Merge pull request #923 from kernc/Preprocessors,unite!

Learner-providing widgets: minor fixup that didn't make it into 38446d3

Orange/widgets/classify/owclassificationtree.py

@@ -12,17 +12,16 @@ class OWClassificationTree(OWProvidesLearner, widget.OWWidget):
     description = "Classification tree algorithm with forward pruning."
     priority = 30
 
-    inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
+    LEARNER = TreeLearner
 
+    inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
     outputs = [
-        ("Learner", TreeLearner),
+        ("Learner", LEARNER),
         ("Tree", TreeClassifier)
     ]
     want_main_area = False
     resizing_enabled = False
 
-    LEARNER = TreeLearner
-
     model_name = Setting("Classification Tree")
     attribute_score = Setting(0)
     limit_min_leaf = Setting(True)

Orange/widgets/classify/owmajority.py

@@ -12,8 +12,10 @@ class OWMajority(OWProvidesLearner, widget.OWWidget):
     priority = 20
     icon = "icons/Majority.svg"
 
+    LEARNER = MajorityLearner
+
     inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
-    outputs = [("Learner", MajorityLearner),
+    outputs = [("Learner", LEARNER),
                ("Classifier", ConstantModel)]
 
     learner_name = Setting("Majority")
@@ -43,8 +45,6 @@ def set_data(self, data):
         if data is not None:
             self.apply()
 
-    LEARNER = MajorityLearner
-
     def apply(self):
         learner = self.LEARNER(
             preprocessors=self.preprocessors

Orange/widgets/classify/owsvmclassification.py

@@ -18,8 +18,10 @@ class OWSVMClassification(OWProvidesLearner, widget.OWWidget):
                   "selection of kernels."
     icon = "icons/SVM.svg"
 
+    LEARNER = SVMLearner
+
     inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
-    outputs = [("Learner", SVMLearner, widget.Default),
+    outputs = [("Learner", LEARNER, widget.Default),
                ("Classifier", SVMClassifier),
                ("Support vectors", Table)]
 
@@ -131,8 +133,6 @@ def set_data(self, data):
         if data is not None:
             self.apply()
 
-    LEARNER = SVMLearner  # OWProvidesLearner uses this
-
     def apply(self):
         kernel = ["linear", "poly", "rbf", "sigmoid"][self.kernel_type]
         common_args = dict(

Orange/widgets/regression/owknnregression.py

@@ -17,8 +17,10 @@ class OWKNNRegression(OWProvidesLearner, widget.OWWidget):
     icon = "icons/kNearestNeighbours.svg"
     priority = 20
 
+    LEARNER = KNNRegressionLearner
+
     inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
-    outputs = [("Learner", KNNRegressionLearner),
+    outputs = [("Learner", LEARNER),
                ("Predictor", SklModel)]
 
     want_main_area = False
@@ -62,8 +64,6 @@ def set_data(self, data):
         if data is not None:
             self.apply()
 
-    LEARNER = KNNRegressionLearner
-
     def apply(self):
         """
         Construct the learner and apply it on the training data if available.

Orange/widgets/regression/owlinearregression.py

@@ -19,8 +19,10 @@ class OWLinearRegression(OWProvidesLearner, widget.OWWidget):
                   "regularization."
     icon = "icons/LinearRegression.svg"
 
+    LEARNER = LinearRegressionLearner
+
     inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
-    outputs = [("Linear Regression", LinearRegressionLearner),
+    outputs = [("Linear Regression", LEARNER),
                ("Model", LinearModel),
                ("Coefficients", Table)]
 
@@ -97,8 +99,6 @@ def _alpha_changed(self):
         self._set_alpha_label()
         self.commit()
 
-    LEARNER = LinearRegressionLearner  # OWProvidesLearner uses this
-
     def apply(self):
         return self.commit()
 

Orange/widgets/regression/owmean.py

@@ -10,8 +10,11 @@ class OWMean(OWProvidesLearner, widget.OWWidget):
     description = "Regression to the average class value from the training set."
     icon = "icons/Mean.svg"
 
+    LEARNER = MeanLearner
+
     inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
-    outputs = [("Learner", MeanLearner), ("Predictor", MeanModel)]
+    outputs = [("Learner", LEARNER),
+               ("Predictor", MeanModel)]
 
     learner_name = settings.Setting("Mean Learner")
 
@@ -35,8 +38,6 @@ def set_data(self, data):
         if data is not None:
             self.apply()
 
-    LEARNER = MeanLearner
-
     def apply(self):
         learner = self.LEARNER(preprocessors=self.preprocessors)
         learner.name = self.learner_name

Orange/widgets/regression/owrandomforestregression.py

@@ -10,10 +10,10 @@ class OWRandomForestRegression(OWRandomForest):
     name = "Random Forest Regression"
     description = "Random forest regression algorithm."
 
