'CPAN::Mini::Extract' provides a base for implementing systems that
download "all" of CPAN, extract the dists and then process the files
within.

It provides the same synchronisation functionality as CPAN::Mini
except that it also maintains a parallel directory tree that contains a
directory located at an identical path to each archive file, with a
controllable subset of the files in the archive extracted below.

How does it work
'CPAN::Mini::Extract' starts with a CPAN::Mini local
mirror, which it will optionally update before each run. Once the
CPAN::Mini directory is current, it will scan both directory
trees, extracting any new archives and removing any extracted archives
no longer in the minicpan mirror.