class OWScatterPlot(OWWidget):
"""Scatterplot visualization with explorative analysis and intelligent
data visualization enhancements."""
name = 'Scatter Plot'
description = "Interactive scatter plot visualization with " \
"intelligent data visualization enhancements."
icon = "icons/ScatterPlot.svg"
priority = 140
class Inputs:
data = Input("Data", Table, default=True)
data_subset = Input("Data Subset", Table)
features = Input("Features", AttributeList)
class Outputs:
selected_data = Output("Selected Data", Table, default=True)
annotated_data = Output(ANNOTATED_DATA_SIGNAL_NAME, Table)
features = Output("Features", Table, dynamic=False)
settingsHandler = DomainContextHandler()
auto_send_selection = Setting(True)
auto_sample = Setting(True)
toolbar_selection = Setting(0)
attr_x = ContextSetting(None)
attr_y = ContextSetting(None)
selection = Setting(None, schema_only=True)
graph = SettingProvider(OWScatterPlotGraph)
jitter_sizes = [0, 0.1, 0.5, 1, 2, 3, 4, 5, 7, 10]
graph_name = "graph.plot_widget.plotItem"
class Information(OWWidget.Information):
sampled_sql = Msg("Large SQL table; showing a sample.")
def __init__(self):
super().__init__()
box = gui.vBox(self.mainArea, True, margin=0)
self.graph = OWScatterPlotGraph(self, box, "ScatterPlot")
box.layout().addWidget(self.graph.plot_widget)
plot = self.graph.plot_widget
axispen = QPen(self.palette().color(QPalette.Text))
axis = plot.getAxis("bottom")
axis.setPen(axispen)
axis = plot.getAxis("left")
axis.setPen(axispen)
self.data = None # Orange.data.Table
self.subset_data = None # Orange.data.Table
self.data_metas_X = None # self.data, where primitive metas are moved to X
self.sql_data = None # Orange.data.sql.table.SqlTable
self.attribute_selection_list = None # list of Orange.data.Variable
self.__timer = QTimer(self, interval=1200)
self.__timer.timeout.connect(self.add_data)
common_options = dict(
labelWidth=50, orientation=Qt.Horizontal, sendSelectedValue=True,
valueType=str)
box = gui.vBox(self.controlArea, "Axis Data")
dmod = DomainModel
self.xy_model = DomainModel(dmod.MIXED, valid_types=dmod.PRIMITIVE)
self.cb_attr_x = gui.comboBox(
box, self, "attr_x", label="Axis x:", callback=self.update_attr,
model=self.xy_model, **common_options)
self.cb_attr_y = gui.comboBox(
box, self, "attr_y", label="Axis y:", callback=self.update_attr,
model=self.xy_model, **common_options)
vizrank_box = gui.hBox(box)
gui.separator(vizrank_box, width=common_options["labelWidth"])
self.vizrank, self.vizrank_button = ScatterPlotVizRank.add_vizrank(
vizrank_box, self, "Find Informative Projections", self.set_attr)
gui.separator(box)
gui.valueSlider(
box, self, value='graph.jitter_size', label='Jittering: ',
values=self.jitter_sizes, callback=self.reset_graph_data,
labelFormat=lambda x:
"None" if x == 0 else ("%.1f %%" if x < 1 else "%d %%") % x)
gui.checkBox(
gui.indentedBox(box), self, 'graph.jitter_continuous',
'Jitter numeric values', callback=self.reset_graph_data)
self.sampling = gui.auto_commit(
self.controlArea, self, "auto_sample", "Sample", box="Sampling",
callback=self.switch_sampling, commit=lambda: self.add_data(1))
self.sampling.setVisible(False)
g = self.graph.gui
g.point_properties_box(self.controlArea)
self.models = [self.xy_model] + g.points_models
box = gui.vBox(self.controlArea, "Plot Properties")
g.add_widgets([g.ShowLegend, g.ShowGridLines], box)
gui.checkBox(
box, self, value='graph.tooltip_shows_all',
label='Show all data on mouse hover')
self.cb_class_density = gui.checkBox(
box, self, value='graph.class_density', label='Show class density',
callback=self.update_density)
self.cb_reg_line = gui.checkBox(
box, self, value='graph.show_reg_line',
label='Show regression line', callback=self.update_regression_line)
gui.checkBox(
box, self, 'graph.label_only_selected',
'Label only selected points', callback=self.graph.update_labels)
self.zoom_select_toolbar = g.zoom_select_toolbar(
gui.vBox(self.controlArea, "Zoom/Select"), nomargin=True,
buttons=[g.StateButtonsBegin, g.SimpleSelect, g.Pan, g.Zoom,
g.StateButtonsEnd, g.ZoomReset]
)
buttons = self.zoom_select_toolbar.buttons
buttons[g.Zoom].clicked.connect(self.graph.zoom_button_clicked)
buttons[g.Pan].clicked.connect(self.graph.pan_button_clicked)
buttons[g.SimpleSelect].clicked.connect(self.graph.select_button_clicked)
buttons[g.ZoomReset].clicked.connect(self.graph.reset_button_clicked)
self.controlArea.layout().addStretch(100)
self.icons = gui.attributeIconDict
p = self.graph.plot_widget.palette()
self.graph.set_palette(p)
gui.auto_commit(self.controlArea, self, "auto_send_selection",
"Send Selection", "Send Automatically")
def zoom(s):
"""Zoom in/out by factor `s`."""
