I am looking for a web based rss/rdf/atom aggregator, but since I could not find anything matching my requirements small and must export xml files, I decided to write my own.
On CPAN I found XML::Feed, which does exactly what I want namely repepresent feeds as perl hashes:
The goal of XML::Feed is to provide a unified API for parsing and using the various syndication formats. The different flavors of RSS and Atom handle data in different ways: date handling; summaries and content; escaping and quoting; etc. This module attempts to remove those differences by providing a wrapper around the formats and the classes implementing those formats (XML::RSS and XML::Atom::Feed). For example, dates are handled differently in each of the above formats. To provide a unified API for date handling, XML::Feed converts all date formats transparently into DateTime objects, which it then returns to the caller.
The module has a list of dependencies, but with the help of dh-make-perl it is pretty easy to create debian packages directly from cpan modules.