lei export-kw: new command to export keywords to Maildirs

IMAP will eventually be supported.

6 files changed, 228 insertions, 2 deletions

diff --git a/lib/PublicInbox/LEI.pm b/lib/PublicInbox/LEI.pm
index 15680fe3..628908b5 100644
--- a/lib/PublicInbox/LEI.pm
+++ b/lib/PublicInbox/LEI.pm
@@ -243,6 +243,10 @@ our %CMD = ( # sorted in order of importance/use:
         qw(stdin| offset=i recursive|r exclude=s include|I=s
         lock=s@ in-format|F=s kw! verbose|v+ incremental! mail-sync!),
         qw(no-torsocks torsocks=s), PublicInbox::LeiQuery::curl_opt(), @c_opt ],
+
+'export-kw' => [ 'LOCATION...|--all',
+        'one-time export of keywords of sync sources',
+        qw(all:s mode=s), @c_opt ],
 'convert' => [ 'LOCATION...|--stdin',
         'one-time conversion from URL or filesystem to another format',
         qw(stdin| in-format|F=s out-format|f=s output|mfolder|o=s lock=s@ kw!),
diff --git a/lib/PublicInbox/LeiExportKw.pm b/lib/PublicInbox/LeiExportKw.pm
new file mode 100644
index 00000000..db4f7441
--- /dev/null
+++ b/lib/PublicInbox/LeiExportKw.pm
@@ -0,0 +1,180 @@
+# Copyright (C) 2021 all contributors <meta@public-inbox.org>
+# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
+
+# front-end for the "lei export-kw" sub-command
+package PublicInbox::LeiExportKw;
+use strict;
+use v5.10.1;
+use parent qw(PublicInbox::IPC PublicInbox::LeiInput);
+use Errno qw(EEXIST ENOENT);
+
+sub export_kw_md { # LeiMailSync->each_src callback
+        my ($oidbin, $id, $self, $mdir) = @_;
+        my $oidhex = unpack('H*', $oidbin);
+        my $sto_kw = $self->{lse}->oid_keywords($oidhex) or return;
+        my $bn = $$id;
+        my ($md_kw, $unknown, @try);
+        if ($bn =~ s/:2,([a-zA-Z]*)\z//) {
+                ($md_kw, $unknown) = PublicInbox::MdirReader::flags2kw($1);
+                @try = qw(cur new);
+        } else {
+                $unknown = [];
+                @try = qw(new cur);
+        }
+        if ($self->{-merge_kw} && $md_kw) { # merging keywords is the default
+                @$sto_kw{keys %$md_kw} = values(%$md_kw);
+        }
+        $bn .= ':2,'.
+                PublicInbox::LeiToMail::kw2suffix([keys %$sto_kw], @$unknown);
+        my $dst = "$mdir/cur/$bn";
+        my @fail;
+        for my $d (@try) {
+                my $src = "$mdir/$d/$$id";
+                next if $src eq $dst;
+
+                # we use link(2) + unlink(2) since rename(2) may
+                # inadvertently clobber if the "uniquefilename" part wasn't
+                # actually unique.
+                if (link($src, $dst)) { # success
+                        # unlink(2) may ENOENT from parallel invocation,
+                        # ignore it, but not other serious errors
+                        if (!unlink($src) and $! != ENOENT) {
+                                $self->{lei}->child_error(1,
+                                                        "E: unlink($src): $!");
+                        }
+                        $self->{lms}->mv_src("maildir:$mdir",
+                                                $oidbin, $id, $bn) or die;
+                        return; # success anyways if link(2) worked
+                }
+                if ($! == ENOENT && !-e $src) { # some other process moved it
+                        $self->{lms}->clear_src("maildir:$mdir", $id);
+                        next;
+                }
+                push @fail, $src if $! != EEXIST;
+        }
+        return unless @fail;
+        # both tries failed
+        my $e = $!;
+        my $orig = '['.join('|', @fail).']';
+        $self->{lei}->child_error(1, "link($orig, $dst) ($oidhex): $e");
+}
+
+# overrides PublicInbox::LeiInput::input_path_url
+sub input_path_url {
+        my ($self, $input, @args) = @_;
+        my $lms = $self->{lms} //= $self->{lse}->lms;
+        $lms->lms_begin;
+        if ($input =~ s/\Amaildir://i) {
+                require PublicInbox::LeiToMail; # kw2suffix
+                $lms->each_src("maildir:$input", \&export_kw_md, $self, $input);
+        }
+        $lms->lms_commit;
+}
+
+sub lei_export_kw {
+        my ($lei, @folders) = @_;
+        my $sto = $lei->_lei_store or return $lei->fail(<<EOM);
+lei/store uninitialized, see lei-import(1)
+EOM
+        my $lse = $sto->search;
