<rss><channel>
<item><title>this</title><description>more about
this</description></item>
<item><title>that</title><description>more about
that</description></item>
</channel></rss>
Each item is a separate "document", and at any given time a single file
contains a collection of items. Over time, I'll accumulate many
instances of a given feed, which will have overlapping items with
previous instances.
At first glance, I think I need to use AI::Categorizer::Document::XML
and maybe AI::Categorizer::Collection::SingleFile. Or do I have to
roll my own AI::Categorizer::Collection::XML class and maybe a special
document class to work with it?
Any advice before I start code-diving would be appreciated.