Refactor Parser/pgen and add documentation and explanations

[pablogsal]

To improve the readability and maintainability of the parser
generator perform the following transformations:

* Separate the metagrammar parser in its own class to simplify
  the parser generator logic.

* Create separate classes for DFAs and NFAs and move methods that
  act exclusively on them from the parser generator to these
  classes.

* Add docstrings and comment docummenting the process to go from
  the grammar file into NFAs and then DFAs. Detail some of the
  algorithms and give some background explanations of some concepts
  that will helps readers not familiar with the parser generation
  process.

* Select more descriptive names for some variables and variables.

* PEP8 formatting and quote-style homogenization.

The output of the parser generator remains the same (Include/graminit.h
and Python/graminit.c remain untached by running the new parser generator).

[pablogsal]

@willingc After a couple of days of work I finished this PR that is doing some refactors to modernize the code and make it less coupled and more maintainable (not very important - the generated parser is untouched) but also is adding multiple docstrings and comments explaining how everything works, some background concepts and clarifications that I think they are very important for any reader not very familiar with LL(1) parsing.

If you have some time, could you review these docstrings and comments to make sure is clear enough and there is nothing you think is missing? Although I am happy with the current form your input will surely make a difference :)

[willingc]

@pablogsal This is mergeable as is or with the suggested wording changes. Great work.

[pablogsal]

Thank you very much @willingc for the thorough review!

I have applied all your suggestions and rephrasing and the comments and docstrings read much clearer now. Thanks!