Transform: Values of continuous variables as feature names

Author: VesnaT

Orange/data/table.py

@@ -20,7 +20,7 @@
 )
 from Orange.data.util import SharedComputeValue, vstack, hstack, \
     assure_array_dense, assure_array_sparse, \
-    assure_column_dense, assure_column_sparse
+    assure_column_dense, assure_column_sparse, get_unique_names_duplicates
 from Orange.statistics.util import bincount, countnans, contingency, \
     stats as fast_stats, sparse_has_implicit_zeros, sparse_count_implicit_zeros, \
     sparse_implicit_zero_weights
@@ -1604,8 +1604,8 @@ def _compute_contingency(self, col_vars=None, row_var=None):
         return contingencies, unknown_rows
 
     @classmethod
-    def transpose(cls, table, feature_names_column="", meta_attr_name="Feature name",
-                  feature_name="Feature"):
+    def transpose(cls, table, feature_names_column="",
+                  meta_attr_name="Feature name", feature_name="Feature"):
         """
         Transpose the table.
 
@@ -1614,6 +1614,7 @@ def transpose(cls, table, feature_names_column="", meta_attr_name="Feature name"
             use for feature names
         :param meta_attr_name: str - name of new meta attribute into which
             feature names are mapped
+        :param feature_name: str - default feature name prefix
         :return: Table - transposed table
         """
 
@@ -1625,10 +1626,14 @@ def transpose(cls, table, feature_names_column="", meta_attr_name="Feature name"
         # - classes and metas to attributes of attributes
         # - arbitrary meta column to feature names
         self.X = table.X.T
-        attributes = [ContinuousVariable(str(row[feature_names_column]))
-                      for row in table] if feature_names_column else \
-            [ContinuousVariable(feature_name + " " + str(i + 1).zfill(
-                int(np.ceil(np.log10(n_cols))))) for i in range(n_cols)]
+        if feature_names_column:
+            names = [str(row[feature_names_column]) for row in table]
+            names = get_unique_names_duplicates(names)
+            attributes = [ContinuousVariable(name) for name in names]
+        else:
+            places = int(np.ceil(np.log10(n_cols)))
+            attributes = [ContinuousVariable(f"{feature_name} {i:0{places}}")
+                          for i in range(1, n_cols + 1)]
         if old_domain is not None and feature_names_column:
             for i, _ in enumerate(attributes):
                 if attributes[i].name in old_domain:

Orange/data/tests/test_util.py

@@ -1,7 +1,7 @@
 import unittest
 
 from Orange.data import Domain, ContinuousVariable
-from Orange.data.util import get_unique_names
+from Orange.data.util import get_unique_names, get_unique_names_duplicates
 
 
 class TestGetUniqueNames(unittest.TestCase):
@@ -46,6 +46,14 @@ def test_get_unique_names_with_domain(self):
         self.assertEqual(get_unique_names(domain, "foo"), "foo (1)")
         self.assertEqual(get_unique_names(domain, "baz"), "baz (4)")
 
+    def test_get_unique_names_from_duplicates(self):
+        self.assertEqual(
+            get_unique_names_duplicates(["foo", "bar", "baz"]),
+            ["foo", "bar", "baz"])
+        self.assertEqual(
+            get_unique_names_duplicates(["foo", "bar", "baz", "bar"]),
+            ["foo", "bar (1)", "baz", "bar (2)"])
+
 
 if __name__ == "__main__":
     unittest.main()

Orange/data/util.py

@@ -2,6 +2,7 @@
 Data-manipulation utilities.
 """
 import re
+from collections import Counter
 from itertools import chain
 
 import numpy as np
@@ -196,3 +197,19 @@ def get_unique_names(names, proposed):
         return proposed
     max_index = max(map(max, indicess), default=0) + 1
     return [f"{name} ({max_index})" for name in proposed]
+
+
+def get_unique_names_duplicates(proposed: list) -> list:
+    """
+    Returns list of unique names. If a name is duplicated, the
+    function appends an index in parentheses.
+    """
+    counter = Counter(proposed)
+    temp_counter = Counter()
+    names = []
+    for name in proposed:
+        if counter[name] > 1:
+            temp_counter.update([name])
+            name = f"{name} ({temp_counter[name]})"
+        names.append(name)
+    return names

Orange/tests/test_table.py

@@ -2751,6 +2751,15 @@ def test_transpose_class_metas_attributes(self):
         self.assertDictEqual(data.domain.attributes[0].attributes,
                              {"attr1": "a1", "attr2": "aa1"})
 