+        my $lms = $lse->lms or return $lei->fail(<<EOM);
+lei mail_sync uninitialized, see lei-import(1)
+EOM
+        my $opt = $lei->{opt};
+        my $all = $opt->{all};
+        my @all = $lms->folders;
+        if (defined $all) { # --all=<local|remote>
+                my %x = map { $_ => $_ } split(/,/, $all);
+                my @ok = grep(defined, delete(@x{qw(local remote), ''}));
+                my @no = keys %x;
+                if (@no) {
+                        @no = (join(',', @no));
+                        return $lei->fail(<<EOM);
+--all=@no not accepted (must be `local' and/or `remote')
+EOM
+                }
+                my (%seen, @inc);
+                for my $ok (@ok) {
+                        if ($ok eq 'local') {
+                                @inc = grep(!m!\A[a-z0-9\+]+://!i, @all);
+                        } elsif ($ok eq 'remote') {
+                                @inc = grep(m!\A[a-z0-9\+]+://!i, @all);
+                        } elsif ($ok ne '') {
+                                return $lei->fail("--all=$all not understood");
+                        } else {
+                                @inc = @all;
+                        }
+                        for (@inc) {
+                                push(@folders, $_) unless $seen{$_}++;
+                        }
+                }
+                return $lei->fail(<<EOM) if !@folders;
+no --mail-sync folders known to lei
+EOM
+        } else {
+                my %all = map { $_ => 1 } @all;
+                my @no;
+                for (@folders) {
+                        next if $all{$_}; # ok
+                        if (-d "$_/new" && -d "$_/cur") {
+                                my $d = 'maildir:'.$lei->rel2abs($_);
+                                push(@no, $_) unless $all{$d};
+                                $_ = $d;
+                        } else {
+                                push @no, $_;
+                        }
+                }
+                my $no = join("\n\t", @no);
+                return $lei->fail(<<EOF) if @no;
+No sync information for: $no
+Run `lei ls-mail-sync' to display valid choices
+EOF
+        }
+        my $self = bless { lse => $lse }, __PACKAGE__;
+        $lei->{opt}->{'mail-sync'} = 1; # for prepare_inputs
+        $self->prepare_inputs($lei, \@folders) or return;
+        my $j = $opt->{jobs} // scalar(@{$self->{inputs}}) || 1;
+        if (my @ro = grep(!/\A(?:maildir|imaps?):/, @folders)) {
+                return $lei->fail("cannot export to read-only folders: @ro");
+        }
+        if (my $net = $lei->{net}) {
+                require PublicInbox::NetWriter;
+                bless $net, 'PublicInbox::NetWriter';
+        }
+        undef $lms;
+        my $m = $opt->{mode} // 'merge';
+        if ($m eq 'merge') { # default
+                $self->{-merge_kw} = 1;
+        } elsif ($m eq 'set') {
+        } else {
+                return $lei->fail(<<EOM);
+--mode=$m not supported (`set' or `merge')
+EOM
+        }
+        my $ops = {};
+        $lei->{auth}->op_merge($ops, $self) if $lei->{auth};
+        $self->{-wq_nr_workers} = $j // 1; # locked
+        (my $op_c, $ops) = $lei->workers_start($self, $j, $ops);
+        $lei->{wq1} = $self;
+        $lei->{-err_type} = 'non-fatal';
+        net_merge_all_done($self) unless $lei->{auth};
+        $op_c->op_wait_event($ops); # calls net_merge_all_done if $lei->{auth}
+}
+
+sub _complete_export_kw {
+        my ($lei, @argv) = @_;
+        my $sto = $lei->_lei_store or return;
+        my $lms = $sto->search->lms or return;
+        my $match_cb = $lei->complete_url_prepare(\@argv);
+        map { $match_cb->($_) } $lms->folders;
+}
+
+no warnings 'once';
+
+*ipc_atfork_child = \&PublicInbox::LeiInput::input_only_atfork_child;
+*net_merge_all_done = \&PublicInbox::LeiInput::input_only_net_merge_all_done;
+
+# the following works even when LeiAuth is lazy-loaded
+*net_merge_all = \&PublicInbox::LeiAuth::net_merge_all;
+
+1;
diff --git a/lib/PublicInbox/LeiMailSync.pm b/lib/PublicInbox/LeiMailSync.pm
index 3bada42d..32e17c65 100644
--- a/lib/PublicInbox/LeiMailSync.pm
+++ b/lib/PublicInbox/LeiMailSync.pm
@@ -138,6 +138,16 @@ DELETE FROM blob2num WHERE fid = ? AND uid = ?
         $sth->execute($fid, $id);
 }
 
