Run cascaded parsing within Java

[eponvert]

And remove chunk.py

This will facilitate operating upparse as a library