diff --git a/dev/.documenter-siteinfo.json b/dev/.documenter-siteinfo.json index 70d9fc1..1c8767f 100644 --- a/dev/.documenter-siteinfo.json +++ b/dev/.documenter-siteinfo.json @@ -1 +1 @@ -{"documenter":{"julia_version":"1.6.7","generation_timestamp":"2024-12-23T21:38:23","documenter_version":"1.8.0"}} \ No newline at end of file +{"documenter":{"julia_version":"1.6.7","generation_timestamp":"2024-12-23T21:40:42","documenter_version":"1.8.0"}} \ No newline at end of file diff --git a/dev/docstring/index.html b/dev/docstring/index.html index 2550c2f..7c425ba 100644 --- a/dev/docstring/index.html +++ b/dev/docstring/index.html @@ -1,31 +1,31 @@ -Docstrings · EltypeExtensions
GitHub

Docstrings

EltypeExtensions._to_precisiontype — Method
_to_precisiontype(T::Type, S::Type)

Convert type S to have the precisiontype of T. An exception is that if T<:Integer, then Rational will also be unwrapped.

Examples

julia> _to_precisiontype(Float64, Complex{Rational{Int}})
+Docstrings · EltypeExtensions
GitHub
Docstrings
EltypeExtensions._to_precisiontype — Method
_to_precisiontype(T::Type, S::Type)
Convert type S to have the precisiontype of T. An exception is that if T<:Integer, then Rational will also be unwrapped.
Examples
julia> _to_precisiontype(Float64, Complex{Rational{Int}})
 ComplexF64 (alias for Complex{Float64})
 
 julia> _to_precisiontype(BigFloat, Matrix{Complex{Bool}})
 Matrix{Complex{BigFloat}} (alias for Array{Complex{BigFloat}, 2})
 
 julia> _to_precisiontype(Int, Complex{Rational{BigInt}})
-Complex{Rational{Int64}}
source
EltypeExtensions.basetype — Method
basetype(T::Type)
Recursively apply eltype to T until convergence.
Examples
julia> basetype(Matrix{BitArray})
+Complex{Rational{Int64}}
source
EltypeExtensions.basetype — Method
basetype(T::Type)
Recursively apply eltype to T until convergence.
Examples
julia> basetype(Matrix{BitArray})
 Bool
 
 julia> basetype(Vector{Set{Complex{Float64}}})
 ComplexF64 (alias for Complex{Float64})
 
 julia> basetype([1:n for n in 1:10])
-Int64
source
EltypeExtensions.convert_eltype — Method
convert_eltype(T, A)
Similar to convert(T, A), but T refers to the eltype. See also _to_eltype.
Examples
julia> convert_eltype(Float64, 1:10)
 1.0:1.0:10.0
 
 julia> typeof(convert_eltype(Float64, rand(Int, 3, 3)))
-Matrix{Float64} (alias for Array{Float64, 2})
source
EltypeExtensions.convert_precisiontype — Method
convert_precisiontype(T::Type, A, prec)
Convert A to have the precisiontype of T. prec is optional.
  • When T has static precision (e.g. Float64), prec has no effect.
  • When T has dynamic precision (e.g. BigFloat), prec specifies the precision of conversion. When prec is not provided, the precision is decided by the external setup from T.
  • When T is an integer, the conversion will dig into Rational as well. In contrast, since Rational as a whole is more "precise" than an integer, precisiontype doesn't unwrap Rational.
Examples
julia> convert_precisiontype(BigFloat, 1//3+im, 128)
+Matrix{Float64} (alias for Array{Float64, 2})
source
EltypeExtensions.convert_precisiontype — Method
convert_precisiontype(T::Type, A, prec)
Convert A to have the precisiontype of T. prec is optional.
  • When T has static precision (e.g. Float64), prec has no effect.
  • When T has dynamic precision (e.g. BigFloat), prec specifies the precision of conversion. When prec is not provided, the precision is decided by the external setup from T.
  • When T is an integer, the conversion will dig into Rational as well. In contrast, since Rational as a whole is more "precise" than an integer, precisiontype doesn't unwrap Rational.
Examples
julia> convert_precisiontype(BigFloat, 1//3+im, 128)
 0.3333333333333333333333333333333333333338 + 1.0im
 
