fix tests

examples/cpp/benchmark_dense_qp.cpp

@@ -43,7 +43,7 @@ Setup Time consumption(dense): 0.0592621s
 Solve Time consumption(dense): 0.101507s
 
 */
-
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp>
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp>
 

examples/cpp/estimate_nonconvex_eigenvalue.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/init_dense_qp.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/init_dense_qp_with_box.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/init_dense_qp_with_other_options.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/init_dense_qp_with_timings.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/init_with_default_options.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/initializing_with_none.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include "proxsuite/proxqp/dense/dense.hpp"
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/initializing_with_none_without_api.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include "proxsuite/proxqp/dense/dense.hpp"
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/loading_dense_qp.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include "proxsuite/proxqp/dense/dense.hpp"
 
 using namespace proxsuite::proxqp;

examples/cpp/loading_dense_qp_with_box_ineq.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include "proxsuite/proxqp/dense/dense.hpp"
 
 using namespace proxsuite::proxqp;

examples/cpp/loading_dense_qp_with_different_backend_choice.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include "proxsuite/proxqp/dense/dense.hpp"
 
 using namespace proxsuite::proxqp;

examples/cpp/overview-simple.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp>
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/solve_dense_qp.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/solve_dense_qp_with_setting.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/update_dense_qp.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

examples/cpp/update_dense_qp_ws_previous_result.cpp

@@ -1,3 +1,4 @@
+#include <iostream>
 #include <proxsuite/proxqp/dense/dense.hpp> // load the dense solver backend
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp> // used for generating a random convex qp
 

test/src/cvxpy.cpp

@@ -1,6 +1,7 @@
 //
 // Copyright (c) 2022 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <proxsuite/proxqp/dense/dense.hpp>

test/src/dense_backward.cpp

@@ -1,6 +1,7 @@
 //
 // Copyright (c) 2023 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <proxsuite/proxqp/dense/dense.hpp>
@@ -84,7 +85,7 @@ DOCTEST_TEST_CASE("proxqp::dense: test compute backward for b (feasible QP)")
   utils::rand::set_seed(1);
   dense::isize dim = 10;
 
-  dense::isize n_eq(5), n_in(2);
+  dense::isize n_eq(5), n_in(0);
   T strong_convexity_factor(1.e-2);
   proxqp::dense::Model<T> random_qp = proxqp::utils::dense_strongly_convex_qp(
     dim, n_eq, n_in, sparsity_factor, strong_convexity_factor);
@@ -93,9 +94,9 @@ DOCTEST_TEST_CASE("proxqp::dense: test compute backward for b (feasible QP)")
   Eigen::Matrix<T, 10, 1> g = random_qp.g;
   Eigen::Matrix<T, 5, 10> A = random_qp.A;
   Eigen::Matrix<T, 5, 1> b = random_qp.b;
-  Eigen::Matrix<T, 2, 10> C = random_qp.C;
-  Eigen::Matrix<T, 2, 1> l = random_qp.l;
-  Eigen::Matrix<T, 2, 1> u = random_qp.u;
+  // Eigen::Matrix<T, 2, 10> C = random_qp.C;
+  // Eigen::Matrix<T, 2, 1> l = random_qp.l;
+  // Eigen::Matrix<T, 2, 1> u = random_qp.u;
 
   dense::QP<T> qp{ dim, n_eq, n_in }; // creating QP object
   qp.settings.eps_abs = eps_abs;

test/src/dense_qp_eq.cpp

@@ -1,13 +1,13 @@
 //
 // Copyright (c) 2022 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <Eigen/Cholesky>
 #include <proxsuite/proxqp/dense/dense.hpp>
 #include <proxsuite/linalg/veg/util/dbg.hpp>
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp>
-#include <iostream>
 
 using T = double;
 using namespace proxsuite;

test/src/dense_qp_solve.cpp

@@ -1,6 +1,7 @@
 //
 // Copyright (c) 2022 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <Eigen/Cholesky>

test/src/dense_qp_solve.py

@@ -35,7 +35,7 @@ def generate_mixed_qp(n, seed=1):
     s = max(np.absolute(np.linalg.eigvals(P)))
     P += (abs(s) + 1e-02) * spa.eye(n)
     P = spa.coo_matrix(P)
-    print("sparsity of P : {}".format((P.nnz) / (n**2)))
+    # print("sparsity of P : {}".format((P.nnz) / (n**2)))
     q = np.random.randn(n)
     A = spa.random(m, n, density=0.15, data_rvs=np.random.randn, format="csc").toarray()
     v = np.random.randn(n)  # Fictitious solution

test/src/dense_qp_with_eq_and_in.cpp

@@ -1,6 +1,7 @@
 //
 // Copyright (c) 2022 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <Eigen/Cholesky>

test/src/dense_qp_wrapper.cpp

@@ -1,6 +1,7 @@
 //
 // Copyright (c) 2022 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <Eigen/Cholesky>

