Allow for the configuration of the scrypt work factor used for encryption

[sidit77]

I made some experimental changes that allow the use to have more control over what work factor is use for scrypt during encryption.
The documentation is currently lacking, but I wanted to see if you're generally on board with these changes before putting more effort into it.

[str4d]

Thanks for the PR! I'm generally interested in this change (and supportive of enabling this kind of tunability in the library API), but I suspect it will conflict with a major refactor that needs doing, to remove the split between the passphrase and recipient encryption APIs (due to how Go age implements passphrase support and wants to extend it). I've been putting off this refactor because it's not going to be fun 😣 but I'll look at working on it in the coming months.

[sidit77]

Don't worry. I can just keep using my fork until you have a clearer picture about the redesign. I really hope that everything end up clicking into place as painlessly as possible.