Refactoring all of AWE_Components

[duckduckdoof]

As of right now, there are numerous tech-debt/readability-related issues with AWE_Components:

• documentation of the full set of Doc/Token features does not exist; there are fragments here and there in both AWE_Workbench and WritingObserver.
• need for clean, cohesive, decoupled code for the 5 extensions + utility function modules
• it is not fully clear where old models have to be used for existing features, and/or if they can be replaced by 1) a new/current spacy feature, or 2) a newer model
• wordseqprobServer needs to be integrated or replaced
• AWE_Info is not cleanly written, nor are there clear errors associated with it when things don't work out
• the use of transformations, filters, and summaryFeats is worked into AWE_Info in an awkward way (this should be redesigned)

This issue covers overarching problems with AWE_Components; further sub-issues will be made to address distinct problems listed above.

[duckduckdoof]

In quite a few import sections of the 5 components, there are imports which are never used. One notable find is that lexica isn't used in lexicalFeatures.py (which I believe is the only place where awe_lexica is imported).