# Copyright (C) 2015-2018 all contributors
# License: AGPL-3.0+
# based on notmuch, but with no concept of folders, files or flags
#
# Indexes mail with Xapian and our (SQLite-based) ::Msgmap for use
# with the web and NNTP interfaces. This index maintains thread
# relationships for use by PublicInbox::SearchThread.
# This writes to the search index.
package PublicInbox::SearchIdx;
use strict;
use warnings;
use Fcntl qw(:flock :DEFAULT);
use PublicInbox::MIME;
use Email::MIME::ContentType;
$Email::MIME::ContentType::STRICT_PARAMS = 0;
use base qw(PublicInbox::Search);
use PublicInbox::MID qw/mid_clean id_compress mid_mime/;
use PublicInbox::MsgIter;
use Carp qw(croak);
use POSIX qw(strftime);
require PublicInbox::Git;
use constant {
MAX_MID_SIZE => 244, # max term size - 1 in Xapian
PERM_UMASK => 0,
OLD_PERM_GROUP => 1,
OLD_PERM_EVERYBODY => 2,
PERM_GROUP => 0660,
PERM_EVERYBODY => 0664,
BATCH_BYTES => 1_000_000,
DEBUG => !!$ENV{DEBUG},
};
my %GIT_ESC = (
a => "\a",
b => "\b",
f => "\f",
n => "\n",
r => "\r",
t => "\t",
v => "\013",
);
sub git_unquote ($) {
my ($s) = @_;
return $s unless ($s =~ /\A"(.*)"\z/);
$s = $1;
$s =~ s/\\([abfnrtv])/$GIT_ESC{$1}/g;
$s =~ s/\\([0-7]{1,3})/chr(oct($1))/ge;
$s;
}
sub new {
my ($class, $inbox, $creat) = @_;
my $git_dir = $inbox;
my $altid;
if (ref $inbox) {
$git_dir = $inbox->{mainrepo};
$altid = $inbox->{altid};
if ($altid) {
require PublicInbox::AltId;
$altid = [ map {
PublicInbox::AltId->new($inbox, $_);
} @$altid ];
}
}
require Search::Xapian::WritableDatabase;
my $self = bless { git_dir => $git_dir, -altid => $altid }, $class;
my $perm = $self->_git_config_perm;
my $umask = _umask_for($perm);
$self->{umask} = $umask;
$self->{lock_path} = "$git_dir/ssoma.lock";
$self->{git} = PublicInbox::Git->new($git_dir);
$self->{creat} = ($creat || 0) == 1;
$self;
}
sub _xdb_release {
my ($self) = @_;
my $xdb = delete $self->{xdb} or croak 'not acquired';
$xdb->close;
_lock_release($self) if $self->{creat};
undef;
}
sub _xdb_acquire {
my ($self) = @_;
croak 'already acquired' if $self->{xdb};
my $dir = PublicInbox::Search->xdir($self->{git_dir});
my $flag = Search::Xapian::DB_OPEN;
if ($self->{creat}) {
require File::Path;
_lock_acquire($self);
File::Path::mkpath($dir);
$flag = Search::Xapian::DB_CREATE_OR_OPEN;
}
$self->{xdb} = Search::Xapian::WritableDatabase->new($dir, $flag);
