Different sorting methods for sample set dictionaries

**Current Problem**
When a `sampleset` is output by the CQM solver, the dictionary is sorted in lexicographical order. Thus reading the values with `sampleset.first.sample.values` may not agree with the order of variables input into the problem, creating confusion if we want to independently verify the feasibility of a solution.

**Proposed Solution**
An optional argument to `sample_cqm()` which sets the order of the sample set's dictionaries. Alternatively we could instead just keep the original order the variables were input into the model; maybe something like `maintain_order = True`

**Alternatives Considered**
A naive workaround is to simply store the order of variables from the original input and create a new sample set dictionary using the lexicographical sample set's dictionary as a key.

**Additional context**
Several customers have run into this in the past