re-enable most test

.github/workflows/cancel.yml

@@ -0,0 +1,16 @@
+name: Cancel
+
+on: [push]
+
+jobs:
+  cancel:
+    name: 'Cancel Previous Runs'
+    runs-on: ubuntu-latest
+    timeout-minutes: 3
+    steps:
+      - uses: styfle/cancel-workflow-action@0.4.0
+        with:
+          # Ids to cancel op/judi
+          # https://api.github.com/repos/slimgroup/JUDI.jl/actions/workflows
+          workflow_id: 2306091
+          access_token: ${{ github.token }}

.github/workflows/ci-test.yml

@@ -0,0 +1,54 @@
+name: CI-SetIntersectionProjection
+
+on:
+  # Trigger the workflow on push to master or pull request
+  # to be merged in master
+  push:
+    branches:
+      - master
+  pull_request:
+    branches:
+      - master
+
+jobs:
+  test:
+    name: Julia ${{ matrix.version }} - ${{ matrix.os }}
+    runs-on: ${{ matrix.os }}
+
+    strategy:
+      fail-fast: false
+
+      matrix:
+        version: ['1.1', '1.2', '1.3', '1.4', '1.5', 'nightly']
+        os: [ubuntu-latest]
+
+        include:
+          - version: 1.3
+            os: macos-latest
+
+          - version: 1.4
+            os: macos-latest
+
+    steps:
+      - name: Checkout SetIntersectionProjection
+        uses: actions/checkout@v2
+
+      - name: Setup julia
+        uses: julia-actions/setup-julia@v1
+        with:
+          version: ${{ matrix.version }}
+          arch: x64
+
+      - name: Add SLIM registery
+        run: |
+           julia --color=yes --check-bounds=yes -e 'using Pkg;Pkg.update();Pkg.Registry.add(RegistrySpec(url="https://github.com/slimgroup/SLIMregistryJL.git"))'
+
+      - name: Build SetIntersectionProjection
+        uses: julia-actions/julia-buildpkg@latest
+
+      - name: Run tests
+        uses: julia-actions/julia-runtest@latest
+
+      - uses: julia-actions/julia-uploadcodecov@latest
+        env:
+          CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}

Project.toml

@@ -1,21 +1,20 @@
-
 name = "SetIntersectionProjection"
 uuid = "335f7d24-6316-57dd-9c3a-df470f2b739e"
 authors = ["Bas Peters <1.bas.peters@gmail.com>"]
-version = "0.0.0"
+version = "0.1.0"
 
 [deps]
 Distributed = "8ba89e20-285c-5b6f-9357-94700520ee1b"
+DistributedArrays = "aaf54ef3-cdf8-58ed-94cc-d582ad619b94"
+FFTW = "7a1cc6ca-52ef-59f5-83cd-3a7055c09341"
+Interpolations = "a98d9a8b-a2ab-59e6-89dd-64a1c18fca59"
+JOLI = "bb331ad6-a1cf-11e9-23da-9bcb53c69f6f"
 LinearAlgebra = "37e2e46d-f89d-539d-b4ee-838fcccc9c8e"
+Parameters = "d96e819e-fc66-5662-9728-84c9c7592b0a"
 Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
 Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
+SortingAlgorithms = "a2af1166-a08f-5f64-846c-94a0d3cef48c"
 SparseArrays = "2f01184e-e22b-5df5-ae63-d93ebab69eaf"
 Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
-FFTW = "7a1cc6ca-52ef-59f5-83cd-3a7055c09341"
-JOLI = "bb331ad6-a1cf-11e9-23da-9bcb53c69f6f"
 Wavelets = "29a6e085-ba6d-5f35-a997-948ac2efa89a"
-SortingAlgorithms = "a2af1166-a08f-5f64-846c-94a0d3cef48c"
-Parameters = "d96e819e-fc66-5662-9728-84c9c7592b0a"
-Interpolations = "a98d9a8b-a2ab-59e6-89dd-64a1c18fca59"
-DistributedArrays = "aaf54ef3-cdf8-58ed-94cc-d582ad619b94"