 julia> convert_precisiontype(Float16, [[m/n for n in 1:3] for m in 1:3])
 3-element Vector{Vector{Float16}}:
  [1.0, 0.5, 0.3333]
  [2.0, 1.0, 0.6665]
- [3.0, 1.5, 1.0]
source
EltypeExtensions.precisiontype — Method
precisiontype(T::Type)
Returns the type that decides the precision of T. The difference from basetype is that precisiontype unwraps composite basetypes such as Complex and that precisiontype is not generalised.
Examples
julia> precisiontype(Complex{Float32})
+ [3.0, 1.5, 1.0]
source
EltypeExtensions.precisiontype — Method
precisiontype(T::Type)
Returns the type that decides the precision of T. The difference from basetype is that precisiontype unwraps composite basetypes such as Complex and that precisiontype is not generalised.
Examples
julia> precisiontype(Complex{Float32})
 Float32
 
 julia> precisiontype(Matrix{ComplexF64})
-Float64
source

+Float64
source
diff --git a/dev/index.html b/dev/index.html index b1dc288..e7e9cf2 100644 --- a/dev/index.html +++ b/dev/index.html @@ -1,2 +1,2 @@ -EltypeExtensions.jl · EltypeExtensions
GitHub

EltypeExtensions.jl

EltypeExtensions.jl is a mini toolbox for eltype-related conversions. The motivation of this package comes from manipulating (nested) arrays with different eltypes. However if you have any reasonable idea that works on other instances, feel free to write an issue/pull request.

We note that this package has some overlap with TypeUtils.jl and Unitless.jl.

Introduction

convert_eltype and _to_eltype

convert_eltype(T, x) works like convert(T, x), except that T refers to the eltype of the result. This can be useful for generic codes.

It should be always true that convert_eltype(T, x) isa _to_eltype(T, typeof(x)). However, since convert_eltype and _to_eltype use different routines, it's possible that the equality doesn't hold for some types. Please submit an issue or PR if that happens.

If typeof(x) is not in Base or stdlib, the package who owns the type should implement corresponding _to_eltype or convert_eltype. convert_eltype has fallbacks, in which case it could be unnecessary:

  • For a subtype of AbstractArray, convert_eltype calls the constructor AbstractArray{T} and _to_eltype returns Array.
  • For a subtype of AbstractUnitRange, convert_eltype calls the constructor AbstractUnitRange{T}.
  • For a subtype of AbstractRange, convert_eltype uses broadcast through map.
  • For a Tuple, convert_eltype uses dot broadcast.
  • For other types, convert_eltype calls convert and _to_eltype.

However, _to_eltype must be implemented for each type to support convert_basetype and convert_precisiontype. The following types from Base and stdlib are explicitly supported by _to_eltype:

AbstractArray, AbstractDict, AbstractSet, Adjoint, Bidiagonal, BitArray, CartesianIndices, Diagonal, Dict, Hermitian, Set, StepRangeLen, Symmetric, SymTridiagonal, Transpose, TwicePrecision, UnitRange

basetype and precisiontype

The basetype is used for nested collections, where eltype is repeatedly applied until the bottom. precisiontype has a similar idea, but goes deeper when possible. precisiontype is used to manipulate the accuracy of (nested) collections.

julia> basetype(Set{Matrix{Vector{Matrix{Complex{Rational{Int}}}}}})Complex{Rational{Int64}}
julia> precisiontype(Set{Matrix{Vector{Matrix{Complex{Rational{Int}}}}}})Rational{Int64}

Method naming convention

  • sometype(T) gets the sometype of type T.
  • sometype(x) = sometype(typeof(x)) is also provided for convenience.
  • _to_sometype(T,S) converts the type S to have the sometype of T.
  • convert_sometype(T,A) converts A to have the sometype of T.

where some can be el, base and precision.

On convert_precisiontype

convert_precisiontype accepts an optional third argument prec.

  • When T has static precision, prec has no effect.
  • When T has dynamic precision, prec specifies the precision of conversion. When prec is not provided, the precision is decided by the external setup from T. The difference is significant when convert_precisiontype is called by another function:
    julia> precision(BigFloat)256
    julia> f(x) = convert_precisiontype(BigFloat, x, 256)f (generic function with 1 method)
    julia> g(x) = convert_precisiontype(BigFloat, x)g (generic function with 1 method)
    julia> setprecision(128)128
    julia> f(π) # static precision3.141592653589793238462643383279502884197169399375105820974944592307816406286198
    julia> g(π) # precision varies with the global setting3.141592653589793238462643383279502884195
  • When T is an integer, the conversion will dig into Rational as well. In contrast, since Rational as a whole is more "precise" than an integer, precisiontype doesn't unwrap Rational.
    julia> precisiontype(convert_precisiontype(Int128, Int8(1)//Int8(2)))Rational{Int128}