viewbox = plot.getViewBox()
# scaleBy scales the view's bounds (the axis range)
viewbox.scaleBy((1 / s, 1 / s))
def fit_to_view():
viewbox = plot.getViewBox()
viewbox.autoRange()
zoom_in = QAction(
"Zoom in", self, triggered=lambda: zoom(1.25)
)
zoom_in.setShortcuts([QKeySequence(QKeySequence.ZoomIn),
QKeySequence(self.tr("Ctrl+="))])
zoom_out = QAction(
"Zoom out", self, shortcut=QKeySequence.ZoomOut,
triggered=lambda: zoom(1 / 1.25)
)
zoom_fit = QAction(
"Fit in view", self,
shortcut=QKeySequence(Qt.ControlModifier | Qt.Key_0),
triggered=fit_to_view
)
self.addActions([zoom_in, zoom_out, zoom_fit])
def keyPressEvent(self, event):
super().keyPressEvent(event)
self.graph.update_tooltip(event.modifiers())
def keyReleaseEvent(self, event):
super().keyReleaseEvent(event)
self.graph.update_tooltip(event.modifiers())
# def settingsFromWidgetCallback(self, handler, context):
# context.selectionPolygons = []
# for curve in self.graph.selectionCurveList:
# xs = [curve.x(i) for i in range(curve.dataSize())]
# ys = [curve.y(i) for i in range(curve.dataSize())]
# context.selectionPolygons.append((xs, ys))
# def settingsToWidgetCallback(self, handler, context):
# selections = getattr(context, "selectionPolygons", [])
# for (xs, ys) in selections:
# c = SelectionCurve("")
# c.setData(xs,ys)
# c.attach(self.graph)
# self.graph.selectionCurveList.append(c)
def reset_graph_data(self, *_):
if self.data is not None:
self.graph.rescale_data()
self.update_graph()
@Inputs.data
def set_data(self, data):
self.clear_messages()
self.Information.sampled_sql.clear()
self.__timer.stop()
self.sampling.setVisible(False)
self.sql_data = None
if isinstance(data, SqlTable):
if data.approx_len() < 4000:
data = Table(data)
else:
self.Information.sampled_sql()
self.sql_data = data
data_sample = data.sample_time(0.8, no_cache=True)
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.sampling.setVisible(True)
if self.auto_sample:
self.__timer.start()
if data is not None and (len(data) == 0 or len(data.domain) == 0):
data = None
if self.data and data and self.data.checksum() == data.checksum():
return
self.closeContext()
same_domain = (self.data and data and
data.domain.checksum() == self.data.domain.checksum())
self.data = data
self.data_metas_X = self.move_primitive_metas_to_X(data)
if not same_domain:
self.init_attr_values()
self.vizrank.initialize()
self.vizrank.attrs = self.data.domain.attributes if self.data is not None else []
self.vizrank_button.setEnabled(
self.data is not None and not self.data.is_sparse() and
self.data.domain.class_var is not None and
len(self.data.domain.attributes) > 1 and len(self.data) > 1)
if self.data is not None and self.data.domain.class_var is None \
and len(self.data.domain.attributes) > 1 and len(self.data) > 1:
self.vizrank_button.setToolTip(
"Data with a class variable is required.")
