Let me just also add that my standard for accuracy is CPAN. If the module can successfully parse 99.9% of the code in CPAN, I consider it accurate enough, although improvements to accuracy are welcome.

With that in mind I would love it if someone could suggest a way of getting hold of all of the Perl code in CPAN without having to install/compile etc...

It's a big ask, but I'm not the sort to back down from a challenge.