Merge branches 'master' and 'master' of github.com:madsjulia/GeostatInversion.jl

Author: montyvesselinov

.github/workflows/CompatHelper.yml

@@ -0,0 +1,45 @@
+name: CompatHelper
+on:
+  schedule:
+    - cron: 0 0 * * *
+  workflow_dispatch:
+permissions:
+  contents: write
+  pull-requests: write
+jobs:
+  CompatHelper:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check if Julia is already available in the PATH
+        id: julia_in_path
+        run: which julia
+        continue-on-error: true
+      - name: Install Julia, but only if it is not already available in the PATH
+        uses: julia-actions/setup-julia@v1
+        with:
+          version: '1'
+          # arch: ${{ runner.arch }}
+        if: steps.julia_in_path.outcome != 'success'
+      - name: "Add the General registry via Git"
+        run: |
+          import Pkg
+          ENV["JULIA_PKG_SERVER"] = ""
+          Pkg.Registry.add("General")
+        shell: julia --color=yes {0}
+      - name: "Install CompatHelper"
+        run: |
+          import Pkg
+          name = "CompatHelper"
+          uuid = "aa819f21-2bde-4658-8897-bab36330d9b7"
+          version = "3"
+          Pkg.add(; name, uuid, version)
+        shell: julia --color=yes {0}
+      - name: "Run CompatHelper"
+        run: |
+          import CompatHelper
+          CompatHelper.main()
+        shell: julia --color=yes {0}
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+          COMPATHELPER_PRIV: ${{ secrets.DOCUMENTER_KEY }}
+          # COMPATHELPER_PRIV: ${{ secrets.COMPATHELPER_PRIV }}

.github/workflows/TagBot.yml

@@ -0,0 +1,33 @@
+name: TagBot
+on:
+  issue_comment:
+    types:
+      - created
+  workflow_dispatch:
+    inputs:
+      lookback:
+        default: 3
+permissions:
+  actions: read
+  checks: read
+  contents: write
+  deployments: read
+  issues: read
+  discussions: read
+  packages: read
+  pages: read
+  pull-requests: read
+  repository-projects: read
+  security-events: read
+  statuses: read
+jobs:
+  TagBot:
+    if: github.event_name == 'workflow_dispatch' || github.actor == 'JuliaTagBot'
+    runs-on: ubuntu-latest
+    steps:
+      - uses: JuliaRegistries/TagBot@v1
+        with:
+          token: ${{ secrets.GITHUB_TOKEN }}
+          # Edit the following line to reflect the actual name of the GitHub Secret containing your private key
+          ssh: ${{ secrets.DOCUMENTER_KEY }}
+          # ssh: ${{ secrets.NAME_OF_MY_SSH_PRIVATE_KEY_SECRET }}

Project.toml

@@ -1,7 +1,7 @@
 name = "GeostatInversion"
 uuid = "acd30bff-8734-59b1-86d1-c3ef2efb63d6"
 authors = ["Daniel O'Malley <[email protected]>", "Velimir V Vesselinov <[email protected]>"]
-version = "1.0.0"
+version = "1.2.0"
 
 [deps]
 Distributed = "8ba89e20-285c-5b6f-9357-94700520ee1b"
@@ -15,10 +15,11 @@ Statistics = "10745b16-79ce-11e8-11f9-7d13ad32a3b2"
 
 [compat]
 FFTW = "1"
-IterativeSolvers = "0.8, 0.9"
-Interpolations = "0.12, 0.13, 0.14, 0.15"
+IterativeSolvers = "0.9"
+Interpolations = "0.14, 0.15"
 RobustPmap = "1"
-julia = "1.5"
+Statistics = "1"
+julia = "1.11"
 
 [extras]
 Distributed = "8ba89e20-285c-5b6f-9357-94700520ee1b"

README.md

@@ -20,7 +20,7 @@ Randomized Geostatistical Approach (RGA) references:
 Two versions of PCGA are implemented in this package
 
 - `pcgadirect`, which uses full matrices and direct solvers during iterations
-- `pcgalsqr`, which uses low rank representations of the matrices combined with iterative solvers during iterations
+- `pcgalsqr`, which uses low-rank representations of the matrices combined with iterative solvers during iterations
 
 The RGA method, `rga`, can use either of these approaches using the keyword argument. That is, by doing `rga(...; pcgafunc=GeostatInversion.pcgadirect)` or `rga(...; pcgafunc=GeostatInversion.pcgalsqr)`.
 
@@ -54,9 +54,9 @@ MADS can execute a wide range of data- and model-based analyses:
 * Machine Learning and Blind Source Separation
 * Decision Analysis and Support
 
-MADS has been tested to perform HPC simulations on a wide-range multi-processor clusters and parallel environments (Moab, Slurm, etc.).
+MADS has been tested to perform HPC simulations on a wide-range of multi-processor clusters and parallel environments (Moab, Slurm, etc.).
 MADS utilizes adaptive rules and techniques which allows the analyses to be performed with a minimum user input.
-The code provides a series of alternative algorithms to execute each type of data- and model-based analyses.
+The code provides a series of alternative algorithms to execute each type of data- and model-based analysis.
 
