diff options
Diffstat (limited to 'lib/PublicInbox/Search.pm')
-rw-r--r-- | lib/PublicInbox/Search.pm | 721 |
1 files changed, 548 insertions, 173 deletions
diff --git a/lib/PublicInbox/Search.pm b/lib/PublicInbox/Search.pm index 7f901125..fbdb48a3 100644 --- a/lib/PublicInbox/Search.pm +++ b/lib/PublicInbox/Search.pm @@ -1,44 +1,122 @@ -# Copyright (C) 2015-2020 all contributors <meta@public-inbox.org> +# Copyright (C) all contributors <meta@public-inbox.org> # License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt> # based on notmuch, but with no concept of folders, files or flags # # Read-only search interface for use by the web and NNTP interfaces package PublicInbox::Search; use strict; -use warnings; +use v5.10.1; +use parent qw(Exporter); +our @EXPORT_OK = qw(retry_reopen int_val get_pct xap_terms); +use List::Util qw(max); +use POSIX qw(strftime); +use Carp (); +our $XHC = 0; # defined but false + +# values for searching, changing the numeric value breaks +# compatibility with old indices (so don't change them it) +use constant { + TS => 0, # Received: in Unix time (IMAP INTERNALDATE, JMAP receivedAt) + YYYYMMDD => 1, # redundant with DT below + DT => 2, # Date: YYYYMMDDHHMMSS (IMAP SENT*, JMAP sentAt) + + # added for public-inbox 1.6.0+ + BYTES => 3, # IMAP RFC822.SIZE + UID => 4, # IMAP UID == NNTP article number == Xapian docid + THREADID => 5, # RFC 8474, RFC 8621 + + # TODO + # REPLYCNT => ?, # IMAP ANSWERED -# values for searching -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 + # SCHEMA_VERSION history + # 0 - initial + # 1 - subject_path is lower-cased + # 2 - subject_path is id_compress in the index, only + # 3 - message-ID is compressed if it includes '%' (hack!) + # 4 - change "Re: " normalization, avoid circular Reference ghosts + # 5 - subject_path drops trailing '.' + # 6 - preserve References: order in document data + # 7 - remove references and inreplyto terms + # 8 - remove redundant/unneeded document data + # 9 - disable Message-ID compression (SHA-1) + # 10 - optimize doc for NNTP overviews + # 11 - merge threads when vivifying ghosts + # 12 - change YYYYMMDD value column to numeric + # 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 + # + # v1.6.0 adds BYTES, UID and THREADID values + SCHEMA_VERSION => 15, +}; -use PublicInbox::SearchMsg; -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; +use PublicInbox::Smsg; +eval { require PublicInbox::Over }; +our $QP_FLAGS; +our %X = map { $_ => 0 } qw(BoolWeight Database Enquire QueryParser Stem Query); +our $Xap; # 'Xapian' or 'Search::Xapian' +our $NVRP; # '$Xap::'.('NumberValueRangeProcessor' or 'NumberRangeProcessor') + +# ENQ_DESCENDING and ENQ_ASCENDING weren't in SWIG Xapian.pm prior to 1.4.16, +# let's hope the ABI is stable +our $ENQ_DESCENDING = 0; +our $ENQ_ASCENDING = 1; +our @MAIL_VMAP = ( + [ YYYYMMDD, 'd:'], + [ TS, 'rt:' ], + # these are undocumented for WWW, but lei and IMAP use them + [ DT, 'dt:' ], + [ BYTES, 'z:' ], + [ UID, 'uid:' ] +); +our @MAIL_NRP; + +# Getopt::Long spec, only short options for portability in C++ implementation +our @XH_SPEC = ( + 'a', # ascending sort + 'c', # code search + 'd=s@', # shard dirs + 'g=s', # git dir (with -c) + 'k=i', # sort column (like sort(1)) + 'm=i', # maximum number of results + 'o=i', # offset + 'r', # 1=relevance then column + 't', # collapse threads + 'A=s@', # prefixes + 'K=i', # timeout kill after i seconds + 'O=s', # eidx_key + 'T=i', # threadid +); sub load_xapian () { return 1 if defined $Xap; - for my $x (qw(Search::Xapian Xapian)) { + # n.b. PI_XAPIAN is intended for development use only + for my $x (($ENV{PI_XAPIAN} // 'Xapian'), 'Search::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(); + # `version_string' was added in Xapian 1.1 + my $xver = eval('v'.eval($x.'::version_string()')) // + eval('v'.eval($x.'::xapian_version_string()')); - # for SearchMsg: - *PublicInbox::SearchMsg::sortable_unserialise = - $Xap.'::sortable_unserialise'; + # NumberRangeProcessor was added in Xapian 1.3.6, + # NumberValueRangeProcessor was removed for 1.5.0+, + # continue with the older /Value/ variant for now... + $NVRP = $x.'::'.($x eq 'Xapian' && $xver ge v1.5 ? + 'NumberRangeProcessor' : 'NumberValueRangeProcessor'); + $X{$_} = $Xap.'::'.$_ for (keys %X); + + *sortable_serialise = $x.'::sortable_serialise'; + *sortable_unserialise = $x.'::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 @@ -47,6 +125,7 @@ sub load_xapian () { # or make indexlevel=medium as default $QP_FLAGS = FLAG_PHRASE() | FLAG_BOOLEAN() | FLAG_LOVEHATE() | FLAG_WILDCARD(); + @MAIL_NRP = map { $NVRP->new(@$_) } @MAIL_VMAP; return 1; } undef; @@ -56,74 +135,62 @@ sub load_xapian () { # a prefix common in patch emails our $LANG = 'english'; -use constant { - # SCHEMA_VERSION history - # 0 - initial - # 1 - subject_path is lower-cased - # 2 - subject_path is id_compress in the index, only - # 3 - message-ID is compressed if it includes '%' (hack!) - # 4 - change "Re: " normalization, avoid circular Reference ghosts - # 5 - subject_path drops trailing '.' - # 6 - preserve References: order in document data - # 7 - remove references and inreplyto terms - # 8 - remove redundant/unneeded document data - # 9 - disable Message-ID compression (SHA-1) - # 10 - optimize doc for NNTP overviews - # 11 - merge threads when vivifying ghosts - # 12 - change YYYYMMDD value column to numeric - # 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 - SCHEMA_VERSION => 15, -}; - -my %bool_pfx_external = ( - mid => 'Q', # Message-ID (full/exact), this is mostly uniQue +our %PATCH_BOOL_COMMON = ( dfpre => 'XDFPRE', dfpost => 'XDFPOST', dfblob => 'XDFPRE XDFPOST', + patchid => 'XDFID', ); -my $non_quoted_body = 'XNQ XDFN XDFA XDFB XDFHH XDFCTX XDFPRE XDFPOST'; -my %prob_prefix = ( - # for mairix compatibility +# 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 <> + %PATCH_BOOL_COMMON +); + +# for mairix compatibility +our $NON_QUOTED_BODY = 'XNQ XDFN XDFA XDFB XDFHH XDFCTX XDFPRE XDFPOST XDFID'; +our %PATCH_PROB_COMMON = ( s => 'S', - 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 => $non_quoted_body . ' XQUOT', - bs => $non_quoted_body . ' XQUOT S', + b => $NON_QUOTED_BODY . ' XQUOT', + bs => $NON_QUOTED_BODY . ' XQUOT S', n => 'XFN', q => 'XQUOT', - nq => $non_quoted_body, + nq => $NON_QUOTED_BODY, dfn => 'XDFN', dfa => 'XDFA', dfb => 'XDFB', dfhh => 'XDFHH', dfctx => 'XDFCTX', +); +my %prob_prefix = ( + m => 'XM', # 'mid:' (bool) is exact, 'm:' (prob) can do partial + l => 'XL', # 'lid:' (bool) is exact, 'l:' (prob) can do partial + t => 'XTO', + tc => 'XTO XCC', + c => 'XCC', + tcf => 'XTO XCC A', + a => 'XTO XCC A', + %PATCH_PROB_COMMON, # default: - '' => 'XM S A XQUOT XFN ' . $non_quoted_body, + '' => 'XM S A XQUOT XFN ' . $NON_QUOTED_BODY, ); # not documenting m: and mid: for now, the using the URLs works w/o Xapian +# not documenting lid: for now, either, it is probably redundant with l:, +# especially since we don't offer boolean searches for To/Cc/From +# headers, either our @HELP = ( 's:' => 'match within Subject e.g. s:"a quick brown fox"', 'd:' => <<EOF, -date range as YYYYMMDD e.g. d:19931002..20101002 -Open-ended ranges such as d:19931002.. and d:..20101002 -are also supported -EOF - 'dt:' => <<EOF, -date-time range as YYYYMMDDhhmmss (e.g. dt:19931002011000..19931002011200) +match date-time range, git "approxidate" formats supported +Open-ended ranges such as `d:last.week..' and +`d:..2.days.ago' are supported EOF 'b:' => 'match within message body, including text attachments', 'nq:' => 'match non-quoted text within message body', @@ -134,6 +201,7 @@ EOF 'f:' => 'match within the From header', 'a:' => 'match within the To, Cc, and From headers', 'tc:' => 'match within the To and Cc headers', + 'l:' => 'match contents of the List-Id header', 'bs:' => 'match within the Subject and body', 'dfn:' => 'match filename from diff', 'dfa:' => 'match diff removed (-) lines', @@ -143,65 +211,90 @@ EOF 'dfpre:' => 'match pre-image git blob ID', 'dfpost:' => 'match post-image git blob ID', 'dfblob:' => 'match either pre or post-image git blob ID', + 'patchid:' => "match `git patch-id --stable' output", + 'rt:' => <<EOF, +match received time, like `d:' if sender's clock was correct +EOF ); chomp @HELP; sub xdir ($;$) { my ($self, $rdonly) = @_; - if ($self->{ibx_ver} == 1) { - "$self->{inboxdir}/public-inbox/xapian" . SCHEMA_VERSION; - } else { - my $dir = "$self->{inboxdir}/xap" . SCHEMA_VERSION; - return $dir if $rdonly; - - my $shard = $self->{shard}; - defined $shard or die "shard not given"; - $dir .= "/$shard"; + if ($rdonly || !defined($self->{shard})) { + $self->{xpfx}; + } else { # v2, extindex, cindex only: + "$self->{xpfx}/$self->{shard}"; } } -sub _xdb ($) { +# returns shard directories as an array of strings, does not verify existence +sub shard_dirs ($) { my ($self) = @_; - my $dir = xdir($self, 1); - my ($xdb, $slow_phrase); - my $qpf = \($self->{qp_flags} ||= $QP_FLAGS); - if ($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; - } - $slow_phrase ||= -f "$shard/iamchert"; - } - } else { - $slow_phrase = -f "$dir/iamchert"; - $xdb = $X{Database}->new($dir); + my $xpfx = $self->{xpfx}; + if ($xpfx =~ m!/xapian[0-9]+\z!) { # v1 inbox + ($xpfx); + } else { # v2 inbox, eidx, cidx + opendir(my $dh, $xpfx) or return (); # not initialized yet + # We need numeric sorting so shard[0] is first for reading + # Xapian metadata, if needed + my $last = max(grep(/\A[0-9]+\z/, readdir($dh))) // return (); + map { "$xpfx/$_" } (0..