Merge pull request #3231 from ales-erjavec/oweditdomain-record-transformations

[ENH] Edit Domain: Record transformations

Author: lanzagar

Orange/widgets/data/oweditdomain.py

@@ -1,82 +1,57 @@
 # Test methods with long descriptive names can omit docstrings
-# pylint: disable=missing-docstring
-
+# pylint: disable=all
 from unittest import TestCase
 import numpy as np
+from numpy.testing import assert_array_equal
 
 from AnyQt.QtCore import QModelIndex, QItemSelectionModel, Qt
-
-from Orange.data import ContinuousVariable, DiscreteVariable, \
-    StringVariable, TimeVariable, Table, Domain
-from Orange.widgets.data.oweditdomain import EditDomainReport, OWEditDomain, \
-    ContinuousVariableEditor, DiscreteVariableEditor, VariableEditor, \
-    TimeVariableEditor
+from AnyQt.QtWidgets import QAction
+from AnyQt.QtTest import QTest
+
+from Orange.data import (
+    ContinuousVariable, DiscreteVariable, StringVariable, TimeVariable,
+    Table, Domain
+)
+from Orange.preprocess.transformation import Identity, Lookup
+from Orange.widgets.data.oweditdomain import (
+    OWEditDomain,
+    ContinuousVariableEditor, DiscreteVariableEditor, VariableEditor,
+    TimeVariableEditor, Categorical, Real, Time, String,
+    Rename, Annotate, CategoriesMapping, report_transform,
+    apply_transform
+)
 from Orange.widgets.data.owcolor import OWColor, ColorRole
 from Orange.widgets.tests.base import WidgetTest, GuiTest
 
-SECTION_NAME = "NAME"
 
-
-class TestEditDomainReport(TestCase):
-    # This tests test private methods
-    # pylint: disable=protected-access
-
-    def setUp(self):
-        self.report = EditDomainReport([], [])
-
-    def test_section_yields_nothing_for_no_changes(self):
-        result = self.report._section(SECTION_NAME, [])
-        self.assertEmpty(result)
-
-    def test_section_yields_header_for_changes(self):
-        result = self.report._section(SECTION_NAME, ["a"])
-        self.assertTrue(any(SECTION_NAME in item for item in result))
-
-    def test_value_changes_yields_nothing_for_no_change(self):
-        a = DiscreteVariable("a", values="abc")
-        self.assertEmpty(self.report._value_changes(a, a))
-
-    def test_value_changes_yields_nothing_for_continuous_variables(self):
-        v1, v2 = ContinuousVariable("a"), ContinuousVariable("b")
-        self.assertEmpty(self.report._value_changes(v1, v2))
-
-    def test_value_changes_yields_changed_values(self):
-        v1, v2 = DiscreteVariable("a", "ab"), DiscreteVariable("b", "ac")
-        self.assertNotEmpty(self.report._value_changes(v1, v2))
-
-    def test_label_changes_yields_nothing_for_no_change(self):
-        v1 = ContinuousVariable("a")
-        v1.attributes["a"] = "b"
-        self.assertEmpty(self.report._value_changes(v1, v1))
-
-    def test_label_changes_yields_added_labels(self):
-        v1 = ContinuousVariable("a")
-        v2 = v1.copy(None)
-        v2.attributes["a"] = "b"
-        self.assertNotEmpty(self.report._label_changes(v1, v2))
-
-    def test_label_changes_yields_removed_labels(self):
-        v1 = ContinuousVariable("a")
-        v1.attributes["a"] = "b"
-        v2 = v1.copy(None)
-        del v2.attributes["a"]
-        self.assertNotEmpty(self.report._label_changes(v1, v2))
-
-    def test_label_changes_yields_modified_labels(self):
-        v1 = ContinuousVariable("a")
-        v1.attributes["a"] = "b"
-        v2 = v1.copy(None)
-        v2.attributes["a"] = "c"
-        self.assertNotEmpty(self.report._label_changes(v1, v2))
-
-    def assertEmpty(self, iterable):
-        self.assertRaises(StopIteration, lambda: next(iter(iterable)))
-
-    def assertNotEmpty(self, iterable):
-        try:
-            next(iter(iterable))
-        except StopIteration:
-            self.fail("Iterator did not produce any lines")
+class TestReport(TestCase):
+    def test_rename(self):
+        var = Real("X", (-1, ""), ())
+        tr = Rename("Y")
+        val = report_transform(var, [tr])
+        self.assertIn("X", val)
+        self.assertIn("Y", val)
+
+    def test_annotate(self):
+        var = Real("X", (-1, ""), (("a", "1"), ("b", "z")))
+        tr = Annotate((("a", "2"), ("j", "z")))
+        r = report_transform(var, [tr])
+        self.assertIn("a", r)
+        self.assertIn("b", r)
+
+    def test_categories_mapping(self):
+        var = Categorical("C", ("a", "b", "c"), None, ())
+        tr = CategoriesMapping(
+            (("a", "aa"),
+             ("b", None),
+             ("c", "cc"),
+             (None, "ee")),
+        )
+        r = report_transform(var, [tr])
+        self.assertIn("a", r)
+        self.assertIn("aa", r)
+        self.assertIn("b", r)
+        self.assertIn("<s>", r)
 
 
 class TestOWEditDomain(WidgetTest):
@@ -137,7 +112,7 @@ def test_list_attributes_remain_lists(self):
         idx = editor.labels_model.index(0, 1)
         editor.labels_model.setData(idx, "[1, 2, 4]", Qt.EditRole)
 