 Documentation
 =============
@@ -73,7 +73,7 @@ Pkg.add("GeostatInversion")
 Installation behind a firewall
 ------------------------------
 
-Julia uses git for the package management.
+Julia uses git for package management.
 To install Julia packages behind a firewall, add the following lines in the `.gitconfig` file in your home directory:
 
 ```git

test/testfd.jl

@@ -11,14 +11,14 @@ import Random
 	end
 	testjacobianf = GeostatInversion.FDDerivatives.makejacobian(f)
 	x = [1., 2.]
-	@Test.test isapprox(jacobianf(x), testjacobianf(x), atol=1e-4)
+	Test.@test isapprox(jacobianf(x), testjacobianf(x), atol=1e-4)
 	for i = 1:10000
 		x = 10 * randn(2)
-		@Test.test isapprox(jacobianf(x), testjacobianf(x), atol=1e-4)
+		Test.@test isapprox(jacobianf(x), testjacobianf(x), atol=1e-4)
 	end
 end
 
 Random.seed!(2017)
-@Test.testset "Basic" begin
+Test.@testset "Basic" begin
 	basictest()
 end

test/testfftrf.jl

@@ -9,9 +9,9 @@ import Statistics
 	dk = rand()
 	beta = -2 - rand()
 	k = GeostatInversion.FFTRF.powerlaw_structuredgrid(Ns, k0, dk, beta)
-	@Test.test Statistics.mean(k) ≈ k0
-	@Test.test Statistics.std(k) ≈ dk
-	@Test.test collect(size(k)) == Ns
+	Test.@test Statistics.mean(k) ≈ k0
+	Test.@test Statistics.std(k) ≈ dk
+	Test.@test collect(size(k)) == Ns
 end
 
 @stderrcapture function testunstructured(N)
@@ -21,11 +21,11 @@ end
 	dk = rand()
 	beta = -2 - rand()
 	k = GeostatInversion.FFTRF.powerlaw_unstructuredgrid(points, Ns, k0, dk, beta)
-	@Test.test length(k) == size(points, 2)
+	Test.@test length(k) == size(points, 2)
 end
 
 Random.seed!(2017)
-@Test.testset "FTRF" begin
+Test.@testset "FTRF" begin
 	for i = 1:10
 		testunstructured(2)
 		testunstructured(3)

test/testrmf.jl

@@ -11,11 +11,11 @@ end
 @stderrcapture function test_rangefinder(n, m)
 	A = makeA(n, m)
 	Q = GeostatInversion.RandMatFact.rangefinder(A)
-	@Test.test isapprox(m, size(Q, 2), atol=1)
-	@Test.test isapprox(LinearAlgebra.norm(A - Q * Q' * A), 0., atol=1e-8)
+	Test.@test isapprox(m, size(Q, 2), atol=1)
+	Test.@test isapprox(LinearAlgebra.norm(A - Q * Q' * A), 0., atol=1e-8)
 	Q = GeostatInversion.RandMatFact.rangefinder(A, m, 2)
-	@Test.test isapprox(m, size(Q, 2), atol=1)
-	@Test.test isapprox(LinearAlgebra.norm(A - Q * Q' * A), 0., atol=1e-8)
+	Test.@test isapprox(m, size(Q, 2), atol=1)
+	Test.@test isapprox(LinearAlgebra.norm(A - Q * Q' * A), 0., atol=1e-8)
 end
 
 @stderrcapture function test_eig_nystrom()
@@ -25,10 +25,10 @@ end
 	Sigma = LinearAlgebra.Diagonal(Sigmavec)
 	Lambda = Sigma * Sigma
 	eigvals, eigvecs = LinearAlgebra.eigen(A)
-	@Test.test isapprox(LinearAlgebra.norm(sort(eigvals, rev=true) - LinearAlgebra.diag(Lambda)), 0., atol=1e-8)
+	Test.@test isapprox(LinearAlgebra.norm(sort(eigvals, rev=true) - LinearAlgebra.diag(Lambda)), 0., atol=1e-8)
 end
 
-@Test.testset "RMF" begin
+Test.@testset "RMF" begin
 	test_rangefinder(10, 2)
 	test_rangefinder(10, 5)
 	test_rangefinder(100, 5)

test/testrpcga.jl

@@ -1,9 +1,9 @@
 import GeostatInversion
+import Random
 import RobustPmap
 import Test
 import SparseArrays
 import LinearAlgebra
-import Random
 using Distributed
 using SparseArrays
 
@@ -12,7 +12,7 @@ using SparseArrays
 	HX = rand(10)
 	R = spzeros(10, 10)
 	A = GeostatInversion.PCGALowRankMatrix(etas, HX, R)
-	@Test.test size(A) == (size(A, 1), size(A, 2))
+	Test.@test size(A) == (size(A, 1), size(A, 2))
 end
 
 @stderrcapture function simplepcgalowranktest(numetas=10, numobs=20)
@@ -36,7 +36,7 @@ end
 				for i = 1:numobs + 1
 					x = zeros(numobs + 1)
 					x[i] = 1.
-					@Test.test isapprox(bigA * x, lrbigA * x)
+					Test.@test isapprox(bigA * x, lrbigA * x)
 				end
 			end
 		end
