X-Git-Url: http://www.git.stargrave.org/?a=blobdiff_plain;f=lib%2FPublicInbox%2FSearch.pm;h=a4e2498eb596d1cd9346b6620ffa464aa5a1ce6d;hb=107123dd03f38b97579ce5976f31ae3fd5f22903;hp=24cb26670754b2f4a21b62e7d22c3cdc76098a5a;hpb=53e8cfbe4e06e5ae6ad61fb7e9bd45804c253a72;p=public-inbox.git
diff --git a/lib/PublicInbox/Search.pm b/lib/PublicInbox/Search.pm
index 24cb2667..a4e2498e 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
@@ -8,15 +8,16 @@ use strict;
use warnings;
# values for searching
-use constant TS => 0; # timestamp
+use constant DS => 0; # Date: header in Unix time
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 => 4; # Received: header in Unix time
+use constant YYYYMMDD => 5; # for searching in the WWW UI
use Search::Xapian qw/:standard/;
use PublicInbox::SearchMsg;
-use Email::MIME;
+use PublicInbox::MIME;
use PublicInbox::MID qw/mid_clean id_compress/;
# This is English-only, everything else is non-standard and may be confused as
@@ -39,7 +40,10 @@ use constant {
# 10 - optimize doc for NNTP overviews
# 11 - merge threads when vivifying ghosts
# 12 - change YYYYMMDD value column to numeric
- SCHEMA_VERSION => 12,
+ # 13 - fix threading for empty References/In-Reply-To
+ # (commit 83425ef12e4b65cdcecd11ddcb38175d4a91d5a0)
+ # 14 - fix ghost root vivification
+ SCHEMA_VERSION => 14,
# n.b. FLAG_PURE_NOT is expensive not suitable for a public website
# as it could become a denial-of-service vector
@@ -53,30 +57,37 @@ my %bool_pfx_internal = (
);
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
@@ -89,7 +100,7 @@ are also supported
EOF
'b:' => 'match within message body, including text attachments',
'nq:' => 'match non-quoted text within message body',
- 'quot:' => 'match quoted text within message body',
+ 'q:' => 'match quoted text within message body',
'n:' => 'match filename of attachment(s)',
't:' => 'match within the To header',
'c:' => 'match within the Cc header',
@@ -97,38 +108,74 @@ 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]} }
-
-our %PFX2TERM_RMAP;
-my %meta_pfx = (mid => 1, thread => 1, path => 1);
-while (my ($k, $v) = each %all_pfx) {
- $PFX2TERM_RMAP{$v} = $k if $meta_pfx{$k};
-}
-
-my $mail_query = Search::Xapian::Query->new(xpfx('type') . 'mail');
+my $mail_query = Search::Xapian::Query->new('T' . '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,
+ }, $class;
+ if ($version >= 2) {
+ my $dir = "$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;
+ $self->{skel} = Search::Xapian::Database->new("$dir/skel");
+ } else {
+ $self->{xdb} = Search::Xapian::Database->new($self->xdir);
+ }
+ $self;
}
-sub reopen { $_[0]->{xdb}->reopen }
+sub reopen {
+ my ($self) = @_;
+ $self->{xdb}->reopen;
+ if (my $skel = $self->{skel}) {
+ $skel->reopen;
+ }
+ $self; # make chaining easier
+}
# read-only
sub query {
@@ -146,14 +193,14 @@ sub query {
sub get_thread {
my ($self, $mid, $opts) = @_;
- my $smsg = eval { $self->lookup_message($mid) };
+ my $smsg = retry_reopen($self, sub { lookup_skeleton($self, $mid) });
return { total => 0, msgs => [] } unless $smsg;
- my $qtid = Search::Xapian::Query->new(xpfx('thread').$smsg->thread_id);
+ my $qtid = Search::Xapian::Query->new('G' . $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);
+ my $qsub = Search::Xapian::Query->new('XPATH' . $path);
$qtid = Search::Xapian::Query->new(OP_OR, $qtid, $qsub);
}
$opts ||= {};
@@ -162,7 +209,7 @@ sub get_thread {
# always sort threads by timestamp, this makes life easier
# for the threading algorithm (in SearchThread.pm)
$opts->{asc} = 1;
-
+ $opts->{enquire} = enquire_skel($self);
_do_enquire($self, $qtid, $opts);
}
@@ -177,6 +224,7 @@ sub retry_reopen {
if (ref($@) eq 'Search::Xapian::DatabaseModifiedError') {
reopen($self);
} else {
+ warn "ref: ", ref($@), "\n";
die;
}
}
@@ -189,7 +237,7 @@ sub _do_enquire {
sub _enquire_once {
my ($self, $query, $opts) = @_;
- my $enquire = $self->enquire;
+ my $enquire = $opts->{enquire} || enquire($self);
if (defined $query) {
$query = Search::Xapian::Query->new(OP_AND,$query,$mail_query);
} else {
@@ -271,38 +319,123 @@ sub num_range_processor {
sub query_xover {
my ($self, $beg, $end, $offset) = @_;
my $qp = Search::Xapian::QueryParser->new;
- $qp->set_database($self->{xdb});
+ $qp->set_database($self->{skel} || $self->{xdb});
$qp->add_valuerangeprocessor($self->num_range_processor);
my $query = $qp->parse_query("$beg..$end", QP_FLAGS);
- _do_enquire($self, $query, {num => 1, limit => 200, offset => $offset});
+ my $opts = {
+ enquire => enquire_skel($self),
+ num => 1,
+ limit => 200,
+ offset => $offset,
+ };
+ _do_enquire($self, $query, $opts);
+}
+
+sub query_ts {
+ my ($self, $ts, $opts) = @_;
+ my $qp = $self->{qp_ts} ||= eval {
+ my $q = Search::Xapian::QueryParser->new;
+ $q->set_database($self->{skel} || $self->{xdb});
+ $q->add_valuerangeprocessor(
+ Search::Xapian::NumberValueRangeProcessor->new(TS));
+ $q
+ };
+ my $query = $qp->parse_query($ts, QP_FLAGS);
+ $opts->{enquire} = enquire_skel($self);
+ _do_enquire($self, $query, $opts);
+}
+
+sub lookup_skeleton {
+ my ($self, $mid) = @_;
+ my $skel = $self->{skel} or return lookup_message($self, $mid);
+ $mid = mid_clean($mid);
+ my $term = 'Q' . $mid;
+ my $smsg;
+ my $beg = $skel->postlist_begin($term);
+ if ($beg != $skel->postlist_end($term)) {
+ my $doc_id = $beg->get_docid;
+ if (defined $doc_id) {
+ # raises on error:
+ my $doc = $skel->get_document($doc_id);
+ $smsg = PublicInbox::SearchMsg->wrap($doc, $mid);
+ $smsg->{doc_id} = $doc_id;
+ }
+ }
+ $smsg;
}
sub lookup_message {
my ($self, $mid) = @_;
$mid = mid_clean($mid);
- my $doc_id = $self->find_unique_doc_id('mid', $mid);
+ my $doc_id = $self->find_first_doc_id('Q' . $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);
+ $smsg->{doc_id} = $doc_id;
}
$smsg;
}
sub lookup_mail { # no ghosts!