}
# we only acquire the flock if creating or reindexing;
# PublicInbox::Import already has the lock on its own.
sub _lock_acquire {
my ($self) = @_;
croak 'already locked' if $self->{lockfh};
sysopen(my $lockfh, $self->{lock_path}, O_WRONLY|O_CREAT) or
die "failed to open lock $self->{lock_path}: $!\n";
flock($lockfh, LOCK_EX) or die "lock failed: $!\n";
$self->{lockfh} = $lockfh;
}
sub _lock_release {
my ($self) = @_;
my $lockfh = delete $self->{lockfh} or croak 'not locked';
flock($lockfh, LOCK_UN) or die "unlock failed: $!\n";
close $lockfh or die "close failed: $!\n";
}
sub add_val ($$$) {
my ($doc, $col, $num) = @_;
$num = Search::Xapian::sortable_serialise($num);
$doc->add_value($col, $num);
}
sub add_values ($$$) {
my ($smsg, $bytes, $num) = @_;
my $ts = $smsg->ts;
my $doc = $smsg->{doc};
add_val($doc, &PublicInbox::Search::TS, $ts);
defined($num) and add_val($doc, &PublicInbox::Search::NUM, $num);
defined($bytes) and add_val($doc, &PublicInbox::Search::BYTES, $bytes);
add_val($doc, &PublicInbox::Search::LINES,
$smsg->{mime}->body_raw =~ tr!\n!\n!);
my $yyyymmdd = strftime('%Y%m%d', gmtime($ts));
add_val($doc, PublicInbox::Search::YYYYMMDD, $yyyymmdd);
}
sub index_users ($$) {
my ($tg, $smsg) = @_;
my $from = $smsg->from;
my $to = $smsg->to;
my $cc = $smsg->cc;
$tg->index_text($from, 1, 'A'); # A - author
$tg->increase_termpos;
$tg->index_text($to, 1, 'XTO') if $to ne '';
$tg->increase_termpos;
$tg->index_text($cc, 1, 'XCC') if $cc ne '';
$tg->increase_termpos;
}
sub index_text_inc ($$$) {
my ($tg, $text, $pfx) = @_;
$tg->index_text($text, 1, $pfx);
$tg->increase_termpos;
}
sub index_old_diff_fn {
my ($tg, $seen, $fa, $fb) = @_;
# no renames or space support for traditional diffs,
# find the number of leading common paths to strip:
my @fa = split('/', $fa);
my @fb = split('/', $fb);
while (scalar(@fa) && scalar(@fb)) {
$fa = join('/', @fa);
$fb = join('/', @fb);
if ($fa eq $fb) {
index_text_inc($tg, $fa,'XDFN') unless $seen->{$fa}++;
return 1;
}
shift @fa;
shift @fb;
}
0;
}
sub index_diff ($$$) {
my ($tg, $lines, $doc) = @_;
my %seen;
my $in_diff;
foreach (@$lines) {
if ($in_diff && s/^ //) { # diff context
index_text_inc($tg, $_, 'XDFCTX');
} elsif (/^-- $/) { # email signature begins
$in_diff = undef;
} elsif (m!^diff --git ("?a/.+) ("?b/.+)\z!) {
my ($fa, $fb) = ($1, $2);
my $fn = (split('/', git_unquote($fa), 2))[1];
index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
$fn = (split('/', git_unquote($fb), 2))[1];
index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
$in_diff = 1;
# traditional diff:
} elsif (m/^diff -(.+) (\S+) (\S+)$/) {
my ($opt, $fa, $fb) = ($1, $2, $3);
# only support unified:
next unless $opt =~ /[uU]/;
$in_diff = index_old_diff_fn($tg, \%seen, $fa, $fb);
} elsif (m!^--- ("?a/.+)!) {
my $fn = (split('/', git_unquote($1), 2))[1];
index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
$in_diff = 1;
} elsif (m!^\+\+\+ ("?b/.+)!) {
my $fn = (split('/', git_unquote($1), 2))[1];
index_text_inc($tg, $fn, 'XDFN') unless $seen{$fn}++;
$in_diff = 1;
} elsif (/^--- (\S+)/) {
$in_diff = $1;
} elsif (defined $in_diff && /^\+\+\+ (\S+)/) {
$in_diff = index_old_diff_fn($tg, \%seen, $in_diff, $1);
} elsif ($in_diff && s/^\+//) { # diff added
index_text_inc($tg, $_, 'XDFB');
} elsif ($in_diff && s/^-//) { # diff removed
index_text_inc($tg, $_, 'XDFA');
} elsif (m!^index ([a-f0-9]+)\.\.([a-f0-9]+)!) {
my ($ba, $bb) = ($1, $2);
index_git_blob_id($doc, 'XDFPRE', $ba);
index_git_blob_id($doc, 'XDFPOST', $bb);
$in_diff = 1;
} elsif (/^@@ (?:\S+) (?:\S+) @@\s*$/) {
# traditional diff w/o -p
} elsif (/^@@ (?:\S+) (?:\S+) @@\s*(\S+.*)$/) {
# hunk header context
index_text_inc($tg, $1, 'XDFHH');
# ignore the following lines:
} elsif (/^(?:dis)similarity index/) {
} elsif (/^(?:old|new) mode/) {
} elsif (/^(?:deleted|new) file mode/) {
} elsif (/^(?:copy|rename) (?:from|to) /) {
} elsif (/^(?:dis)?similarity index /) {
} elsif (/^\\ No newline at end of file/) {
} elsif (/^Binary files .* differ/) {
} elsif ($_ eq '') {
$in_diff = undef;
} else {
warn "non-diff line: $_\n" if DEBUG && $_ ne '';
$in_diff = undef;
}
}
}
sub index_body ($$$) {
my ($tg, $lines, $doc) = @_;
my $txt = join("\n", @$lines);
$tg->index_text($txt, !!$doc, $doc ? 'XNQ' : 'XQUOT');
$tg->increase_termpos;
