Convert inv to literal pow

[MasonProtter]

Should accomplish the goal "make sure inv does not get introduced, only negative powers do." in #133

I also made a random global I noticed a global const. Does MPOLY_CLEANUP even need to be a global, or can it just be internal?

[shashi]

Okay I think this could work... Although we should run reverse CI with ModelingToolkit. Because this is the weird part about not having invs:

julia> g(f)=f(2,-2)
g (generic function with 1 method)

julia> g(^)
ERROR: DomainError with -2:
Cannot raise an integer x to a negative power -2.
Make x or -2 a float by adding a zero decimal (e.g., 2.0^-2 or 2^-2.0 instead of 2^-2), or write 1/x^2, float(x)^-2, x^float(-2) or (x//1)^-2
Stacktrace:
 [1] ^(::Int64, ::Int64) at ./intfuncs.jl:273
 [2] g(::typeof(^)) at ./REPL[19]:1
 [3] top-level scope at REPL[20]:1
 [4] run_repl(::REPL.AbstractREPL, ::Any) at /build/julia/src/julia-1.5.2/usr/share/julia/stdlib/v1.5/REPL/src/REPL.jl:288

inv converts the base into float first.

[shashi]

When I recently looked into this, invs were actually gone without this fix... I think it could have been because we made things Real by default (in MTK at least -- which is what I was testing, but I'm not sure how that should change anything)

[MasonProtter]

inv was mostly gone, but appears through Base.literal_pow before this PR. I'm surprised CI passed, it failed for me locally because of a problem with literal_pow in the fuzzer. I'll make a new PR with the fix.