improve geocoding process flow and modularity

[kollerbud]

current workflow to add new addresses format processor is a bit convoluted. Example, changes need to be made to process
"W BRAYTON ST & S STATE ST (0 E) & W 126 PL (120 W)", in order to geocode:

1. location_format_processing.py: potentially new Regex pattern, then add new type to "LocationFormat", new function to break down to string to individual streets
2. location_geocoding.py: update LocationGeocoder to accept the new LocationFormat type, then function to output (potentially) new Geometry
3. geocoder_api/geocoder_local: also output Geometry shape

Idea:
consolidate responsibilities between scripts, consistent with data types in location_structures.py:

• location_format_processing.py: only responsible to breakdown string formats to a list of individual streets/intersections:
  • [Intersection("W BRAYTON ST", "S STATE ST"), Street("120 W BRAYTON ST")] or [Intersection(""W BRAYTON ST & S STATE ST", Intersection(" S STATE ST & W 126 PL ")]
• this list then send to geocoder_api/local.py to get list of coordinates,
• coordinates then send to location_geocoding.py to output proper Geometry object