format

Author: andrewrosemberg

src/NonLinearProgram/NonLinearProgram.jl

@@ -434,7 +434,13 @@ function _cache_evaluator!(model::Model)
     return model.cache
 end
 
-function DiffOpt.forward_differentiate!(model::Model; tol = 1e-6, st = 1e-6, max_corrections = 50, allow_inertia_correction = true)
+function DiffOpt.forward_differentiate!(
+    model::Model;
+    tol = 1e-6,
+    st = 1e-6,
+    max_corrections = 50,
+    allow_inertia_correction = true,
+)
     model.diff_time = @elapsed begin
         cache = _cache_evaluator!(model)
         form = model.model
@@ -448,7 +454,13 @@ function DiffOpt.forward_differentiate!(model::Model; tol = 1e-6, st = 1e-6, max
         end
 
         # Compute Jacobian
-        Δs = compute_sensitivity(model; tol = tol, st = st, max_corrections = max_corrections, allow_inertia_correction = allow_inertia_correction)
+        Δs = compute_sensitivity(
+            model;
+            tol = tol,
+            st = st,
+            max_corrections = max_corrections,
+            allow_inertia_correction = allow_inertia_correction,
+        )
 
         # Extract primal and dual sensitivities
         primal_Δs = Δs[1:length(model.cache.primal_vars), :] * Δp # Exclude slacks
@@ -462,13 +474,25 @@ function DiffOpt.forward_differentiate!(model::Model; tol = 1e-6, st = 1e-6, max
     return nothing
 end
 
-function DiffOpt.reverse_differentiate!(model::Model; tol = 1e-6, st = 1e-6, max_corrections = 50, allow_inertia_correction = true)
+function DiffOpt.reverse_differentiate!(
+    model::Model;
+    tol = 1e-6,
+    st = 1e-6,
+    max_corrections = 50,
+    allow_inertia_correction = true,
+)
     model.diff_time = @elapsed begin
         cache = _cache_evaluator!(model)
         form = model.model
 
         # Compute Jacobian
-        Δs = compute_sensitivity(model; tol = tol, st = st, max_corrections = max_corrections, allow_inertia_correction = allow_inertia_correction)
+        Δs = compute_sensitivity(
+            model;
+            tol = tol,
+            st = st,
+            max_corrections = max_corrections,
+            allow_inertia_correction = allow_inertia_correction,
+        )
         num_primal = length(cache.primal_vars)
         Δx = zeros(num_primal)
         for (i, var_idx) in enumerate(cache.primal_vars)
@@ -506,7 +530,8 @@ function DiffOpt.reverse_differentiate!(model::Model; tol = 1e-6, st = 1e-6, max
         Δp = Δs' * Δw
 
         # Order by ConstraintIndex
-        varorder = sort(collect(keys(form.var2ci)); by = x -> form.var2ci[x].value)
+        varorder =
+            sort(collect(keys(form.var2ci)); by = x -> form.var2ci[x].value)
         Δp = [Δp[form.var2param[var_idx].value] for var_idx in varorder]
 
         model.back_grad_cache = ReverseCache(; Δp = Δp)
@@ -546,7 +571,7 @@ function MOI.get(
     ::DiffOpt.ReverseConstraintSet,
     ci::MOI.ConstraintIndex{MOI.VariableIndex,MOI.Parameter{T}},
 ) where {T}
-    return MOI.Parameter{T}(model.back_grad_cache.Δp[ci.value],)
+    return MOI.Parameter{T}(model.back_grad_cache.Δp[ci.value])
 end
 
 end # module NonLinearProgram

src/NonLinearProgram/nlp_utilities.jl

@@ -436,7 +436,12 @@ end
 
 Inertia correction for the factorization of the KKT matrix. Dense version.
 """
-function inertia_corrector_factorization(M; st = 1e-6, max_corrections = 50, allow_inertia_correction = true)
+function inertia_corrector_factorization(
+    M;
+    st = 1e-6,
+    max_corrections = 50,
+    allow_inertia_correction = true,
+)
     num_c = 0
     if cond(M) > 1 / st
         @assert allow_inertia_correction "Inertia correction needed but not allowed"
@@ -476,7 +481,9 @@ function compute_derivatives_no_relax(
     ineq_locations::Vector{Z},
     has_up::Vector{Z},
     has_low::Vector{Z};
-    st = 1e-6, max_corrections = 50, allow_inertia_correction = true
+    st = 1e-6,
+    max_corrections = 50,
+    allow_inertia_correction = true,
 ) where {Z<:Integer}
     M, N = build_sensitivity_matrices(
         model,
@@ -497,9 +504,13 @@ function compute_derivatives_no_relax(
     num_vars = get_num_primal_vars(model)
     num_cons = get_num_constraints(model)
     num_ineq = length(ineq_locations)
-    K = inertia_corrector_factorization(M, num_vars + num_ineq, num_cons; 
-        st = st, max_corrections = max_corrections,
-        allow_inertia_correction = allow_inertia_correction
+    K = inertia_corrector_factorization(
+        M,
+        num_vars + num_ineq,
+        num_cons;
+        st = st,
+        max_corrections = max_corrections,
+        allow_inertia_correction = allow_inertia_correction,
     ) # Factorization
     if isnothing(K)
         return zeros(size(M, 1), size(N, 2)), K, N
@@ -519,7 +530,13 @@ sense_mult(model::Model) = objective_sense(model) == MOI.MIN_SENSE ? 1.0 : -1.0
 