else:
self.vizrank_button.setToolTip("")
self.openContext(self.data)
def findvar(name, iterable):
"""Find a Orange.data.Variable in `iterable` by name"""
for el in iterable:
if isinstance(el, Orange.data.Variable) and el.name == name:
return el
return None
# handle restored settings from < 3.3.9 when attr_* were stored
# by name
if isinstance(self.attr_x, str):
self.attr_x = findvar(self.attr_x, self.xy_model)
if isinstance(self.attr_y, str):
self.attr_y = findvar(self.attr_y, self.xy_model)
if isinstance(self.graph.attr_label, str):
self.graph.attr_label = findvar(
self.graph.attr_label, self.graph.gui.label_model)
if isinstance(self.graph.attr_color, str):
self.graph.attr_color = findvar(
self.graph.attr_color, self.graph.gui.color_model)
if isinstance(self.graph.attr_shape, str):
self.graph.attr_shape = findvar(
self.graph.attr_shape, self.graph.gui.shape_model)
if isinstance(self.graph.attr_size, str):
self.graph.attr_size = findvar(
self.graph.attr_size, self.graph.gui.size_model)
def add_data(self, time=0.4):
if self.data and len(self.data) > 2000:
return self.__timer.stop()
data_sample = self.sql_data.sample_time(time, no_cache=True)
if data_sample:
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.data = Table.concatenate((self.data, data), axis=0)
self.data_metas_X = self.move_primitive_metas_to_X(self.data)
self.handleNewSignals()
def switch_sampling(self):
self.__timer.stop()
if self.auto_sample and self.sql_data:
self.add_data()
self.__timer.start()
def move_primitive_metas_to_X(self, data):
if data is not None:
new_attrs = [a for a in data.domain.attributes + data.domain.metas
if a.is_primitive()]
new_metas = [m for m in data.domain.metas if not m.is_primitive()]
new_domain = Domain(new_attrs, data.domain.class_vars, new_metas)
data = data.transform(new_domain)
return data
@Inputs.data_subset
def set_subset_data(self, subset_data):
self.warning()
if isinstance(subset_data, SqlTable):
if subset_data.approx_len() < AUTO_DL_LIMIT:
subset_data = Table(subset_data)
else:
self.warning("Data subset does not support large Sql tables")
subset_data = None
self.subset_data = self.move_primitive_metas_to_X(subset_data)
self.controls.graph.alpha_value.setEnabled(subset_data is None)
# called when all signals are received, so the graph is updated only once
def handleNewSignals(self):
self.graph.new_data(self.sparse_to_dense(self.data_metas_X),
self.sparse_to_dense(self.subset_data))
if self.attribute_selection_list and self.graph.domain and \
all(attr in self.graph.domain
for attr in self.attribute_selection_list):
self.attr_x = self.attribute_selection_list[0]
self.attr_y = self.attribute_selection_list[1]
self.attribute_selection_list = None
self.update_graph()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
self.cb_reg_line.setEnabled(self.graph.can_draw_regresssion_line())
self.apply_selection()
self.unconditional_commit()
def prepare_data(self):
"""
Only when dealing with sparse matrices.
GH-2152
"""
self.graph.new_data(self.sparse_to_dense(self.data_metas_X),
self.sparse_to_dense(self.subset_data),
new=False)
def sparse_to_dense(self, input_data=None):
if input_data is None or not input_data.is_sparse():
return input_data
keys = []
attrs = {self.attr_x,
self.attr_y,
self.graph.attr_color,
self.graph.attr_shape,
self.graph.attr_size,
self.graph.attr_label}
for i, attr in enumerate(input_data.domain):
if attr in attrs:
keys.append(i)
new_domain = input_data.domain.select_columns(keys)
dmx = input_data.transform(new_domain)
dmx.X = dmx.X.toarray()
# TODO: remove once we make sure Y is always dense.
if sp.issparse(dmx.Y):
dmx.Y = dmx.Y.toarray()
return dmx
def apply_selection(self):
"""Apply selection saved in workflow."""
if self.data is not None and self.selection is not None:
self.graph.selection = np.zeros(len(self.data), dtype=np.uint8)
self.selection = [x for x in self.selection if x < len(self.data)]
self.graph.selection[self.selection] = 1
self.graph.update_colors(keep_colors=True)
@Inputs.features
def set_shown_attributes(self, attributes):
if attributes and len(attributes) >= 2:
self.attribute_selection_list = attributes[:2]
else:
self.attribute_selection_list = None
def get_shown_attributes(self):
return self.attr_x, self.attr_y
def init_attr_values(self):
domain = self.data and self.data.domain
for model in self.models:
model.set_domain(domain)
self.attr_x = self.xy_model[0] if self.xy_model else None
self.attr_y = self.xy_model[1] if len(self.xy_model) >= 2 \
else self.attr_x
self.graph.attr_color = domain and self.data.domain.class_var or None
self.graph.attr_shape = None
self.graph.attr_size = None
self.graph.attr_label = None
def set_attr(self, attr_x, attr_y):
self.attr_x, self.attr_y = attr_x, attr_y
self.update_attr()
def update_attr(self):
self.prepare_data()
self.update_graph()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
self.cb_reg_line.setEnabled(self.graph.can_draw_regresssion_line())