# does it look like a diff?
if ($doc && $txt =~ /^(?:diff|---|\+\+\+) /ms) {
$txt = undef;
index_diff($tg, $lines, $doc);
}
@$lines = ();
}
sub add_message {
my ($self, $mime, $bytes, $num, $blob) = @_; # mime = Email::MIME object
my $db = $self->{xdb};
my ($doc_id, $old_tid);
my $mid = mid_clean(mid_mime($mime));
eval {
die 'Message-ID too long' if length($mid) > MAX_MID_SIZE;
my $smsg = $self->lookup_message($mid);
if ($smsg) {
# convert a ghost to a regular message
# it will also clobber any existing regular message
$doc_id = $smsg->{doc_id};
$old_tid = $smsg->thread_id;
}
$smsg = PublicInbox::SearchMsg->new($mime);
my $doc = $smsg->{doc};
$doc->add_term('Q' . $mid);
my $subj = $smsg->subject;
if ($subj ne '') {
my $path = $self->subject_path($subj);
$doc->add_term('XPATH' . id_compress($path));
}
add_values($smsg, $bytes, $num);
my $tg = $self->term_generator;
$tg->set_document($doc);
$tg->index_text($subj, 1, 'S') if $subj;
$tg->increase_termpos;
index_users($tg, $smsg);
msg_iter($mime, sub {
my ($part, $depth, @idx) = @{$_[0]};
my $ct = $part->content_type || 'text/plain';
my $fn = $part->filename;
if (defined $fn && $fn ne '') {
$tg->index_text($fn, 1, 'XFN');
}
return if $ct =~ m!\btext/x?html\b!i;
my $s = eval { $part->body_str };
if ($@) {
if ($ct =~ m!\btext/plain\b!i) {
# Try to assume UTF-8 because Alpine
# seems to do wacky things and set
# charset=X-UNKNOWN
$part->charset_set('UTF-8');
$s = eval { $part->body_str };
$s = $part->body if $@;
}
}
defined $s or return;
my (@orig, @quot);
my $body = $part->body;
my @lines = split(/\n/, $body);
while (defined(my $l = shift @lines)) {
if ($l =~ /^>/) {
index_body($tg, \@orig, $doc) if @orig;
push @quot, $l;
} else {
index_body($tg, \@quot, 0) if @quot;
push @orig, $l;
}
}
index_body($tg, \@quot, 0) if @quot;
index_body($tg, \@orig, $doc) if @orig;
});
link_message($self, $smsg, $old_tid);
$tg->index_text($mid, 1, 'XMID');
$doc->set_data($smsg->to_doc_data($blob));
if (my $altid = $self->{-altid}) {
foreach my $alt (@$altid) {
my $id = $alt->mid2alt($mid);
next unless defined $id;
$doc->add_term($alt->{xprefix} . $id);
}
}
if (defined $doc_id) {
$db->replace_document($doc_id, $doc);
} else {
$doc_id = $db->add_document($doc);
}
};
if ($@) {
warn "failed to index message <$mid>: $@\n";
return undef;
}
$doc_id;
}
# returns deleted doc_id on success, undef on missing
sub remove_message {
my ($self, $mid) = @_;
my $db = $self->{xdb};
my $doc_id;
$mid = mid_clean($mid);
eval {
$doc_id = $self->find_unique_doc_id('Q' . $mid);
if (defined $doc_id) {
$db->delete_document($doc_id);
} else {
warn "cannot remove non-existent <$mid>\n";
}
};
if ($@) {
warn "failed to remove message <$mid>: $@\n";
return undef;
}
$doc_id;
}
sub term_generator { # write-only
my ($self) = @_;
my $tg = $self->{term_generator};
return $tg if $tg;
$tg = Search::Xapian::TermGenerator->new;
$tg->set_stemmer($self->stemmer);
$self->{term_generator} = $tg;
