sign of simplices blocking type inference

[lukem12345]

          > Tracing back through the function calls, the `sign` function, which returns a vector of positive and negative ones corresponding to the sign of a simplex, is returning `::Any`.

This is because the type of the orientation is encoded as a type parameter to EmbeddedDeltaDualComplex2D.

One way to work around this is to hard-code Bool a la

CombinatorialSpaces.jl/src/DiscreteExteriorCalculus.jl

Line 1101 in ab1ea66

lazy_edge_orient .= one(Bool)

Originally posted by @lukem12345 in #78 (comment)

[lukem12345]

We could also standardize on storing +1 or -1 directly, as small floats or ints. There are many instances in this codebase where we check orientation simply to decide whether to return +1 or -1.
E.g. explicitly here:

CombinatorialSpaces.jl/src/FastDEC.jl

Line 399 in 7354e5b

e_orient[i] = (e_orient[i] == 1 ? 1 : -1)

Further, since only the orientation of top-level simplices matters for DEC purposes, we could go ahead and refactor the ACSet schemas to only store such. This would eliminate many calls to sign outright.

[GeorgeR227]

I think we should just store Bool since all of the meshes use Bool and changing that is breaking. I get what you're saying about the +1 and -1 but we can just write a helper function to convert between those.

I think changing the schema here is doable and all current mesh generation functions seem to have a check to see if the orientation column exists. Although some DEC operators may check for lower-level orientations, I think exterior derivative is one, so those may break.