New hypergraph (and hypergraph vertex) properties

[espottesmith]

This PR implements the quad clustering coefficient for undirected hypergraphs (DOI: 10.1063/5.0188246; see also https://github.com/Gyeong-GyunHa/qch) and defines the diameter of a hypergraph as the maximum distance between any vertices.

[pszufe]

Thanks for this implementation!
Here are the comments:

• It looks that you use Matrix{Bool} as the major type throughout your code. Matrix is non-abstract type and such code seriously limit interoperability - AbstractMatrix should be used instead
• Matrix{Bool} is a dense matrix while the incidence matrices are sparse. This will be a huge performance bottlenecks for larger hypergraphs.
• Hypegraphs.jl distinguishes vertex-hyperedge incidence value weight = 0 from a situation when there is no incidence. In the Bool representation you assume that h = Hypergraph(Matrix{Union{Bool,Nothing}}(zeros(Bool, 2,2))) will yield a matrix of ones(Bool, 1, 1) while this might make sense - it should be documented in order to to surprise the user.

In order to provide good performance I could propose (for your consideration):

• implement SparseHypergraphView{Bool} <: SparseArrays.AbstractSparseMatrixCSC{Bool, Int64}. (it could be perhaps SparseHypergraphView{T} where T <: Real)
• have your methods accept argument AbstractMatrix{Bool} and iterate the arrays elements differently depending how issparse(arg) resolves.

If we have such implementation this should be both fast with large hypergraphs and flexible.