Replacing an HTML element with multiple elements using HTML::TreeBuilder

mldvx4 has asked for the wisdom of the Perl Monks concerning the following question:

I'd like to replace selected paragraphs in place but am not figuring out the right way of doing so. Specifically I'd like to replace a specific, single P with multiple P at the same point in the tree. I've tried many dozen variations of the below code, but what I have gives an error, "the target node's parent has no content!?"

I don't understand. The replace_with_content or push_content methods should have established something for postinsert to add to. Clearly I have missed something? What though?

#!/usr/bin/perl

use HTML::TreeBuilder::XPath;

use warnings;
use strict;

&readfile;

exit(0);

sub readfile {
    my ($file)= (@_);

    my $xhtml = HTML::TreeBuilder::XPath->new;
    $xhtml->implicit_tags(1);
    $xhtml->no_space_compacting(1);
    $xhtml->parse_file(\*DATA) or die();

    # find double-spaced paragraphs inside blockquotes and expand them
    for my $p ($xhtml->findnodes('//blockquote/p')) {
    my $text = $p->as_text();
    
    $text =~  s/^\s+//; $text =~ s/\s+$//;
    next unless($text =~/\n\s*\n\s*/);
    
    my @paragraphs = split(/\s*\n\s*/, $text);
    print qq(\t\@paragraphs=),join(',',@paragraphs),qq(\n);
    
    if ($#paragraphs >= 0) {
        my $pp = shift(@paragraphs);
        print qq(\t\tpp1=$pp\n);
        $p->replace_with_content();
        $p->push_content(['p',,$pp]);
        
        print qq(Identified :\n);
        print qq(«),$p->as_XML_indented,qq(»\n);

        foreach $pp (@paragraphs) {
        print qq(\t\tpp2=$pp\n);
        $p->postinsert(['p',,$pp]);
        }
    }
    }

    print qq(\n),qq(-)x30,qq(\n);
    
    my ($body) = $xhtml->findnodes('//body');
    print qq(\n);
    print $body->as_XML_indented;
    $xhtml->delete;
    return (1);
}


__DATA__

<body>
<blockquote id="one">
aaa

bbb

ccc
</blockquote>

<blockquote id="two">
<p>
ddd

eee

fff
</p>
</blockquote>

<blockquote id="three">
<p>
ggg
 </p> <p>
hhh
 </p> <p>
iii
</p>
</blockquote>

<blockquote id="four">
<p>
jjj
</p>
</blockquote>

</body>
[download]

The expected output would be for BLOCKQUOTE number two to contain three separate paragraphs instead of one (or four). The other P in the other BLOCKQUOTE elements should continue to be left alone, as the script currently does.

Comment on Replacing an HTML element with multiple elements using HTML::TreeBuilder Download Code

Replies are listed 'Best First'.
Re: Replacing an HTML element with multiple elements using HTML::TreeBuilder by marto (Cardinal) on Jul 03, 2019 at 09:23 UTC
Thanks for the example, it'd be easier to understand what you want if you showed us the HTML output you require, based on the input in your example.	[reply]
Re^2: Replacing an HTML element with multiple elements using HTML::TreeBuilder by mldvx4 (Hermit) on Jul 03, 2019 at 13:33 UTC
I'm aiming for some output like the following, not counting whitespace: `<blockquote id="two"> <p>ddd</p> <p>eee</p> <p>fff</p> </blockquote>` [download] The one P would be replaced with three new P elements.	[reply] [d/l]
Re: Replacing an HTML element with multiple elements using HTML::TreeBuilder by choroba (Cardinal) on Jul 03, 2019 at 14:20 UTC
Here's how to do your task in XML::XSH2, a wrapper around XML::LibXML I happen to maintain: `open file.xml ; for my $b in //blockquote[count(p)=1] { my $texts = xsh:split("\n\n", $b/p) ; if (count($texts) > 1) { for my $text in $texts { my $p := insert element p append $b ; insert text normalize-space($text) into $p ; } rm $b/p[1] ; } } save :b ;` [download] `map{substr$_->[0],$_->[1]\|\|0,1}[\\|\|{},3],[[]],[ref qr-1,-,-1],[{}],[sub{}^ARGV,3]`	[reply] [d/l] [select]
Re: Replacing an HTML element with multiple elements using HTML::TreeBuilder by tangent (Parson) on Jul 03, 2019 at 16:08 UTC
Not sure if I understand your requirements fully, but this works for me: `for my $p ($xhtml->findnodes('//blockquote/p')) { my $text = $p->as_text(); $text =~ s/^\s+//; $text =~ s/\s+$//; if ( $text =~/\n\s\n\s/ ) { my @paragraphs = split(/\s\n\s/, $text); my @new_elems; for my $para (@paragraphs) { my $new = HTML::Element->new('p'); $new->push_content($para); push(@new_elems, $new); } $p->replace_with(@new_elems); } }` [download]	[reply] [d/l]
Re^2: Replacing an HTML element with multiple elements using HTML::TreeBuilder by mldvx4 (Hermit) on Jul 03, 2019 at 17:52 UTC
Thanks. I had tried several approaches using `replace_with`, but each met with various kinds of failure. Your example works and gives me a bit more of an idea how to use the methods.	[reply] [d/l]
Re: Replacing an HTML element with multiple elements using HTML::TreeBuilder by daxim (Curate) on Jul 03, 2019 at 12:49 UTC
Treebuilder API is cancer. Use something better. `use Web::Query::LibXML qw(); my $w = Web::Query->new_from_html(<<'HTML'); <body> … HTML $w->find('blockquote p')->each(sub { my @parts = split ' ', $_->text; if (@parts > 1) { for my $p (reverse @parts) { $_->after("<p>$p</p>"); } $_->remove; } }); print $w->as_html; __END__ … <blockquote id="two"><p>ddd</p><p>eee</p><p>fff</p></blockquote> …` [download]	[reply] [d/l]
Re^2: Replacing an HTML element with multiple elements using HTML::TreeBuilder by Anonymous Monk on Jul 04, 2019 at 04:01 UTC
daxim: Treebuilder API is cancer. Use something better. No it isn't. It is neither damaging nor does it spread. Sugar is nice, unless it rots your attitude :P	[reply]
Re: Replacing an HTML element with multiple elements using HTML::TreeBuilder by skleblan (Sexton) on Jul 03, 2019 at 14:19 UTC
I’m still working on understanding your code, but doesn’t `\s` include newline and carriage return characters? Whitespace	[reply] [d/l]
Re^2: Replacing an HTML element with multiple elements using HTML::TreeBuilder by mldvx4 (Hermit) on Jul 03, 2019 at 14:46 UTC
Yes, \s covers newlines as well as other whitespace. That part of the code works for my needs. The part that I do not know how to do is getting the whole original P element replaced by multiple new P elements. Perhaps I should write a more focused example.	[reply]