Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Fix range implementation on Julia master and resolve method ambiguities #514

Merged
merged 19 commits into from
Jan 26, 2022
Merged

Fix range implementation on Julia master and resolve method ambiguities #514

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
19 commits
Select commit Hold shift + click to select a range
f73f5c6
_range(start, ::Nothing, ::Nothing, len)
sostock Jan 12, 2022
1e40e64
_range(::Nothing, ::Nothing, stop, len)
sostock Jan 13, 2022
2bfcf3a
Add tests for ranges with complex elements
sostock Jan 13, 2022
8215952
Use modern syntax
sostock Jan 14, 2022
bdd15d6
Remove unnecessary `<:Real`
sostock Jan 14, 2022
e40fb20
Fix range(start::Quantity{BigFloat}; step::Quantity{BigFloat}; length)
sostock Jan 14, 2022
b0f507e
Fix for missing _rangestyle on 1.8
sostock Jan 15, 2022
325fd3f
Remove `real`
sostock Jan 15, 2022
3e7cf5e
Use `step` everywhere (instead of `st`)
sostock Jan 15, 2022
9ae54eb
Fix range(;step, stop, len) on Julia master
sostock Jan 19, 2022
8a5ff23
Fix range(;stop, step, length) for mixed number/quantity
sostock Jan 19, 2022
388bc5a
Add some newlines and comments
sostock Jan 20, 2022
8c6afc9
Rename _range_step_stop_length → _unitful_step_stop_length
sostock Jan 20, 2022
984ae08
Refactor _range(start, ::Nothing, stop, length)
sostock Jan 20, 2022
b8944b8
Refactor _range(start, step, ::Nothing, length)
sostock Jan 20, 2022
c2c64f3
Resolve method ambiguity on Julia ≥ 1.7
sostock Jan 20, 2022
fb95218
Simplify implementation
sostock Jan 24, 2022
02029bf
Use modern `where` syntax for `colon`
sostock Jan 25, 2022
9a32ed3
Add two-argument colon for dimensionless quantities
sostock Jan 26, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
124 changes: 84 additions & 40 deletions src/range.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,46 +7,79 @@ import Base.Broadcast: DefaultArrayStyle, broadcasted
*(y::Units, r::AbstractRange) = *(r,y)
*(r::AbstractRange, y::Units, z::Units...) = *(r, *(y,z...))

Base._range(start::Quantity{<:Real}, ::Nothing, stop, len::Integer) =
_range(promote(start, stop)..., len)
Base._range(start, ::Nothing, stop::Quantity{<:Real}, len::Integer) =
_range(promote(start, stop)..., len)
Base._range(start::Quantity{<:Real}, ::Nothing, stop::Quantity{<:Real}, len::Integer) =
_range(promote(start, stop)..., len)
(Base._range(start::T, ::Nothing, stop::T, len::Integer) where (T<:Quantity{<:Real})) =
LinRange{T}(start, stop, len)
(Base._range(start::T, ::Nothing, stop::T, len::Integer) where (T<:Quantity{<:Integer})) =
Base._linspace(Float64, ustrip(start), ustrip(stop), len, 1)*unit(T)
function Base._range(start::T, ::Nothing, stop::T, len::Integer) where (T<:Quantity{S}
where S<:Union{Float16,Float32,Float64})
range(ustrip(start), stop=ustrip(stop), length=len) * unit(T)
end
function _range(start::Quantity{T}, stop::Quantity{T}, len::Integer) where {T}
# start, stop, length
Base._range(start::Quantity, ::Nothing, stop, len::Integer) =
_unitful_start_stop_length(start, stop, len)
Base._range(start, ::Nothing, stop::Quantity, len::Integer) =
_unitful_start_stop_length(start, stop, len)
Base._range(start::Quantity, ::Nothing, stop::Quantity, len::Integer) =
_unitful_start_stop_length(start, stop, len)
function _unitful_start_stop_length(start, stop, len)
dimension(start) != dimension(stop) && throw(DimensionError(start, stop))
Base._range(start, nothing, stop, len)
a, b = promote(start, stop)
Base._range(a, nothing, b, len)
end
function Base._range(a::T, st::T, ::Nothing, len::Integer) where (T<:Quantity{S}
where S<:Union{Float16,Float32,Float64})
return Base._range(ustrip(a), ustrip(st), nothing, len) * unit(T)
Base._range(start::T, ::Nothing, stop::T, len::Integer) where {T<:Quantity} =
LinRange{T}(start, stop, len)
Base._range(start::T, ::Nothing, stop::T, len::Integer) where {T<:Quantity{<:Integer}} =
Base._linspace(Float64, ustrip(start), ustrip(stop), len, 1)*unit(T)
Base._range(start::T, ::Nothing, stop::T, len::Integer) where {T<:Quantity{<:Base.IEEEFloat}} =
Base._range(ustrip(start), nothing, ustrip(stop), len) * unit(T)

