add `skip` argument to initializeRecombinationRateFromFile

[petrelharp]

Currently, initializeRecombinationRateFromFile has

No column header line should be present; the file should start immediately with numerical data.

However, given the many ways to go wrong with intervals and genetic map data (is the "position" column a starting or ending position, for instance?), it's pretty important to have a header in the file. If there was a "skip" argument (that means "skip this many lines before looking for numerical data"), then we could allow headers.

I also see that if there's a row in the file for which the position is the same as the lastPosition argument, then we get an error ("requires ends to be in strictly ascending order"). It's also good practice for files to contain the endpoint, so maybe there can be an exception: if the last position is the same as lastPosition, then the error doesn't occur?

[bhaller]

Do you care whether I fix this now (i.e., for the SLiM 6 beta) or later?

[petrelharp]

Nope.