LogTransformer raises error if transform is called before it's fitted

src/transforms.py

@@ -1,6 +1,8 @@
 import numpy as np
 from sklearn.base import BaseEstimator, TransformerMixin
 import pandas as pd
+from .exceptions import DataProcessorError
+from sklearn.exceptions import NotFittedError
 
 
 class LogTransformer(BaseEstimator, TransformerMixin):
@@ -9,13 +11,18 @@ class LogTransformer(BaseEstimator, TransformerMixin):
     replace the values in the column with a logarithm: X[i, col] = np.log(1 + X[i, col] - min(X[:, col]))
     """
     def __init__(self, threshold=1e5):
+        self.fitted = False
         self.threshold = threshold
 
+    def __str__(self):
+        return 'LogTransformer(threshold={})'.format(self.threshold)
+
     def _reset(self):
         if hasattr(self, 'columns_'):
             del self.columns_
             del self.column_names_
             del self.min_vals_
+            self.fitted = False
 
     def fit(self, X):
         self._reset()
@@ -50,15 +57,20 @@ def fit(self, X):
         if type(X) == pd.DataFrame:
             self.column_names_ = [X.columns[i] for i in self.columns_]
 
+        self.fitted = True
+
         return self
 
     def transform(self, X):
-        if type(X) == pd.DataFrame:
-            Y = X.values
-        else:
-            Y = X
+        if self.fitted:
+            if type(X) == pd.DataFrame:
+                Y = X.values
+            else:
+                Y = X
 
-        for i, col in enumerate(self.columns_):
-            Y[:, col] = np.log1p(Y[:, col] - self.min_vals_[i])
+            for i, col in enumerate(self.columns_):
+                Y[:, col] = np.log1p(Y[:, col] - self.min_vals_[i])
 
-        return X
+            return X
+        else:
+            raise NotFittedError('This LogTransformer has not been fitted yet')