update on result storage:

- spectral data stored to hdf5 format
- added joblib-launcher support for parallel batch processing
- added file lock for parallel batch processing
- adapted single value storage
- new hydra callbacks for various purposes + generic template in callbacks.py

- misc./clean-up