openscvx 0.5.3.dev52__tar.gz → 0.5.3.dev53__tar.gz

{openscvx-0.5.3.dev52/openscvx.egg-info → openscvx-0.5.3.dev53}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openscvx
-Version: 0.5.3.dev52
+Version: 0.5.3.dev53
 Summary: A general Python-based successive convexification implementation which uses a JAX backend.
 Author-email: Chris Hayner and Griffin Norris <haynec@uw.edu>
 License: Apache Software License

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/openscvx/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '0.5.3.dev52'
-__version_tuple__ = version_tuple = (0, 5, 3, 'dev52')
+__version__ = version = '0.5.3.dev53'
+__version_tuple__ = version_tuple = (0, 5, 3, 'dev53')
 
-__commit_id__ = commit_id = 'gf2ea6fcdb'
+__commit_id__ = commit_id = 'gcd86fe74e'

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/openscvx/discretization/discretize_linearize.py

@@ -129,7 +129,7 @@ class VectorizeDiscretizeLinearize(Discretizer):
             u_next: np.ndarray,
             params: dict,
         ) -> jnp.ndarray:
-            beta = tau * N * foh_mask
+            beta = tau * (N - 1) * foh_mask
 
             x = x.reshape(N - 1, n_x)
             u = u_cur + beta * (u_next - u_cur)
@@ -360,7 +360,7 @@ class DiscretizeLinearizeVectorize(Discretizer):
             node: int,
             params: dict,
         ) -> jnp.ndarray:
-            beta = tau * N * foh_mask
+            beta = tau * (N - 1) * foh_mask
 
             u = u_cur + beta * (u_next - u_cur)
             F = single_state_dot(x, u, node, params)

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/openscvx/discretization/linearize_discretize.py

@@ -236,8 +236,8 @@ def _dVdt(
     # Unflatten V
     V = V.reshape(-1, i4)
 
-    # Per-control interpolation weights: beta_i = tau*N for FOH, 0 for ZOH
-    beta = tau * N * foh_mask
+    # Per-control interpolation weights: beta_i = tau*(N-1) for FOH, 0 for ZOH.
+    beta = tau * (N - 1) * foh_mask
     alpha = 1 - beta
 
     # Interpolate the control input

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/openscvx/discretization/linearize_discretize_sparse.py

@@ -182,7 +182,7 @@ def _dVdt_sparse(
     V = V.reshape(-1, aug_dim)
 
     # Per-control interpolation weights
-    beta = tau * N * foh_mask
+    beta = tau * (N - 1) * foh_mask
     alpha = 1 - beta
 
     u = u_cur + beta * (u_next - u_cur)

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/openscvx/propagation/propagation.py

@@ -61,7 +61,7 @@ def prop_aug_dy(
     """
     x = x[None, :]
 
-    beta = (tau - tau_init) * N * foh_mask
+    beta = (tau - tau_init) * (N - 1) * foh_mask
     u = u_current + beta * (u_next - u_current)
 
     return state_dot(x, u, node, params).squeeze()

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53/openscvx.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: openscvx
-Version: 0.5.3.dev52
+Version: 0.5.3.dev53
 Summary: A general Python-based successive convexification implementation which uses a JAX backend.
 Author-email: Chris Hayner and Griffin Norris <haynec@uw.edu>
 License: Apache Software License

openscvx-0.5.3.dev53/openscvx.egg-info/scm_version.json

@@ -0,0 +1,8 @@
+{
+  "tag": "0.5.2",
+  "distance": 53,
+  "node": "gcd86fe74e40856d31e6e4d67202bdadd321cde4e",
+  "dirty": false,
+  "branch": "main",
+  "node_date": "2026-08-26"
+}

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/tests/e2e/test_brachistochrone.py

@@ -19,15 +19,27 @@ import pytest
 from tests.brachistochrone_analytical import compare_trajectory_to_analytical
 
 _BRACHISTOCHRONE_ALGORITHM = {
-    "autotuner": "ConstantProximalWeight",
-    "lam_prox": 1e0,
-    "lam_cost": 6e-1,
-    "ep_tr": 1e-5,
-    "ep_vb": 1e-5,
-    "ep_vc": 1e-9,
+    "lam_prox": 1e-1,
+    "lam_cost": 1e0,
+    "lam_vc": 1e2,
+    "k_max": 100,
 }
 
