Add a best effort internal profiler

Accuracy is not important here. just subtracting nanotime should be good enough. We don't care about IO blocking or wall vs cpu time etc.

I would like to time the following:

Plain Traversal (Total and Per File)
Generation (Total, Per Generator, Per File)
FQN conversions (Total, and Per File)

This is related to, but not the same as #5

This will be targeted at the user, not the library.

For example, It would be nice to get an output such as

Scalagen Timings.
Total: 5s
Traversal: 3s (200 files/s)
Generation: 1.5s 
 - org.simulacrum.typeclass: 900ms
 - org.scalameta.ast: 100ms
 - org.freesio.free: 500ms
FQN: 0.5s

These would be outputted to SBT.

This would allow the Author's to optimize easily. It would also allow us to get an idea of how scalagen affects various projects.

scalameta / scalagen

Add a best effort internal profiler #33