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

WIP: Add dropdims(f, args..; dims, kwargs..) for reductions to drop dims #4

Closed
wants to merge 3 commits into from
Closed

WIP: Add dropdims(f, args..; dims, kwargs..) for reductions to drop dims #4

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
d9cbea9
Add squeeze(f, A, dims) for reductions to drop dims
mbauman Aug 29, 2017
012d910
Add NEWS
mbauman Oct 12, 2017
29945f3
Allow `dropdims` with reduction to take mutliple args and kwargs
nickrobinson251 Aug 31, 2019
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
4 changes: 2 additions & 2 deletions NEWS.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,14 +26,14 @@ New library functions
* The `tempname` function now takes a `cleanup::Bool` keyword argument defaulting to `true`, which causes the process to try to ensure that any file or directory at the path returned by `tempname` is deleted upon process exit ([#33090]).
* The `readdir` function now takes a `join::Bool` keyword argument defaulting to `false`, which when set causes `readdir` to join its directory argument with each listed name ([#33113]).


Standard library changes
------------------------

* The methods of `mktemp` and `mktempdir` which take a function body to pass temporary paths to no longer throw errors if the path is already deleted when the function body returns ([#33091]).

#### Libdl
* A new `dropdims(f, args...; dims, kwargs...)` method computes the reduction `f` over the region described by `dims` and then drops those dimensions from the result ([#23500]).

#### Libdl

#### LinearAlgebra

Expand Down
28 changes: 28 additions & 0 deletions base/abstractarraymath.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -84,8 +84,36 @@ function _dropdims(A::AbstractArray, dims::Dims)
end
reshape(A, d::typeof(_sub(axes(A), dims)))
end

_dropdims(A::AbstractArray, dim::Integer) = _dropdims(A, (Int(dim),))

"""
dropdims(f, args...; dims, kwargs...)

Compute reduction `f` over dimensions `dims` and drop those dimensions from the result.

# Examples
```jldoctest
julia> a = [3.0 2.0 6.0 8.0
6.0 1.0 4.0 2.0
3.0 0.0 7.0 6.0];
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure on expected style here... just wanted something short and clear

Also, using Floats not Ints since Int is different on 32 and 64 bit systems


julia> dropdims(sum, a, dims=1)
4-element Array{Float64,1}:
12.0
3.0
17.0
16.0

julia> dropdims(sum, abs2, a, dims=2)
3-element Array{Float64,1}:
113.0
57.0
94.0
```
"""
dropdims(f, args...; dims, kwargs...) = dropdims(f(args...; kwargs..., dims=dims), dims=dims)

## Unary operators ##

conj(x::AbstractArray{<:Real}) = x
Expand Down
31 changes: 31 additions & 0 deletions test/arrayops.jl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -303,6 +303,37 @@ end
@test_throws ArgumentError dropdims(a, dims=4)
@test_throws ArgumentError dropdims(a, dims=6)

# dropdims with reductions. issue #16606
@test (@inferred(dropdims(sum, a, dims=1)) ==
@inferred(dropdims(sum, a, dims=(1,))) ==
reshape(sum(a, dims=1), (1, 8, 8, 1)))
@test (@inferred(dropdims(sum, a, dims=3)) ==
@inferred(dropdims(sum, a, dims=(3,))) ==
reshape(sum(a, dims=3), (1, 1, 8, 1)))
@test (@inferred(dropdims(sum, a, dims=4)) ==
@inferred(dropdims(sum, a, dims=(4,))) ==
reshape(sum(a, dims=4), (1, 1, 8, 1)))
@test (@inferred(dropdims(sum, a, dims=(1, 5))) ==
dropdims(sum, a, dims=(5, 1)) ==
reshape(sum(a, (5, 1)), (1, 8, 8)))
@test (@inferred(dropdims(sum, a, dims=(1, 2, 5))) ==
dropdims(sum, a, dims=(5, 2, 1)) ==
reshape(sum(a, dims=(5, 2, 1)), (8, 8)))
@test (@inferred(dropdims(sum, abs2, a, dims=1)) ==
@inferred(dropdims(sum, abs2, a, dims=(1,))) ==
reshape(sum(a, dims=1), (1, 8, 8, 1)))
_sumplus(x; dims, plus) = sum(x; dims=dims) .+ plus # reduction with keywords
@test (@inferred(dropdims(_sumplus, a, dims=4, plus=1)) ==
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Couldn't think of / find a Base function that takes both dims and another keyword. Any suggestions?

(Of course, plenty of function outside Base do, e.g. Statistics.std takes dims and corrected keywords)

Copy link
Member

@oxinabox oxinabox Aug 31, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more. 

sort(A::AbstractArray; dims, alg, lt, by, rev, order) in Base.Sort at sort.jl:972

But idk that it is one that you would use dropdims with

Copy link
Author

@nickrobinson251 nickrobinson251 Aug 31, 2019 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

thanks,! good find ... but yeah, agree - not sure that one will make the tests here nicer...

@inferred(dropdims(_sumplus, a, dims=(4,), plus=1)) ==
reshape(sum(a, dims=4) .+ 1, (1, 1, 8, 1)))
@test_throws UndefKeywordError dropdims(sum, a)
@test_throws UndefKeywordError dropdims(sum, a, 1)
@test_throws ArgumentError dropdims(sum, a, dims=0)
@test_throws ArgumentError dropdims(sum, a, dims=(1, 1))
@test_throws ArgumentError dropdims(sum, a, dims=(1, 2, 1))
@test_throws ArgumentError dropdims(sum, a, dims=(1, 1, 2))
@test_throws ArgumentError dropdims(sum, a, dims=6)

sz = (5,8,7)
A = reshape(1:prod(sz),sz...)
@test A[2:6] == [2:6;]
Expand Down