-    outputs = [("Learner", RandomForestRegressionLearner),
+    LEARNER = RandomForestRegressionLearner
+    outputs = [("Learner", LEARNER),
                ("Model", RandomForestRegressor)]
 
-    LEARNER = RandomForestRegressionLearner
     learner_name = settings.Setting("RF Regression Learner")
     n_estimators = settings.Setting(10)
     max_features = settings.Setting(5)

Orange/widgets/regression/owregressiontree.py

@@ -8,9 +8,11 @@ class OWRegressionTree(OWClassificationTree):
     icon = "icons/RegressionTree.svg"
     description = "Regression tree algorithm with forward pruning."
 
-    outputs = [("Learner", TreeRegressionLearner), ("Tree", TreeRegressor)]
-
     LEARNER = TreeRegressionLearner
+
+    outputs = [("Learner", LEARNER),
+               ("Tree", TreeRegressor)]
+
     model_name = Setting("Regression Tree")
     attribute_score = Setting(0)
     limit_min_leaf = Setting(True)

Orange/widgets/regression/owsgdregression.py

@@ -16,8 +16,10 @@ class OWSGDRegression(OWProvidesLearner, widget.OWWidget):
     description = "Stochastic gradient descent algorithm for regression."
     icon = "icons/SGDRegression.svg"
 
+    LEARNER = SGDRegressionLearner
+
     inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
-    outputs = [("Learner", SGDRegressionLearner),
+    outputs = [("Learner", LEARNER),
                ("Predictor", LinearModel)]
 
     learner_name = settings.Setting("SGD Regression")
@@ -136,8 +138,6 @@ def set_data(self, data):
         if data is not None:
             self.apply()
 
-    LEARNER = SGDRegressionLearner
-
     def apply(self):
         loss = ["squared_loss", "huber", "epsilon_insensitive", "squared_epsilon_insensitive"][self.loss_function]
         penalty = ["l1", "l2", "elasticnet"][self.penalty_type]

Orange/widgets/regression/owsvmregression.py

@@ -16,8 +16,11 @@ class OWSVMRegression(OWProvidesLearner, widget.OWWidget):
     name = "SVM Regression"
     description = "Support vector machine regression algorithm."
     icon = "icons/SVMRegression.svg"
+
+    LEARNER = SVRLearner
+
     inputs = [("Data", Table, "set_data")] + OWProvidesLearner.inputs
-    outputs = [("Learner", SVRLearner, widget.Default),
+    outputs = [("Learner", LEARNER, widget.Default),
                ("Predictor", SklModel),
                ("Support vectors", Table)]
 
@@ -150,8 +153,6 @@ def set_data(self, data):
         if data is not None:
             self.apply()
 
-    LEARNER = SVRLearner  # OWProvidesLearner uses this
-
     def apply(self):
         kernel = ["linear", "poly", "rbf", "sigmoid"][self.kernel_type]
         common_args = dict(

Orange/widgets/regression/owunivariateregression.py

@@ -22,9 +22,11 @@ class OWUnivariateRegression(OWProvidesLearner, widget.OWWidget):
     description = "Univariate regression with polynomial expansion."
     icon = "icons/UnivariateRegression.svg"
 
+    LEARNER = LinearRegressionLearner
+
     inputs = [("Data", Table, "set_data", widget.Default),
               ("Learner", Learner, "set_learner")] + OWProvidesLearner.inputs
-    outputs = [("Learner", Learner),
+    outputs = [("Learner", LEARNER),
                ("Predictor", LinearModel)]
 
     learner_name = settings.Setting("Univariate Regression")
@@ -180,8 +182,6 @@ def plot_regression_line(self, x_data, y_data):
         self.plotview.addItem(self.plot_item)
         self.plotview.replot()
 
-    LEARNER = LinearRegressionLearner
-
     def apply(self):
         learner = self.learner
         predictor = None