 Compute the sensitivity of the solution given sensitivity of the parameters (Δp).
 """
-function compute_sensitivity(model::Model; tol = 1e-6, st = 1e-6, max_corrections = 50, allow_inertia_correction = true)
+function compute_sensitivity(
+    model::Model;
+    tol = 1e-6,
+    st = 1e-6,
+    max_corrections = 50,
+    allow_inertia_correction = true,
+)
     # Solution and bounds
     X,
     V_L,
@@ -547,8 +564,9 @@ function compute_sensitivity(model::Model; tol = 1e-6, st = 1e-6, max_correction
         ineq_locations,
         has_up,
         has_low;
-        st = st, max_corrections = max_corrections,
-        allow_inertia_correction = allow_inertia_correction
+        st = st,
+        max_corrections = max_corrections,
+        allow_inertia_correction = allow_inertia_correction,
     )
     ## Adjust signs based on JuMP convention
     num_vars = get_num_primal_vars(model)

src/jump_moi_overloads.jl

@@ -362,11 +362,17 @@ end
 
 # MOI.Utilities
 
-function reverse_differentiate!(model::MOI.Utilities.CachingOptimizer; kwargs...)
+function reverse_differentiate!(
+    model::MOI.Utilities.CachingOptimizer;
+    kwargs...,
+)
     return reverse_differentiate!(model.optimizer; kwargs...)
 end
 
-function forward_differentiate!(model::MOI.Utilities.CachingOptimizer; kwargs...)
+function forward_differentiate!(
+    model::MOI.Utilities.CachingOptimizer;
+    kwargs...,
+)
     return forward_differentiate!(model.optimizer; kwargs...)
 end
 
@@ -377,11 +383,17 @@ end
 
 # MOIB
 
-function reverse_differentiate!(model::MOI.Bridges.AbstractBridgeOptimizer; kwargs...)
+function reverse_differentiate!(
+    model::MOI.Bridges.AbstractBridgeOptimizer;
+    kwargs...,
+)
     return reverse_differentiate!(model.model; kwargs...)
 end
 
-function forward_differentiate!(model::MOI.Bridges.AbstractBridgeOptimizer; kwargs...)
+function forward_differentiate!(
+    model::MOI.Bridges.AbstractBridgeOptimizer;
+    kwargs...,
+)
     return forward_differentiate!(model.model; kwargs...)
 end
 

test/nlp_program.jl

@@ -217,7 +217,12 @@ function test_compute_derivatives_Analytical(;
         optimize!(model)
         @assert is_solved_and_feasible(model)
         # Set pertubations
-        MOI.set.(model, DiffOpt.ForwardConstraintSet(), ParameterRef.(params), Parameter.(Δp))
+        MOI.set.(
+            model,
+            DiffOpt.ForwardConstraintSet(),
+            ParameterRef.(params),
+            Parameter.(Δp),
+        )
         # Compute derivatives
         DiffOpt.forward_differentiate!(model)
         # Test sensitivities primal_vars
@@ -397,7 +402,12 @@ function test_compute_derivatives_Finite_Diff(;
         optimize!(model)
         @assert is_solved_and_feasible(model)
         # Set pertubations
-        MOI.set.(model, DiffOpt.ForwardConstraintSet(), ParameterRef.(params), Parameter.(Δp))
+        MOI.set.(
+            model,
+            DiffOpt.ForwardConstraintSet(),
+            ParameterRef.(params),
+            Parameter.(Δp),
+        )
         # Compute derivatives
         DiffOpt.forward_differentiate!(model)
         Δx = [
@@ -468,12 +478,20 @@ function test_differentiating_non_trivial_convex_qp_jump()
     db = grads_actual[6]
 
     for (i, ci) in enumerate(c_le)
-        @test -dh[i] ≈ -MOI.get(model, DiffOpt.ReverseConstraintSet(), ParameterRef(p_le[i])).value atol =
-            1e-2 rtol = 1e-2
+        @test -dh[i] ≈
+              -MOI.get(
+            model,
+            DiffOpt.ReverseConstraintSet(),
+            ParameterRef(p_le[i]),
+        ).value atol = 1e-2 rtol = 1e-2
     end
     for (i, ci) in enumerate(c_eq)
-        @test -db[i] ≈ -MOI.get(model, DiffOpt.ReverseConstraintSet(), ParameterRef(p_eq[i])).value atol =
-            1e-2 rtol = 1e-2
+        @test -db[i] ≈
+              -MOI.get(
+            model,
+            DiffOpt.ReverseConstraintSet(),
+            ParameterRef(p_eq[i]),
+        ).value atol = 1e-2 rtol = 1e-2
     end
 
     return