(cat::Category)(x)::eltype(codomain(cat))

Evaluate the function cat at point x. This is written as a function call cat(x). See also evaluate.

evaluate(cat::Category, x)::eltype(codomain(cat))

Evaluate the function cat at point x. This can also be written as a function call cat(x).

integrate(cat::Category)::eltype(codomain(cat))

Integrate the function cat over its domain.

project(dst::T, src::Category)::T where {T<:Category}

Project the function src onto the representation given by dst. This can be used e.g. to sample Julia functions on a given grid, or to resample functiont on a different grid. dst can be a skeleton representation.

JuliaFunction{DS,S,DT,T}(name::AbstractString, domain::Box{DS,S}, codomain::Box{DT,T})
JuliaFunction(name::AbstractString, domain::Box{DS,S}, codomain::Box{DT,T})

Create a zero-valued Julia function, i.e. a Julia function that is zero everywhere. The codomain can be a zero-valued domain created via Box{DT,T}().

GridFunction1D{S,T}(name::AbstractString, domain::Interval{S}, codomain::Interval{T}, npoints::Int)
GridFunction1D(name::AbstractString, domain::Interval{S}, codomain::Interval{T}, npoints::Int)

Create a zero-valued grid function, i.e. a grid function that is zero everywhere. The codomain can be a zero-valued domain created via Interval{T}().

npoints specifies the resolution of the grid function. This makes it convenient to use zero-valued grid functions as "templates" or "skeletons" when creating other grid functions.

This zero-valued grid function is represented efficiently and does not store one element per grid point.

GridFunction{DS,S,DT,T}(name::AbstractString, domain::Box{DS,S}, codomain::Box{DT,T}, grid_size::SVector{DS,Int})
GridFunction(name::AbstractString, domain::Box{DS,S}, codomain::Box{DT,T}, grid_size::SVector{DS,Int})

Create a zero-valued grid function, i.e. a grid function that is zero everywhere. The codomain can be a zero-valued domain created via Box{DT,T}().

grid_size specifies the resolution of the grid function. This makes it convenient to use zero-valued grid functions as "templates" or "skeletons" when creating other grid functions.

This zero-valued grid function is represented efficiently and does not store one element per grid point.

KernelFunction{DS,S,DT,T}(name::AbstractString, domain::Box{DS,S}, codomain::Box{DT,T}, grid_size::SVector{DS,Int})
KernelFunction(name::AbstractString, domain::Box{DS,S}, codomain::Box{DT,T}, grid_size::SVector{DS,Int})

Create a zero-valued Kernel function, i.e. a Kernel function that is zero everywhere. The codomain can be a zero-valued domain created via Box{DT,T}().

grid_size specifies the resolution of the Kernel function. This makes it convenient to use zero-valued Kernel functions as "templates" or "skeletons" when creating other Kernel functions.

==(dom1::Domain, dom2::Domain)::Bool

Equality for domains. Domains are equal if the have the same number of dimensions (see ndims), and the same lower and upper bounds (see first and last). The element types might be different if they can be compared for equality, e.g. Float64 and BigFloat.

Incompatible domains (living in different number of dimensions) are treated as unequal.

last(dom::Domain{T})::T

Return the upper bound of the domain. (At the moment domains are assumed to be cuboid, i.e. to form a box or hyper-rectangle.) Domain bounds can be infinity.

See also first.

isdisjoint(dom1::Domain, dom2::Domain)::Bool

Test whether dom1 is disjoint from dom2. Only compatible domains (which live in the same number of dimensions) can be compared.

eltype(dom::Domain)::Type

Return the element type of a Domain.

ndims(dom::Domain)::Int

Return the number of dimensions of a Domain. Scalar domains are one-dimensional. Zero-dimensional domains consist of a single point only.

issubset(dom1::Domain, dom2::Domain)::Bool

Test whether dom1 is a subset of dom2. Only compatible domains (which live in the same number of dimensions) can be compared.

in(elem, dom::Domain)::Bool

Test whether elem is inside domain dom.

first(dom::Domain{T})::T

Return the lower bound of the domain. (At the moment domains are assumed to be cuboid, i.e. to form a box or hyper-rectangle.) Domain bounds can be infinity.

See also last.