$last); } - $$qpf |= FLAG_PHRASE() unless $slow_phrase; - $xdb; +} + +# returns all shards as separate Xapian::Database objects w/o combining +sub xdb_shards_flat ($) { + my ($self) = @_; + load_xapian(); + $self->{qp_flags} //= $QP_FLAGS; + my $slow_phrase; + my @xdb = map { + $slow_phrase ||= -f "$_/iamchert"; + $X{Database}->new($_); # raises if missing + } shard_dirs($self); + $self->{qp_flags} |= FLAG_PHRASE() if !$slow_phrase; + @xdb; +} + +# v2 Xapian docids don't conflict, so they're identical to +# NNTP article numbers and IMAP UIDs. +# https://trac.xapian.org/wiki/FAQ/MultiDatabaseDocumentID +sub mdocid { + my ($nshard, $mitem) = @_; + my $docid = $mitem->get_docid; + int(($docid - 1) / $nshard) + 1; +} + +sub docids_to_artnums { + my $nshard = shift->{nshard}; + # XXX does array vs arrayref make a difference in modern Perls? + map { int(($_ - 1) / $nshard) + 1 } @_; +} + +sub mset_to_artnums { + my ($self, $mset) = @_; + my $nshard = $self->{nshard}; + [ map { mdocid($nshard, $_) } $mset->items ]; } sub xdb ($) { my ($self) = @_; - $self->{xdb} ||= do { - load_xapian(); - _xdb($self); + $self->{xdb} // do { + my @xdb = $self->xdb_shards_flat or return; + $self->{nshard} = scalar(@xdb); + my $xdb = shift @xdb; + $xdb->add_database($_) for @xdb; + $self->{xdb} = $xdb; }; } sub new { my ($class, $ibx) = @_; ref $ibx or die "BUG: expected PublicInbox::Inbox object: $ibx"; - my $self = bless { - 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"); + my $xap = $ibx->version > 1 ? 'xap' : 'public-inbox/xapian'; + my $xpfx = "$ibx->{inboxdir}/$xap".SCHEMA_VERSION; + my $self = bless { xpfx => $xpfx }, $class; + $self->{altid} = $ibx->{altid} if defined($ibx->{altid}); $self; } @@ -213,99 +306,276 @@ sub reopen { $self; # make chaining easier } -# read-only -sub query { - my ($self, $query_string, $opts) = @_; - $opts ||= {}; - if ($query_string eq '' && !$opts->{mset}) { - $self->{over_ro}->recent($opts); - } else { - my $qp = qp($self); - my $qp_flags = $self->{qp_flags}; - my $query = $qp->parse_query($query_string, $qp_flags); - $opts->{relevance} = 1 unless exists $opts->{relevance}; - _do_enquire($self, $query, $opts); +# Convert git "approxidate" ranges to something usable with our +# Xapian indices. At the moment, Xapian only offers a C++-only API +# and neither the SWIG nor XS bindings allow us to use custom code +# to parse dates (and libgit2 doesn't expose git__date_parse, either, +# so we're running git-rev-parse(1)). +# This replaces things we need to send to $git->date_parse with +# "\0".$strftime_format.['+'|$idx]."\0" placeholders +sub date_parse_prepare { + my ($to_parse, $pfx, $range) = @_; + # are we inside a parenthesized statement? + my $end = $range =~ s/([\)\s]*)\z// ? $1 : ''; + my @r = split(/\.\./, $range, 2); + + # expand "dt:2010-10-02" => "dt:2010-10-02..2010-10-03" and like + # n.b. git doesn't do YYYYMMDD w/o '-', it needs YYYY-MM-DD + # We upgrade "d:" to "dt:" unconditionally + if ($pfx eq 'd') { + $pfx = 'dt'; + # upgrade YYYYMMDD to YYYYMMDDHHMMSS + $_ .= ' 00:00:00' for (grep(m!\A[0-9]{4}[^[:alnum:]] + [0-9]{2}[^[:alnum:]] + [0-9]{2}\z!x, @r)); + $_ .