Merge branch 'master' into dmw/adtypes

Author: devmotion

Project.toml

@@ -2,10 +2,8 @@ name = "Optim"
 uuid = "429524aa-4258-5aef-a3af-852621145aeb"
 version = "1.14.0"
 
-
 [deps]
 ADTypes = "47edcb42-4c32-4615-8424-f2b9edc5f35b"
-Compat = "34da2185-b29b-5c13-b0c7-acf172513d20"
 EnumX = "4e289a0a-7415-4d19-859d-a7e5c4648b56"
 FillArrays = "1a297f60-69ca-5386-bcde-b61e274b549b"
 ForwardDiff = "f6369f11-7733-5829-9624-2563aa707210"
@@ -26,27 +24,37 @@ MathOptInterface = "b8f27783-ece8-5eb3-8dc8-9495eed66fee"
 OptimMOIExt = "MathOptInterface"
 
 [compat]
-ADTypes = "1.11.0"
-Compat = "3.2.0, 3.3.0, 3.4.0, 3.5.0, 3.6.0, 4"
+ADTypes = "1.11"
+Aqua = "0.8.14"
+Distributions = "0.25.122"
 EnumX = "1.0.4"
+ExplicitImports = "1.13.2"
 FillArrays = "0.6.2, 0.7, 0.8, 0.9, 0.10, 0.11, 0.12, 0.13, 1"
 ForwardDiff = "0.10, 1"
+JET = "0.9, 0.10"
 LineSearches = "7.4.0"
 LinearAlgebra = "<0.0.1, 1.6"
 MathOptInterface = "1.17"
+Measurements = "2.14.1"
 NLSolversBase = "7.9.0"
 NaNMath = "0.3.2, 1"
 OptimTestProblems = "2.0.3"
 PositiveFactorizations = "0.2.2"
 Printf = "<0.0.1, 1.6"
 Random = "<0.0.1, 1.6"
+RecursiveArrayTools = "3.39"
+ReverseDiff = "1.16.1"
 SparseArrays = "<0.0.1, 1.6"
+StableRNGs = "1.0.3"
 StatsBase = "0.29, 0.30, 0.31, 0.32, 0.33, 0.34"
 Test = "<0.0.1, 1.6"
 julia = "1.10"
 
 [extras]
+Aqua = "4c88cf16-eb10-579e-8560-4a9242c79595"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
+ExplicitImports = "7d51a73a-1435-4ff3-83d9-f097790105c7"
+JET = "c3a54625-cd67-489e-a8e7-0a5a0ff4e31b"
 LineSearches = "d3d80556-e9d4-5f37-9878-2ab0fcc64255"
 MathOptInterface = "b8f27783-ece8-5eb3-8dc8-9495eed66fee"
 Measurements = "eff96d63-e80a-5855-80a2-b1b0885c5ab7"
@@ -60,4 +68,4 @@ StableRNGs = "860ef19b-820b-49d6-a774-d7a799459cd3"
 Test = "8dfed614-e22c-5e08-85e1-65c5234f0b40"
 
 [targets]
-test = ["Test", "Distributions", "MathOptInterface", "Measurements", "OptimTestProblems", "Random", "RecursiveArrayTools", "StableRNGs", "LineSearches", "NLSolversBase", "PositiveFactorizations", "ReverseDiff"]
+test = ["Test", "Aqua", "Distributions", "ExplicitImports", "JET", "MathOptInterface", "Measurements", "OptimTestProblems", "Random", "RecursiveArrayTools", "StableRNGs", "LineSearches", "NLSolversBase", "PositiveFactorizations", "ReverseDiff"]

README.md

@@ -4,6 +4,7 @@
 [![](https://img.shields.io/badge/docs-latest-blue.svg)](https://julianlsolvers.github.io/Optim.jl/dev)
 [![Build Status](https://github.com/JuliaNLSolvers/Optim.jl/actions/workflows/CI.yml/badge.svg?branch=master)](https://github.com/JuliaNLSolvers/Optim.jl/actions/workflows/CI.yml?query=branch%3Amaster)
 [![Codecov branch](https://img.shields.io/codecov/c/github/JuliaNLSolvers/Optim.jl/master.svg)](https://codecov.io/gh/JuliaNLSolvers/Optim.jl)
+[![Aqua QA](https://raw.githubusercontent.com/JuliaTesting/Aqua.jl/master/badge.svg)](https://github.com/JuliaTesting/Aqua.jl)
 [![JOSS](http://joss.theoj.org/papers/10.21105/joss.00615/status.svg)](https://doi.org/10.21105/joss.00615)
 
