NAME
HTML::HTML5::ToText - convert HTML to plain text
SYNOPSIS
my $dom = HTML::HTML5::Parser->load_html(IO => \*STDIN);
print HTML::HTML5::ToText
->with_traits(qw/ShowLinks ShowImages RenderTables/)
->new()
->process($dom);
DESCRIPTION
The HTML::HTML5::ToText module itself produces a pretty boring
conversion of HTML to text, but thanks to Moose and MooseX::Traits it
can easily be composed with "traits" that improve the output.
Compositor
"with_traits(@traits)"
This class method creates a new class that composes
"HTML::HTML5::ToText" with each trait given, returning the name of
that class. That class will be a subclass of "HTML::HTML5::ToText".
Traits are taken to be in the "HTML::HTML5::ToText::Trait" namespace
unless overridden by prefixing the trait with "+".
Constructors
* "new(%attrs)"
Creates a new instance of the class.
* "new_with_traits(traits => \@traits, %attrs)"
Shortcut for:
HTML::HTML5::ToText->with_traits(@traits)->new(%attrs)
Attributes
As per usual for Moose classes, accessor methods are provided for each
attribute, and attributes may be set in the constructor.
"HTML::HTML5::ToText" does not actually provide any attributes, but some
traits may.
Methods
* "process($node)"
Processes an XML::LibXML::Node and returns a string. May be called
as a class or object method.
Because "process" likes to perform some alterations to the DOM tree,
as a first stage it makes a clone of the DOM tree (so that it can
leave the original intact). If you don't care about any changes to
the tree, and want to save a bit of CPU, then you can suppress the
cloning by passing a true value as a second argument to "process".
HTML::HTML5::ToText->process($node, 'no_clone')
* "process_string($string)"
As per "process", but first parses the string with
HTML::HTML5::Parser. The second argument (for cloning) does not
exist as cloning is not needed in this case.
There are also methods named (in upper-case) after every element defined
in HTML5: "STRONG($node)", "DL($node)", "IMG($node)" and so on, which
"process($node)" delegates to; and a "textnode($node)" method which is
the equivalent for text nodes. These are the methods which traits tend
to modify.
EXTENDING
MooseX::Traits makes it pretty easy to cleanly extend this module. Say
for example, we want to add the feature where the HTML "<del>" element
is output as the empty string. (The default behavious treats it rather
like "<div>".)
{
package Local::SkipDEL;
use Moose::Role;
override DEL => sub { '' };
}
print HTML::HTML5::ToText
-> with_traits(qw/ShowLinks ShowImages +Local::SkipDEL/)
-> process_string($html);
Or maybe we want to force "<big>" elements into uppercase?
{
package Local::Embiggen;
use Moose::Role;
around BIG => sub
{
my ($orig, $self, $elem) = @_;
return uc $self->$orig($elem);
};
}
print HTML::HTML5::ToText
-> with_traits(qw/+Local::Embiggen/)
-> process_string($html);
Share your examples of extending HTML::HTML5::ToText at
<https://bitbucket.org/tobyink/p5-html-html5-totext/wiki/Extending>.
BUGS
Please report any bugs to
<http://rt.cpan.org/Dist/Display.html?Queue=HTML-HTML5-ToText>.
SEE ALSO
HTML::HTML5::Parser, HTML::HTML5::Table.
HTML::HTML5::ToText::Trait::RenderTables,
HTML::HTML5::ToText::Trait::ShowImages,
HTML::HTML5::ToText::Trait::ShowLinks,
HTML::HTML5::ToText::Trait::TextFormatting.
Similar Modules on CPAN
* HTML::FormatText
About 15 years old, and still maintained, this falls into the
"mature" category. This module is based on HTML::Tree, so its HTML
parser may not behave as closely to modern browsers as
HTML::HTML5::Parser's parsing, but its conversion to text seems
somewhat better than HTML::HTML5::ToText's default output (i.e. with
no traits applied).
At the time of writing, its bug queue on rt.cpan.org lists eight
bugs, some quite serious. However, since being taken over by its
latest maintainer, there seems to be progress being made on them.
Fairly extensible, but not in the mix-and-match traits way allowed
by HTML::HTML5::ToText.
* HTML::FormatText::WithLinks
An extension of HTML::FormatText.
* HTML::FormatText::WithLinks::AndTables
An extension of HTML::FormatText::WithLinks.
The code that deals with tables is pretty crude compared with
HTML::HTML5::ToText::Trait::RenderTables. It doesn't support
"colspan", "rowspan", or the "<th>" element.
* LEOCHARRE::HTML::Text
Very basic conversion; basically just tag stripping using regular
expressions.
* HTML::FormatExternal
Passes HTML through external command-line tools such as `lynx`.
Obviously this has limited portability.
AUTHOR
Toby Inkster <tobyink@cpan.org>.
THANKS
Everyone behind Moose. No way I could have done all this in a few hours
without Moose's strange brand of meta-programming!
COPYRIGHT AND LICENCE
This software is copyright (c) 2012-2013 by Toby Inkster.
This is free software; you can redistribute it and/or modify it under
the same terms as the Perl 5 programming language system itself.
DISCLAIMER OF WARRANTIES
THIS PACKAGE IS PROVIDED "AS IS" AND WITHOUT ANY EXPRESS OR IMPLIED
WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF
MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE.