Fusion Categories from F-Symbols

In most literature concerning fusion categories they are characterized by so called $F$-symbols. Often only those and the fusion rules are provided or of interest. Thus we provide a datatype that allows to get a workable fusion category from provided `6j-symbols.

$F$-Symbols

Let $\mathcal C$ be a locally finite semisimple multitensor category. Then, if $\{X_i\mid i \in \mathcal I\}$ is a collection of the non-isomorphic simple objects, there is an equivalence of abelian categories

\[F \colon \mathcal C \cong \bigoplus\limits_{i \in \mathcal I} \mathrm{Vec}_k\]

given by

\[X \mapsto \mathrm{Hom}(X_i,X).\]

We define $H_{ij}^k := \mathrm{Hom}(X_k, X_i\otimes X_j)$ to be the multiplicity spaces. Now considering the image of a tensor product $X_i \otimes X_j$ of two simple objects we obtain

\[X_i \otimes X_j \mapsto \bigoplus\limits_{k \in \mathcal I} H_{ij}^k\]

After fixing a natural isomorphism

\[(X_i \otimes X_j) \otimes X_k \cong X_i \otimes (X_j \otimes X_k)\]

we obtain morphisms

\[\bigoplus\limits_{k ∈ I} H_{ij}^k \xrightarrow\]

Hecke.complex_embeddingMethod
complex_embedding(C::SixJCategory)
complex_embedding(C::SixJCategory, e::AbsSimpleNumFieldEmbedding)

Return the complex embedding of C if an embedding of the ground field is specified or given.

source
Hecke.dualMethod
dual(X::SixJObject)

Return the dual object of $X$. An error is thrown if $X$ is not rigid.

source
TensorCategories.set_associator!Method
set_associator!(F::SixJCategory, ass::Array{MatElem,4})
set_associator!(F::SixJCategory, i::Int, j::Int, k::Int, ass::Vector{<:MatElem})
set_associator!(F::SixJCategory, i::Int, j::Int, k::Int, l::Int, ass::Array{T,N}) where {T,N}
set_associator!(F::SixJCategory, i::Int, j::Int, k::Int, l::Int, m::Int, n::Int, v::RingElem)

Set the $F$-symbols of $F$.

source
TensorCategories.six_j_categoryFunction
six_j_category(F::Ring, mult::Array{Int,3}, [names::Vector{String}])
six_j_category(F::Ring, names::Vector{String})

Initialize a fusion category. Associativity constraints are all set to 1, i.e. are most likely false.

source