# start, step, length
Base._range(a::T, step::T, ::Nothing, len::Integer) where {T<:Quantity{<:Base.IEEEFloat}} =
Base._range(ustrip(a), ustrip(step), nothing, len) * unit(T)
Base._range(a::T, step::T, ::Nothing, len::Integer) where {T<:Quantity{<:AbstractFloat}} =
StepRangeLen{typeof(step*len),typeof(a),typeof(step)}(a, step, len)
Base._range(a::T, step::T, ::Nothing, len::Integer) where {T<:Quantity} =
@static if VERSION ≥ v"1.8.0-DEV"
Base.range_start_step_length(a, step, len)
else
Base._rangestyle(OrderStyle(a), ArithmeticStyle(a), a, step, len)
end
Base._range(a::Quantity{<:Real}, step::Quantity{<:AbstractFloat}, ::Nothing, len::Integer) =
_unitful_start_step_length(float(a), step, len)
Base._range(a::Quantity{<:AbstractFloat}, step::Quantity{<:Real}, ::Nothing, len::Integer) =
_unitful_start_step_length(a, float(step), len)
Base._range(a::Quantity{<:AbstractFloat}, step::Quantity{<:AbstractFloat}, ::Nothing, len::Integer) =
_unitful_start_step_length(a, step, len)
Base._range(a, step::Quantity, ::Nothing, len::Integer) =
_unitful_start_step_length(a, step, len)
Base._range(a::Quantity, step, ::Nothing, len::Integer) =
_unitful_start_step_length(a, step, len)
Base._range(a::Quantity, step::Quantity, ::Nothing, len::Integer) =
_unitful_start_step_length(a, step, len)
function _unitful_start_step_length(start, step, len)
dimension(start) != dimension(step) && throw(DimensionError(start,step))
Base._range(promote(start, uconvert(unit(start), step))..., nothing, len)
end
Base._range(a::Quantity{<:Real}, st::Quantity{<:AbstractFloat}, ::Nothing, len::Integer) =
Base._range(float(a), st, nothing, len)
Base._range(a::Quantity{<:AbstractFloat}, st::Quantity{<:Real}, ::Nothing, len::Integer) =
Base._range(a, float(st), nothing, len)
function Base._range(a::Quantity{<:AbstractFloat}, st::Quantity{<:AbstractFloat}, ::Nothing, len::Integer)
dimension(a) != dimension(st) && throw(DimensionError(a, st))
Base._range(promote(a, uconvert(unit(a), st))..., nothing, len)

# start, length (step defaults to 1)
Base._range(a::Quantity, ::Nothing, ::Nothing, len::Integer) =
Base._range(a, one(a), nothing, len)

# step, stop, length
@static if VERSION ≥ v"1.7"
Base._range(::Nothing, step, stop::Quantity, len::Integer) =
_unitful_step_stop_length(step, stop, len)
Base._range(::Nothing, step::Quantity, stop, len::Integer) =
_unitful_step_stop_length(step, stop, len)
Base._range(::Nothing, step::Quantity, stop::Quantity, len::Integer) =
_unitful_step_stop_length(step, stop, len)
Base._range(::Nothing, step::Quantity, ::Nothing, len::Integer) =
Base.range_error(nothing, step, nothing, len)
function _unitful_step_stop_length(step, stop, len)
dimension(stop) != dimension(step) && throw(DimensionError(stop,step))
Base.range_step_stop_length(promote(uconvert(unit(stop), step), stop)..., len)
end
end
Base._range(a::Quantity, st::Real, ::Nothing, len::Integer) =
Base._range(promote(a, uconvert(unit(a), st))..., nothing, len)
Base._range(a::Real, st::Quantity, ::Nothing, len::Integer) =
Base._range(promote(a, uconvert(unit(a), st))..., nothing, len)
# the following is needed to give sane error messages when doing e.g. range(1°, 2V, 5)
function Base._range(a::Quantity, step, ::Nothing, len::Integer)
dimension(a) != dimension(step) && throw(DimensionError(a,step))
_a, _step = promote(a, uconvert(unit(a), step))
return Base._rangestyle(OrderStyle(_a), ArithmeticStyle(_a), _a, _step, len)

# stop, length (step defaults to 1)
@static if VERSION ≥ v"1.7"
Base._range(::Nothing, ::Nothing, stop::Quantity, len::Integer) =
Base._range(nothing, one(stop), stop, len)
end

