Format

Author: michel2323

src/batch.jl

@@ -34,7 +34,7 @@ fully independent.
 - `nnzj_per_scenario::Int` : Jacobian nonzeros per scenario
 - `nnzh_per_scenario::Int` : Hessian nonzeros per scenario
 """
-struct BatchExaModel{T, VT <: AbstractVector{T}, M <: ExaModel{T, VT}} <: NLPModels.AbstractNLPModel{T,VT}
+struct BatchExaModel{T, VT <: AbstractVector{T}, M <: ExaModel{T, VT}} <: NLPModels.AbstractNLPModel{T, VT}
     model::M
     ns::Int
     nv::Int

src/two_stage.jl

@@ -46,7 +46,7 @@ Evaluates all scenarios in ONE kernel launch.
 - Global variable layout: [v₁; v₂; ...; vₛ; d]
 - Global constraint layout: [c₁; c₂; ...; cₛ]
 """
-struct TwoStageExaModel{T, VT <: AbstractVector{T}, M <: ExaModel{T, VT}} <: NLPModels.AbstractNLPModel{T,VT}
+struct TwoStageExaModel{T, VT <: AbstractVector{T}, M <: ExaModel{T, VT}} <: NLPModels.AbstractNLPModel{T, VT}
     model::M
     ns::Int
     nv::Int

test/BatchTest/BatchTest.jl

@@ -5,15 +5,15 @@ using ExaModels
 import NLPModels
 import NLPModels: obj, cons!, cons_nln!, grad!, jac_coord!, hess_coord!, jac_structure!, hess_structure!
 import ExaModels: num_scenarios, num_vars_per_scenario, num_constraints_per_scenario, total_vars,
-                  total_cons, set_scenario_parameters!, set_all_scenario_parameters!, var_indices,
-                  cons_block_indices, grad_indices,
-                  extract_vars!, extract_cons_block!,
-                  global_var_index, global_con_index, get_model
+    total_cons, set_scenario_parameters!, set_all_scenario_parameters!, var_indices,
+    cons_block_indices, grad_indices,
+    extract_vars!, extract_cons_block!,
+    global_var_index, global_con_index, get_model
 
 import NLPModelsIpopt: ipopt
 
 function runtests()
-    @testset "BatchExaModel" begin
+    return @testset "BatchExaModel" begin
 
         @testset "Construction and dimensions" begin
             ns, nv = 3, 2
@@ -379,11 +379,11 @@ function runtests()
             # Each scenario's variable should equal its parameter
             for (i, θ) in enumerate(θ_vals)
                 v_sol = x_sol[var_indices(model, i)]
-                @test v_sol[1] ≈ θ atol = 1e-5
+                @test v_sol[1] ≈ θ atol = 1.0e-5
             end
 
             # Optimal objective should be 0
-            @test result.objective ≈ 0.0 atol = 1e-8
+            @test result.objective ≈ 0.0 atol = 1.0e-8
         end
 
         @testset "Ipopt solver - multiple variables per scenario" begin
@@ -416,12 +416,12 @@ function runtests()
             x_sol = result.solution
 
             # Scenario 1: v = [1, 3]
-            @test x_sol[var_indices(model, 1)] ≈ [1.0, 3.0] atol = 1e-5
+            @test x_sol[var_indices(model, 1)] ≈ [1.0, 3.0] atol = 1.0e-5
 
             # Scenario 2: v = [2, 2]
-            @test x_sol[var_indices(model, 2)] ≈ [2.0, 2.0] atol = 1e-5
+            @test x_sol[var_indices(model, 2)] ≈ [2.0, 2.0] atol = 1.0e-5
 
-            @test result.objective ≈ 0.0 atol = 1e-8
+            @test result.objective ≈ 0.0 atol = 1.0e-8
         end
 
         @testset "Variable bounds with start values" begin
@@ -431,7 +431,8 @@ function runtests()
 
             v_start_vals = [0.1, 0.2, 0.3, 0.4]  # ns * nv = 4 values
 
-            model = BatchExaModel(nv, ns, θ_sets;
+            model = BatchExaModel(
+                nv, ns, θ_sets;
                 v_start = v_start_vals,
                 v_lvar = 0.0,
                 v_uvar = 10.0