hyperopt fix (#257)

Co-authored-by: Gabriel Hurtado <[email protected]>

Co-authored-by: Balazs Kegl <[email protected]>

Author: gabriel-hurtado

rampwf/hyperopt/hyperopt.py

@@ -147,7 +147,7 @@ def python_repr(self):
                 The string representation of the hyperparameter.
         """
         repr = '{} = Hyperparameter(\n'.format(self.name)
-        repr += '\tdtype={}'.format(str(self.dtype))
+        repr += "\tdtype='{}'".format(str(self.dtype))
         repr += ', default={}'.format(self.default_repr)
         repr += ', values={})\n'.format(self.values_repr)
         return repr
@@ -341,7 +341,8 @@ def next_hyperparameter_indices(self, df_scores, n_folds):
             next_value_indices = [
                 h.get_index(v) for h, v
                 in zip(self.hyperparameters, next_values)]
-            fold_i = incomplete_folds.iloc[0]['fold_i'] % n_folds
+            # for some reason iloc converts int to float
+            fold_i = int(incomplete_folds.iloc[0]['fold_i']) % n_folds
         # Otherwise select hyperparameter values from those that haven't
         # been selected yet, using also prior
         else:
@@ -461,8 +462,12 @@ def _save_best_model(self):
         else:
             best_defaults = official_scores.idxmax()
         print('Best hyperparameters: ', best_defaults)
-        for bd, h in zip(best_defaults, self.hyperparameters):
-            h.set_default(bd)
+        try:
+            for bd, h in zip(best_defaults, self.hyperparameters):
+                h.set_default(bd)
+        except(TypeError):
+            # single hyperparameter
+            self.hyperparameters[0].set_default(best_defaults)
         # Overwrite the submission with the best hyperparameter values
         write_hyperparameters(
             self.submission_dir, self.submission_dir,

rampwf/hyperopt/tests/interfaces/header_in_files/titanic/submissions/one_hyper_kit/classifier.py

@@ -0,0 +1,26 @@
+from sklearn.linear_model import LogisticRegression
+from sklearn.impute import SimpleImputer
+from sklearn.pipeline import Pipeline
+from sklearn.base import BaseEstimator
+from rampwf.hyperopt import Hyperparameter
+
+# test with only one hyperparameter
+# RAMP START HYPERPARAMETERS
+logreg_C = Hyperparameter(
+    dtype='float', default=1.0, values=[0.01, 0.1, 0.9, 1.0])
+# RAMP END HYPERPARAMETERS
+
+
+class Classifier(BaseEstimator):
+    def __init__(self):
+        self.clf = Pipeline([
+            ('imputer',
+             SimpleImputer(strategy='median')),
+            ('classifier', LogisticRegression(C=float(logreg_C)))
+        ])
+
+    def fit(self, X, y):
+        self.clf.fit(X, y)
+
+    def predict_proba(self, X):
+        return self.clf.predict_proba(X)

rampwf/hyperopt/tests/interfaces/header_in_files/titanic/submissions/one_hyper_kit/feature_extractor.py

@@ -0,0 +1,22 @@
+import pandas as pd
+
+
+class FeatureExtractor():
+    def __init__(self):
+        pass
+
+    def fit(self, X_df, y):
+        pass
+
+    def transform(self, X_df):
+        X_df_new = pd.concat(
+            [X_df.get(['Fare', 'Age', 'SibSp', 'Parch']),
+                pd.get_dummies(X_df.Sex, prefix='Sex', drop_first=True),
+                pd.get_dummies(X_df.Pclass, prefix='Pclass', drop_first=True),
+                pd.get_dummies(
+                    X_df.Embarked, prefix='Embarked', drop_first=True)],
+            axis=1)
+
+        X_df_new = X_df_new.fillna(-1)
+        XX = X_df_new.values
+        return XX

rampwf/hyperopt/tests/test_hyperparameter.py

@@ -35,10 +35,11 @@ def test_hyperparameter():
     assert str(e.value) == 'Default must be among values.'
 
 
-def test_hyperopt():
+@pytest.mark.parametrize("submission", ['starting_kit', 'one_hyper_kit'])
+def test_hyperopt(submission):
     ramp_kit_dir = os.path.join(
         PATH, 'interfaces', 'header_in_files', 'titanic')
-    submission = 'starting_kit'
     run_hyperopt(
-        ramp_kit_dir, ramp_kit_dir, os.path.join(ramp_kit_dir, 'submissions'),
+        ramp_kit_dir, ramp_kit_dir,
+        os.path.join(ramp_kit_dir, 'submissions'),
         submission, 'random', 64, is_cleanup=True)