add target encodingt to API, fixes with RAF algo

fedot_ind/api/main.py

@@ -69,6 +69,10 @@ def __init__(self, **kwargs):
         self.preprocessing = kwargs.get('industrial_preprocessing', False)
         self.backend_method = kwargs.get('backend', 'cpu')
         self.RAF_workers = kwargs.get('RAF_workers', None)
+        self.path_to_composition_results = kwargs.get('history_dir', None)
+        prefix = './composition_results' if self.path_to_composition_results is None \
+            else self.path_to_composition_results
+        Path(prefix).mkdir(parents=True, exist_ok=True)
 
         if self.output_folder is None:
             self.output_folder = default_path_to_save_results
@@ -93,9 +97,11 @@ def __init__(self, **kwargs):
         self.predicted_probs = None
         self.predict_data = None
         self.config_dict = kwargs
+        self.config_dict['history_dir'] = prefix
         self.config_dict['available_operations'] = kwargs.get('available_operations',
                                                               default_industrial_availiable_operation(
                                                                   self.config_dict['problem']))
+
         self.config_dict['optimizer'] = kwargs.get(
             'optimizer', IndustrialEvoOptimizer)
         self.config_dict['initial_assumption'] = kwargs.get('initial_assumption',
@@ -176,8 +182,10 @@ def fit(self,
         """
         self.train_data = deepcopy(
             input_data)  # we do not want to make inplace changes
-        self.train_data = DataCheck(
-            input_data=self.train_data, task=self.config_dict['problem']).check_input_data()
+        input_preproc = DataCheck(
+            input_data=self.train_data, task=self.config_dict['problem'])
+        self.train_data = input_preproc.check_input_data()
+        self.target_encoder = input_preproc.get_target_encoder()
         self.solver = self.__init_solver()
         if self.preprocessing:
             self._preprocessing_strategy(self.train_data)
@@ -205,8 +213,8 @@ def predict(self,
         elif isinstance(self.solver, list):
             predict = self._predict_raf_ensemble()
         else:
-            predict = self.solver.predict(self.predict_data, 'labels').predict
-        self.predicted_labels = predict
+            predict = self.solver.predict(self.predict_data, 'labels')
+        self.predicted_labels = predict if self.target_encoder is None else self.target_encoder.transform(predict)
         return self.predicted_labels
 
     def predict_proba(self,
@@ -229,7 +237,7 @@ def predict_proba(self,
         elif isinstance(self.solver, list):
             return self.predicted_probs if self.predicted_probs is not None else self._predict_raf_ensemble()
         else:
-            predict = self.solver.predict(self.predict_data, 'probs').predict
+            predict = self.solver.predict(self.predict_data, 'probs')
         self.predicted_probs = predict
         return self.predicted_probs
 

fedot_ind/api/utils/checkers_collections.py

@@ -34,6 +34,7 @@ def __init__(self,
         self.input_data = input_data
         self.task = task
         self.task_dict = FEDOT_TASK
+        self.label_encoder = None
 
     def __check_features_and_target(self, X, y):
         multi_features, X = check_multivariate_data(X)
@@ -72,9 +73,11 @@ def _init_input_data(self) -> None:
             features, is_multivariate_data, target = self.__check_features_and_target(
                 X, y)
 
-        if y is not None and type(y[0]) is np.str_ and self.task == 'classification':
-            label_encoder = LabelEncoder()
-            target = label_encoder.fit_transform(target)
+        if self.label_encoder is None and type(y[0]) is np.str_ and self.task == 'classification':
+            self.label_encoder = LabelEncoder()
+            target = self.label_encoder.fit_transform(target)
+        else:
+            self.label_encoder = self.label_encoder
 
         if is_multivariate_data:
             self.input_data = InputData(idx=np.arange(len(X)),
@@ -146,3 +149,6 @@ def check_input_data(self) -> InputData:
         self._check_input_data_features()
         self._check_input_data_target()
         return self.input_data
+
+    def get_target_encoder(self):
+        return self.label_encoder

fedot_ind/core/ensemble/random_automl_forest.py

@@ -1,11 +1,15 @@
+from copy import deepcopy
+
 from fedot.core.data.data import InputData
 from fedot.core.data.multi_modal import MultiModalData
 from fedot.core.pipelines.pipeline import Pipeline
 from fedot.core.pipelines.pipeline_builder import PipelineBuilder
 from fedot.core.repository.dataset_types import DataTypesEnum
 from sklearn.model_selection import StratifiedKFold, StratifiedGroupKFold, train_test_split
 from fedot_ind.core.architecture.settings.computational import backend_methods as np
-from fedot_ind.core.repository.constanst_repository import FEDOT_ATOMIZE_OPERATION, FEDOT_HEAD_ENSEMBLE, FEDOT_TASK
+from fedot_ind.core.repository.constanst_repository import FEDOT_ATOMIZE_OPERATION, FEDOT_HEAD_ENSEMBLE, FEDOT_TASK, \
+    FEDOT_ENSEMBLE_ASSUMPTIONS
+from fedot_ind.core.repository.model_repository import SKLEARN_CLF_MODELS, SKLEARN_REG_MODELS
 
