From: Dobrica Pavlinusic <dpavlin@rot13.org>
Date: Fri, 12 Nov 2010 19:58:59 +0000 (+0100)
Subject: export published items from reblog
X-Git-Url: http://git.rot13.org/?a=commitdiff_plain;h=e4b48f19744c93b86610c016b0936d95a9f5ee1f;hp=-c;p=NoSQL-toys.git

export published items from reblog
---

e4b48f19744c93b86610c016b0936d95a9f5ee1f
diff --git a/reblog/reblog-published.pl b/reblog/reblog-published.pl
new file mode 100755
index 0000000..947dd01
--- /dev/null
+++ b/reblog/reblog-published.pl
@@ -0,0 +1,102 @@
+#!/usr/bin/perl -w
+
+use strict;
+use DBI;
+use Data::Dump qw/dump/;
+
+$|++;
+
+my $debug = @ARGV ? 1 : 0;
+
+my $database = 'reblog';
+
+my $dbi = "DBI:mysql:database=$database";
+$dbi .= ";host=127.0.0.1;port=13306";	# XXX over ssh
+
+my $dbh = DBI->connect($dbi,"","",{ RaiseError => 1 });
+
+$dbh->do(qq{
+	create temporary table published_items as
+	select
+		item_id
+	from
+		items_userdata
+	where
+		label = 'published' and
+		value_numeric = 1
+});
+
+my $sql = qq{
+	select
+		i.id as item_id,
+--		i.guid as _id,
+		i.link as _id,
+		i.*,
+		f.url as feed_url,
+		f.title as feed_title,
+		f.link as feed_link,
+		f.description as feed_description
+	from items i
+	join published_items p on i.id = p.item_id
+	join feeds f on i.feed_id = f.id
+	where i.id > ?
+	order by i.id asc
+	limit 1000
+};
+
+my $sql_tags = qq{
+select
+	items_userdata.item_id,
+	value_long as tags,
+	timestamp
+from items_userdata
+join published_items p
+	on items_userdata.item_id = p.item_id and label='tags'
+where
+	items_userdata.item_id > ?
+order by items_userdata.item_id asc
+};
+
+my $last_row = 0; # FIXME
+$last_row = 0 if $debug;
+
+print "Fetching items from $dbi id > $last_row\n";
+
+my $sth = $dbh->prepare($sql);
+$sth->execute( $last_row );
+
+my @columns = @{ $sth->{NAME} };
+warn dump( @columns );
+
+my @feed = grep { s/^feed_// } @columns;
+
+print "found ",$sth->rows," items to process...\n";
+
+my $sth_tags = $dbh->prepare($sql_tags);
+$sth_tags->execute( $last_row );
+print "found ",$sth_tags->rows, " tags found...\n";
+
+my $count = 0;
+
+my $row_tags = $sth_tags->fetchrow_hashref();
+
+while (my $row = $sth->fetchrow_hashref() ) {
+	my $_id = $row->{_id} || "c$count";
+	$_id =~ s{\W+}{_}g;
+	$_id =~ s{_+$}{};
+
+	my $feed;
+	$feed->{$_} = delete $row->{ "feed_$_" } foreach @feed;
+	$row->{feed} = $feed;
+
+	while ( $row_tags && $row_tags->{item_id} < $row->{item_id} ) {
+		$row_tags = $sth_tags->fetchrow_hashref();
+		warn "## got tags: ",dump( $row_tags ) if $debug;
+		$row->{tags} = [ split(/\s+/, $row_tags->{tags} ) ];
+	}
+
+	warn dump($row);
+
+}
+
+