 
+def _brachistochrone_algorithm(**overrides):
+    algorithm = dict(_BRACHISTOCHRONE_ALGORITHM)
+    algorithm.update(overrides)
+    return algorithm
+
+
+_BRACHISTOCHRONE_CONSTANT_ALGORITHM = _brachistochrone_algorithm(
+    autotuner="ConstantProximalWeight",
+    ep_tr=1e-5,
+    ep_vb=1e-5,
+    ep_vc=1e-9,
+)
+
+
 def _make_brachistochrone_problem(solver):
     """Build the example brachistochrone problem with a custom solver configuration."""
     from examples.abstract.brachistochrone import (
@@ -48,7 +60,7 @@ def _make_brachistochrone_problem(solver):
         constraints=constraint_exprs,
         N=n,
         float_dtype="float64",
-        algorithm=_BRACHISTOCHRONE_ALGORITHM,
+        algorithm=_BRACHISTOCHRONE_CONSTANT_ALGORITHM,
         solver=solver,
     )
 
@@ -266,7 +278,7 @@ def test_monolithic():
         constraints=constraint_exprs,
         N=n,
         licq_max=1e-8,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_brachistochrone_algorithm(),
     )
 
     problem.settings.prp.dt = 0.01
@@ -386,7 +398,7 @@ def test_backend(backend, constraint_type, with_parameters):
         constraints=constraint_exprs,
         N=n,
         licq_max=1e-8,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_brachistochrone_algorithm(),
         solver={"backend": backend},
         # QPAX / Moreau consume the global JAX dtype, so float32 caps the
         # QP's condition number aggressively. The CVXPy backend goes
@@ -683,7 +695,7 @@ def test_constraint_types(constraint_type):
         constraints=constraint_exprs,
         N=n,
         licq_max=1e-8,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_brachistochrone_algorithm(),
     )
 
     problem.settings.prp.dt = 0.01
@@ -823,7 +835,7 @@ def test_autotuning(autotuner_spec):
         constraints=constraint_exprs,
         N=n,
         licq_max=1e-8,
-        algorithm={"autotuner": autotuner, "lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_brachistochrone_algorithm(autotuner=autotuner),
     )
 
     problem.settings.prp.dt = 0.01
@@ -965,7 +977,7 @@ def test_cross_nodal(test_case):
         constraints=constraint_exprs,
         N=n,
         licq_max=1e-8,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0, "k_max": 50},
+        algorithm=_brachistochrone_algorithm(),
     )
 
     problem.settings.prp.dt = 0.01
@@ -1097,14 +1109,11 @@ def test_parameters():
         constraints=constraint_exprs,
         N=n,
         licq_max=1e-8,
-        algorithm={
-            "autotuner": ox.ConstantProximalWeight(),
-            "lam_prox": 1e-1,
-            "lam_cost": 6e-2,
-            "lam_vc": 1e0,
-            "ep_tr": 1e-5,
-            "ep_vc": 1e-8,
-        },
+        algorithm=_brachistochrone_algorithm(
+            autotuner=ox.ConstantProximalWeight(),
+            ep_tr=1e-5,
+            ep_vc=1e-8,
+        ),
     )
 
     problem.solver.solver_args = {"abstol": 1e-6, "reltol": 1e-9, "warm_start": False}
@@ -1300,7 +1309,7 @@ def test_propagation():
         dynamics_prop=dynamics_prop_extra,  # Only extra states
         states_prop=states_prop_extra,  # Only extra states
         algebraic_prop=algebraic_prop,  # Algebraic outputs
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_brachistochrone_algorithm(),
     )
 
     problem.settings.prp.dt = 0.01
@@ -1697,7 +1706,7 @@ def test_byof(byof_mode):
         N=n,
         licq_max=1e-8,
         byof=byof,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_brachistochrone_algorithm(),
     )
 
     problem.settings.prp.dt = 0.01

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/tests/test_loader.py

@@ -35,6 +35,13 @@ X0, Y0 = 0.0, 10.0
 X1, Y1 = 10.0, 5.0
 G = 9.81
 
+_BRACHISTOCHRONE_LOADER_ALGORITHM = {
+    "lam_prox": 1e-1,
+    "lam_cost": 1e0,
+    "lam_vc": 1e2,
+    "k_max": 100,
+}
+
 