+    def test_transpose_duplicate_feature_names(self):
+        table = Table("iris")
+        domain = table.domain
+        attrs, metas = domain.attributes[:3], domain.attributes[3:]
+        table = table.transform(Domain(attrs, domain.class_vars, metas))
+        transposed = Table.transpose(table, domain.attributes[3].name)
+        names = [f.name for f in transposed.domain.attributes]
+        self.assertEqual(len(names), len(set(names)))
+
     def test_transpose(self):
         zoo = Table("zoo")
         t1 = Table.transpose(zoo)

Orange/widgets/data/owtranspose.py

@@ -1,4 +1,4 @@
-from Orange.data import Table, StringVariable
+from Orange.data import Table, ContinuousVariable, StringVariable
 from Orange.widgets.settings import (Setting, ContextSetting,
                                      DomainContextHandler)
 from Orange.widgets.utils.itemmodels import DomainModel
@@ -21,7 +21,7 @@ class Inputs:
     class Outputs:
         data = Output("Data", Table, dynamic=False)
 
-    GENERIC, FROM_META_ATTR = range(2)
+    GENERIC, FROM_VAR = range(2)
 
     resizing_enabled = False
     want_main_area = False
@@ -34,6 +34,11 @@ class Outputs:
     feature_names_column = ContextSetting(None)
     auto_apply = Setting(True)
 
+    class Warning(OWWidget.Warning):
+        duplicate_names = Msg("Values are not unique.\nTo avoid multiple "
+                              "features with the same name, values \nof "
+                              "'{}' have been augmented with indices.")
+
     class Error(OWWidget.Error):
         value_error = Msg("{}")
 
@@ -53,10 +58,10 @@ def __init__(self):
             placeholderText="Type a prefix ...", toolTip="Custom feature name")
         edit.editingFinished.connect(self._apply_editing)
 
-        self.meta_button = gui.appendRadioButton(box, "From meta attribute:")
+        self.meta_button = gui.appendRadioButton(box, "From variable:")
         self.feature_model = DomainModel(
-            order=DomainModel.METAS, valid_types=StringVariable,
-            alphabetical=True)
+            valid_types=(ContinuousVariable, StringVariable),
+            alphabetical=False)
         self.feature_combo = gui.comboBox(
             gui.indentedBox(box, gui.checkButtonOffsetHint(button)), self,
             "feature_names_column", contentsLength=12,
@@ -75,7 +80,7 @@ def _apply_editing(self):
         self.apply()
 
     def _feature_combo_changed(self):
-        self.feature_type = self.FROM_META_ATTR
+        self.feature_type = self.FROM_VAR
         self.apply()
 
     @Inputs.data
@@ -95,7 +100,7 @@ def set_controls(self):
         self.meta_button.setEnabled(bool(self.feature_model))
         if self.feature_model:
             self.feature_names_column = self.feature_model[0]
-            self.feature_type = self.FROM_META_ATTR
+            self.feature_type = self.FROM_VAR
         else:
             self.feature_names_column = None
 
@@ -104,10 +109,15 @@ def apply(self):
         transposed = None
         if self.data:
             try:
+                variable = self.feature_type == self.FROM_VAR and \
+                           self.feature_names_column
                 transposed = Table.transpose(
-                    self.data,
-                    self.feature_type == self.FROM_META_ATTR and self.feature_names_column,
+                    self.data, variable,
                     feature_name=self.feature_name or self.DEFAULT_PREFIX)
+                if variable:
+                    names = self.data.get_column_view(variable)[0]
+                    if len(names) != len(set(names)):
+                        self.Warning.duplicate_names(variable)
             except ValueError as e:
                 self.Error.value_error(e)
         self.Outputs.data.send(transposed)
@@ -116,7 +126,7 @@ def send_report(self):
         if self.feature_type == self.GENERIC:
             names = self.feature_name or self.DEFAULT_PREFIX
         else:
-            names = "from meta attribute"
+            names = "from variable"
             if self.feature_names_column:
                 names += "  '{}'".format(self.feature_names_column.name)
         self.report_items("", [("Feature names", names)])
@@ -125,4 +135,4 @@ def send_report(self):
 
 
 if __name__ == "__main__":  # pragma: no cover
-    WidgetPreview(OWTranspose).run(Table("zoo"))
+    WidgetPreview(OWTranspose).run(Table("iris"))

