conflict resolved

.gitlab-ci.yml

@@ -32,6 +32,8 @@ check:
 test:
   stage: test
   script:
+    - python3 -m pip install cmake --user
+    - python3 -m pip install swig --user
     - tox -e clean,py38,report
   artifacts:
     reports:

src/dfba_sampling/metamodel.py

@@ -175,7 +175,7 @@ class Metamodel(object):
         index_col=None,
         exp=None,
         metamodel_config=None,
-        relative_path='./',
+        relative_path="./",
     ):
         self.kernel_type = kernel_type
         self.bact = bact
@@ -201,10 +201,11 @@ class Metamodel(object):
                 "selected_hyperparameters" in self.metamodel_yml.keys()
                 and "identity" in self.metamodel_yml["selected_hyperparameters"]
             ):
-                tmp_path=relative_path+self.metamodel_yml["selected_hyperparameters"]["identity"]
-                with open(
-                    tmp_path, "r"
-                ) as file:  # Load the user parameter yml
+                tmp_path = (
+                    relative_path
+                    + self.metamodel_yml["selected_hyperparameters"]["identity"]
+                )
+                with open(tmp_path, "r") as file:  # Load the user parameter yml
                     identity = yaml.safe_load(file)
                     for spec in identity:
                         self.metamodel_yml[spec]["identity"] = identity[spec]
@@ -239,14 +240,15 @@ class Metamodel(object):
             ]
             self.C_learn = self.C_learn[col_subset]
             self.constraint_ID_in_Y = np.array([self.exp.y_id(i) for i in col_subset])
-            if (self.metamodel_yml is not None
+            if (
+                self.metamodel_yml is not None
                 and "selected_hyperparameters" in self.metamodel_yml.keys()
                 and "identity" in self.metamodel_yml["selected_hyperparameters"]
             ):
                 self.constraint_Id_in_Y_identity = [
-                        self.exp.y_id(comp)
-                        for comp in self.metamodel_yml[self.bact]["identity"]
-                    ]                
+                    self.exp.y_id(comp)
+                    for comp in self.metamodel_yml[self.bact]["identity"]
+                ]
         else:
             self.constraint_ID_in_Y = np.arange(self.C_learn.shape[1])
         self.substrate_metabolites = list(self.C_learn.columns)
@@ -315,14 +317,15 @@ class Metamodel(object):
             ]
             self.Flux_learn = self.Flux_learn[col_subset]
             self.Flux_ID_in_Y = np.array([self.exp.y_id(i) for i in col_subset])
-            if (self.metamodel_yml is not None
+            if (
+                self.metamodel_yml is not None
                 and "selected_hyperparameters" in self.metamodel_yml.keys()
                 and "identity" in self.metamodel_yml["selected_hyperparameters"]
             ):
                 self.Flux_Id_in_output_identity = [
-                        col_subset.index(comp)
-                        for comp in self.metamodel_yml[self.bact]["identity"]
-                    ]
+                    col_subset.index(comp)
+                    for comp in self.metamodel_yml[self.bact]["identity"]
+                ]
         else:
             self.Flux_ID_in_Y = np.arange(self.Flux_learn.shape[1])
         self.compound_name = list(self.Flux_learn.columns)
@@ -608,18 +611,26 @@ class Metamodel(object):
                     )
             if self.Scaler_fluxes_flag != "None":
                 output = self.Scaler_fluxes.inverse_transform(output)
-            if (self.metamodel_yml is not None
+            if (
+                self.metamodel_yml is not None
                 and "selected_hyperparameters" in self.metamodel_yml.keys()
                 and "identity" in self.metamodel_yml["selected_hyperparameters"]
             ):
-                output[:,self.Flux_Id_in_output_identity] = c_in[:,self.constraint_Id_in_Y_identity]
+                output[:, self.Flux_Id_in_output_identity] = c_in[
+                    :, self.constraint_Id_in_Y_identity
+                ]
             output = output.clip(
                 self.Clipping.data_min_,
                 self.Clipping.data_max_,
             )
         else:
             output = 0.0
-        output_y0[:, self.Flux_ID_in_Y] = output        
+        output_y0[:, self.Flux_ID_in_Y] = output
+        # clipping by constraint c_in since output cannot be lower than c_in input by definition of FBA constraints
+        # output_y0 is ordered as Y.
+        tmp_output_for_constraints = output_y0[:, self.constraint_ID_in_Y]
+        tmp_output_for_constraints = np.maximum(tmp_output_for_constraints, c)
+        output_y0[:, self.constraint_ID_in_Y] = tmp_output_for_constraints
         return output_y0
 
     def Gram_matrix_sqrt(self, eps=1e-8):