Update and apply formatter

.dev/Project.toml

@@ -2,4 +2,4 @@
 JuliaFormatter = "98e50ef6-434e-11e9-1051-2b60c6c9e899"
 
 [compat]
-JuliaFormatter = "0.3"
+JuliaFormatter = "1"

src/transform_fourier.jl

@@ -36,8 +36,8 @@ function truncate_modes(tr::FourierTransform{N}, coeff) where {N}
 
  # indices for the spectral coefficients that we need to retain
  inds = [
- vcat(collect(1:(m)), collect((s - m + 1):s))
- for (s, m) in zip(size(coeff)[2:(end - 1)], tr.modes)
+ vcat(collect(1:(m)), collect((s - m + 1):s)) for
+ (s, m) in zip(size(coeff)[2:(end - 1)], tr.modes)
  ]
 
  # we need to handle the first dimension of the real Fourier transform
@@ -63,8 +63,8 @@ function pad_modes(
  size_padded = (size(coeff)[1], size_pad..., size(coeff)[end])
  coeff_padded = zeros(eltype(coeff), size_padded)
  inds = [
- vcat(collect(1:(div(m, 2) + 1)), collect((s - div(m, 2) + 2):s))
- for (s, m) in zip(size_pad, size(coeff)[2:(end - 1)])
+ vcat(collect(1:(div(m, 2) + 1)), collect((s - div(m, 2) + 2):s)) for
+ (s, m) in zip(size_pad, size(coeff)[2:(end - 1)])
  ]
 
  # we need to handle the first dimension of the real Fourier transform

src/utils.jl

@@ -3,11 +3,8 @@ mview(c, inds, ::Val{2}) = view(c, :, inds[1], inds[2], :)
 mview(c, inds, ::Val{3}) = view(c, :, inds[1], inds[2], inds[3], :)
 mview(c, inds, ::Val{4}) = view(c, :, inds[1], inds[2], inds[3], inds[4], :)
 
-tensor_contraction(
- A,
- B,
- ::Val{1},
-) = @tullio C[o, a, b] := A[i, a, o] * B[i, a, b]
+tensor_contraction(A, B, ::Val{1}) =
+ @tullio C[o, a, b] := A[i, a, o] * B[i, a, b]
 tensor_contraction(A, B, ::Val{2}) =
  @tullio C[o, a₁, a₂, b] := A[i, a₁, a₂, o] * B[i, a₁, a₂, b]
 tensor_contraction(A, B, ::Val{3}) =
@@ -16,21 +13,15 @@ tensor_contraction(A, B, ::Val{4}) = @tullio C[o, a₁, a₂, a₃, a₄, b] :=
  A[i, a₁, a₂, a₃, a₄, o] * B[i, a₁, a₂, a₃, a₄, b]
 
 sparse_mean(w, c, ::Val{1}) = @tullio μ[o, a, b] := w[i, o] * c[i, a, b]
-sparse_mean(
- w,
- c,
- ::Val{2},
-) = @tullio μ[o, a₁, a₂, b] := w[i, o] * c[i, a₁, a₂, b]
+sparse_mean(w, c, ::Val{2}) =
+ @tullio μ[o, a₁, a₂, b] := w[i, o] * c[i, a₁, a₂, b]
 sparse_mean(w, c, ::Val{3}) =
  @tullio μ[o, a₁, a₂, a₃, b] := w[i, o] * c[i, a₁, a₂, a₃, b]
 sparse_mean(w, c, ::Val{4}) =
  @tullio μ[o, a₁, a₂, a₃, a₄, b] := w[i, o] * c[i, a₁, a₂, a₃, a₄, b]
 
-sparse_covariance(
- w,
- c,
- ::Val{1},
-) = @tullio μ[o, a, r, b] := w[i, r, o] * c[i, a, b]
+sparse_covariance(w, c, ::Val{1}) =
+ @tullio μ[o, a, r, b] := w[i, r, o] * c[i, a, b]
 sparse_covariance(w, c, ::Val{2}) =
  @tullio μ[o, a₁, a₂, r, b] := w[i, r, o] * c[i, a₁, a₂, b]
 sparse_covariance(w, c, ::Val{3}) =

test/transform_fourier.jl

@@ -42,8 +42,8 @@ using LinearAlgebra
  trafo = FourierTransform(modes = (12, 5))
  c = rand(27, 16, 14, 1)
  inds = [
- [(1:(div(m, 2) + 1))..., ((s - div(m, 2) + 2):s)...]
- for (s, m) in zip((M, 14), size(c)[2:(end - 1)])
+ [(1:(div(m, 2) + 1))..., ((s - div(m, 2) + 2):s)...] for
+ (s, m) in zip((M, 14), size(c)[2:(end - 1)])
  ]
  inds[1] = collect(1:16)
  @test all(OperatorFlux.pad_modes(trafo, c, (32, 14))[:, inds..., :] .== c)