1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
|
from itertools import chain
from AnyQt.QtWidgets import QComboBox, QCheckBox
from orangewidget import gui
from orangewidget.settings import Setting
from orangewidget.widget import Msg
from Orange.data import Variable, Table
from Orange.modelling.column import (
ColumnModel, ColumnLearner, valid_value_sets, valid_prob_range)
from Orange.widgets.widget import OWWidget, Input, Output
from Orange.widgets.utils.itemmodels import VariableListModel
from Orange.widgets.utils.widgetpreview import WidgetPreview
class OWFeatureAsPredictor(OWWidget):
name = "Feature as Predictor"
description = "Use a column as probabilities or predictions"
icon = "icons/FeatureAsPredictor.svg"
priority = 1000
keywords = "column predictor"
want_main_area = False
resizing_enabled = False
class Inputs:
data = Input("Data", Table)
class Outputs:
learner = Output("Learner", ColumnLearner)
model = Output("Model", ColumnModel)
class Error(OWWidget.Error):
no_class = Msg("Data has no target variable.")
no_variables = Msg("No useful variables")
column_hint: Variable = Setting(None, schema_only=True)
# Stores the last user setting.
# apply_transformation tells what will actually happens;
# checkbox may be disabled and set to reflect apply_transformation.
apply_transformation_setting = Setting(False)
auto_apply = Setting(True)
def __init__(self):
super().__init__()
self.data = None
self.column = None
self.apply_transformation = False
self.pars_to_report = (False, False)
box = gui.vBox(self.controlArea, True)
self.column_combo = combo = QComboBox()
combo.setModel(VariableListModel())
box.layout().addWidget(combo)
@combo.activated.connect
def on_column_changed(index):
self.column = combo.model()[index]
self.column_hint = self.column.name
self._update_controls()
self.commit.deferred()
self.cb_transformation = cb = QCheckBox("", self)
box.layout().addWidget(cb)
@cb.clicked.connect
def on_apply_transformation_changed(checked):
self.apply_transformation_setting \
= self.apply_transformation = checked
self.commit.deferred()
gui.auto_apply(self.controlArea, self, "auto_apply")
self._update_controls()
def _update_controls(self):
cb = self.cb_transformation
data = self.data
if data is None or self.column is None:
cb.setChecked(self.apply_transformation_setting)
cb.setDisabled(False)
return
if self.column.is_discrete:
self.apply_transformation = False
cb.setChecked(False)
cb.setDisabled(True)
elif (data.domain.class_var.is_discrete
and not valid_prob_range(data.get_column(self.column))):
self.apply_transformation = True
cb.setChecked(True)
cb.setDisabled(True)
else:
self.apply_transformation = self.apply_transformation_setting
cb.setChecked(self.apply_transformation_setting)
cb.setDisabled(False)
shape = "logistic" if data.domain.class_var.is_discrete else "linear"
cb.setText(f"Transform through {shape} function")
cb.setToolTip(f"Use {shape} regression to fit the model's coefficients")
@Inputs.data
def set_data(self, data):
self._set_data(data)
self._update_controls()
self.commit.now()
def _set_data(self, data):
column_model: VariableListModel = self.column_combo.model()
self.Error.clear()
column_model.clear()
self.column = None
self.data = None
if data is None:
return
class_var = data.domain.class_var
if class_var is None:
self.Error.no_class()
return
allow_continuous = (class_var.is_continuous
or len(class_var.values) == 2)
column_model[:] = (
var
for var in chain(data.domain.attributes, data.domain.metas)
if (var.is_continuous and allow_continuous
or (var.is_discrete and class_var.is_discrete
and valid_value_sets(class_var, var))
)
)
if not column_model:
self.Error.no_variables()
return
self.data = data
if self.column_hint \
and self.column_hint in self.data.domain \
and (var := self.data.domain[self.column_hint]) in column_model:
self.column = var
self.column_combo.setCurrentIndex(column_model.indexOf(self.column))
else:
self.column = column_model[0]
self.column_combo.setCurrentIndex(0)
self.column_hint = self.column.name
@gui.deferred
def commit(self):
self.pars_to_report = (False, False)
if self.column is None:
self.Outputs.learner.send(None)
self.Outputs.model.send(None)
return
learner = ColumnLearner(
self.data.domain.class_var, self.column, self.apply_transformation)
model = learner(self.data)
self.Outputs.learner.send(learner)
self.Outputs.model.send(model)
if self.apply_transformation:
self.pars_to_report = (model.intercept, model.coefficient)
def send_report(self):
if self.column is None:
return
self.report_items((
("Predict values from", self.column.name),
("Applied transformation",
self.apply_transformation and self.data is not None and
("logistic" if self.data.domain.class_var.is_discrete else "linear")),
("Intercept", self.pars_to_report[0]),
("Coefficient", self.pars_to_report[1])
))
if __name__ == "__main__": # pragma: no cover
WidgetPreview(OWFeatureAsPredictor).run(Table("heart_disease"))
|