*(r::AbstractRange, y::Units) = range(first(r)*y, step=step(r)*y, length=length(r))

# first promote start and stop, leaving step alone
Expand All @@ -70,7 +103,7 @@ end
OrderStyle(::Type{<:AbstractQuantity{T}}) where T = OrderStyle(T)
ArithmeticStyle(::Type{<:AbstractQuantity{T}}) where T = ArithmeticStyle(T)

(colon(start::T, step::T, stop::T) where T <: Quantity{<:Real}) =
colon(start::T, step::T, stop::T) where {T<:Quantity{<:Real}} =
_colon(OrderStyle(T), ArithmeticStyle(T), start, step, stop)
_colon(::Ordered, ::Any, start::T, step, stop::T) where {T} = StepRange(start, step, stop)
_colon(::Ordered, ::ArithmeticRounds, start::T, step, stop::T) where {T} =
Expand All @@ -83,10 +116,20 @@ _colon(::Any, ::Any, start::T, step, stop::T) where {T} =
*(x::Base.TwicePrecision, y::Quantity) = (x * ustrip(y)) * unit(y)
uconvert(y, x::Base.TwicePrecision) = Base.TwicePrecision(uconvert(y, x.hi), uconvert(y, x.lo))

function colon(start::T, step::T, stop::T) where (T<:Quantity{S}
where S<:Union{Float16,Float32,Float64})
# This will always return a StepRangeLen
return colon(ustrip(start), ustrip(step), ustrip(stop)) * unit(T)
colon(start::T, step::T, stop::T) where {T<:Quantity{<:Base.IEEEFloat}} =
colon(ustrip(start), ustrip(step), ustrip(stop)) * unit(T) # This will always return a StepRangeLen

# two-argument colon
colon(start, stop::Quantity) = _unitful_start_stop(start, stop)
colon(start::Quantity, stop) = _unitful_start_stop(start, stop)
colon(start::Quantity, stop::Quantity) = _unitful_start_stop(start, stop)
function _unitful_start_stop(start, stop)
dimension(start) != dimension(stop) && throw(DimensionError(start, stop))
colon(promote(start, stop)...)
end
function colon(start::T, stop::T) where {T<:Quantity}
step = uconvert(unit(start), one(start))
colon(promote(start, step, stop)...)
end

