add comments for sensing SDP test + remove redundant data

can reconstruct measurement vector from the original matrix
plus the sensing matrices.

src/mlpack/tests/lrsdp_test.cpp

@@ -184,58 +184,7 @@ BOOST_AUTO_TEST_CASE(ErdosRenyiRandomGraphMaxCutSDP)
   BOOST_REQUIRE_CLOSE(finalValue, -3672.7, 1e-1);
 }
 
-BOOST_AUTO_TEST_CASE(GaussianMatrixSensingSDP)
-{
-  arma::mat Xorig, A, bmat;
-
-  data::Load("sensing_X.csv", Xorig, true, false);
-  data::Load("sensing_A.csv", A, true, false);
-  data::Load("sensing_b.csv", bmat, true, false);
-
-  arma::vec b = bmat;
-
-  const size_t m = Xorig.n_rows;
-  const size_t n = Xorig.n_cols;
-  const size_t p = b.n_elem;
-
-  assert(A.n_rows == p);
-  assert(A.n_cols == m * m);
-
-  float r = 0.5 + sqrt(0.25 + 2 * p);
-  if (ceil(r) > m + n)
-    r = m + n;
-
-  std::srand(0x0FAE8543);
-  arma::mat coordinates = arma::randu<arma::mat>(m + n, ceil(r));
-  std::srand(std::time(nullptr));
-
-  LRSDP sensing(0, p, coordinates);
-  sensing.SparseC().eye(m + n, m + n);
-  sensing.DenseB() = 2. * b;
-  for (size_t i = 0; i < p; ++i)
-  {
-    const auto rows = arma::span(0, m - 1);
-    const auto cols = arma::span(m, m + n - 1);
-    const auto Ai = arma::reshape(A.row(i), n, m);
-    sensing.DenseA()[i].zeros(m + n, m + n);
-    sensing.DenseA()[i](rows, cols) = trans(Ai);
-    sensing.DenseA()[i](cols, rows) = Ai;
-  }
-
-  double finalValue = sensing.Optimize(coordinates);
-  BOOST_REQUIRE_CLOSE(finalValue, 44.7550132629, 1e-1);
-
-  const arma::mat rrt = coordinates * trans(coordinates);
-  for (size_t i = 0; i < p; ++i)
-  {
-    const auto rows = arma::span(0, m - 1);
-    const auto cols = arma::span(m, m + n - 1);
-    const auto Ai = arma::reshape(A.row(i), n, m);
-    BOOST_REQUIRE_CLOSE(arma::dot(trans(Ai), rrt(rows, cols)), b(i), 1e-3);
-  }
-}
-
-/**
+/*
  * Test a nuclear norm minimization SDP.
  *
  * Specifically, fix an unknown m x n matrix X. Our goal is to recover X from p