= '000000' for (grep(m!\A[0-9]{8}\z!, @r)); + } + if ($pfx eq 'dt') { + if (!defined($r[1])) { # git needs gaps and not /\d{14}/ + if ($r[0] =~ /\A([0-9]{4})([0-9]{2})([0-9]{2}) + ([0-9]{2})([0-9]{2})([0-9]{2})\z/x) { + push @$to_parse, "$1-$2-$3 $4:$5:$6"; + } else { + push @$to_parse, $r[0]; + } + $r[0] = "\0%Y%m%d%H%M%S$#$to_parse\0"; + $r[1] = "\0%Y%m%d%H%M%S+\0"; + } else { + for my $x (@r) { + next if $x eq '' || $x =~ /\A[0-9]{14}\z/; + push @$to_parse, $x; + $x = "\0%Y%m%d%H%M%S$#$to_parse\0"; + } + } + } else { # (rt|ct), let git interpret "YYYY", deal with Y10K later :P + for my $x (@r) { + next if $x eq '' || $x =~ /\A[0-9]{5,}\z/; + push @$to_parse, $x; + $x = "\0%s$#$to_parse\0"; + } + $r[1] //= "\0%s+\0"; # add 1 day + } + "$pfx:".join('..', @r).$end; +} + +sub date_parse_finalize { + my ($git, $to_parse) = @_; + # git-rev-parse can handle any number of args up to system + # limits (around (4096*32) bytes on Linux). + my @r = $git->date_parse(@$to_parse); + # n.b. git respects TZ, times stored in SQLite/Xapian are always UTC, + # and gmtime doesn't seem to do the right thing when TZ!=UTC + my ($i, $t); + $_[2] =~ s/\0(%[%YmdHMSs]+)([0-9\+]+)\0/ + $t = $2 eq '+' ? ($r[$i]+86400) : $r[$i=$2+0]; + $1 eq '%s' ? $t : strftime($1, gmtime($t))/sge; +} + +# n.b. argv never has NUL, though we'll need to filter it out +# if this $argv isn't from a command execution +sub query_argv_to_string { + my (undef, $git, $argv) = @_; + my $to_parse; + my $tmp = join(' ', map {; + if (s!\b(d|rt|dt):(\S+)\z!date_parse_prepare( + $to_parse //= [], $1, $2)!sge) { + $_; + } elsif (/\s/) { + s/(.*?)\b(\w+:)// ? qq{$1$2"$_"} : qq{"$_"}; + } else { + $_ + } + } @$argv); + date_parse_finalize($git, $to_parse, $tmp) if $to_parse; + $tmp +} + +# this is for the WWW "q=" query parameter and "lei q --stdin" +# it can't do d:"5 days ago", but it will do d:5.days.ago +sub query_approxidate { + my (undef, $git) = @_; # $_[2] = $query_string (modified in-place) + my $DQ = qq<"\x{201c}\x{201d}>; # Xapian can use curly quotes + $_[2] =~ tr/\x00/ /; # Xapian doesn't do NUL, we use it as a placeholder + my ($terms, $phrase, $to_parse); + $_[2] =~ s{([^$DQ]*)([$DQ][^$DQ]*[$DQ])?}{ + ($terms, $phrase) = ($1, $2); + $terms =~ s!\b(d|rt|dt):(\S+)! + date_parse_prepare($to_parse //= [], $1, $2)!sge; + $terms.($phrase // ''); + }sge; + date_parse_finalize($git, $to_parse, $_[2]) if $to_parse; +} + +# read-only, for mail only (codesearch has different rules) +sub mset { + my ($self, $qry_str, $opt) = @_; + my $qp = $self->{qp} //= $self->qparse_new; + my $qry = $qp->parse_query($qry_str, $self->{qp_flags}); + if (defined(my $eidx_key = $opt->{eidx_key})) { + $qry = $X{Query}->new(OP_FILTER(), $qry, 'O'.