Orange/widgets/data/tests/test_owtranspose.py

@@ -5,7 +5,7 @@
 
 import numpy as np
 
-from Orange.data import Table, Domain
+from Orange.data import Table
 from Orange.widgets.data.owtranspose import OWTranspose
 from Orange.widgets.tests.base import WidgetTest
 
@@ -32,18 +32,11 @@ def test_feature_type(self):
         widget = self.widget
         data = Table("conferences.tab")
         metas = data.domain.metas
-        domain = data.domain
-        # Put one non-string column to metas, so widget must skip it
-        domain2 = Domain(domain.attributes[:-1],
-                         domain.class_vars,
-                         (domain.attributes[0], ) + domain.metas)
-        data2 = Table(domain2, data)
-
         widget.feature_type = widget.GENERIC
-        self.send_signal(widget.Inputs.data, data2)
+        self.send_signal(widget.Inputs.data, data)
 
-        # By default, the widget switches from GENERIC to the first string meta
-        self.assertEqual(widget.feature_type, widget.FROM_META_ATTR)
+        # By default, the widget switches from GENERIC to the first meta
+        self.assertEqual(widget.feature_type, widget.FROM_VAR)
         self.assertIs(widget.feature_names_column, metas[0])
         output = self.get_output(widget.Outputs.data)
         self.assertListEqual(
@@ -54,9 +47,9 @@ def test_feature_type(self):
         widget.feature_names_column = metas[1]
         widget.apply()
         output = self.get_output(widget.Outputs.data)
-        self.assertListEqual(
-            [a.name for a in output.domain.attributes],
-            [metas[1].to_val(m) for m in data.metas[:, 1]])
+        self.assertTrue(
+            all(a.name.startswith(metas[1].to_val(m))
+                for a, m in zip(output.domain.attributes, data.metas[:, 1])))
 
         # Switch to generic
         self.assertEqual(widget.DEFAULT_PREFIX, "Feature")
@@ -84,13 +77,13 @@ def test_feature_type(self):
 
     def test_send_report(self):
         widget = self.widget
-        widget.feature_type = widget.FROM_META_ATTR
+        widget.feature_type = widget.FROM_VAR
         widget.report_button.click()
         widget.feature_type = widget.GENERIC
         widget.report_button.click()
 
         self.send_signal(widget.Inputs.data, self.zoo)
-        widget.feature_type = widget.FROM_META_ATTR
+        widget.feature_type = widget.FROM_VAR
         widget.report_button.click()
         widget.feature_type = widget.GENERIC
         widget.report_button.click()
@@ -110,7 +103,7 @@ def test_gui_behaviour(self):
         # Data with metas: default type is meta, radio enabled
         self.send_signal(widget.Inputs.data, self.zoo)
         self.assertTrue(widget.meta_button.isEnabled())
-        self.assertEqual(widget.feature_type, widget.FROM_META_ATTR)
+        self.assertEqual(widget.feature_type, widget.FROM_VAR)
         self.assertIs(widget.feature_names_column, widget.feature_model[0])
         self.assertTrue(widget.apply.called)
 
@@ -123,7 +116,7 @@ def test_gui_behaviour(self):
         # Changing combo changes the radio button to meta
         widget.apply.reset_mock()
         widget.feature_combo.activated.emit(0)
-        self.assertEqual(widget.feature_type, widget.FROM_META_ATTR)
+        self.assertEqual(widget.feature_type, widget.FROM_VAR)
         self.assertTrue(widget.apply.called)
 
     def test_all_whitespace(self):
@@ -141,5 +134,19 @@ def test_error(self):
         self.send_signal(widget.Inputs.data, self.zoo)
         self.assertFalse(widget.Error.value_error.is_shown())
 
+    def test_feature_names_from_cont_vars(self):
+        table = Table("iris")
+        self.send_signal(self.widget.Inputs.data, table)
+        self.assertListEqual(self.widget.feature_model[:],
+                             list(table.domain.attributes))
+        self.widget.feature_combo.activated.emit(3)
+        output = self.get_output(self.widget.Outputs.data)
+        self.assertListEqual([f.name for f in output.domain.attributes[:10]],
+                             ["0.2 (1)", "0.2 (2)", "0.2 (3)", "0.2 (4)",
+                              "0.2 (5)", "0.4 (1)", "0.3 (1)", "0.2 (6)",
+                              "0.2 (7)", "0.1 (1)"])
+        self.assertTrue(self.widget.Warning.duplicate_names.is_shown())
+
+
 if __name__ == "__main__":
     unittest.main()