}
# increments last_thread_id counter
# returns a 64-bit integer represented as a hex string
sub next_thread_id {
my ($self) = @_;
my $db = $self->{xdb};
my $last_thread_id = int($db->get_metadata('last_thread_id') || 0);
$db->set_metadata('last_thread_id', ++$last_thread_id);
$last_thread_id;
}
sub link_message {
my ($self, $smsg, $old_tid) = @_;
my $doc = $smsg->{doc};
my $mid = $smsg->mid;
my $mime = $smsg->{mime};
my $hdr = $mime->header_obj;
# last References should be IRT, but some mail clients do things
# out of order, so trust IRT over References iff IRT exists
my @refs = (($hdr->header_raw('References') || '') =~ /<([^>]+)>/g);
push(@refs, (($hdr->header_raw('In-Reply-To') || '') =~ /<([^>]+)>/g));
my $tid;
if (@refs) {
my %uniq = ($mid => 1);
my @orig_refs = @refs;
@refs = ();
# prevent circular references via References: here:
foreach my $ref (@orig_refs) {
if (length($ref) > MAX_MID_SIZE) {
warn "References: <$ref> too long, ignoring\n";
}
next if $uniq{$ref};
$uniq{$ref} = 1;
push @refs, $ref;
}
}
if (@refs) {
$smsg->{references} = '<'.join('> <', @refs).'>';
# first ref *should* be the thread root,
# but we can never trust clients to do the right thing
my $ref = shift @refs;
$tid = $self->_resolve_mid_to_tid($ref);
$self->merge_threads($tid, $old_tid) if defined $old_tid;
# the rest of the refs should point to this tid:
foreach $ref (@refs) {
my $ptid = $self->_resolve_mid_to_tid($ref);
merge_threads($self, $tid, $ptid);
}
} else {
$tid = defined $old_tid ? $old_tid : $self->next_thread_id;
}
$doc->add_term('G' . $tid);
}
sub index_blob {
my ($self, $mime, $bytes, $num, $blob) = @_;
$self->add_message($mime, $bytes, $num, $blob);
}
sub index_git_blob_id {
my ($doc, $pfx, $objid) = @_;
my $len = length($objid);
for (my $len = length($objid); $len >= 7; ) {
$doc->add_term($pfx.$objid);
$objid = substr($objid, 0, --$len);
}
}
sub unindex_blob {
my ($self, $mime) = @_;
my $mid = eval { mid_clean(mid_mime($mime)) };
$self->remove_message($mid) if defined $mid;
}
sub index_mm {
my ($self, $mime) = @_;
my $mid = mid_clean(mid_mime($mime));
my $mm = $self->{mm};
my $num = $mm->mid_insert($mid);
# fallback to num_for since filters like RubyLang set the number
defined $num ? $num : $mm->num_for($mid);
}
sub unindex_mm {
my ($self, $mime) = @_;
$self->{mm}->mid_delete(mid_clean(mid_mime($mime)));
}
sub index_mm2 {
my ($self, $mime, $bytes, $blob) = @_;
my $num = $self->{mm}->num_for(mid_clean(mid_mime($mime)));
index_blob($self, $mime, $bytes, $num, $blob);
}
sub unindex_mm2 {
my ($self, $mime) = @_;
$self->{mm}->mid_delete(mid_clean(mid_mime($mime)));
unindex_blob($self, $mime);
}
sub index_both {
my ($self, $mime, $bytes, $blob) = @_;
my $num = index_mm($self, $mime);
index_blob($self, $mime, $bytes, $num, $blob);
}
sub unindex_both {
my ($self, $mime) = @_;
unindex_blob($self, $mime);
unindex_mm($self, $mime);
}
sub do_cat_mail {
my ($git, $blob, $sizeref) = @_;
