Add support for `reim`

[hyrodium]

The method Base.imag(q::Quaternion) should not be defined because the output type must be changed, but Base.reim(c::Complex) is a Tuple, so Base.reim(q::Quaternion) can be defined with return type Tuple.

[codecov[bot]]

Codecov Report

Merging #88 (0088456) into master (4cf7f2d) will increase coverage by 0.01%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master      #88      +/-   ##
==========================================
+ Coverage   97.21%   97.22%   +0.01%     
==========================================
  Files           3        3              
  Lines         395      397       +2     
==========================================
+ Hits          384      386       +2     
  Misses         11       11              

Impacted Files	Coverage Δ
src/Octonion.jl	96.66% <100.00%> (+0.03%)	:arrow_up:
src/Quaternion.jl	98.51% <100.00%> (+<0.01%)	:arrow_up:

:mega: Codecov can now indicate which changes are the most critical in Pull Requests. Learn more

[hyrodium]

Oh, I just noticed reim(a::Real) is a NTuple{2, Real}.

julia> reim(3)
(3, 0)

Should we close this PR, or rename Base.reim(q::Quaternion) with something like Quaternions.reim_quat(q::Quaternion)?

[sethaxen]

Oh, I just noticed reim(a::Real) is a NTuple{2, Real}.

julia> reim(3)
(3, 0)

Should we close this PR, or rename Base.reim(q::Quaternion) with something like Quaternions.reim_quat(q::Quaternion)?

I agree, I would prefer a different function name than reim, but let me ask, what's the benefit of including any such function at all? Are there places where it substantially simplifies/reduces the code?

[hyrodium]

but let me ask, what's the benefit of including any such function at all? Are there places where it substantially simplifies/reduces the code?

Slightly improvement, but we can reduce these two lines to one line. https://github.com/JuliaGeometry/Rotations.jl/blob/524f61d48396f3f604e90d422917efc247b9f58a/src/unitquaternion.jl#L126-L127

When I made this PR, I thought adding a method to reim would be better than imag_part because imag_part is not a function in Base. I now agree with avoiding adding method to reim, and It seems okay with real and imag_part instead of reim.