src/cg.jl

@@ -3,7 +3,7 @@ export cg
 
 function cg(A::SparseMatrixCSC{T1,Int},b::Array{T2,1}; kwargs...) where {T1,T2}
 	x = zeros(promote_type(T1,T2),size(A,2)) # pre-allocate
-	return cg(v -> At_mul_B!(1.0,A,v,0.0,x),b;kwargs...) # multiply with transpose of A for efficiency
+	return cg(v -> mul!(x, transpose(A), v, 1.0, 0.0),b; kwargs...) # multiply with transpose of A for efficiency
 end
 
 cg(A,b::Vector;kwargs...) = cg(x -> A*x,b;kwargs...)

test/runtests.jl

@@ -12,7 +12,7 @@ end
 @assert nprocs() > 3
 @assert nworkers() >= 3
 
-@everywhere using DistributedArrays
+@everywhere using DistributedArrays, SparseArrays
 @everywhere using JOLI
 
 @everywhere using SetIntersectionProjection
@@ -23,32 +23,31 @@ end
 end
 
 
-
 @testset "SetIntersectionProjection" begin
 
-include("test_projectors.jl")
-include("test_setup_constraints.jl")
-
-#still need to port the stuff below to Julia 1.1
-
-# include("test_TD_OPs.jl")
-# include("test_prox_l2s!.jl")
-# include("test_argmin_x.jl")
-# include("test_update_y_l.jl")
-#
-# #parallel scripts
-# include("test_update_y_l_parallel.jl")
-# include("test_adapt_rho_gamma_parallel.jl")
-#
-# #linear algebra subroutines
-# include("test_cg.jl")
-# include("test_CDS_Mvp.jl")
-# include("test_CDS_scaled_add.jl")
-# include("test_Q_update.jl")
-#
-# #test full algorithms
-# include("test_PARSDMM.jl")
-# include("test_PARSDMM_parallel.jl")
-# include("test_PARSDMM_multilevel.jl")
+  include("test_projectors.jl")
+  include("test_setup_constraints.jl")
+
+  #still need to port the stuff below to Julia 1.1
+
+  include("test_TD_OPs.jl")
+  include("test_prox_l2s!.jl")
+  include("test_argmin_x.jl")
+  include("test_update_y_l.jl")
+
+  #parallel scripts
+  include("test_update_y_l_parallel.jl")
+  include("test_adapt_rho_gamma_parallel.jl")
+
+  #linear algebra subroutines
+  include("test_cg.jl")
+  include("test_CDS_Mvp.jl")
+  include("test_CDS_scaled_add.jl")
+  include("test_Q_update.jl")
+
+  #test full algorithms
+  # include("test_PARSDMM.jl")
+  # include("test_PARSDMM_parallel.jl")
+  # include("test_PARSDMM_multilevel.jl")
 
 end

test/test_PARSDMM.jl

@@ -5,11 +5,10 @@ options=PARSDMM_options()
 options.FL                    = Float64
 options.parallel              = false
 options.zero_ini_guess        = true
-options.linear_inv_prob_flag  = false
 
 x=randn(100,201)
 
-comp_grid=compgrid((1.0, 1.0),(100, 201))
+comp_grid = compgrid((1.0, 1.0),(100, 201))
 x         = vec(x)
 m2        = deepcopy(x)
 x_ini2    = deepcopy(x)
@@ -18,8 +17,8 @@ constraint= Dict()
 #check if input returns if the input satisfies the constraints
   #bound constraints
   constraint["use_bounds"]=true
-  constraint["m_min"]=minimum(vec(x))
-  constraint["m_max"]=maximum(vec(x))
+  constraint["min"]=minimum(vec(x))
+  constraint["max"]=maximum(vec(x))
 
