Split Datasets class from Experiments

benchmark/Arguments.py

@@ -1,6 +1,6 @@
 import sys
 import argparse
-from .Experiments import Models
+from .Models import Models
 from .Utils import Files, NO_ENV
 
 ALL_METRICS = (

benchmark/Datasets.py

@@ -0,0 +1,103 @@
+import os
+import pandas as pd
+from scipy.io import arff
+from .Utils import Files
+from .Arguments import EnvData
+
+
+class Diterator:
+    def __init__(self, data):
+        self._stack = data.copy()
+
+    def __next__(self):
+        if len(self._stack) == 0:
+            raise StopIteration()
+        return self._stack.pop(0)
+
+
+class DatasetsArff:
+    @staticmethod
+    def dataset_names(name):
+        return f"{name}.arff"
+
+    @staticmethod
+    def folder():
+        return "datasets"
+
+    def load(self, name, class_name="class"):
+        file_name = os.path.join(self.folder(), self.dataset_names(name))
+        data = arff.loadarff(file_name)
+        df = pd.DataFrame(data[0])
+        y = df[class_name]
+        X = data.drop(class_name, axis=1).to_numpy()
+        y = data[class_name].to_numpy()
+        return X, y
+
+
+class DatasetsTanveer:
+    @staticmethod
+    def dataset_names(name):
+        return f"{name}_R.dat"
+
+    @staticmethod
+    def folder():
+        return "data"
+
+    def load(self, name):
+        file_name = os.path.join(self.folder(), self.dataset_names(name))
+        data = pd.read_csv(
+            file_name,
+            sep="\t",
+            index_col=0,
+        )
+        X = data.drop("clase", axis=1).to_numpy()
+        y = data["clase"].to_numpy()
+        return X, y
+
+
+class DatasetsSurcov:
+    @staticmethod
+    def dataset_names(name):
+        return f"{name}.csv"
+
+    @staticmethod
+    def folder():
+        return "datasets"
+
+    def load(self, name):
+        file_name = os.path.join(self.folder(), self.dataset_names(name))
+        data = pd.read_csv(
+            file_name,
+            index_col=0,
+        )
+        data.dropna(axis=0, how="any", inplace=True)
+        self.columns = data.columns
+        col_list = ["class"]
+        X = data.drop(col_list, axis=1).to_numpy()
+        y = data["class"].to_numpy()
+        return X, y
+
+
+class Datasets:
+    def __init__(self, dataset_name=None):
+        envData = EnvData.load()
+        class_name = getattr(
+            __import__(__name__),
+            f"Datasets{envData['source_data']}",
+        )
+        self.dataset = class_name()
+        if dataset_name is None:
+            file_name = os.path.join(self.dataset.folder(), Files.index)
+            with open(file_name) as f:
+                self.data_sets = f.read().splitlines()
+        else:
+            self.data_sets = [dataset_name]
+
+    def load(self, name):
+        try:
+            return self.dataset.load(name)
+        except FileNotFoundError:
+            raise ValueError(f"Unknown dataset: {name}")
+
+    def __iter__(self) -> Diterator:
+        return Diterator(self.data_sets)

benchmark/Experiments.py

@@ -6,7 +6,6 @@ import time
 from datetime import datetime
 from tqdm import tqdm
 import numpy as np
-import pandas as pd
 from sklearn.model_selection import (
     StratifiedKFold,
     KFold,
@@ -14,93 +13,14 @@ from sklearn.model_selection import (
     cross_validate,
 )
 from .Utils import Folders, Files, NO_RESULTS
+from .Datasets import Datasets
 from .Models import Models
-from .Arguments import EnvData
 
 
 class Randomized:
     seeds = [57, 31, 1714, 17, 23, 79, 83, 97, 7, 1]
 
 
-class Diterator:
-    def __init__(self, data):
-        self._stack = data.copy()
-
-    def __next__(self):
-        if len(self._stack) == 0:
-            raise StopIteration()
-        return self._stack.pop(0)
-
-
-class DatasetsTanveer:
-    @staticmethod
-    def dataset_names(name):
-        return f"{name}_R.dat"
-
-    @staticmethod
-    def folder():
-        return "data"
-
-    def load(self, name):
-        file_name = os.path.join(self.folder(), self.dataset_names(name))
-        data = pd.read_csv(
-            file_name,
-            sep="\t",
-            index_col=0,
-        )
-        X = data.drop("clase", axis=1).to_numpy()
-        y = data["clase"].to_numpy()
-        return X, y
-
-
-class DatasetsSurcov:
-    @staticmethod
-    def dataset_names(name):
-        return f"{name}.csv"
-
-    @staticmethod
-    def folder():
-        return "datasets"
-
-    def load(self, name):
-        file_name = os.path.join(self.folder(), self.dataset_names(name))
-        data = pd.read_csv(
-            file_name,
-            index_col=0,
-        )
-        data.dropna(axis=0, how="any", inplace=True)
-        self.columns = data.columns
-        col_list = ["class"]
-        X = data.drop(col_list, axis=1).to_numpy()
-        y = data["class"].to_numpy()
-        return X, y
-
-
-class Datasets:
-    def __init__(self, dataset_name=None):
-        envData = EnvData.load()
-        class_name = getattr(
-            __import__(__name__),
-            f"Datasets{envData['source_data']}",
-        )
-        self.dataset = class_name()
-        if dataset_name is None:
-            file_name = os.path.join(self.dataset.folder(), Files.index)
-            with open(file_name) as f:
-                self.data_sets = f.read().splitlines()
-        else:
-            self.data_sets = [dataset_name]
-
-    def load(self, name):
-        try:
-            return self.dataset.load(name)
-        except FileNotFoundError:
-            raise ValueError(f"Unknown dataset: {name}")
-
-    def __iter__(self) -> Diterator:
-        return Diterator(self.data_sets)
-
-
 class BestResults:
     def __init__(self, score, model, datasets, quiet=False):
         self.score_name = score