$eidx_key); + } + if (defined(my $uid_range = $opt->{uid_range})) { + my $range = $X{Query}->new(OP_VALUE_RANGE(), UID, + sortable_serialise($uid_range->[0]), + sortable_serialise($uid_range->[1])); + $qry = $X{Query}->new(OP_FILTER(), $qry, $range); + } + if (defined(my $tid = $opt->{threadid})) { + $tid = sortable_serialise($tid); + $qry = $X{Query}->new(OP_FILTER(), $qry, + $X{Query}->new(OP_VALUE_RANGE(), THREADID, $tid, $tid)); + } + do_enquire($self, $qry, $opt, TS); +} + +sub xhc_start_maybe (@) { + require PublicInbox::XapClient; + my $xhc = PublicInbox::XapClient::start_helper(@_); + require PublicInbox::XhcMset if $xhc; + $xhc; +} + +sub xh_opt ($) { + my ($opt) = @_; + my $lim = $opt->{limit} || 50; + my @ret; + push @ret, '-o', $opt->{offset} if $opt->{offset}; + push @ret, '-m', $lim; + my $rel = $opt->{relevance} // 0; + if ($rel == -2) { # ORDER BY docid/UID (highest first) + push @ret, '-k', '-1'; + } elsif ($rel == -1) { # ORDER BY docid/UID (lowest first) + push @ret, '-k', '-1'; + push @ret, '-a'; + } elsif ($rel == 0) { + push @ret, '-k', $opt->{sort_col} // TS; + push @ret, '-a' if $opt->{asc}; + } else { # rel > 0 + push @ret, '-r'; + push @ret, '-k', $opt->{sort_col} // TS; + push @ret, '-a' if $opt->{asc}; + } + push @ret, '-t' if $opt->{threads}; + push @ret, '-T', $opt->{threadid} if defined $opt->{threadid}; + push @ret, '-O', $opt->{eidx_key} if defined $opt->{eidx_key}; + @ret; +} + +# returns a true value if actually handled asynchronously, +# and a falsy value if handled synchronously +sub async_mset { + my ($self, $qry_str, $opt, $cb, @args) = @_; + if ($XHC) { # unconditionally retrieving pct + rank for now + xdb($self); # populate {nshards} + my @margs = ($self->xh_args, xh_opt($opt)); + my $ret = eval { + my $rd = $XHC->mkreq(undef, 'mset', @margs, $qry_str); + PublicInbox::XhcMset->maybe_new($rd, $self, $cb, @args); + }; + $cb->(@args, undef, $@) if $@; + $ret; + } else { # synchronous + my $mset = $self->mset($qry_str, $opt); + $cb->(@args, $mset); + undef; } } +sub do_enquire { # shared with CodeSearch + my ($self, $qry, $opt, $col) = @_; + my $enq = $X{Enquire}->new(xdb($self)); + $enq->set_query($qry); + my $rel = $opt->{relevance} // 0; + if ($rel == -2) { # ORDER BY docid/UID (highest first) + $enq->set_weighting_scheme($X{BoolWeight}->new); + $enq->set_docid_order($ENQ_DESCENDING); + } elsif ($rel == -1) { # ORDER BY docid/UID (lowest first) + $enq->set_weighting_scheme($X{BoolWeight}->new); + $enq->set_docid_order($ENQ_ASCENDING); + } elsif ($rel == 0) { + $enq->set_sort_by_value_then_relevance($col, !$opt->{asc}); + } else { # rel > 0 + $enq->set_sort_by_relevance_then_value($col, !$opt->{asc}); + } + + # `lei q -t / --threads' or JMAP collapseThreads; but don't collapse + # on `-tt' ({threads} > 1) which sets the Flagged|Important keyword + (($opt->{threads} // 0) == 1 && has_threadid($self)) and + $enq->set_collapse_key(THREADID); + retry_reopen($self, \&enquire_once, $enq, + $opt->{offset} || 0, $opt->{limit} || 50); +} + sub retry_reopen { - my ($self, $cb, $arg) = @_; + my ($self, $cb, @arg) = @_; for my $i (1..10) { if (wantarray) { - my @ret; - eval { @ret = $cb->($arg) }; + my @ret = eval { $cb->($self, @arg) }; return @ret unless $@; } else { - my $ret; - eval { $ret = $cb->($arg) }; + my $ret = eval { $cb->($self, @arg) }; return $ret unless $@; } # Exception: The revision being read has been discarded - # you should call Xapian::Database::reopen() if (ref($@) =~ /\bDatabaseModifiedError\b/) { - warn "reopen try #$i on $@\n"; reopen($self); } else { # let caller decide