Micro patch 0.4.1.2 (#126)

* small fixes for 1 sample prediction cases. Excluded IOPT tuner * patch version --------- Co-authored-by: v1docq <[email protected]> Co-authored-by: technocreep <[email protected]>
aimclub · Mar 7, 2024 · 4b271c1 · 4b271c1
1 parent c53e963
commit 4b271c1
Show file tree

Hide file tree

Showing 5 changed files with 16 additions and 9 deletions.
diff --git a/fedot_ind/api/main.py b/fedot_ind/api/main.py
@@ -366,11 +366,12 @@ def load(self, path):
 
         """
         self.repo = IndustrialModels().setup_repository()
-
-        dir_list = os.listdir(path)
-        p = [x for x in dir_list if x.__contains__('pipeline_saved')][0]
-        path = f'{path}/{p}'
+        if not path.__contains__('pipeline_saved'):
+            dir_list = os.listdir(path)
+            p = [x for x in dir_list if x.__contains__('pipeline_saved')][0]
+            path = f'{path}/{p}'
         dir_list = os.listdir(path)
+
         if 'fitted_operations' in dir_list:
             self.solver = Pipeline().load(path)
         else:
@@ -395,6 +396,7 @@ def save_best_model(self):
             for idx, p in enumerate(self.solver.ensemble_branches):
                 Pipeline(p).save(f'./raf_ensemble/{idx}_ensemble_branch', create_subdir=True)
             Pipeline(self.solver.ensemble_head).save(f'./raf_ensemble/ensemble_head', create_subdir=True)
+            self.solver.current_pipeline.save(f'./raf_ensemble/ensemble_composed', create_subdir=True)
 
     def plot_fitness_by_generation(self, **kwargs):
         """Plot prediction of the model"""

diff --git a/fedot_ind/core/architecture/preprocessing/data_convertor.py b/fedot_ind/core/architecture/preprocessing/data_convertor.py
@@ -411,8 +411,10 @@ def output_mode_converter(self, output_mode, n_classes):
             return self.probs_prediction_converter(output_mode, n_classes)
 
     def probs_prediction_converter(self, output_mode, n_classes):
-        prediction = self.operation_implementation.predict_proba(
-            self.train_data.features)
+        try:
+            prediction = self.operation_implementation.predict_proba(self.train_data.features)
+        except Exception:
+            prediction = self.operation_implementation.predict_proba(self.train_data.features.T)
         if n_classes < 2:
             raise ValueError(
                 'Data set contain only 1 target class. Please reformat your data.')

diff --git a/fedot_ind/core/operation/transformation/basis/eigen_basis.py b/fedot_ind/core/operation/transformation/basis/eigen_basis.py
@@ -66,8 +66,11 @@ def _channel_decompose(self, features):
         return predict
 
     def _convert_basis_to_predict(self, basis, input_data):
-        self.predict = basis
 
+        if input_data.features.shape[0] == 1 and len(input_data.features.shape) == 3:
+            self.predict = basis[np.newaxis, :, :]
+        else:
+            self.predict = basis
         if input_data.task.task_params is None:
             input_data.task.task_params = self.__repr__()
         elif input_data.task.task_params not in [self.__repr__(), 'LargeFeatureSpace']:

diff --git a/fedot_ind/core/repository/constanst_repository.py b/fedot_ind/core/repository/constanst_repository.py
@@ -192,7 +192,7 @@ class FedotOperationConstant(Enum):
                             'regression': RegressionMetricsEnum.RMSE}
     FEDOT_TUNER_STRATEGY = {'sequential': partial(SequentialTuner, inverse_node_order=True),
                             'simultaneous': SimultaneousTuner,
-                            'IOptTuner': IOptTuner,
+                          #  'IOptTuner': IOptTuner,
                             'optuna': OptunaTuner}
     FEDOT_HEAD_ENSEMBLE = {'regression': 'treg',
                            'classification': 'logit'}

diff --git a/setup.py b/setup.py
@@ -8,7 +8,7 @@
 
 # The text of the README file
 NAME = 'fedot_ind'
-VERSION = '0.4.1'
+VERSION = '0.4.1.2'
 AUTHOR = 'NSS Lab'
 AUTHOR_EMAIL = '[email protected]'
 SHORT_DESCRIPTION = 'Automated machine learning framework for time series analysis'