# No need to confuse things by changing the type once units are on there,
Expand All @@ -95,6 +138,7 @@ end
StepRangeLen{typeof(zero(eltype(r))*y)}(r.ref*y, r.step*y, length(r), r.offset)
*(r::LinRange, y::Units) = LinRange(r.start*y, r.stop*y, length(r))
*(r::StepRange, y::Units) = StepRange(r.start*y, r.step*y, r.stop*y)
*(r::AbstractUnitRange, y::Units) = StepRange(first(r)*y, oneunit(first(r))*y, last(r)*y)
function /(x::Base.TwicePrecision, v::Quantity)
x / Base.TwicePrecision(oftype(ustrip(x.hi)/ustrip(v)*unit(v), v))
end
Expand Down
67 changes: 67 additions & 0 deletions test/runtests.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1146,6 +1146,7 @@ end
@test (-2.0Hz:1.0Hz:2.0Hz)/1.0Hz == -2.0:1.0:2.0 # issue 160
@test (range(0, stop=2, length=5) * u"°")[2:end] ==
range(0.5, stop=2, length=4) * u"°" # issue 241
@test range(big(1.0)m, step=big(1.0)m, length=5) == (big(1.0):big(1.0):big(5.0))*m
end
@testset ">> LinSpace" begin
# Not using Compat.range for these because kw args don't infer in julia 0.6.2
Expand Down Expand Up @@ -1207,6 +1208,72 @@ end
@test @inferred((1:2:5) .* cm .|> mm .|> ustrip) === 10:20:50
@test @inferred((1f0:2f0:5f0) .* cm .|> mm .|> ustrip) === 10f0:20f0:50f0
end
@testset ">> quantities and non-quantities" begin
@test range(1, step=1m/mm, length=5) == 1:1000:4001
@test range(1, step=1mm/m, length=5) == (1//1):(1//1000):(251//250)
@test eltype(range(1, step=1m/mm, length=5)) == Int
@test eltype(range(1, step=1mm/m, length=5)) == Rational{Int}
@test range(1m/mm, step=1, length=5) == ((1//1):(1//1000):(251//250)) * m/mm
@test range(1mm/m, step=1, length=5) == (1:1000:4001) * mm/m
@test eltype(range(1m/mm, step=1, length=5)) == typeof((1//1)m/mm)
@test eltype(range(1mm/m, step=1, length=5)) == typeof(1mm/m)
end
@testset ">> complex" begin
@test range((1+2im)m, step=(1+2im)m, length=5) == range(1+2im, step=1+2im, length=5) * m
@test range((1+2im)m, step=(1+2im)mm, length=5) == range(1//1+(2//1)im, step=1//1000+(1//500)im, length=5) * m
@test range((1.0+2.0im)m, stop=(3.0+4.0im)m, length=5) == LinRange(1.0+2.0im, 3.0+4.0im, 5) * m
@test range((1.0+2.0im)mm, stop=(3.0+4.0im)m, length=3) == LinRange(0.001+0.002im, 3.0+4.0im, 3) * m
end
@testset ">> step defaults to 1" begin
@test range(1.0mm/m, length=5) == (1.0mm/m):(1000.0mm/m):(4001.0mm/m)
@test range((1+2im)mm/m, length=5) == range(1+2im, step=1000, length=5)*mm/m
@test_throws DimensionError range(1.0m, length=5)
@test_throws DimensionError range((1+2im)m, length=5)
@test (1mm/m):(5001mm/m) == (1:1000:5001) * mm/m
@test (1m/mm):(5m/mm) == (1//1:1//1000:5//1) * m/mm
@test (1mm/m):(1m/mm) == 1//1000:999001//1000
@test (1m/mm):(1mm/m) == 1000//1:999//1
@test (1.0mm/m):(5001mm/m) == (1.0:1000.0:5001.0) * mm/m
@test (1m/mm):(5.0m/mm) == (1.0:0.001:5.0) * m/mm
@test (1.0mm/m):(1m/mm) == 0.001:999.001
@test (1m/mm):(1.0mm/m) == 1000.0:1.0:999.0
@test_throws DimensionError (1m):(1m)
@test_throws DimensionError (1m):(1000cm)
@test_throws DimensionError (1m):(1s)
@test (1m/cm):1 == 100:99
@test (1m/cm):1000 == 100:1000
@test (1m/cm):1.0 == 100.0:99.0
@test (1.0m/cm):1000 == 100.0:1000.0
@test_throws DimensionError (1m):1
@test 1:(1m/mm) == 1:1000
@test 1000:(1m/mm) == 1000:1000
@test 1.0:(1m/mm) == 1.0:1000.0
@test 1000:(1.0m/mm) == 1000.0:1000.0
@test_throws DimensionError 1:(1m)
end
@static if VERSION ≥ v"1.7"
@testset ">> no start argument" begin
@test range(stop=1.0m, step=2.0m, length=5) == -7.0m:2.0m:1.0m
@test range(stop=1.0mm, step=1.0m, length=5) == -3999.0mm:1000.0mm:1.0mm
@test range(stop=(1.0+2.0im)mm, step=(1.0+1.0im)m, length=5) == range(stop=1.0+2.0im, step=(1000+1000im), length=5)*mm
@test range(stop=1.0mm/m, length=5) == (-3999.0mm/m):(1000.0mm/m):(1.0mm/m)
@test range(stop=(1+2im)mm/m, length=5) == range(stop=1+2im, step=1000, length=5)*mm/m
@test range(stop=1.0mm/m, step=1, length=5) == (-3999.0mm/m):(1000.0mm/m):(1.0mm/m)
@test_throws DimensionError range(stop=1.0m, step=1V, length=5)
@test_throws DimensionError range(stop=(1+2im)m, step=1V, length=5)
@test_throws DimensionError range(stop=1.0m, length=5)
@test_throws DimensionError range(stop=(1+2im)m, length=5)
@test range(stop=1, step=1m/mm, length=5) == -3999:1000:1
@test range(stop=1, step=1mm/m, length=5) == (249//250):(1//1000):(1//1)
@test eltype(range(stop=1, step=1m/mm, length=5)) == Int
@test eltype(range(stop=1, step=1mm/m, length=5)) == Rational{Int}
@test range(stop=1m/mm, step=1, length=5) == ((249//250):(1//1000):(1//1)) * m/mm
@test range(stop=1mm/m, step=1, length=5) == (-3999:1000:1) * mm/m
@test eltype(range(stop=1m/mm, step=1, length=5)) == typeof((1//1)m/mm)
@test eltype(range(stop=1mm/m, step=1, length=5)) == typeof(1mm/m)
@test_throws ArgumentError range(step=1m, length=5)
end
end
end
@testset "> Arrays" begin
@testset ">> Array multiplication" begin
Expand Down