 
 class RAFensembler:
@@ -81,7 +85,13 @@ def _raf_ensemble(self, features, target, n_splits):
                 branch_idx=i)
             data_dict.update({f'data_source_img/{i}': train_fold})
         train_multimodal = MultiModalData(data_dict)
+        head_automl_params = deepcopy(self.atomized_automl_params)
+        head_automl_params['available_operations'] = [operation for operation
+                                                      in head_automl_params['available_operations']
+                                                      if operation in list(SKLEARN_CLF_MODELS.keys())
+                                                      or operation in list(SKLEARN_REG_MODELS.keys())]
+        head_automl_params['initial_assumption'] = FEDOT_ENSEMBLE_ASSUMPTIONS[self.atomized_automl_params['problem']].build()
         raf_ensemble = raf_ensemble.join_branches(self.head,
-                                                  params=self.atomized_automl_params).build()
+                                                  params=head_automl_params).build()
         raf_ensemble.fit(input_data=train_multimodal)
         return raf_ensemble

fedot_ind/core/repository/constanst_repository.py

@@ -39,7 +39,7 @@ class ComputationalConstant(Enum):
     }
     BATCH_SIZE_FOR_FEDOT_WORKER = 1000
     FEDOT_WORKER_NUM = 5
-    FEDOT_WORKER_TIMEOUT_PARTITION = 2
+    FEDOT_WORKER_TIMEOUT_PARTITION = 4
     PATIENCE_FOR_EARLY_STOP = 15
 
 
@@ -186,6 +186,11 @@ class FedotOperationConstant(Enum):
         'ts_forecasting': PipelineBuilder().add_node('lagged').add_node('ridge')
     }
 
+    FEDOT_ENSEMBLE_ASSUMPTIONS = {
+        'classification': PipelineBuilder().add_node('logit'),
+        'regression': PipelineBuilder().add_node('treg')
+    }
+
 
 class ModelCompressionConstant(Enum):
     ENERGY_THR = [0.9, 0.95, 0.99, 0.999]
@@ -405,6 +410,7 @@ class BenchmarkDatasets(Enum):
 FEDOT_TUNING_METRICS = FedotOperationConstant.FEDOT_TUNING_METRICS.value
 FEDOT_ASSUMPTIONS = FedotOperationConstant.FEDOT_ASSUMPTIONS.value
 FEDOT_API_PARAMS = FedotOperationConstant.FEDOT_API_PARAMS.value
+FEDOT_ENSEMBLE_ASSUMPTIONS = FedotOperationConstant.FEDOT_ENSEMBLE_ASSUMPTIONS.value
 
 CPU_NUMBERS = ComputationalConstant.CPU_NUMBERS.value
 BATCH_SIZE_FOR_FEDOT_WORKER = ComputationalConstant.BATCH_SIZE_FOR_FEDOT_WORKER.value

requirements.txt

@@ -7,16 +7,36 @@ MKLpy==0.6
 PyMonad==2.4.0
 PyWavelets==1.4.1
 
-fastcore
-fastai
-distributed
-datasetsforecast
+fastcore~=1.5.29
+fastai~=2.7.14
+distributed~=2024.1.1
+datasetsforecast~=0.0.8
 
 tensorly==0.8.1
-torch==2.0.0
+torch~=2.2.0
 torchmetrics==0.11.4
-torchvision==0.15.1
+torchvision~=0.17.0
 tensorboard>=2.12.0
 statsforecast==1.5.0
 
-chardet
+chardet~=5.2.0
+sphinx~=7.2.6
+numpy~=1.24.4
+pytest~=8.0.0
+matplotlib~=3.8.2
+pandas~=1.5.3
+fedot~=0.7.3
+scipy~=1.12.0
+typing~=3.7.4.3
+scikit-learn~=1.2.2
+PyYAML~=6.0.1
+setuptools~=60.2.0
+pillow~=10.2.0
+tqdm~=4.65.2
+seaborn~=0.13.2
+joblib~=1.3.2
+hyperopt~=0.2.7
+ripser~=0.6.4
+statsmodels~=0.14.1
+xgboost~=2.0.3
+sktime~=0.16.1