-        self.widget.unconditional_commit()
+        self.widget.commit()
         t2 = self.get_output(self.widget.Outputs.data)
         self.assertEqual(t2.domain["a"].attributes["list"], [1, 2, 4])
 
@@ -156,15 +131,20 @@ def test_duplicate_names(self):
         self.widget.domain_view.setCurrentIndex(idx)
         editor = self.widget.editor_stack.findChild(ContinuousVariableEditor)
 
-        editor.name_edit.setText("iris")
-        editor.commit()
+        def enter_text(widget, text):
+            # type: (QLineEdit, str) -> None
+            widget.selectAll()
+            QTest.keyClick(widget, Qt.Key_Delete)
+            QTest.keyClicks(widget, text)
+            QTest.keyClick(widget, Qt.Key_Return)
+
+        enter_text(editor.name_edit, "iris")
         self.widget.commit()
         self.assertTrue(self.widget.Error.duplicate_var_name.is_shown())
         output = self.get_output(self.widget.Outputs.data)
         self.assertIsNone(output)
 
-        editor.name_edit.setText("sepal height")
-        editor.commit()
+        enter_text(editor.name_edit, "sepal height")
         self.widget.commit()
         self.assertFalse(self.widget.Error.duplicate_var_name.is_shown())
         output = self.get_output(self.widget.Outputs.data)
@@ -176,10 +156,12 @@ def test_time_variable_preservation(self):
         self.send_signal(self.widget.Inputs.data, table)
         output = self.get_output(self.widget.Outputs.data)
         self.assertEqual(str(table[0, 4]), str(output[0, 4]))
+        view = self.widget.variables_view
+        view.setCurrentIndex(view.model().index(4))
 
         editor = self.widget.editor_stack.findChild(TimeVariableEditor)
         editor.name_edit.setText("Date")
-        editor.commit()
+        editor.variable_changed.emit()
         self.widget.commit()
         output = self.get_output(self.widget.Outputs.data)
         self.assertEqual(str(table[0, 4]), str(output[0, 4]))
@@ -188,54 +170,68 @@ def test_time_variable_preservation(self):
 class TestEditors(GuiTest):
     def test_variable_editor(self):
         w = VariableEditor()
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
 
-        v = StringVariable(name="S")
-        v.attributes.update({"A": 1, "B": "b"},)
-        w.set_data(v)
+        v = String("S", (("A", "1"), ("B", "b")))
+        w.set_data(v, [])
 
         self.assertEqual(w.name_edit.text(), v.name)
-        self.assertEqual(w.labels_model.get_dict(), v.attributes)
-        self.assertTrue(w.is_same())
+        self.assertEqual(w.labels_model.get_dict(),
+                         {"A": "1", "B": "b"})
+        self.assertEqual(w.get_data(), (v, []))
 
         w.set_data(None)
         self.assertEqual(w.name_edit.text(), "")
         self.assertEqual(w.labels_model.get_dict(), {})
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
+
+        w.set_data(v, [Rename("T"), Annotate((("a", "1"), ("b", "2")))])
+        self.assertEqual(w.name_edit.text(), "T")
+        self.assertEqual(w.labels_model.rowCount(), 2)
+        add = w.findChild(QAction, "action-add-label")
+        add.trigger()
+        remove = w.findChild(QAction, "action-delete-label")
+        remove.trigger()
 
     def test_continuous_editor(self):
         w = ContinuousVariableEditor()
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
 
-        v = ContinuousVariable("X", number_of_decimals=5)
-        v.attributes.update({"A": 1, "B": "b"})
-        w.set_data(v)
+        v = Real("X", (-1, ""), (("A", "1"), ("B", "b")))
+        w.set_data(v, [])
 
         self.assertEqual(w.name_edit.text(), v.name)
-        self.assertEqual(w.labels_model.get_dict(), v.attributes)
-        self.assertTrue(w.is_same())
+        self.assertEqual(w.labels_model.get_dict(), dict(v.annotations))
 
         w.set_data(None)
         self.assertEqual(w.name_edit.text(), "")
         self.assertEqual(w.labels_model.get_dict(), {})
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
 
     def test_discrete_editor(self):
         w = DiscreteVariableEditor()
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
 
