Documented configuration file as the default way of running experiments; other small documentation changes and corrections

script/fusion_llr.py was not getting installed with setuptools.

Fixed load_resource to be less critical and always choose one of the registered resources, if possible; updated setup.py

Added single scripts to run preprocessing, extraction, enrollment and scoring separately; Fixed command line use None by default

First version of (locally) running verify script and some dummy implementations of database, preprocessor, extractor and algorithm