my $mime = eval {
my $str = $git->cat_file($blob, $sizeref);
# fixup bugs from import:
$$str =~ s/\A[\r\n]*From [^\r\n]*\r?\n//s;
PublicInbox::MIME->new($str);
};
$@ ? undef : $mime;
}
sub index_sync {
my ($self, $opts) = @_;
with_umask($self, sub { $self->_index_sync($opts) });
}
sub batch_adjust ($$$$) {
my ($max, $bytes, $batch_cb, $latest) = @_;
$$max -= $bytes;
if ($$max <= 0) {
$$max = BATCH_BYTES;
$batch_cb->($latest, 1);
}
}
sub rlog {
my ($self, $log, $add_cb, $del_cb, $batch_cb) = @_;
my $hex = '[a-f0-9]';
my $h40 = $hex .'{40}';
my $addmsg = qr!^:000000 100644 \S+ ($h40) A\t${hex}{2}/${hex}{38}$!;
my $delmsg = qr!^:100644 000000 ($h40) \S+ D\t${hex}{2}/${hex}{38}$!;
my $git = $self->{git};
my $latest;
my $bytes;
my $max = BATCH_BYTES;
local $/ = "\n";
my $line;
while (defined($line = <$log>)) {
if ($line =~ /$addmsg/o) {
my $blob = $1;
my $mime = do_cat_mail($git, $blob, \$bytes) or next;
batch_adjust(\$max, $bytes, $batch_cb, $latest);
$add_cb->($self, $mime, $bytes, $blob);
} elsif ($line =~ /$delmsg/o) {
my $blob = $1;
my $mime = do_cat_mail($git, $blob, \$bytes) or next;
batch_adjust(\$max, $bytes, $batch_cb, $latest);
$del_cb->($self, $mime);
} elsif ($line =~ /^commit ($h40)/o) {
$latest = $1;
}
}
$batch_cb->($latest, 0);
}
sub _msgmap_init {
my ($self) = @_;
$self->{mm} = eval {
require PublicInbox::Msgmap;
PublicInbox::Msgmap->new($self->{git_dir}, 1);
};
}
sub _git_log {
my ($self, $range) = @_;
$self->{git}->popen(qw/log --reverse --no-notes --no-color
--raw -r --no-abbrev/, $range);
}
# indexes all unindexed messages
sub _index_sync {
my ($self, $opts) = @_;
my $tip = $opts->{ref} || 'HEAD';
my $reindex = $opts->{reindex};
my ($mkey, $last_commit, $lx, $xlog);
$self->{git}->batch_prepare;
my $xdb = _xdb_acquire($self);
$xdb->begin_transaction;
do {
$xlog = undef;
$mkey = 'last_commit';
$last_commit = $xdb->get_metadata('last_commit');
$lx = $last_commit;
if ($reindex) {
$lx = '';
$mkey = undef if $last_commit ne '';
}
$xdb->cancel_transaction;
$xdb = _xdb_release($self);
# ensure we leak no FDs to "git log"
my $range = $lx eq '' ? $tip : "$lx..$tip";
$xlog = _git_log($self, $range);
$xdb = _xdb_acquire($self);
$xdb->begin_transaction;
} while ($xdb->get_metadata('last_commit') ne $last_commit);
my $mm = _msgmap_init($self);
my $dbh = $mm->{dbh} if $mm;
my $mm_only;
my $cb = sub {
my ($commit, $more) = @_;
if ($dbh) {
$mm->last_commit($commit) if $commit;
$dbh->commit;
}
if (!$mm_only) {
$xdb->set_metadata($mkey, $commit) if $mkey && $commit;
$xdb->commit_transaction;
$xdb = _xdb_release($self);
}
# let another process do some work... <
if ($more) {
if (!$mm_only) {
$xdb = _xdb_acquire($self);
$xdb->begin_transaction;
}
$dbh->begin_work if $dbh;
}
};
if ($mm) {
$dbh->begin_work;
my $lm = $mm->last_commit || '';
if ($lm eq $lx) {
# Common case is the indexes are synced,
# we only need to run git-log once:
rlog($self, $xlog, *index_both, *unindex_both, $cb);
