Merge branch 'develop' into select_class_models

.github/workflows/publish_dockerhub.yml

@@ -18,7 +18,7 @@ jobs:
         uses: actions/checkout@v4
 
       - name: Log in to Docker Hub
-        uses: docker/login-action@v2
+        uses: docker/login-action@v3
         with:
           username: ${{ secrets.DOCKER_USERNAME }}
           password: ${{ secrets.DOCKER_TOKEN }}

pypesto/hierarchical/relative/solver.py

@@ -238,6 +238,8 @@ def apply_inner_parameters_to_rdatas(
         """
         sim = [rdata["y"] for rdata in rdatas]
         sigma = [rdata["sigmay"] for rdata in rdatas]
+        inner_parameters = copy.deepcopy(inner_parameters)
+        inner_parameters = scale_back_value_dict(inner_parameters, problem)
 
         # apply offsets, scalings and sigmas
         for x in problem.get_xs_for_type(InnerParameterType.SCALING):

pypesto/optimize/optimizer.py

@@ -549,7 +549,7 @@ def get_default_options(self):
 
 
 class IpoptOptimizer(Optimizer):
-    """Use IpOpt (https://pypi.org/project/ipopt/) for optimization."""
+    """Use Ipopt (https://pypi.org/project/cyipopt/) for optimization."""
 
     def __init__(self, options: dict = None):
         """
@@ -560,6 +560,9 @@ def __init__(self, options: dict = None):
         options:
             Options are directly passed on to `cyipopt.minimize_ipopt`, except
             for the `approx_grad` option, which is handled separately.
+
+            For a list of available options, see the Ipopt documentation
+            (https://coin-or.github.io/Ipopt/OPTIONS.html).
         """
         super().__init__()
         self.approx_grad = False
@@ -599,7 +602,7 @@ def minimize(
             jac = objective.get_grad
         else:
             raise ValueError(
-                "For IPOPT, the objective must either be able to return "
+                "For Ipopt, the objective must either be able to return "
                 "gradients or the `approx_grad` must be set to True."
             )
 

pypesto/store/read_from_hdf5.py

@@ -144,6 +144,7 @@ def read(self, objective: ObjectiveBase = None) -> Problem:
         problem.x_fixed_vals = [float(val) for val in problem.x_fixed_vals]
         problem.x_fixed_indices = [int(ix) for ix in problem.x_fixed_indices]
         problem.x_names = [name.decode() for name in problem.x_names]
+        problem.x_scales = [scale.decode() for scale in problem.x_scales]
 
         return problem
 

test/util.py

@@ -40,6 +40,8 @@ def obj_for_sensi(fun, grad, hess, max_sensi_order, integrated, x):
     ret: dict
         With fields obj, max_sensi_order, x, fval, grad, hess.
     """
+    x = np.asarray(x)
+
     if integrated:
         if max_sensi_order == 2: