Seed recoording functionality

[aburke1605]

Added TNamed object to write the seed which was used to randomly sample the ntuple to the output files (rather than having to remember the seed).

[cofitzpa]

Need to have a think about this one:

• Typically I store the seed by piping the stdout to a text file. The current tool prints all its input arguments and output info for this reason. I consider the seed to be part of the metadata rather than the data itself.
• If you sample the output of a previously sampled ntuple, this loses the prior seed and adds the new one. Is that intended?