Different mutators

[karnawhat]

Similar to how we have many parsers to parse different files (i.e. CSV, JSON, or SDF) or different preprocessors to parse the dataset into different features (i.e. sequence-baaed or mol-based), we should allow for different mutators to mutate a structure/sequence.

For example, it makes sense to have a BaseMutator and classes that extend this such as SequenceMutator and StructureMutator to mutate sequence and structures, respectively. This would allow for changes in using different mutators without specifying the exact type within the PDBMutator instance parameters (as shown below)

https://github.com/ayushkarnawat/profit/blob/31f3ef7e8910af90c86a74ef5dc5376ce02355dc/profit/dataset/preprocessing/mutator.py#L88-L90

Rather, we can call the exact type of mutator from a list of available mutators. See #7 for more information on how this could also support the mutations of arbitrary sequences/structures.

Note: Idea is still WIP...
Additionally, this is likely a breaking change since it affects how the dataset will be mutated within the parsers.

[karnawhat]

Note that it is important to list all available mutator within preprocessing/__init__.py file. This would make implementation of calling different mutators similar to how different preprocessors are handled:

https://github.com/ayushkarnawat/profit/blob/31f3ef7e8910af90c86a74ef5dc5376ce02355dc/profit/dataset/preprocessors/__init__.py#L15-L18

For brevity, this will also affect how we handle which preprocessors can work with different mutators.
https://github.com/ayushkarnawat/profit/blob/31f3ef7e8910af90c86a74ef5dc5376ce02355dc/profit/dataset/parsers/base_parser.py#L24-L31

It will also affect how the call is made when trying to load the data (to be updated...)