 # =============================================================================
 # Helpers
@@ -181,7 +188,7 @@ def test_load_file(fmt):
     kwargs = loader(FIXTURE_DIR / filename)
     problem = Problem(
         **kwargs,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_BRACHISTOCHRONE_LOADER_ALGORITHM,
     )
     result = _configure_and_solve(problem)
     _validate_result(result, problem, f"Brachistochrone {fmt.upper()}")
@@ -243,7 +250,7 @@ def test_constraint_types(constraint_type):
     kwargs = load_dict(data)
     problem = Problem(
         **kwargs,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_BRACHISTOCHRONE_LOADER_ALGORITHM,
     )
     result = _configure_and_solve(problem)
     _validate_result(result, problem, f"YAML {constraint_type}")
@@ -285,7 +292,7 @@ def test_propagation():
     kwargs = load_dict(data)
     problem = Problem(
         **kwargs,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0},
+        algorithm=_BRACHISTOCHRONE_LOADER_ALGORITHM,
     )
     result = _configure_and_solve(problem)
 
@@ -350,7 +357,7 @@ def test_cross_nodal(feasible):
     kwargs = load_dict(data)
     problem = Problem(
         **kwargs,
-        algorithm={"lam_prox": 1e0, "lam_cost": 1e-1, "lam_vc": 1e0, "k_max": 50},
+        algorithm=_BRACHISTOCHRONE_LOADER_ALGORITHM,
     )
 
     problem.settings.prp.dt = 0.01

{openscvx-0.5.3.dev52 → openscvx-0.5.3.dev53}/tests/test_propagation.py

@@ -43,13 +43,13 @@ def _attach_sim_u_and_n_controls(p, n_controls: int):
     p.sim.u = SimpleNamespace(foh_mask=None)
 
 
-@pytest.mark.parametrize("dis_type,beta_expected", [("ZOH", 0.0), ("FOH", 1.0)])
+@pytest.mark.parametrize("dis_type,beta_expected", [("ZOH", 0.0), ("FOH", 0.8)])
 def test_prop_aug_dy_linear(dis_type, beta_expected):
     """
     prop_aug_dy should compute:
       u = u_cur + beta*(u_next - u_cur)
       return state_dot(x_batch, u).squeeze()
-    for both ZOH (beta=0) and FOH (beta=(tau-tau_init)*N).
+    for both ZOH (beta=0) and FOH (beta=(tau-tau_init)*(N-1)).
     Time-dilation is already included in state_dot symbolically.
     """
     tau = 0.2
@@ -64,7 +64,7 @@ def test_prop_aug_dy_linear(dis_type, beta_expected):
     # Per-control foh_mask: same hold on both components (scalar beta per column)
     foh_scalar = 0.0 if dis_type == "ZOH" else 1.0
     foh_mask = np.array([foh_scalar, foh_scalar], dtype=float)
-    beta = (tau - tau_init) * N * foh_scalar
+    beta = (tau - tau_init) * (N - 1) * foh_scalar
     assert pytest.approx(beta) == beta_expected
 
     # manually compute expected
@@ -98,7 +98,7 @@ def test_prop_aug_dy_mixed_foh_zoh_per_control():
     u_cur = np.array([[0.0, 3.0]])
     u_next = np.array([[2.0, 5.0]])
     foh_mask = np.array([1.0, 0.0], dtype=float)
-    beta0 = (tau - tau_init) * N * 1.0
+    beta0 = (tau - tau_init) * (N - 1) * 1.0
     beta1 = 0.0
     u_expected = u_cur + np.array([[beta0 * (2.0 - 0.0), beta1 * (5.0 - 3.0)]])
     expected = u_expected[:, 1] * (x + u_expected[:, 0])

openscvx-0.5.3.dev52/openscvx.egg-info/scm_version.json

@@ -1,8 +0,0 @@
-{
-  "tag": "0.5.2",
-  "distance": 52,
-  "node": "gf2ea6fcdbfc2ea919b47c1a4c30ed2010b0d0219",
-  "dirty": false,
-  "branch": "main",
-  "node_date": "2026-08-22"
-}