Allow for condition ids to be defined inside array files (SciML problems only)

petab/v2/converters.py

@@ -249,6 +249,14 @@ def _convert_experiment(self, experiment: Experiment) -> None:
             #  single-period experiments.
             if i_period == 0:
                 exp_ind_id = self.get_experiment_indicator(experiment.id)
+                # Skip if condition ids are set by array data
+                # importers handle this
+                if self._new_problem.extensions.sciml:
+                    if set(period.condition_ids).issubset(
+                        self._new_problem.extensions.sciml._get_array_data_condition_ids()
+                    ):
+                        continue
+
                 for change in self._new_problem.get_changes_for_period(period):
                     period0_assignments.setdefault(
                         change.target_id, []

petab/v2/extensions/sciml.py

@@ -303,3 +303,12 @@ def from_config(
             hybridization_tables=hybridization_tables,
             array_data_files=array_data_files,
         )
+
+    def _get_array_data_condition_ids(self) -> set[str]:
+        """Get the set of condition IDs that are referenced in the array data
+        files."""
+        condition_ids = set()
+        for array_data in self.array_data_files:
+            for input in array_data.inputs.values():
+                condition_ids.update(input.keys())
+        return condition_ids

petab/v2/lint.py

@@ -532,6 +532,13 @@ class CheckExperimentConditionsExist(ValidationTask):
     def run(self, problem: Problem) -> ValidationIssue | None:
         messages = []
         available_conditions = {c.id for c in problem.conditions}
+        if problem.extensions.sciml:
+            available_conditions |= {
+                key
+                for array_data in problem.extensions.sciml.array_data_files
+                for input_array in array_data.inputs.values()
+                for key in input_array.keys()
+            }
         for experiment in problem.experiments:
             missing_conditions = (
                 set(

tests/v2/test_sciml.py

@@ -48,9 +48,9 @@ def _get_test_problem():
       -> B; gamma_;
     end
     """)
-    problem.add_experiment("e1", 0, "")
+    problem.add_experiment("e1", 0, "cond1")
     problem.add_mapping("net1_input1", "net1.inputs[0][0]")
-    problem.add_mapping("net1_input2", "net1.inputs[0][1]")
+    problem.add_mapping("net1_input2", "net1.inputs[1]")
     problem.add_mapping("net1_output1", "net1.outputs[0][0]")
     problem.add_mapping("net1_ps", "net1.parameters")
     problem.add_measurement("B_obs", time=1, measurement=1, experiment_id="e1")
@@ -62,7 +62,7 @@ def _get_test_problem():
         "net1_ps", estimate=True, lb=-np.inf, ub=np.inf, nominal_value="array"
     )
     problem.extensions.sciml.add_hybridization("net1_input1", "A")
-    problem.extensions.sciml.add_hybridization("net1_input2", "B")
+    problem.extensions.sciml.add_hybridization("net1_input2", "array")
     problem.extensions.sciml.add_hybridization("gamma_", "net1_output1")
     problem.extensions.sciml.add_neural_network_from_dict(
         "net1",
@@ -106,7 +106,11 @@ def _get_test_problem():
     problem.extensions.sciml.add_array_data_from_dict(
         {
             "metadata": {"pytorch_format": True},
-            "inputs": {},
+            "inputs": {
+                "net1_input2": {
+                    "cond1": np.random.randn(2),
+                }
+            },
             "parameters": {
                 "net1": {
                     "layer1": {