Massive refactor in python

[mnlevy1981]

I like the schema for the YAML files that describe the analysis to be done (e.g. CORE_vs_JRA.yml) but the python code that parses it (specifically analysis_class.py, data_source_classes.py, and generic_classes.py) all have remnants of previous iterations of the YAML file scattered throughout. I think a top-down rewrite should focus on the following:

1. better class names (already a ticket via Better names for classes in code #2 and I'll comment on that issue with some thoughts)

2. AnalysisCategory._open_datasets() should be far slimmer. No need for code like

        if 'climo' in self.operation:
            if 'ann_climo' in self.operation:
                AnalysisElement.climo = 'ann_climo'
            elif 'mon_climo' in self.operation:
                AnalysisElement.climo = 'mon_climo'
            else:
                raise ValueError("'{}' is not a valid operation".format(self.operation))
   

   when we can manage all of that in an Operators class (see how xcollection handles operators for motivation)