Describe sample groups

The default behavior places all samples in the same group.  There already is support for having multiple groups, where the network similarity constraints are only applied within a group.  This needs to be documented, and an example dataset should be included.

See the `LoadTerminalFiles` function in `ConstrainedMultiSample.py`
