Trouble Replacing Text in HTML fragment using Mojo::DOM - perl

Question

I need to scan through html fragments looking for certain strings in text (not within element attributes) and wrapping those matching strings with a <span></span>. Here's a sample attempt with output:
use v5.10;
use Mojo::DOM;
my $body = qq|
<div>
<p>Boring Text:</p>
<p>
Highlight Cool whenever we see it.
but not here.
<code>
sub Cool {
print "Foo\n";
}
</code>
And here is more Cool.
</p>
</div>
|;
my $dom = Mojo::DOM->new($body);
foreach my $e ($dom->find('*')->each) {
my $text = $e->text;
say "e text is: $text ";
if ($text =~ /Cool/) {
(my $newtext = $text ) =~ s/Cool/<span class="fun">Cool<\/span>/g;
$e->replace_content($newtext);
}
}
say $dom->root;
the output:
e text is:
e text is: Boring Text:
e text is: Highlight Cool whenever we see it. but not. And here is more Cool.
e text is: here
e text is: sub Cool { print "Foo "; }
<div>
<p>Boring Text:</p>
<p>Highlight <span class="fun">Cool</span> whenever we see it. but not. And here is more <span class="fun">Cool</span>.</p>
</div>
Close but what I really want to see is something like the following:
<div>
<p>Boring Text:</p>
<p>Highlight <span class="fun">Cool</span> whenever we see it. but not here.
<code>
sub <span class="fun">Cool<span> {
print "Foo\n";
}
</code>
And here is more <span class="fun">Cool</span>.</p>
</div>
Any help / pointers would be greatly appreciated.
Thanks,
Todd

Answer 1

Having looked into XML::Twig I'm not so sure it's the correct tool. It's surprising how awkward such a simple task can be.
This is a working program that uses HTML::TreeBuilder. Unfortunately it doesn't produce formatted output so I've added some whitespace myself.
use strict;
use warnings;
use HTML::TreeBuilder;
my $html = HTML::TreeBuilder->new_from_content(<<__HTML__);
<div>
<p>Boring Text:</p>
<p>
Highlight Cool whenever we see it.
but not here.
<code>
sub Cool {
print "Foo\n";
}
</code>
And here is more Cool.
</p>
</div>
__HTML__
$html->objectify_text;
for my $text_node ($html->look_down(_tag => '~text')) {
my $text = $text_node->attr('text');
if (my #replacement = process_text($text)) {
my $old_node = $text_node->replace_with(#replacement);
$old_node->delete;
}
}
$html->deobjectify_text;
print $html->guts->as_XML;
sub process_text {
my #nodes = split /\bCool\b/, shift;
return unless #nodes > 1;
my $span = HTML::Element->new('span', class => 'fun');
$span->push_content('Cool');
for (my $i = 1; $i < #nodes; $i += 2) {
splice #nodes, $i, 0, $span->clone;
}
$span->delete;
#nodes;
}
output
<div>
<p>Boring Text:</p>
<p>
Highlight <span class="fun">Cool</span> whenever we see it.
but not here.
<code> sub <span class="fun">Cool</span> { print "Foo "; } </code>
And here is more <span class="fun">Cool</span>.
</p>
</div>

Answer 2

Here is a start using XML::Twig. One issue is the literal newline inside <code> tag. I guess that the parser cannot see the difference between it and a normal one. Perhaps it would help to encode it as &#10 or use CDATA sections. Otherwise I don't know how to handle it:
Content of script.pl:
#!/usr/bin/env perl
use warnings;
use strict;
use XML::Twig;
my $body = qq|
<div>
<p>Boring Text:</p>
<p>
Highlight Cool whenever we see it.
but not here.
<code>
sub Cool {
print "Foo\n";
}
</code>
And here is more Cool.
</p>
</div>
|;
XML::Twig::Elt::set_replaced_ents(q{});
my $elt = XML::Twig::Elt->new( 'span' => { class => 'fun' }, 'Cool' );
my $twig = XML::Twig->new( pretty_print => 'nice' )->parse( $body );
$twig->subs_text( 'Cool', $elt->sprint );
$twig->print;
Running it like:
perl script.pl
It yields:
<div>
<p>Boring Text:</p>
<p>
Highlight <span class="fun">Cool</span>
whenever we see it.
but not here.
<code>
sub <span class="fun">Cool</span>
{
print "Foo
";
}
</code>
And here is more <span class="fun">Cool</span>
.
</p>
</div>