my ($self, $mid) = @_;
- my $smsg = lookup_message($self, $mid) or return;
- PublicInbox::SearchMsg->load_doc($smsg->{doc});
+ retry_reopen($self, sub {
+ my $smsg = lookup_skeleton($self, $mid) or return;
+ $smsg->load_expand;
+ });
+}
+
+sub lookup_article {
+ my ($self, $num) = @_;
+ my $term = 'XNUM'.$num;
+ my $smsg;
+ eval {
+ retry_reopen($self, sub {
+ my $db = $self->{skel} || $self->{xdb};
+ my $head = $db->postlist_begin($term);
+ return if $head == $db->postlist_end($term);
+ my $doc_id = $head->get_docid;
+ return unless defined $doc_id;
+ # raises on error:
+ my $doc = $db->get_document($doc_id);
+ $smsg = PublicInbox::SearchMsg->wrap($doc);
+ $smsg->load_expand;
+ $smsg->{doc_id} = $doc_id;
+ });
+ };
+ $smsg;
+}
+
+sub each_smsg_by_mid {
+ my ($self, $mid, $cb) = @_;
+ # XXX retry_reopen isn't necessary for V2Writable, but the PSGI
+ # interface will need it...
+ my $db = $self->{skel} || $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;
+ }
}
sub find_unique_doc_id {
- my ($self, $term, $value) = @_;
+ my ($self, $termval) = @_;
- my ($begin, $end) = $self->find_doc_ids($term, $value);
+ my ($begin, $end) = $self->find_doc_ids($termval);
return undef if $begin->equal($end); # not found
@@ -310,26 +443,30 @@ sub find_unique_doc_id {
# sanity check
$begin->inc;
- $begin->equal($end) or die "Term '$term:$value' is not unique\n";
+ $begin->equal($end) or die "Term '$termval' is not unique\n";
$rv;
}
# returns begin and end PostingIterator
sub find_doc_ids {
- my ($self, $term, $value) = @_;
+ my ($self, $termval) = @_;
+ my $db = $self->{xdb};
- $self->find_doc_ids_for_term(xpfx($term) . $value);
+ ($db->postlist_begin($termval), $db->postlist_end($termval));
}
-# returns begin and end PostingIterator
-sub find_doc_ids_for_term {
- my ($self, $term) = @_;
- my $db = $self->{xdb};
+sub find_first_doc_id {
+ my ($self, $termval) = @_;
+
+ my ($begin, $end) = $self->find_doc_ids($termval);
+
+ return undef if $begin->equal($end); # not found
- ($db->postlist_begin($term), $db->postlist_end($term));
+ $begin->get_docid;
}
# normalize subjects so they are suitable as pathnames for URLs
+# XXX: consider for removal
sub subject_path {
my $subj = pop;
$subj = subject_normalized($subj);
@@ -347,37 +484,20 @@ sub subject_normalized {
$subj;
}
-# for doc data
-sub subject_summary {
- my $subj = pop;
- my $max = 68;
- if (length($subj) > $max) {
- my @subj = split(/\s+/, $subj);
- $subj = '';
- my $l;
-
- while ($l = shift @subj) {
- my $new = $subj . $l . ' ';
- last if length($new) >= $max;
- $subj = $new;
- }
- if ($subj ne '') {
- my $r = scalar @subj ? ' ...' : '';
- $subj =~ s/ \z/$r/s;
- } else {
- # subject has one REALLY long word, and NOT spam? wtf
- @subj = ($l =~ /\A(.{1,72})/);
- $subj = $subj[0] . ' ...';
- }
- }
- $subj;
-}
-
sub enquire {
my ($self) = @_;
$self->{enquire} ||= Search::Xapian::Enquire->new($self->{xdb});
}
+sub enquire_skel {
+ my ($self) = @_;
+ if (my $skel = $self->{skel}) {
+ $self->{enquire_skel} ||= Search::Xapian::Enquire->new($skel);
+ } else {
+ enquire($self);
+ }
+}
+
sub help {
my ($self) = @_;
$self->qp; # parse altids