benchmark/Results.py

@@ -7,7 +7,8 @@ import shutil
 import subprocess
 import xlsxwriter
 import numpy as np
-from .Experiments import Datasets, BestResults
+from .Experiments import BestResults
+from .Datasets import Datasets
 from .Utils import (
     Folders,
     Files,

benchmark/__init__.py

@@ -1,4 +1,5 @@
-from .Experiments import Experiment, Datasets, DatasetsSurcov, DatasetsTanveer
+from .Datasets import Datasets, DatasetsSurcov, DatasetsTanveer
+from .Experiments import Experiment
 from .Results import Report, Summary
 
 __author__ = "Ricardo Montañana Gómez"

benchmark/scripts/be_build_best.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python
 from benchmark.Results import ReportBest
-from benchmark.Experiments import Datasets, BestResults
+from benchmark.Experiments import BestResults
+from benchmark.Datasets import Datasets
 from benchmark.Arguments import Arguments
 
 """Build a json file with the best results of a model and its hyperparameters

benchmark/scripts/be_grid.py

@@ -1,5 +1,6 @@
 #!/usr/bin/env python
-from benchmark.Experiments import GridSearch, Datasets
+from benchmark.Experiments import GridSearch
+from benchmark.Datasets import Datasets
 from benchmark.Arguments import Arguments
 
 """Do experiment and build result file, optionally print report with results

benchmark/scripts/be_main.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python
 import os
-from benchmark.Experiments import Experiment, Datasets
+from benchmark.Experiments import Experiment
+from benchmark.Datasets import Datasets
 from benchmark.Results import Report
 from benchmark.Arguments import Arguments
 

benchmark/scripts/be_print_strees.py

@@ -3,7 +3,7 @@ import os
 import json
 from stree import Stree
 from graphviz import Source
-from benchmark.Experiments import Datasets
+from benchmark.Datasets import Datasets
 from benchmark.Utils import Files, Folders
 from benchmark.Arguments import Arguments
 

benchmark/tests/BestResults_test.py

@@ -1,6 +1,7 @@
 import os
 from .TestBase import TestBase
-from ..Experiments import BestResults, Datasets
+from ..Experiments import BestResults
+from ..Datasets import Datasets
 
 
 class BestResultTest(TestBase):

benchmark/tests/Dataset_test.py

@@ -1,6 +1,7 @@
 import shutil
 from .TestBase import TestBase
-from ..Experiments import Randomized, Datasets
+from ..Experiments import Randomized
+from ..Datasets import Datasets
 
 
 class DatasetTest(TestBase):

benchmark/tests/Experiment_test.py

@@ -1,6 +1,7 @@
 import json
 from .TestBase import TestBase
-from ..Experiments import Experiment, Datasets
+from ..Experiments import Experiment
+from ..Datasets import Datasets
 
 
 class ExperimentTest(TestBase):

benchmark/tests/GridSearch_test.py

@@ -1,6 +1,7 @@
 import json
 from .TestBase import TestBase
-from ..Experiments import GridSearch, Datasets
+from ..Experiments import GridSearch
+from ..Datasets import Datasets
 
 
 class GridSearchTest(TestBase):

benchmark/tests/results/grid_output_accuracy_STree.json

@@ -6,7 +6,7 @@
             "kernel": "liblinear",
             "multiclass_strategy": "ovr"
         },
-        "v. 1.2.4, Computed on Test on 2022-02-22 at 12:00:00 took 1s"
+        "v. 1.3.0, Computed on Test on 2022-02-22 at 12:00:00 took 1s"
     ],
     "balloons": [
         0.625,
@@ -15,6 +15,6 @@
             "kernel": "linear",
             "multiclass_strategy": "ovr"
         },
-        "v. 1.2.4, Computed on Test on 2022-02-22 at 12:00:00 took 1s"
+        "v. 1.3.0, Computed on Test on 2022-02-22 at 12:00:00 took 1s"
     ]
 }

requirements.txt

@@ -1,5 +1,6 @@
 pandas
 scikit-learn
+scipy
 odte
 mufs
 xlsxwriter