Finish converting calcstate to use DataSet/DataFile framework, add data file 'groupby' options

[drroe]

This finishes the conversion to DataSet/DataFile framework for calcstate begun in #629, converting the transitions data into DataSets that can use DataFiles. Since the transitions have a different dimension than the states data (there is more transition data than state data), this PR also adds an option to standard data file writes to group data in different ways.

	groupby <type> : (1D) group data sets by <type>.
		name   : Group by name.
		aspect : Group by aspect.
		idx    : Group by index.
		ens    : Group by ensemble number.
		dim    : Group by dimension.

The groupby option controls what data goes side by side in standard data file output. For example, normally all data sets in a standard data file are printed in columns - if a data set does not contain data for a given index, a default blank value is printed. For example, if running calcstate with 2 states, the state data will have size 3 (the 2 states and Undefined) while the transitions will probably be larger.

calcstate state 1,d1,3.0,4.0 state 2,a1,100,120 out state.dat curveout curve.agr \
  stateout States.dat transout States.dat name d1_a1

The output file may look like so:

#d1_a1[Nlifetimes] d1_a1[Avglife] d1_a1[Maxlife] d1_a1[Name] d1_a1[Xlifetimes] d1_a1[Xavglife] d1_a1[Xmaxlife] d1_a1[Xname]
                21         3.5238             10   Undefined                19          3.4737              10 Undefined->1
                19         1.3158              3           1                 1          1.0000               1 Undefined->2
                 1         1.0000              1           2                19          1.3158               3 1->Undefined
                 0         0.0000              0      NoData                 1          1.0000               1 2->Undefined

Here the state data was size 3, but the transitions data was size 4, so for the last line the state data is all 0 and the state name is NoData (blank string value). This doesn't look nice. It makes more sense in this case to group by dimension.

calcstate state 1,d1,3.0,4.0 state 2,a1,100,120 out state.dat curveout curve.agr \
  stateout States.dat transout States.dat name d1_a1
datafile States.dat groupby dim

Now the output file looks like this:

#d1_a1[Nlifetimes] d1_a1[Avglife] d1_a1[Maxlife] d1_a1[Name]
                21         3.5238             10   Undefined
                19         1.3158              3           1
                 1         1.0000              1           2

#d1_a1[Xlifetimes] d1_a1[Xavglife] d1_a1[Xmaxlife] d1_a1[Xname]
                19          3.4737              10 Undefined->1
                 1          1.0000               1 Undefined->2
                19          1.3158               3 1->Undefined
                 1          1.0000               1 2->Undefined

Much easier to read.