-        v = DiscreteVariable("C", values=["a", "b", "c"])
-        v.attributes.update({"A": 1, "B": "b"})
+        v = Categorical("C", ("a", "b", "c"), None,
+                        (("A", "1"), ("B", "b")))
         w.set_data(v)
 
         self.assertEqual(w.name_edit.text(), v.name)
-        self.assertEqual(w.labels_model.get_dict(), v.attributes)
-        self.assertTrue(w.is_same())
-
+        self.assertEqual(w.labels_model.get_dict(), dict(v.annotations))
+        self.assertEqual(w.get_data(), (v, []))
         w.set_data(None)
         self.assertEqual(w.name_edit.text(), "")
         self.assertEqual(w.labels_model.get_dict(), {})
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
+        mapping = [
+            ("c", "C"),
+            ("a", "A"),
+            ("b", None),
+            (None, "b")
+        ]
+        w.set_data(v, [CategoriesMapping(mapping)])
+        w.grab()  # run delegate paint method
+        self.assertEqual(w.get_data(), (v, [CategoriesMapping(mapping)]))
 
         # test selection/deselection in the view
         w.set_data(v)
@@ -249,21 +245,79 @@ def test_discrete_editor(self):
 
     def test_time_editor(self):
         w = TimeVariableEditor()
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
 
-        v = TimeVariable("T", have_date=1)
-        v.attributes.update({"A": 1, "B": "b"})
-        w.set_data(v)
+        v = Time("T", (("A", "1"), ("B", "b")))
+        w.set_data(v,)
 
         self.assertEqual(w.name_edit.text(), v.name)
-        self.assertEqual(w.labels_model.get_dict(), v.attributes)
-        self.assertTrue(w.is_same())
-
-        var = w.get_data()
-        self.assertTrue(var.have_date)
-        self.assertFalse(var.have_time)
+        self.assertEqual(w.labels_model.get_dict(), dict(v.annotations))
 
         w.set_data(None)
         self.assertEqual(w.name_edit.text(), "")
         self.assertEqual(w.labels_model.get_dict(), {})
-        self.assertIs(w.get_data(), None)
+        self.assertEqual(w.get_data(), (None, []))
+
+
+class TestTransforms(TestCase):
+    def _test_common(self, var):
+        tr = [Rename(var.name + "_copy"), Annotate((("A", "1"),))]
+        XX = apply_transform(var, tr)
+        self.assertEqual(XX.name, var.name + "_copy")
+        self.assertEqual(XX.attributes, {"A": 1})
+        self.assertIsInstance(XX.compute_value, Identity)
+        self.assertIs(XX.compute_value.variable, var)
+
+    def test_continous(self):
+        X = ContinuousVariable("X")
+        self._test_common(X)
+
+    def test_string(self):
+        X = StringVariable("S")
+        self._test_common(X)
+
+    def test_time(self):
+        X = TimeVariable("X")
+        self._test_common(X)
+
+    def test_discrete(self):
+        D = DiscreteVariable("D", values=("a", "b"))
+        self._test_common(D)
+
+    def test_discrete_rename(self):
+        D = DiscreteVariable("D", values=("a", "b"))
+        DD = apply_transform(D, [CategoriesMapping((("a", "A"), ("b", "B")))])
+        self.assertSequenceEqual(DD.values, ["A", "B"])
+        self.assertIs(DD.compute_value.variable, D)
+
+    def test_discrete_reorder(self):
+        D = DiscreteVariable("D", values=("2", "3", "1", "0"))
+        DD = apply_transform(D, [CategoriesMapping((("0", "0"), ("1", "1"),
+                                                    ("2", "2"), ("3", "3")))])
+        self.assertSequenceEqual(DD.values, ["0", "1", "2", "3"])
+        self._assertLookupEquals(
+            DD.compute_value, Lookup(D, np.array([2, 3, 1, 0]))
+        )
+
+    def test_discrete_add_drop(self):
+        D = DiscreteVariable("D", values=("2", "3", "1", "0"), base_value=1)
+        mapping = (
+            ("0", None),
+            ("1", "1"),
+            ("2", "2"),
+            ("3", None),
+            (None, "A"),
+        )
+        tr = [CategoriesMapping(mapping)]
+        DD = apply_transform(D, tr)
+        self.assertSequenceEqual(DD.values, ["1", "2", "A"])
+        self._assertLookupEquals(
+            DD.compute_value, Lookup(D, np.array([1, np.nan, 0, np.nan]))
+        )
+        self.assertEqual(DD.base_value, -1)
+
+    def _assertLookupEquals(self, first, second):
+        self.assertIsInstance(first, Lookup)
+        self.assertIsInstance(second, Lookup)
+        self.assertIs(first.variable, second.variable)
+        assert_array_equal(first.lookup_table, second.lookup_table)