} else {
# Uncommon case, msgmap and xapian are out-of-sync
# do not care for performance (but git is fast :>)
# This happens if we have to reindex Xapian since
# msgmap is a frozen format and our Xapian format
# is evolving.
my $r = $lm eq '' ? $tip : "$lm..$tip";
# first, ensure msgmap is up-to-date:
my $mkey_prev = $mkey;
$mkey = undef; # ignore xapian, for now
my $mlog = _git_log($self, $r);
$mm_only = 1;
rlog($self, $mlog, *index_mm, *unindex_mm, $cb);
$mm_only = $mlog = undef;
# now deal with Xapian
$mkey = $mkey_prev;
$dbh = undef;
rlog($self, $xlog, *index_mm2, *unindex_mm2, $cb);
}
} else {
# user didn't install DBD::SQLite and DBI
rlog($self, $xlog, *index_blob, *unindex_blob, $cb);
}
}
# this will create a ghost as necessary
sub _resolve_mid_to_tid {
my ($self, $mid) = @_;
my $smsg = $self->lookup_message($mid) || $self->create_ghost($mid);
$smsg->thread_id;
}
sub create_ghost {
my ($self, $mid) = @_;
my $tid = $self->next_thread_id;
my $doc = Search::Xapian::Document->new;
$doc->add_term('Q' . $mid);
$doc->add_term('G' . $tid);
$doc->add_term('T' . 'ghost');
my $smsg = PublicInbox::SearchMsg->wrap($doc, $mid);
$self->{xdb}->add_document($doc);
$smsg;
}
sub merge_threads {
my ($self, $winner_tid, $loser_tid) = @_;
return if $winner_tid == $loser_tid;
my $db = $self->{xdb};
my $batch_size = 1000; # don't let @ids grow too large to avoid OOM
while (1) {
my ($head, $tail) = $self->find_doc_ids('G' . $loser_tid);
return if $head == $tail;
my @ids;
for (; $head != $tail && @ids < $batch_size; $head->inc) {
push @ids, $head->get_docid;
}
foreach my $docid (@ids) {
my $doc = $db->get_document($docid);
$doc->remove_term('G' . $loser_tid);
$doc->add_term('G' . $winner_tid);
$db->replace_document($docid, $doc);
}
}
}
sub _read_git_config_perm {
my ($self) = @_;
my @cmd = qw(config core.sharedRepository);
my $fh = PublicInbox::Git->new($self->{git_dir})->popen(@cmd);
local $/ = "\n";
my $perm = <$fh>;
chomp $perm if defined $perm;
$perm;
}
sub _git_config_perm {
my $self = shift;
my $perm = scalar @_ ? $_[0] : _read_git_config_perm($self);
return PERM_GROUP if (!defined($perm) || $perm eq '');
return PERM_UMASK if ($perm eq 'umask');
return PERM_GROUP if ($perm eq 'group');
if ($perm =~ /\A(?:all|world|everybody)\z/) {
return PERM_EVERYBODY;
}
return PERM_GROUP if ($perm =~ /\A(?:true|yes|on|1)\z/);
return PERM_UMASK if ($perm =~ /\A(?:false|no|off|0)\z/);
my $i = oct($perm);
return PERM_UMASK if ($i == PERM_UMASK);
return PERM_GROUP if ($i == OLD_PERM_GROUP);
return PERM_EVERYBODY if ($i == OLD_PERM_EVERYBODY);
if (($i & 0600) != 0600) {
die "core.sharedRepository mode invalid: ".
sprintf('%.3o', $i) . "\nOwner must have permissions\n";
}
($i & 0666);
}
sub _umask_for {
my ($perm) = @_; # _git_config_perm return value
my $rv = $perm;
return umask if $rv == 0;
# set +x bit if +r or +w were set
$rv |= 0100 if ($rv & 0600);
$rv |= 0010 if ($rv & 0060);
$rv |= 0001 if ($rv & 0006);
(~$rv & 0777);
}
sub with_umask {
my ($self, $cb) = @_;
my $old = umask $self->{umask};
my $rv = eval { $cb->() };
my $err = $@;
umask $old;
die $err if $err;
$rv;
}
sub DESTROY {
# order matters for unlocking
$_[0]->{xdb} = undef;
$_[0]->{lockfh} = undef;
}
1;