Generalize StructureAlignment Visualization

[lafita]

Related to #393. The idea is to extract selectors from the alignment data structures in BioJava (AFPChain and MultipleAlignment) so that any molecular viewer can represent structure alignments. A good solution could be to extract ResidueNumbers, which can then be serialized and parsed by the viewers.

[lafita]

Steps to be made:

1. All classes that are Jmol specific should be moved to structure-gui module and must contain the word Jmol in their name.
2. All methods and classes that are not Jmol specific should be moved to the structure module, should never contain the word Jmol in their name and should contain the word Display (to differentiate from Utils and Tools).
3. The structure-gui module should contain as little algorithmic part as possible (only build the BioJava-Jmol connection), and all the current features implemented should be able to be reproduced by calling methods in the Display classes of the structure module.

[sbliven]

StructureIdentifier already provides more-or-less the residue-based selector functionality you're describing. You would just need to write a method that converts a SubstructureIdentifier into a Jmol selection string.

This also nicely complements the long-range plans discussed in #81 for a successor to SubstructureIdentifier that would be capable of arbitrary selections, including additional properties and basic set operations.

[lafita]

Yes, it is currently what we are using for symmetric repeats in internal symmetry.
For alignments it would be a bit more involved to include gaps in the Identifier Ranges, but I like the idea of a method to convert a StructureIdentifier into a Jmol selector (or any kind of selector).

[andreasprlic]

+1