@@ -47,13 +47,13 @@ end
 	samples = Array{Float64, 1}[[-.5, 0., .5], [1., -1., 0.], [-.5, 1., -.5]]
 	lrcm = GeostatInversion.LowRankCovMatrix(samples)
 	fullcm = LinearAlgebra.Matrix{Float64}(LinearAlgebra.I, 3, 3) * lrcm
-	@Test.test isapprox(fullcm, lrcm * LinearAlgebra.Matrix{Float64}(LinearAlgebra.I, 3, 3))
-	@Test.test isapprox(sum(map(x->x * x', samples)) / (length(samples) - 1), fullcm)
-	@Test.test isapprox(sum(map(x->x * x', samples)) / (length(samples) - 1), fullcm)
+	Test.@test isapprox(fullcm, lrcm * LinearAlgebra.Matrix{Float64}(LinearAlgebra.I, 3, 3))
+	Test.@test isapprox(sum(map(x->x * x', samples)) / (length(samples) - 1), fullcm)
+	Test.@test isapprox(sum(map(x->x * x', samples)) / (length(samples) - 1), fullcm)
 	for i = 1:100
 		x = randn(3, 3)
-		@Test.test isapprox(fullcm * x, lrcm * x)
-		@Test.test isapprox(fullcm' * x, lrcm' * x)
+		Test.@test isapprox(fullcm * x, lrcm * x)
+		Test.@test isapprox(fullcm' * x, lrcm' * x)
 	end
 end
 
@@ -73,9 +73,10 @@ end
 	lrcm = GeostatInversion.LowRankCovMatrix(samples)
 	lrcmfull = lrcm * LinearAlgebra.Matrix{Float64}(LinearAlgebra.I, M, M)
 	@Test.test isapprox(LinearAlgebra.norm(lrcmfull - covmatrix, 2), 0.; atol=M ^ 2 / sqrt(N) + 10)
+	# Test.@test isapprox(LinearAlgebra.norm(lrcmfull - covmatrix, 2), 0.; atol=M ^ 2 / sqrt(N))
 	for i = 1:100
 		x = randn(M)
-		@Test.test isapprox(lrcm * x, lrcmfull * x)
+		Test.@test isapprox(lrcm * x, lrcmfull * x)
 	end
 end
 
@@ -96,7 +97,7 @@ end
 		we get from the full matrix.
 		=#
 		#This test is also tricky because the randsvd's in the two getxis calls need to be generating the same random numbers
-		@Test.test isapprox(0., min(LinearAlgebra.norm(fullxis[i] - lrcmxis[i]), LinearAlgebra.norm(fullxis[i] + lrcmxis[i])), atol=1e-6)
+		Test.@test isapprox(0., min(LinearAlgebra.norm(fullxis[i] - lrcmxis[i]), LinearAlgebra.norm(fullxis[i] + lrcmxis[i])), atol=1e-6)
 	end
 end
 
@@ -122,29 +123,30 @@ end
 @stderrcapture function simpletestpcga(M::Int, N::Int, mu::Float64=0.)
 	forward, p0, X, xis, R, yobs, truep = setupsimpletest(M, N, mu)
 	popt = GeostatInversion.pcgadirect(forward, p0, X, xis, R, yobs)
-	@Test.test isapprox(LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep), 0., atol=2e-2)
+	Test.@test isapprox(LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep), 0., atol=2e-2)
 	if M < N / 6
 		popt = GeostatInversion.pcgalsqr(forward, p0, X, xis, R, yobs)
-		@Test.test isapprox(LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep), 0., atol=2e-2)
+		Test.@test isapprox(LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep), 0., atol=2e-2)
 	end
 end
 
 @stderrcapture function simpletestrga(M::Int, N::Int, Nreduced::Int, mu::Float64=0.)
 	forward, p0, X, xis, R, yobs, truep = setupsimpletest(M, N, mu)
 	S = randn(Nreduced, N) * (1 / sqrt(N))
 	popt = GeostatInversion.rga(forward, p0, X, xis, R, yobs, S)
-	@Test.test isapprox(LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep), 0., atol=2e-2)
+	Test.@test isapprox(LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep), 0., atol=2e-2)
 end
 
 #=
 function simpletestpcgalm(M, N, mu=0.)
 	forward, p0, X, xis, R, yobs, truep = setupsimpletest(M, N, mu)
 	popt = GeostatInversion.pcgalm(forward, p0, X, xis, diag(R), yobs)
-	@Test.test_approx_eq_eps LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep) 0. 2e-2
+	Test.@test_approx_eq_eps LinearAlgebra.norm(popt - truep) / LinearAlgebra.norm(truep) 0. 2e-2
 end
 =#
 
-@Test.testset "RPSGA" begin
+Test.@testset "RPSGA" begin
+	@everywhere import Random
 	@everywhere Random.seed!(2017)
 	pcgalowranksize()
 	simplepcgalowranktest()
@@ -168,4 +170,4 @@ end
 		end
 	end
 end
-:passed
+:passed