 Univariate and multivariate optimization in Julia.

docs/src/user/tipsandtricks.md

@@ -86,7 +86,7 @@ that only use the gradient in some iterations but not in others.
 Now we call `optimize` with the following syntax:
 
 ```julia
-Optim.optimize(Optim.only_fg!(fg!), [0., 0.], Optim.LBFGS())
+Optim.optimize(NLSolversBase.only_fg!(fg!), [0., 0.], Optim.LBFGS())
 ```
 
 Similarly, for a computation that requires the Hessian, we can write:
@@ -99,7 +99,7 @@ function fgh!(F, G, H, x)
   nothing
 end
 
-Optim.optimize(Optim.only_fgh!(fgh!), [0., 0.], Optim.Newton())
+Optim.optimize(NLSolversBase.only_fgh!(fgh!), [0., 0.], Optim.Newton())
 ```
 
 ## Provide gradients

ext/OptimMOIExt.jl

@@ -1,13 +1,11 @@
 module OptimMOIExt
 
 using Optim
-using Optim.LinearAlgebra
+using Optim.LinearAlgebra: rmul! 
 import MathOptInterface as MOI
 
 function __init__()
-    @static if isdefined(Base, :get_extension)
-        setglobal!(Optim, :Optimizer, Optimizer)
-    end
+    setglobal!(Optim, :Optimizer, Optimizer)
 end
 
 mutable struct Optimizer{T} <: MOI.AbstractOptimizer
@@ -423,7 +421,7 @@ MOI.get(::Optimizer, ::MOI.DualStatus) = MOI.NO_SOLUTION
 
 function MOI.get(model::Optimizer, attr::MOI.ObjectiveValue)
     MOI.check_result_index_bounds(model, attr)
-    val = minimum(model.results)
+    val = Optim.minimum(model.results)
     if model.sense == MOI.MAX_SENSE
         val = -val
     end

src/Optim.jl

@@ -15,54 +15,55 @@ Besides the help provided at the REPL, it is possible to find help and general
 documentation online at http://julianlsolvers.github.io/Optim.jl/stable/ .
 """
 module Optim
-using Compat
-using NLSolversBase          # for shared infrastructure in JuliaNLSolvers
 
-using PositiveFactorizations # for globalization strategy in Newton
-import PositiveFactorizations: cholesky!, cholesky
+using PositiveFactorizations: Positive # for globalization strategy in Newton
 
-using LineSearches           # for globalization strategy in Quasi-Newton algs
+using LineSearches: LineSearches # for globalization strategy in Quasi-Newton algs
 
 import NaNMath               # for functions that ignore NaNs (no poisoning)
 
-using Printf                 # For printing, maybe look into other options
+using Printf: @printf        # For printing, maybe look into other options
 
-using FillArrays             # For handling scalar bounds in Fminbox
-
-#using Compat                 # for compatibility across multiple julia versions
-
-# for extensions of functions defined in Base.
-import Base: length, push!, show, getindex, setindex!, maximum, minimum
+using FillArrays: Fill       # For handling scalar bounds in Fminbox
 
 # objective and constraints types and functions relevant to them.
-import NLSolversBase:
+using NLSolversBase:
+    NLSolversBase,
+    AbstractObjective,
+    InplaceObjective,
+    NotInplaceObjective,
     NonDifferentiable,
     OnceDifferentiable,
     TwiceDifferentiable,
+    TwiceDifferentiableHV,
+    AbstractConstraints,
+    ConstraintBounds,
+    TwiceDifferentiableConstraints,
     nconstraints,
     nconstraints_x,
-    NotInplaceObjective,
-    InplaceObjective
+    hessian,
+    hessian!,
+    hessian!!,
+    hv_product,
+    hv_product!
 
 # var for NelderMead
 import StatsBase: var
 
 import ADTypes
 
-import LinearAlgebra
-import LinearAlgebra:
+using LinearAlgebra:
+    LinearAlgebra,
     Diagonal,
-    diag,
     Hermitian,
     Symmetric,
     rmul!,
     mul!,
     norm,
-    normalize!,
     diagind,
     eigen,
-    BLAS,
     cholesky,
+    cholesky!,
     Cholesky, # factorizations
     I,