how to spew, because ExtMsg queries # get wonky and trigger: # "something terrible happened at .../Xapian/Enquire.pm" - die; + Carp::croak($@); } } - die "Too many Xapian database modifications in progress\n"; -} - -sub _do_enquire { - my ($self, $query, $opts) = @_; - retry_reopen($self, \&_enquire_once, [ $self, $query, $opts ]); -} - -sub _enquire_once { # retry_reopen callback - my ($self, $query, $opts) = @{$_[0]}; - my $xdb = xdb($self); - my $enquire = $X{Enquire}->new($xdb); - $enquire->set_query($query); - $opts ||= {}; - my $desc = !$opts->{asc}; - if (($opts->{mset} || 0) == 2) { - $enquire->set_docid_order($ENQ_ASCENDING); - $enquire->set_weighting_scheme($X{BoolWeight}->new); - } elsif ($opts->{relevance}) { - $enquire->set_sort_by_relevance_then_value(TS, $desc); - } else { - $enquire->set_sort_by_value_then_relevance(TS, $desc); - } - my $offset = $opts->{offset} || 0; - my $limit = $opts->{limit} || 50; - my $mset = $enquire->get_mset($offset, $limit); - return $mset if $opts->{mset}; - my @msgs = map { PublicInbox::SearchMsg::from_mitem($_) } $mset->items; - return \@msgs unless wantarray; + Carp::croak("Too many Xapian database modifications in progress\n"); +} + +# returns true if all docs have the THREADID value +sub has_threadid ($) { + my ($self) = @_; + (xdb($self)->get_metadata('has_threadid') // '') eq '1'; +} - ($mset->get_matches_estimated, \@msgs) +sub enquire_once { # retry_reopen callback + my (undef, $enq, $offset, $limit) = @_; + $enq->get_mset($offset, $limit); +} + +sub mset_to_smsg { + my ($self, $ibx, $mset) = @_; + my $nshard = $self->{nshard}; + my $i = 0; + my %order = map { mdocid($nshard, $_) => ++$i } $mset->items; + my @msgs = sort { + $order{$a->{num}} <=> $order{$b->{num}} + } @{$ibx->over->get_all(keys %order)}; + wantarray ? ($mset->get_matches_estimated, \@msgs) : \@msgs; } # read-write sub stemmer { $X{Stem}->new($LANG) } -# read-only -sub qp { +sub qp_init_common { my ($self) = @_; - - my $qp = $self->{query_parser}; - return $qp if $qp; - my $xdb = xdb($self); - # new parser - $qp = $X{QueryParser}->new; + my $qp = $X{QueryParser}->new; $qp->set_default_op(OP_AND()); - $qp->set_database($xdb); - $qp->set_stemmer($self->stemmer); + $qp->set_database(xdb($self)); + $qp->set_stemmer(stemmer($self)); $qp->set_stemming_strategy(STEM_SOME()); - $qp->set_max_wildcard_expansion(100); - my $nvrp = $X{NumberValueRangeProcessor}; - $qp->add_valuerangeprocessor($nvrp->new(YYYYMMDD, 'd:')); - $qp->add_valuerangeprocessor($nvrp->new(DT, 'dt:')); + my $cb = $qp->can('set_max_wildcard_expansion') // + $qp->can('set_max_expansion'); # Xapian 1.5.0+ + $cb->($qp, 100); + $qp; +} +# read-only +sub qparse_new { + my ($self) = @_; + my $qp = qp_init_common($self); + my $cb = $qp->can('add_valuerangeprocessor') // + $qp->can('add_rangeprocessor'); # Xapian 1.5.0+ + + $cb->($qp, $_) for @MAIL_NRP; while (my ($name, $prefix) = each %bool_pfx_external) { $qp->add_boolean_prefix($name, $_) foreach split(/ /, $prefix); } @@ -313,9 +583,11 @@ 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, "$pfx:", <<EOF; @@ -330,13 +602,46 @@ EOF while (my ($name, $prefix) = each %prob_prefix) { $qp->add_prefix($name, $_) foreach split(/ /, $prefix); } + $qp; +} + +sub generate_cxx () { # generates snippet for xap_helper.h + my $ret = <<EOM; +# line ${\__LINE__} "${\__FILE__}" +static NRP *mail_nrp[${\scalar(@MAIL_VMAP)}]; +static void mail_nrp_init(void) +{ +EOM + for (0..