Adjusting the parameters for QuantumMDM (class and pipeline) (#230)

* Removing un unused parameter "gen_feature_map" for QuantumMDM
* Adding a new parameter "upper_bound" for QuantumMDM
* Fixed a problem in base class
* Fixed a problem in VotingClassifier
* Tested with QuantumMDM and QuantumSVM

pyriemann_qiskit/classification.py

@@ -182,7 +182,8 @@ def fit(self, X, y):
 
         n_features = get_feature_dimension(self._training_input)
         self._log("Feature dimension = ", n_features)
-        self._feature_map = self.gen_feature_map(n_features)
+        if hasattr(self, "gen_feature_map") and self.gen_feature_map is not None:
+            self._feature_map = self.gen_feature_map(n_features)
         if self.quantum:
             if not hasattr(self, "_backend"):
                 devices = get_devices(self._provider, n_features)
@@ -619,11 +620,10 @@ class QuanticMDM(QuanticClassifierBase):
         If true, will output all intermediate results and logs.
     shots : int (default:1024)
         Number of repetitions of each circuit, for sampling.
-    gen_feature_map : Callable[int, QuantumCircuit | FeatureMap] \
-                      (default : Callable[int, ZZFeatureMap])
-        Function generating a feature map to encode data into a quantum state.
     seed: int | None (default: None)
         Random seed for the simulation
+    upper_bound : int (default: 7)
+        The maximum integer value for matrix normalization.
 
     See Also
     --------
@@ -651,21 +651,22 @@ def __init__(
         q_account_token=None,
         verbose=True,
         shots=1024,
-        gen_feature_map=gen_zz_feature_map(),
         seed=None,
+        upper_bound=7,
     ):
         QuanticClassifierBase.__init__(
-            self, quantum, q_account_token, verbose, shots, gen_feature_map, seed
+            self, quantum, q_account_token, verbose, shots, None, seed
         )
         self.metric = metric
+        self.upper_bound = upper_bound
 
     def _init_algo(self, n_features):
         self._log("Convex MDM initiating algorithm")
         classifier = MDM(metric=self.metric)
         if self.quantum:
             self._log("Using NaiveQAOAOptimizer")
             self._optimizer = NaiveQAOAOptimizer(
-                quantum_instance=self._quantum_instance, upper_bound=7
+                quantum_instance=self._quantum_instance, upper_bound=self.upper_bound
             )
         else:
             self._log("Using ClassicalOptimizer (COBYLA)")

pyriemann_qiskit/pipelines.py

@@ -9,7 +9,7 @@
 from pyriemann.preprocessing import Whitening
 from pyriemann_qiskit.utils.filtering import NoDimRed
 from pyriemann_qiskit.utils.hyper_params_factory import (
-    gen_zz_feature_map,
+    # gen_zz_feature_map,
     gen_x_feature_map,
     gen_two_local,
     get_spsa,
@@ -331,9 +331,8 @@ class QuantumMDMWithRiemannianPipeline(BasePipeline):
         If true, will output all intermediate results and logs.
     shots : int (default:1024)
         Number of repetitions of each circuit, for sampling.
-    gen_feature_map : Callable[int, QuantumCircuit | FeatureMap] \
-                      (default : Callable[int, XFeatureMap])
-        Function generating a feature map to encode data into a quantum state.
+    upper_bound : int (default: 7)
+        The maximum integer value for matrix normalization.
 
     Attributes
     ----------
@@ -357,14 +356,14 @@ def __init__(
         q_account_token=None,
         verbose=True,
         shots=1024,
-        gen_feature_map=gen_zz_feature_map(),
+        upper_bound=7,
     ):
         self.convex_metric = convex_metric
         self.quantum = quantum
         self.q_account_token = q_account_token
         self.verbose = verbose
         self.shots = shots
-        self.gen_feature_map = gen_feature_map
+        self.upper_bound = upper_bound
 
         BasePipeline.__init__(self, "QuantumMDMWithRiemannianPipeline")
 
@@ -390,12 +389,12 @@ def _create_pipe(self):
             filtering = NoDimRed()
 
         clf = QuanticMDM(
-            metric,
-            self.quantum,
-            self.q_account_token,
-            self.verbose,
-            self.shots,
-            self.gen_feature_map,
+            metric=metric,
+            quantum=self.quantum,
+            q_account_token=self.q_account_token,
+            verbose=self.verbose,
+            shots=self.shots,
+            upper_bound=self.upper_bound,
         )
 
         return make_pipeline(covariances, filtering, clf)
@@ -451,13 +450,13 @@ def __init__(
         q_account_token=None,
         verbose=True,
         shots=1024,
-        gen_feature_map=gen_zz_feature_map(),
+        upper_bound=7,
     ):
         self.quantum = quantum
         self.q_account_token = q_account_token
         self.verbose = verbose
         self.shots = shots
-        self.gen_feature_map = gen_feature_map
+        self.upper_bound = upper_bound
         BasePipeline.__init__(self, "QuantumMDMVotingClassifier")
 
     def _create_pipe(self):
@@ -467,15 +466,15 @@ def _create_pipe(self):
             self.q_account_token,
             self.verbose,
             self.shots,
-            self.gen_feature_map,
+            self.upper_bound,
         )
         clf_mean_convex_dist_euclid = QuantumMDMWithRiemannianPipeline(
             "mean",
             self.quantum,
             self.q_account_token,
             self.verbose,
             self.shots,
-            self.gen_feature_map,
+            self.upper_bound,
         )
 
         return make_pipeline(