+# Maildir-only
+sub mv_src {
+        my ($self, $folder, $oidbin, $id, $newbn) = @_;
+        my $fid = $self->{fmap}->{$folder} //= _fid_for($self, $folder, 1);
+        my $sth = $self->{dbh}->prepare_cached(<<'');
+UPDATE blob2name SET name = ? WHERE fid = ? AND oidbin = ? AND name = ?
+
+        $sth->execute($newbn, $fid, $oidbin, $$id);
+}
+
 # read-only, iterates every oidbin + UID or name for a given folder
 sub each_src {
         my ($self, $folder, $cb, @args) = @_;
diff --git a/lib/PublicInbox/LeiSearch.pm b/lib/PublicInbox/LeiSearch.pm
index fb19229f..9297d060 100644
--- a/lib/PublicInbox/LeiSearch.pm
+++ b/lib/PublicInbox/LeiSearch.pm
@@ -27,6 +27,20 @@ sub msg_keywords {
         wantarray ? sort(keys(%$kw)) : $kw;
 }
 
+# returns undef if blob is unknown
+sub oid_keywords {
+        my ($self, $oidhex) = @_;
+        my @num = $self->over->blob_exists($oidhex) or return;
+        my $xdb = $self->xdb; # set {nshard};
+        my %kw;
+        for my $num (@num) { # there should only be one...
+                my $doc = $xdb->get_document(num2docid($self, $num));
+                my $x = xap_terms('K', $doc);
+                %kw = (%kw, %$x);
+        }
+        \%kw;
+}
+
 # lookup keywords+labels for external messages
 sub xsmsg_vmd {
         my ($self, $smsg, $want_label) = @_;
diff --git a/lib/PublicInbox/LeiToMail.pm b/lib/PublicInbox/LeiToMail.pm
index 0cbdff8b..96a1f881 100644
--- a/lib/PublicInbox/LeiToMail.pm
+++ b/lib/PublicInbox/LeiToMail.pm
@@ -243,10 +243,14 @@ sub _rand () {
         sprintf('%x,%x,%x,%x', rand(0xffffffff), time, $$, ++$seq);
 }
 
+sub kw2suffix ($;@) {
+        my $kw = shift;
+        join('', sort(map { $kw2char{$_} // () } @$kw, @_));
+}
+
 sub _buf2maildir {
         my ($dst, $buf, $smsg) = @_;
         my $kw = $smsg->{kw} // [];
-        my $sfx = join('', sort(map { $kw2char{$_} // () } @$kw));
         my $rand = ''; # chosen by die roll :P
         my ($tmp, $fh, $base, $ok);
         my $common = $smsg->{blob} // _rand;
@@ -263,7 +267,7 @@ sub _buf2maildir {
                 $dst .= 'cur/';
                 $rand = '';
                 do {
-                        $base = $rand.$common.':2,'.$sfx
+                        $base = $rand.$common.':2,'.kw2suffix($kw);
                 } while (!($ok = link($tmp, $dst.$base)) && $!{EEXIST} &&
                         ($rand = _rand.','));
                 die "link($tmp, $dst$base): $!" unless $ok;
diff --git a/lib/PublicInbox/MdirReader.pm b/lib/PublicInbox/MdirReader.pm
index 7a0641fb..304be63d 100644
--- a/lib/PublicInbox/MdirReader.pm
+++ b/lib/PublicInbox/MdirReader.pm
@@ -86,4 +86,18 @@ sub maildir_each_eml {
 
 sub new { bless {}, __PACKAGE__ }
 
+sub flags2kw ($) {
+        my @unknown;
+        my %kw;
+        for (split(//, $_[0])) {
+                my $k = $c2kw{$_};
+                if (defined($k)) {
+                        $kw{$k} = 1;
+                } else {
+                        push @unknown, $_;
+                }
+        }
+        (\%kw, \@unknown);
+}
+
 1;