scripts/dupe-finder

#!/usr/bin/perl -w
# Copyright (C) 2018-2021 all contributors <meta@public-inbox.org>
# License: AGPL-3.0+ <https://www.gnu.org/licenses/agpl-3.0.txt>
#
# ad-hoc tool for finding duplicates, unstable!
use strict;
use warnings;
use PublicInbox::Inbox;
use PublicInbox::Over;
use PublicInbox::Search;
use PublicInbox::Config;
my $repo = shift;
my $ibx;
if (index($repo, '@') > 0) {
	$ibx = PublicInbox::Config->new->lookup($repo);
} elsif (-d $repo) {
	$ibx = { inboxdir => $repo, address => 'unnamed@example.com' };
	$ibx = PublicInbox::Inbox->new($ibx);
} else {
	$ibx = PublicInbox::Config->new->lookup_name($repo);
}
$ibx or die "No inbox";
$ibx->search or die "search not available for inbox";
my $over = $ibx->over;
my $dbh = $over->dbh;

sub emit ($) {
	my ($nums) = @_;
	foreach my $n (@$nums) {
		my $smsg = $over->get_art($n) or next;
		print STDERR "$n $smsg->{blob} $smsg->{mid}\n";
		my $msg = $ibx->msg_by_smsg($smsg) or next;
		print "From $smsg->{blob}\@$n Thu Jan  1 00:00:00 1970\n";
		$$msg =~ s/^(>*From )/>$1/gm;
		print $$msg, "\n";
	}
}

my $sth = $dbh->prepare(<<'');
SELECT id,num FROM id2num WHERE num > 0 ORDER BY id

$sth->execute;
my $prev_id = -1;
my ($id, $num, @nums);
while (1) {
	($id, $num) = $sth->fetchrow_array;
	defined $id or last;
	if ($prev_id != $id) {
		emit(\@nums) if scalar(@nums) > 1;
		@nums = ();
	}
	$prev_id = $id;
	push @nums, $num;
}