Parallelalize grew transform

[theatollersrud]

Running grew transform on large treebanks is quite slow at the moment (152 seconds), and I was wondering if it would be possible to implement some kind of batch processing in grew transform to speed things up?

[bguil]

@theatollersrud: Thanks for the suggestion. I agree it would be a useful feature, but I don't plan to work on it any time soon.
I have turned this into a discussion to keep the idea open.

In the short term, I suggest parallelising the execution outside of Grew, for example by calling Grew in parallel on different subcorpora.