test/src/dense_qp_wrapper.py

@@ -6,6 +6,8 @@
 import scipy.sparse as spa
 import unittest
 
+np.printoptions(precision=16)
+
 
 def normInf(x):
     if x.shape[0] == 0:
@@ -33,7 +35,7 @@ def generate_mixed_qp(n, seed=1, reg=0.01):
     s = max(np.absolute(np.linalg.eigvals(P)))
     P += (abs(s) + reg) * spa.eye(n)
     P = spa.coo_matrix(P)
-    print("sparsity of P : {}".format((P.nnz) / (n**2)))
+    # print("sparsity of P : {}".format((P.nnz) / (n**2)))
     q = np.random.randn(n)
     A = spa.random(m, n, density=0.15, data_rvs=np.random.randn, format="csc").toarray()
     v = np.random.randn(n)  # Fictitious solution
@@ -63,7 +65,7 @@ def generate_mixed_qp_with_box(n, seed=1):
     s = max(np.absolute(np.linalg.eigvals(P)))
     P += (abs(s) + 1e-02) * spa.eye(n)
     P = spa.coo_matrix(P)
-    print("sparsity of P : {}".format((P.nnz) / (n**2)))
+    # print("sparsity of P : {}".format((P.nnz) / (n**2)))
     q = np.random.randn(n)
     A = spa.random(m, n, density=0.15, data_rvs=np.random.randn, format="csc").toarray()
     v = np.random.randn(n)  # Fictitious solution
@@ -126,6 +128,13 @@ def test_case_deterministic_behavior(self):
                 u=np.asfortranarray(u),
             )
             qp.solve()
+            print(f"{normInf(x_prev - qp.results.x)=}")
+            print(f"{normInf(y_prev - qp.results.y)=}")
+            print(f"{normInf(z_prev - qp.results.z)=}")
+
+            print(f"{y_prev=}")
+            print(f"{qp.results.y=}")
+
             assert normInf(x_prev - qp.results.x) <= 1e-15
             assert normInf(y_prev - qp.results.y) <= 1e-15
             assert normInf(z_prev - qp.results.z) <= 1e-15

test/src/dense_ruiz_equilibration.cpp

@@ -1,3 +1,7 @@
+//
+// Copyright (c) 2022-2023 INRIA
+//
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <Eigen/Cholesky>
@@ -13,12 +17,12 @@ DOCTEST_TEST_CASE("ruiz preconditioner")
   int dim = 5;
   int n_eq = 6;
   int n_in = 0;
-  // auto sym = proxqp::Symmetry::general; // 0 : upper triangular (by default),
+  auto sym = proxqp::Symmetry::general; // 0 : upper triangular (by default),
   // 1:
-  auto sym = proxqp::Symmetry::lower; // 0 : upper triangular (by default), 1:
-                                      // lower triangular ; else full matrix
+  // auto sym = proxqp::Symmetry::lower; // 0 : upper triangular (by default),
+  // 1: lower triangular ; else full matrix
 
-  Scalar sparsity_factor(0.15);
+  Scalar sparsity_factor(0.75);
   Scalar strong_convexity_factor(0.01);
   proxqp::dense::Model<Scalar> qp_random =
     proxqp::utils::dense_strongly_convex_qp(

test/src/dense_unconstrained_qp.cpp

@@ -1,6 +1,7 @@
 //
 // Copyright (c) 2022 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <Eigen/Core>
 #include <Eigen/Cholesky>

test/src/serialization.cpp

@@ -1,6 +1,7 @@
 //
 // Copyright (c) 2022 INRIA
 //
+#include <iostream>
 #include <doctest.hpp>
 #include <proxsuite/proxqp/dense/dense.hpp>
 #include <proxsuite/proxqp/utils/random_qp_problems.hpp>

test/src/serialization.py

@@ -27,7 +27,7 @@ def generate_mixed_qp(n, seed=1):
     s = max(np.absolute(np.linalg.eigvals(P)))
     P += (abs(s) + 1e-02) * spa.eye(n)
     P = spa.coo_matrix(P)
-    print("sparsity of P : {}".format((P.nnz) / (n**2)))
+    # print("sparsity of P : {}".format((P.nnz) / (n**2)))
     q = np.random.randn(n)
     A = spa.random(m, n, density=0.15, data_rvs=np.random.randn, format="csc").toarray()
     v = np.random.randn(n)  # Fictitious solution

test/src/sparse_maros_meszaros.cpp

@@ -2,6 +2,7 @@
 // Copyright (c) 2022-2023 INRIA
 //
 
+#include <iostream>
 #include <doctest.hpp>
 #include <maros_meszaros.hpp>
 #include <proxsuite/proxqp/sparse/wrapper.hpp>

test/src/sparse_qp_solve.py

@@ -33,7 +33,7 @@ def generate_mixed_qp(n, seed=1):
     s = max(np.absolute(np.linalg.eigvals(P)))
     P += (abs(s) + 1e-02) * spa.eye(n)
     P = spa.coo_matrix(P)
-    print("sparsity of P : {}".format((P.nnz) / (n**2)))
+    # print("sparsity of P : {}".format((P.nnz) / (n**2)))
     q = np.random.randn(n)
     A = spa.random(m, n, density=0.15, data_rvs=np.random.randn, format="csc")
     v = np.random.randn(n)  # Fictitious solution

test/src/sparse_qp_wrapper.py

@@ -33,7 +33,7 @@ def generate_mixed_qp(n, seed=1, reg=0.01):
     s = max(np.absolute(np.linalg.eigvals(P)))
     P += (abs(s) + reg) * spa.eye(n)
     P = spa.coo_matrix(P)
-    print("sparsity of P : {}".format((P.nnz) / (n**2)))
+    # print("sparsity of P : {}".format((P.nnz) / (n**2)))
     q = np.random.randn(n)
     A = spa.random(m, n, density=0.15, data_rvs=np.random.randn, format="csc")
     v = np.random.randn(n)  # Fictitious solution