defined $query) { - $query = Search::Xapian::Query->new(OP_AND,$query,$mail_query); - } else { - $query = $mail_query; - } + my $enquire = enquire($self); + $query = Search::Xapian::Query->new(OP_AND,$query); $enquire->set_query($query); $opts ||= {}; my $desc = !$opts->{asc}; if ($opts->{relevance}) { $enquire->set_sort_by_relevance_then_value(TS, $desc); - } elsif ($opts->{num}) { - $enquire->set_sort_by_value(NUM, 0); } else { $enquire->set_sort_by_value_then_relevance(TS, $desc); } @@ -208,8 +232,9 @@ sub _enquire_once { my @msgs = map { PublicInbox::SearchMsg->load_doc($_->get_document); } $mset->items; + return \@msgs unless wantarray; - { total => $mset->get_matches_estimated, msgs => \@msgs } + ($mset->get_matches_estimated, \@msgs) } # read-write @@ -259,72 +284,69 @@ EOF $self->{query_parser} = $qp; } -sub num_range_processor { - $_[0]->{nrp} ||= Search::Xapian::NumberValueRangeProcessor->new(NUM); -} - # only used for NNTP server sub query_xover { my ($self, $beg, $end, $offset) = @_; - my $qp = Search::Xapian::QueryParser->new; - $qp->set_database($self->{xdb}); - $qp->add_valuerangeprocessor($self->num_range_processor); - my $query = $qp->parse_query("$beg..$end", QP_FLAGS); + $self->{over_ro}->query_xover($beg, $end, $offset); +} - _do_enquire($self, $query, {num => 1, limit => 200, offset => $offset}); +sub query_ts { + my ($self, $ts, $prev) = @_; + $self->{over_ro}->query_ts($ts, $prev); } -sub lookup_message { +sub first_smsg_by_mid { my ($self, $mid) = @_; - $mid = mid_clean($mid); - - my $doc_id = $self->find_unique_doc_id('mid', $mid); my $smsg; - if (defined $doc_id) { - # raises on error: - my $doc = $self->{xdb}->get_document($doc_id); - $smsg = PublicInbox::SearchMsg->wrap($doc, $mid); - $smsg->{doc_id} = $doc_id; - } + retry_reopen($self, sub { + each_smsg_by_mid($self, $mid, sub { $smsg = $_[0]; undef }); + }); $smsg; } -sub lookup_mail { # no ghosts! - my ($self, $mid) = @_; +sub lookup_article { + my ($self, $num) = @_; + my $term = 'XNUM'.$num; + my $db = $self->{xdb}; retry_reopen($self, sub { - my $smsg = lookup_message($self, $mid) or return; - PublicInbox::SearchMsg->load_doc($smsg->{doc}); + 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; + $smsg->load_expand; }); } -sub find_unique_doc_id { - my ($self, $term, $value) = @_; - - my ($begin, $end) = $self->find_doc_ids($term, $value); - - return undef if $begin->equal($end); # not found - - my $rv = $begin->get_docid; - - # sanity check - $begin->inc; - $begin->equal($end) or die "Term '$term:$value' is not unique\n"; - $rv; -} - -# returns begin and end PostingIterator -sub find_doc_ids { - my ($self, $term, $value) = @_; - - $self->find_doc_ids_for_term(xpfx($term) . $value); -} - -# returns begin and end PostingIterator -sub find_doc_ids_for_term { - my ($self, $term) = @_; +sub each_smsg_by_mid { + my ($self, $mid, $cb) = @_; + # XXX retry_reopen isn't necessary for V2Writable, but the PSGI + # interface will need it... my $db = $self->{xdb}; - - ($db->postlist_begin($term), $db->postlist_end($term)); + 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