Concept for "compressed master ref" subset.

[vinjana]

Currently, the Hg37ChrConverter (legacy) and the GenericChrConverter encapsulate some of the handling of the global chromosome index space and the compressed-master ref subspace. Still, the client code has problems:

1. Client code uses use using declarations of unsigned long to model chromosome indices. This means, there is no type checking and a confusion of functions for one set with indices of the other set is always imminent.
2. The client code actually has to deal with this discrimination.
3. The chromosome converters are complex, but it is not clear, whether these two sets are really necessary to be handled as separate index sets for reource reasons (memory/CPU), or not.

To-Do

1. Clarify what is the semantic value of the compressed master refs for the SOPHIA algorithm? Is the concept actually needed?
2. If it is needed, try to encapsulate all mapping logic in a way that client code has not to deal much with it, in particular, if possible not on a line-by-line level. E.g. strongy typing may be sufficient (e.g. 2 classes of indices, or a class that has a "global" and a "compressedMref" view, or a hybrid class that incorporates the mapping between the two domains).