Rewrite how to evaluate booleans in enum creation process

[anth-volk]

• make format && make documentation has been run.

What's changed

Fixes #203. Following the merging of #188, policyengine-uk variables of type enum consistently produced strange behavior when loaded as part of a microsimulation, seemingly being weighted correctly, but against an array populated with default values, as opposed to the expected distribution.

Manual testing found that, confusingly, the code introduced in #188 resulted in every enum being loaded twice - once as a string-type enum that was parsed into a single-element array of [0], and then again as an enum-type array that for some reason extended this array out to the same length as the weighting parameter. This contrasts with the previous behavior, whereby most enums (e.g., the variable gender in policyengine-uk) would be read only as a string-type enum, creating an array of correct length with correct values at that time.

Testing eventually demonstrated that the cause of this was the line from #188 that replaced the prior code's check of whether or not an enum's value array is Boolean-type to a different method of assessing this. When this line was replaced, without changing anything else, the code functioned correctly for policyengine-uk, and appears to function correctly for policyengine-us, as well, though my testing on this was less thorough.

That said, being frank, I don't know why this is the solution. It took quite a while to even find it, considering how seemingly unrelated it is to the problem. I'm also at a bit of a loss as to how to test it more substantively. That said, the code does pass all existing tests. This PR also adds ".env" to the gitignore file.

Bug fix

• Regression test added
• Regression test passing

[MaxGhenis]

We can merge this as a patch, but it does reopen #100

https://github.com/PolicyEngine/policyengine-core/actions/runs/8992822145/job/24703453743?pr=205#step:5:34

/home/runner/work/policyengine-core/policyengine-core/policyengine_core/enums/enum.py:53: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison
if isinstance(array == 0, bool):

Any way to get it working without this regression?

[nikhilwoodruff]

Thanks- merging given the severity of the UK issue and will reopen #100.