wrap union to deal with SLiM metadata

[mufernando]

e.g., by adjusting the slim mutation ids in the metadata, and the generation

[petrelharp]

Things to do here are:

• make SLiM mutation IDs not overlap
• make SLiM individual IDs not overlap (and also adjust genome IDs accordingly)

... and, what exactly needs to happen with the times/generations? Anything else?

[mufernando]

never mind about the generation, this has to do with having to adjust the times if the two simulations were run for different amount of times. we could decide to deal with this within the pyslim.union though.

[petrelharp]

Right - so we need a

• dt argument that allows for shifting of the times if the two ts were run for different amounts of time

[petrelharp]

Also see #105