Implemenetation of partially defined evaluations in ParametricRealField

[ComboProblem]

The now correct branch...

Summary of Changes:

• Implemented methods to support partially defined evaluations in ParametricRealField. The implementation assumes partially defined evaluations are true. Determining if a ParametricRealField has consistent inequalities is partially shifted to the underlying BSA when dealing with variables without a test value.
• Updated tests to reflect changes as needed.
• Currently, the test coverage passing at same rate as in the current version (afik).
• Moved FactorUndetermined to a shared folder as it is used in both spam and igp folders.
• codespell and pycodestyle-minimal linting on parametric.sage and parametric_real_field_element.py

[ComboProblem]

As a note, when I test this with a full sagemath (10.5) installation these changes work as intended.
When I run this branch with the passagemath installation as per the instructions, something fails.

[mkoeppe]

• Moved FactorUndetermined to a shared folder as it is used in both spam and igp folders.

Let's not do that. cutgeneratingfunctionology.spam is already a dependency of .igp via parametric.sage. Just define the exception somewhere in .spam

[mkoeppe]

(Eventually, I want to move other bits from parametric.sage, in particular the ParametricRealField parent class, to .spam.)

[ComboProblem]

I've moved all the parametric field exceptions into a file in .spam and updated some documentation of a test that behavior has changed with the implementation partial evaluations.