   (P_sub,TD_OP,TD_Prop) = setup_constraints(constraint,comp_grid,options.FL)
   (TD_OP,AtA,l,y) = PARSDMM_precompute_distribute(TD_OP,TD_Prop,comp_grid,options)
@@ -33,8 +32,8 @@ constraint= Dict()
 #add a few convex constraints. The projected model should satisfy these constraints
   #bound constraints
   constraint["use_bounds"]=true
-  constraint["m_min"]=0.5*minimum(vec(x))
-  constraint["m_max"]=0.5*maximum(vec(x))
+  constraint["min"]=0.5*minimum(vec(x))
+  constraint["max"]=0.5*maximum(vec(x))
 
   #transform domain bounds
   (TD_OP, AtA_diag, dense, TD_n)=get_TD_operator(comp_grid,"D_z",options.FL)
@@ -182,8 +181,8 @@ constraint= Dict()
   #add a few convex constraints which the model already satisfies
     #bound constraints
     constraint["use_bounds"]=false
-    constraint["m_min"]=1.0*minimum(vec(c_l_solution))
-    constraint["m_max"]=1.456*maximum(vec(c_l_solution))
+    constraint["min"]=1.0*minimum(vec(c_l_solution))
+    constraint["max"]=1.456*maximum(vec(c_l_solution))
 
     #transform domain bounds
     (TD_OP, AtA_diag, dense, TD_n)=get_TD_operator(comp_grid,"D_z",options.FL)
@@ -236,8 +235,8 @@ constraint["TD_nuclear_operator_1"]="identity"
 #add a few convex constraints which the model already satisfies
   #bound constraints
   constraint["use_bounds"]=true
-  constraint["m_min"]=1.0*minimum(x)
-  constraint["m_max"]=0.50*maximum(x)
+  constraint["min"]=1.0*minimum(x)
+  constraint["max"]=0.50*maximum(x)
 
   #transform domain bounds
   (TD_OP, AtA_diag, dense, TD_n)=get_TD_operator(comp_grid,"D_z",options.FL)

test/test_Q_update.jl

@@ -2,7 +2,7 @@
 
 #sparse matrix
   A=sprandn(100,100,0.1)
-  B=Vector{SparseMatrixCSC{Float64,Int64}}(2)
+  B=Vector{SparseMatrixCSC{Float64,Int64}}(undef, 2)
   B[1]=sprandn(100,100,0.1)
   B[2]=sprandn(100,100,0.1)
   #make sure all nonzero diagonals in B are also in A, otherwise CDS doesnt work
@@ -14,15 +14,15 @@
   maxit=100;
   p=2
   log_PARSDMM = log_type_PARSDMM(zeros(maxit,p-1),zeros(maxit,p),zeros(maxit,p),zeros(maxit),zeros(maxit),zeros(maxit),zeros(maxit),zeros(maxit,p),zeros(maxit,p),zeros(maxit),zeros(maxit),0.00,0.00,0.00,0.00,0.00,0.00,0.00);
-  rho=Vector{Float64}(2)
+  rho=Vector{Float64}(undef, 2)
   rho[1]=1.0
   rho[2]=1.0
   log_PARSDMM.rho[1,1]=2.0
   log_PARSDMM.rho[1,2]=3.0
 
   #add A = A+ 1.0*B[1] + 2.0*B[2]
   i=1
-  ind_updated = find(rho .!= log_PARSDMM.rho[i,:])  #::Vector{Integer}# locate changed rho index
+  ind_updated = findall(rho .!= log_PARSDMM.rho[i,:])  #::Vector{Integer}# locate changed rho index
   ind_updated = convert(Array{Int64,1},ind_updated)
   #use explicit solution:
     if isempty(ind_updated) == false
@@ -33,7 +33,11 @@
   #use Q_update:
     TI=Int64
     TF=Float64
-    TD_Prop=set_properties(zeros(99),zeros(99),zeros(99),Vector{Tuple{TI,TI}}(99),Vector{Tuple{String,String}}(99),zeros(99),Vector{Vector{TI}}(99))
+    TD_Prop=set_properties(zeros(Bool, 99), zeros(Bool, 99), zeros(Bool, 99),
+                           fill!(Vector{Tuple{TI,TI}}(undef, 99), (0, 0)),
+                           fill!(Vector{Tuple{String,String, String, String}}(undef, 99), ("", "", "", "")),
+                           zeros(Bool, 99),
+                           Vector{Vector{TI}}(undef,99))
     A2=Q_update!(A2,B2,TD_Prop,rho,ind_updated,log_PARSDMM,i,[])
 
