AmiciObjective: check that sensitivities wrt all relevant parameters …

doc/example/petab_import.ipynb

@@ -178,7 +178,7 @@
     "converter_config = libsbml.SBMLLocalParameterConverter().getDefaultProperties()\n",
     "petab_problem.sbml_document.convert(converter_config)\n",
     "\n",
-    "obj = importer.create_objective()\n",
+    "obj = problem.objective\n",
     "\n",
     "# for some models, hyperparameters need to be adjusted\n",
     "# obj.amici_solver.setMaxSteps(10000)\n",
@@ -200,7 +200,7 @@
    "outputs": [],
    "source": [
     "ret = obj(\n",
-    "    petab_problem.x_nominal_scaled,\n",
+    "    petab_problem.x_nominal_free_scaled,\n",
     "    mode=\"mode_fun\",\n",
     "    sensi_orders=(0, 1),\n",
     "    return_dict=True,\n",
@@ -383,7 +383,7 @@
    "outputs": [],
    "source": [
     "ref = visualize.create_references(\n",
-    "    x=petab_problem.x_nominal_scaled, fval=obj(petab_problem.x_nominal_scaled)\n",
+    "    x=petab_problem.x_nominal_scaled, fval=obj(petab_problem.x_nominal_free_scaled)\n",
     ")\n",
     "\n",
     "visualize.waterfall(result, reference=ref, scale_y=\"lin\")\n",

pypesto/objective/amici/amici.py

@@ -4,6 +4,7 @@
 import tempfile
 from collections import OrderedDict
 from collections.abc import Sequence
+from itertools import chain
 from pathlib import Path
 from typing import TYPE_CHECKING, Optional, Union
 
@@ -233,6 +234,33 @@ def __init__(
         # `set_custom_timepoints` method for more information.
         self.custom_timepoints = None
 
+    def _get_missing_sensitivities(self) -> set[str]:
+        """Get the model parameters for which sensitivities are missing.
+
+        Get the model parameters w.r.t. which sensitivities are required,
+        but aren't available missing.
+        """
+        required_sensitivities = set()
+        objective_parameters = set(self.x_names)
+        # resolve parameter mapping: collect all model parameters that depend
+        #  on the optimization parameters
+        for condition_mapping in self.parameter_mapping:
+            required_sensitivities.update(
+                model_par
+                for model_par, opt_par in chain(
+                    condition_mapping.map_preeq_fix.items(),
+                    condition_mapping.map_sim_fix.items(),
+                    condition_mapping.map_sim_var.items(),
+                )
+                # we assume that the values in the parameter mapping are
+                #  either numeric values or parameter ID strings
+                if opt_par in objective_parameters
+            )
+
+        # All parameters w.r.t. which we can compute sensitivities are the
+        # non-fixed parameters
+        return required_sensitivities - set(self.amici_model.getParameterIds())
+
     def get_config(self) -> dict:
         """Return basic information of the objective configuration."""
         info = super().get_config()
@@ -456,6 +484,19 @@ def call_unprocessed(
         """
         import amici
 
+        # Check that we can compute the requested sensitivities
+        #  based on the supplied model. If not, raise an error.
+        #  This has to be done on every call, since the preprocessor may
+        #  change the parameters w.r.t. which sensitivities are required.
+        # max(0, 0) to work with empty sensi_orders
+        if max(0, 0, *sensi_orders) > 0 and (
+            missing_sensitivities := self._get_missing_sensitivities()
+        ):
+            raise ValueError(
+                f"Requested sensitivities w.r.t. parameters that can't "
+                f"be computed by the current model: {missing_sensitivities}."
+            )
+
         x_dct = self.par_arr_to_dct(x)
 
         # only ask amici to compute required quantities

pypesto/petab/importer.py

@@ -447,6 +447,9 @@ def create_objective(
         Returns
         -------
         A :class:`pypesto.objective.AmiciObjective` for the model and the data.
+        This object is expected to be passed to
+        :class:`PetabImporter.create_problem` to correctly handle fixed
+        parameters.
         """
         # get simulation conditions
         simulation_conditions = petab.get_simulation_conditions(

test/petab/test_petab_import.py

@@ -167,16 +167,19 @@ def test_max_sensi_order():
     """Test that the AMICI objective created via PEtab exposes derivatives
     correctly."""
     model_name = "Boehm_JProteomeRes2014"
-    problem = pypesto.petab.PetabImporter.from_yaml(
+    importer = pypesto.petab.PetabImporter.from_yaml(
         os.path.join(models.MODELS_DIR, model_name, model_name + ".yaml")
     )
+    problem = importer.create_problem()
 
     # define test parameter
-    par = problem.petab_problem.x_nominal_scaled
-    npar = len(par)
+    par = np.asarray(importer.petab_problem.x_nominal_scaled)[
+        problem.x_free_indices
+    ]
+    npar = problem.dim
 
     # auto-computed max_sensi_order and fim_for_hess
-    objective = problem.create_objective()
+    objective = problem.objective
     hess = objective(par, sensi_orders=(2,))
     assert hess.shape == (npar, npar)
     assert (hess != 0).any()
@@ -190,18 +193,24 @@ def test_max_sensi_order():
     )
 
     # fix max_sensi_order to 1
-    objective = problem.create_objective(max_sensi_order=1)
+    objective = importer.create_problem(
+        objective=importer.create_objective(max_sensi_order=1)
+    ).objective
     objective(par, sensi_orders=(1,))
     with pytest.raises(ValueError):
         objective(par, sensi_orders=(2,))
 
     # do not use FIM
-    objective = problem.create_objective(fim_for_hess=False)
+    objective = importer.create_problem(
+        objective=importer.create_objective(fim_for_hess=False)
+    ).objective
     with pytest.raises(ValueError):
         objective(par, sensi_orders=(2,))
 
     # only allow computing function values
-    objective = problem.create_objective(max_sensi_order=0)
+    objective = importer.create_problem(
+        objective=importer.create_objective(max_sensi_order=0)
+    ).objective
     objective(par)
     with pytest.raises(ValueError):
         objective(par, sensi_orders=(1,))