$#MAIL_VMAP) { + my $x = $MAIL_VMAP[$_]; + $ret .= qq{\tmail_nrp[$_] = new NRP($x->[0], "$x->[1]");\n} + } +$ret .= <<EOM; +} - $self->{query_parser} = $qp; +# line ${\__LINE__} "${\__FILE__}" +static void qp_init_mail_search(Xapian::QueryParser *qp) +{ + for (size_t i = 0; i < MY_ARRAY_SIZE(mail_nrp); i++) + qp->ADD_RP(mail_nrp[i]); +EOM + for my $name (sort keys %bool_pfx_external) { + for (split(/ /, $bool_pfx_external{$name})) { + $ret .= qq{\tqp->add_boolean_prefix("$name", "$_");\n} + } + } + # TODO: altid support + for my $name (sort keys %prob_prefix) { + for (split(/ /, $prob_prefix{$name})) { + $ret .= qq{\tqp->add_prefix("$name", "$_");\n} + } + } + $ret .= "}\n"; } sub help { my ($self) = @_; - $self->qp; # parse altids + $self->{qp} //= $self->qparse_new; # parse altids my @ret = @HELP; if (my $user_pfx = $self->{-user_pfx}) { push @ret, @$user_pfx; @@ -344,4 +649,74 @@ sub help { \@ret; } +# always returns a scalar value +sub int_val ($$) { + my ($doc, $col) = @_; + my $val = $doc->get_value($col) or return undef; # undef is '' in Xapian + sortable_unserialise($val) + 0; # PV => IV conversion +} + +sub get_pct ($) { # mset item + # Capped at "99%" since "100%" takes an extra column in the + # thread skeleton view. <xapian/mset.h> says the value isn't + # very meaningful, anyways. + my $n = $_[0]->get_percent; + $n > 99 ? 99 : $n; +} + +sub xap_terms ($$;@) { + my ($pfx, $xdb_or_doc, @docid) = @_; # @docid may be empty () + my $end = $xdb_or_doc->termlist_end(@docid); + my $cur = $xdb_or_doc->termlist_begin(@docid); + $cur->skip_to($pfx); + my (@ret, $tn); + my $pfxlen = length($pfx); + for (; $cur != $end; $cur++) { + $tn = $cur->get_termname; + index($tn, $pfx) ? last : push(@ret, substr($tn, $pfxlen)); + } + wantarray ? @ret : +{ map { $_ => undef } @ret }; +} + +# get combined docid from over.num: +# (not generic Xapian, only works with our sharding scheme for mail) +sub num2docid ($$) { + my ($self, $num) = @_; + my $nshard = $self->{nshard}; + ($num - 1) * $nshard + $num % $nshard + 1; +} + +sub all_terms { + my ($self, $pfx) = @_; + my $cur = xdb($self)->allterms_begin($pfx); + my $end = $self->{xdb}->allterms_end($pfx); + my $pfxlen = length($pfx); + my @ret; + for (; $cur != $end; $cur++) { + push @ret, substr($cur->get_termname, $pfxlen); + } + wantarray ? @ret : +{ map { $_ => undef } @ret }; +} + +sub xh_args { # prep getopt args to feed to xap_helper.h socket + map { ('-d', $_) } shard_dirs($_[0]); +} + +sub docids_by_postlist ($$) { + my ($self, $q) = @_; + my $cur = $self->xdb->postlist_begin($q); + my $end = $self->{xdb}->postlist_end($q); + my @ids; + for (; $cur != $end; $cur++) { push(@ids, $cur->get_docid) }; + @ids; +} + +sub get_doc ($$) { + my ($self, $docid) = @_; + eval { $self->{xdb}->get_document($docid) } // do { + die $@ if $@ && ref($@) !~ /\bDocNotFoundError\b/; + undef; + } +} + 1; |