Fuzzylogic

Higher!
**Important: This release is for Python >= 3.12 only.**
*There are some features (specifically typing-related) that require python 3.12 in this release.
I don't have the time (or inclination) to maintain backwards compatibility. If you really need backwards compatibility, you may drop me a dime and I may consider it. Otherwise, feel free to write a PR.*

Better caching, better singleton handling, better defuzzification, nicer plotting.

* Singletons are now special Set-subclasses to really, positively, handle them correctly.
* Arrays are now cached from the start for better, consistent performance.
* Complete revamp of the defuzzification subsystem. Added all relevant defuzz methods in a defuzz.py module for clear separation of concern. Much cleaner. Also added proper tests for that.
* Plotting now also includes each functions' CoG for convenience and other niceties.

All functionality appears to be stable. No major issues reported and hypothesis tests are all green.

It is now possible to describe complex rules as tables via pandas! Only strings at this moment, but pandas could also parse excel, csv etc., so it should be trivial to add if needed.

Fuzzylogic should now properly work on Python 3.7 - please report any issues.

* All parts of the library follow the same basic intuitive pattern (functions are parametrized then called with a single argument, Domain and Set provide additional methods with logic as abstractions for ease of use) with naming that is consistent and easy to memorize
* All important parts are thoroughly tested with doctests (to ensure docs are up-to-date and provide relevant examples of use), asserts in function headers (to ensure that inputs hold up their promises), hypothesis unit tests (to ensure that functions only return what they promise and that there are no unhandled edge cases) and functional tests (to ensure that critical use cases work)
* It is now possible to do a full roundtrip - that means fuzzification of a measured value, fuzzy logic operations with fuzzy values within a domain, inference of different domains and defuzzification, that is mapping a fuzzy value back into the real world!

Much easier handling and full fuzzification - inference - defuzzification processing thanks to overhauled Domain, Set and Rule classes.