Changelog:
- Refactor estimators_class.py so that all the calculations are more modular and easier to handle.
- Added AWSEM model to proteins.py and implemented fitting for fragment memory potentials and direct contact potentials.
- Added nonlinear model_loader for loading a model with nonlinear parameters.
Future Improvements:
- Test and optimize the new AWSEM loader.
- AWSEM models should be able to handle optimization on multiple parameters.
- Implement ddG fitting.
- Modify how observables are computed during optimization so that it is compatible with ddG optimization. Re-weight frames and throw all trajectories back into the observables object for recomputing the observable. This would allow us to handle more complicated observables, such as ddG, where it isn't as simple as re-weighting each term.