Merge pull request #8 from JuliaOptimizationVariationalAnalysis/auto-juliaformatter-pr

[AUTO] JuliaFormatter.jl run

Author: tmigot

src/model/api.jl

@@ -283,7 +283,11 @@ function hprod_residual! end
     Hop = hess_op_residual(model, x, i)
 Computes the Hessian of the i-th residual at x, in linear operator form.
 """
-function NLPModels.hess_op_residual(model::AbstractVIModel{T, S}, x::AbstractVector{T}, i::Int) where {T, S}
+function NLPModels.hess_op_residual(
+  model::AbstractVIModel{T, S},
+  x::AbstractVector{T},
+  i::Int,
+) where {T, S}
   @lencheck model.meta.nvar x
   Hiv = S(undef, model.meta.nvar)
   return hess_op_residual!(model, x, i, Hiv)

src/solvers/projected_reflected_gradient.jl

@@ -3,30 +3,35 @@
 > SIAM Journal on Optimization, 25(1):502–520, 2015.
 """
 function ProjectedReflectedGradientVI(stp::AbstractStopping; rho0::Float64 = 0.5)
-    xk = stp.current_state.x
-    xkp = similar(xk)
-    yk  = copy(xk)
-    rho = rho0
-    Fx = similar(xk)
-  
-    OK = update_and_start!(stp)
-    while !OK
-      abcresidual!(stp.pb, xk, rho, yk, Fx)
-      project!(stp.pb, Fx, xk) # possible failure here
-      yk .= 2 .* xkp .- xk
-  
-      if norm(xk - xkp, Inf) < stp.meta.atol * rho
-        stp.meta.optimal = true
-      end
-      OK = update_and_stop!(stp, x = xk)
+  xk = stp.current_state.x
+  xkp = similar(xk)
+  yk = copy(xk)
+  rho = rho0
+  Fx = similar(xk)
+
+  OK = update_and_start!(stp)
+  while !OK
+    abcresidual!(stp.pb, xk, rho, yk, Fx)
+    project!(stp.pb, Fx, xk) # possible failure here
+    yk .= 2 .* xkp .- xk
+
+    if norm(xk - xkp, Inf) < stp.meta.atol * rho
+      stp.meta.optimal = true
     end
-  
-    return xk
+    OK = update_and_stop!(stp, x = xk)
   end
-  
+
+  return xk
+end
+
 export ProjectedReflectedGradientVI
 
-function ProjectedReflectedGradientVI(model::AbstractVIModel, x0::AbstractVector; rho0::Float64 = 0.5, kwargs...)
+function ProjectedReflectedGradientVI(
+  model::AbstractVIModel,
+  x0::AbstractVector;
+  rho0::Float64 = 0.5,
+  kwargs...,
+)
   stp = GenericStopping(model, x0; kwargs...)
   return ProjectedReflectedGradientVI(stp, rho0 = rho0)
 end

test/runtests.jl

@@ -23,7 +23,7 @@ end
 
     xr = ones(T, 2)
     @test residual(vi, xr) == T[0; 0]
-    J = T[ 1 0; -20 10]
+    J = T[1 0; -20 10]
     @test jac_residual(vi, xr) == J
     v = ones(T, 2)
     @test jprod_residual(vi, xr, v) == T[1; -10]