     svd,
@@ -71,7 +72,9 @@ import LinearAlgebra:
     ldiv!,
     dot
 
-import SparseArrays: AbstractSparseMatrix
+using SparseArrays: AbstractSparseMatrix
+
+using EnumX: @enumx # Better enums
 
 # exported functions and types
 export optimize,

src/maximize.jl

@@ -36,7 +36,7 @@ function maximize(
     f,
     x0::AbstractArray,
     method::AbstractOptimizer,
-    options = Optim.Options()
+    options = Options()
 )
     fmax = let f=f
         x -> -f(x)
@@ -48,7 +48,7 @@ function maximize(
     g,
     x0::AbstractArray,
     method::AbstractOptimizer,
-    options = Optim.Options()
+    options = Options()
 )
     fmax = let f=f
         x -> -f(x)
@@ -65,7 +65,7 @@ function maximize(
     h,
     x0::AbstractArray,
     method::AbstractOptimizer,
-    options = Optim.Options()
+    options = Options()
 )
     fmax = let f=f
         x -> -f(x)

src/multivariate/solvers/constrained/fminbox.jl

@@ -316,7 +316,7 @@ function optimize(f, l::Number, u::Number, initial_x::AbstractArray; autodiff::A
         Fill(T(u), size(initial_x)...),
         initial_x,
         Fminbox(),
-        Optim.Options(),
+        Options(),
     )
 end
 
@@ -560,7 +560,7 @@ function optimize(
             dfbox.obj.h_calls[1] = 0
         end
         copyto!(x, minimizer(results))
-        boxdist = min(minimum(x - l), minimum(u - x))
+        boxdist = Base.minimum(((xi, li, ui),) -> min(xi - li, ui - xi), zip(x, l, u)) # Base.minimum !== minimum
         if show_trace > 0
             println()
             println("Exiting inner optimizer with x = ", x)

src/multivariate/solvers/constrained/ipnewton/interior.jl

@@ -124,7 +124,7 @@ Base.hash(b::BarrierStateVars, u::UInt) = hash(
     hash(b.λxE, hash(b.λc, hash(b.λx, hash(b.slack_c, hash(b.slack_x, u + bsv_seed))))),
 )
 
-function dot(v::BarrierStateVars, w::BarrierStateVars)
+function LinearAlgebra.dot(v::BarrierStateVars, w::BarrierStateVars)
     dot(v.slack_x, w.slack_x) +
     dot(v.slack_c, w.slack_c) +
     dot(v.λx, w.λx) +
@@ -133,7 +133,7 @@ function dot(v::BarrierStateVars, w::BarrierStateVars)
     dot(v.λcE, w.λcE)
 end
 
-function norm(b::BarrierStateVars, p::Real)
+function LinearAlgebra.norm(b::BarrierStateVars, p::Real)
     norm(b.slack_x, p) +
     norm(b.slack_c, p) +
     norm(b.λx, p) +

src/multivariate/solvers/constrained/ipnewton/ipnewton.jl

@@ -80,7 +80,7 @@ mutable struct IPNewtonState{T,Tx} <: AbstractBarrierState
     constr_c::Vector{T}   # value of the user-supplied constraints at x
     constr_J::Matrix{T}   # value of the user-supplied Jacobian at x
     ev::T                 # equality violation, ∑_i λ_Ei (c*_i - c_i)
-    Optim.@add_linesearch_fields() # x_ls and alpha
+    @add_linesearch_fields() # x_ls and alpha
     b_ls::BarrierLineSearchGrad{T}
     gtilde::Tx
     Htilde::Any               # Positive Cholesky factorization of H from PositiveFactorizations.jl
@@ -179,7 +179,7 @@ function initial_state(
         constr_c,
         constr_J,
         T(NaN),
-        Optim.@initial_linesearch()..., # Maintain a cache for line search results in state.lsr
+        @initial_linesearch()..., # Maintain a cache for line search results in state.lsr
         b_ls,
         gtilde,
         0,

src/multivariate/solvers/constrained/ipnewton/utilities/assess_convergence.jl

@@ -1,7 +1,7 @@
 function assess_convergence(state::IPNewtonState, d, options::Options)
     # We use the whole bstate-gradient `bgrad`
     bgrad = state.bgrad
-    Optim.assess_convergence(
+    assess_convergence(
         state.x,
         state.x_previous,
         state.L,