26 lines
637 B
Plaintext
26 lines
637 B
Plaintext
|
#!/usr/bin/perl
|
||
|
|
|
||
|
|
# missing: Sanitize html
|
||
|
|
|
||
|
|
use strict;
|
||
|
|
use warnings;
|
||
|
|
use Rss2Html::Scrubber;
|
||
|
|
|
||
|
|
use DBI;
|
||
|
|
|
||
|
|
my $dbh = DBI->connect("dbi:SQLite:dbname=rss2html.sqlite", "", "");
|
||
|
|
$dbh->{sqlite_unicode} = 1;
|
||
|
|
my $scrubber = Rss2Html::Scrubber->new();
|
||
|
|
my $item
|
||
|
|
= $dbh->selectrow_hashref(
|
||
|
|
"select link, items.title as item_title, content, items.id as item_id
|
||
|
|
from items
|
||
|
|
where item_id=?
|
||
|
|
",
|
||
|
|
{ Slice => {} }, $ARGV[0]);
|
||
|
|
my $content = $scrubber->scrub($item->{content});
|
||
|
|
open my $fh1, '>:encoding(UTF-8)', "$ARGV[0].raw.html";
|
||
|
|
print $fh1 $item->{content};
|
||
|
|
open my $fh2, '>:encoding(UTF-8)', "$ARGV[0].scrubbed.html";
|
||
|
|
print $fh2 $content;
|