Notable behaviours

Ranges

Ranges in Julia are not consistently processed:

julia> r = StepRange(1,1,5)1:1:5
julia> Float64.(r) |> typeofVector{Float64} (alias for Array{Float64, 1})
julia> map(Float64,r) |> typeofStepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}}

We adapt _to_eltype to the return type of Base.map:

julia> _to_eltype(Float64, StepRange{Int,Int})ERROR: UndefVarError: _to_eltype not defined
+EltypeExtensions.jl · EltypeExtensions
GitHub

EltypeExtensions.jl

EltypeExtensions.jl is a mini toolbox for eltype-related conversions. The motivation of this package comes from manipulating (nested) arrays with different eltypes. However if you have any reasonable idea that works on other instances, feel free to write an issue/pull request.

We note that this package has some overlap with TypeUtils.jl and Unitless.jl.

Introduction

convert_eltype and _to_eltype

convert_eltype(T, x) works like convert(T, x), except that T refers to the eltype of the result. This can be useful for generic codes.

It should be always true that convert_eltype(T, x) isa _to_eltype(T, typeof(x)). However, since convert_eltype and _to_eltype use different routines, it's possible that the equality doesn't hold for some types. Please submit an issue or PR if that happens.

If typeof(x) is not in Base or stdlib, the package who owns the type should implement corresponding _to_eltype or convert_eltype. convert_eltype has fallbacks, in which case it could be unnecessary:

  • For a subtype of AbstractArray, convert_eltype calls the constructor AbstractArray{T} and _to_eltype returns Array.
  • For a subtype of AbstractUnitRange, convert_eltype calls the constructor AbstractUnitRange{T}.
  • For a subtype of AbstractRange, convert_eltype uses broadcast through map.
  • For a Tuple, convert_eltype uses dot broadcast.
  • For other types, convert_eltype calls convert and _to_eltype.

However, _to_eltype must be implemented for each type to support convert_basetype and convert_precisiontype. The following types from Base and stdlib are explicitly supported by _to_eltype:

AbstractArray, AbstractDict, AbstractSet, Adjoint, Bidiagonal, BitArray, CartesianIndices, Diagonal, Dict, Hermitian, Set, StepRangeLen, Symmetric, SymTridiagonal, Transpose, TwicePrecision, UnitRange

basetype and precisiontype

The basetype is used for nested collections, where eltype is repeatedly applied until the bottom. precisiontype has a similar idea, but goes deeper when possible. precisiontype is used to manipulate the accuracy of (nested) collections.

julia> basetype(Set{Matrix{Vector{Matrix{Complex{Rational{Int}}}}}})Complex{Rational{Int64}}
julia> precisiontype(Set{Matrix{Vector{Matrix{Complex{Rational{Int}}}}}})Rational{Int64}

Method naming convention

  • sometype(T) gets the sometype of type T.
  • sometype(x) = sometype(typeof(x)) is also provided for convenience.
  • _to_sometype(T,S) converts the type S to have the sometype of T.
  • convert_sometype(T,A) converts A to have the sometype of T.

where some can be el, base and precision.

On convert_precisiontype

convert_precisiontype accepts an optional third argument prec.

  • When T has static precision, prec has no effect.
  • When T has dynamic precision, prec specifies the precision of conversion. When prec is not provided, the precision is decided by the external setup from T. The difference is significant when convert_precisiontype is called by another function:
    julia> precision(BigFloat)256
    julia> f(x) = convert_precisiontype(BigFloat, x, 256)f (generic function with 1 method)
    julia> g(x) = convert_precisiontype(BigFloat, x)g (generic function with 1 method)
    julia> setprecision(128)128
    julia> f(π) # static precision3.141592653589793238462643383279502884197169399375105820974944592307816406286198
    julia> g(π) # precision varies with the global setting3.141592653589793238462643383279502884195
  • When T is an integer, the conversion will dig into Rational as well. In contrast, since Rational as a whole is more "precise" than an integer, precisiontype doesn't unwrap Rational.
    julia> precisiontype(convert_precisiontype(Int128, Int8(1)//Int8(2)))Rational{Int128}

Notable behaviours

Ranges

Ranges in Julia are not consistently processed:

julia> r = StepRange(1,1,5)1:1:5
julia> Float64.(r) |> typeofVector{Float64} (alias for Array{Float64, 1})
julia> map(Float64,r) |> typeofStepRangeLen{Float64, Base.TwicePrecision{Float64}, Base.TwicePrecision{Float64}}

We adapt _to_eltype to the return type of Base.map:

julia> _to_eltype(Float64, StepRange{Int,Int})ERROR: UndefVarError: _to_eltype not defined