     x=randn(size(A,2))
@@ -43,7 +47,7 @@
   (R_A,offset_A)=mat2CDS(A3)
   (R_B1,offset_B1)=mat2CDS(B[1])
   (R_B2,offset_B2)=mat2CDS(B[2])
-  R_B=Vector{Array{Float64,2}}(2)
+  R_B=Vector{Array{Float64,2}}(undef, 2)
   R_B[1]=R_B1
   R_B[2]=R_B2
   TD_Prop.AtA_offsets[1]=offset_B1

test/test_TD_OPs.jl

@@ -11,8 +11,8 @@
   D2z = get_discrete_Grad(n1,n2,h1,h2,"D_z")
 
   x=zeros(n1,n2) #test on a 'cross' image
-  x[:,3]=1.0
-  x[4,:]=1.0
+  x[:,3].=1.0
+  x[4,:].=1.0
 
   a1 = D2x*vec(x); a1=reshape(a1,n1-1,n2)
   a2 = D2z*vec(x); a2=reshape(a2,n1,n2-1)
@@ -21,17 +21,17 @@
   a3b = reshape(a3b,n1-1,n2)
 
   #this test depends on the values of h1 and h2, as well as the type of derivative
-  @test a1==diff(x,1)./h1
-  @test a2==diff(x,2)./h2
+  @test a1==diff(x, dims=1)./h1
+  @test a2==diff(x, dims=2)./h2
 
   #some more general tests:
-  @test countnz(a1[:,3])==0
+  @test count(!iszero, a1[:,3])==0
   for i in [1 2 4 5 6]
     @test a1[:,1]==a1[:,i]
   end
 
   #some more general tests:
-  @test countnz(a2[4,:])==0
+  @test count(!iszero, a2[4,:])==0
   for i in [1 2 3 5 6 7 8 9]
    @test a2[1,:]==a2[i,:]
   end
@@ -52,32 +52,32 @@
     D3z = get_discrete_Grad(n1,n2,n3,h1,h2,h3,"D_z")
 
     x=zeros(n1,n2,n3) #test on a 'cross' image
-    x[2,:,:]=1.0
-    x[:,4,:]=1.0
-    x[:,:,3]=1.0
+    x[2,:,:].=1.0
+    x[:,4,:].=1.0
+    x[:,:,3].=1.0
 
     a1 = D3x*vec(x); a1=reshape(a1,n1-1,n2,n3)
     a2 = D3y*vec(x); a2=reshape(a2,n1,n2-1,n3)
     a3 = D3z*vec(x); a3=reshape(a3,n1,n2,n3-1)
     for i=1:n2
-      @test a1[:,i,:]==diff(x[:,i,:],1)./h1
+      @test a1[:,i,:]==diff(x[:,i,:], dims=1)./h1
     end
     for i=1:n3
-      @test a1[:,:,i]==diff(x[:,:,i],1)./h1
+      @test a1[:,:,i]==diff(x[:,:,i], dims=1)./h1
     end
 
     for i=1:n1
-      @test a2[i,:,:]==diff(x[i,:,:],1)./h2
+      @test a2[i,:,:]==diff(x[i,:,:], dims=1)./h2
     end
     for i=1:n3
-      @test a2[:,:,i]==diff(x[:,:,i],2)./h2
+      @test a2[:,:,i]==diff(x[:,:,i], dims=2)./h2
     end
 
     for i=1:n1
-      @test a3[i,:,:]==diff(x[i,:,:],2)./h3
+      @test a3[i,:,:]==diff(x[i,:,:], dims=2)./h3
     end
     for i=1:n2
-      @test a3[:,i,:]==diff(x[:,i,:],2)./h3
+      @test a3[:,i,:]==diff(x[:,i,:], dims=2)./h3
     end
 
 end