self.send_features()
def update_colors(self):
self.prepare_data()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
def update_density(self):
self.update_graph(reset_view=False)
def update_regression_line(self):
self.update_graph(reset_view=False)
def update_graph(self, reset_view=True, **_):
self.graph.zoomStack = []
if self.graph.data is None:
return
self.graph.update_data(self.attr_x, self.attr_y, reset_view)
def selection_changed(self):
self.send_data()
@staticmethod
def create_groups_table(data, selection):
if data is None:
return None
names = [var.name for var in data.domain.variables + data.domain.metas]
name = get_next_name(names, "Selection group")
metas = data.domain.metas + (
DiscreteVariable(
name,
["Unselected"] + ["G{}".format(i + 1)
for i in range(np.max(selection))]),
)
domain = Domain(data.domain.attributes, data.domain.class_vars, metas)
table = data.transform(domain)
table.metas[:, len(data.domain.metas):] = \
selection.reshape(len(data), 1)
return table
def send_data(self):
selected = None
selection = None
# TODO: Implement selection for sql data
graph = self.graph
if isinstance(self.data, SqlTable):
selected = self.data
elif self.data is not None:
selection = graph.get_selection()
if len(selection) > 0:
selected = self.data[selection]
if graph.selection is not None and np.max(graph.selection) > 1:
annotated = self.create_groups_table(self.data, graph.selection)
else:
annotated = create_annotated_table(self.data, selection)
self.Outputs.selected_data.send(selected)
self.Outputs.annotated_data.send(annotated)
# Store current selection in a setting that is stored in workflow
if self.selection is not None and len(selection):
self.selection = list(selection)
def send_features(self):
features = None
if self.attr_x or self.attr_y:
dom = Domain([], metas=(StringVariable(name="feature"),))
features = Table(dom, [[self.attr_x], [self.attr_y]])
features.name = "Features"
self.Outputs.features.send(features)
def commit(self):
self.send_data()
self.send_features()
def get_widget_name_extension(self):
if self.data is not None:
return "{} vs {}".format(self.attr_x.name, self.attr_y.name)
def send_report(self):
if self.data is None:
return
def name(var):
return var and var.name
caption = report.render_items_vert((
("Color", name(self.graph.attr_color)),
("Label", name(self.graph.attr_label)),
("Shape", name(self.graph.attr_shape)),
("Size", name(self.graph.attr_size)),
("Jittering", (self.attr_x.is_discrete or
self.attr_y.is_discrete or
self.graph.jitter_continuous) and
self.graph.jitter_size)))
self.report_plot()
if caption:
self.report_caption(caption)
def onDeleteWidget(self):
super().onDeleteWidget()
self.graph.plot_widget.getViewBox().deleteLater()
self.graph.plot_widget.clear()
class OWScatterPlot(OWWidget):
"""Scatterplot visualization with explorative analysis and intelligent
data visualization enhancements."""
name = 'Scatter Plot'
description = "Interactive scatter plot visualization with " \
"intelligent data visualization enhancements."
icon = "icons/ScatterPlot.svg"
priority = 140
class Inputs:
data = Input("Data", Table, default=True)
data_subset = Input("Data Subset", Table)
features = Input("Features", AttributeList)
class Outputs:
selected_data = Output("Selected Data", Table, default=True)
annotated_data = Output(ANNOTATED_DATA_SIGNAL_NAME, Table)
features = Output("Features", AttributeList, dynamic=False)
settings_version = 2
settingsHandler = DomainContextHandler()
auto_send_selection = Setting(True)
auto_sample = Setting(True)
toolbar_selection = Setting(0)
attr_x = ContextSetting(None)
attr_y = ContextSetting(None)
#: Serialized selection state to be restored
selection_group = Setting(None, schema_only=True)
graph = SettingProvider(OWScatterPlotGraph)
jitter_sizes = [0, 0.1, 0.5, 1, 2, 3, 4, 5, 7, 10]
graph_name = "graph.plot_widget.plotItem"
class Information(OWWidget.Information):
sampled_sql = Msg("Large SQL table; showing a sample.")
def __init__(self):
super().__init__()
box = gui.vBox(self.mainArea, True, margin=0)
self.graph = OWScatterPlotGraph(self, box, "ScatterPlot")
box.layout().addWidget(self.graph.plot_widget)
plot = self.graph.plot_widget
axispen = QPen(self.palette().color(QPalette.Text))
axis = plot.getAxis("bottom")
axis.setPen(axispen)
axis = plot.getAxis("left")
axis.setPen(axispen)
self.data = None # Orange.data.Table
self.subset_data = None # Orange.data.Table
self.sql_data = None # Orange.data.sql.table.SqlTable
self.attribute_selection_list = None # list of Orange.data.Variable
self.__timer = QTimer(self, interval=1200)
self.__timer.timeout.connect(self.add_data)
#: Remember the saved state to restore
self.__pending_selection_restore = self.selection_group
self.selection_group = None
common_options = dict(
labelWidth=50, orientation=Qt.Horizontal, sendSelectedValue=True,
valueType=str)
box = gui.vBox(self.controlArea, "Axis Data")
dmod = DomainModel
self.xy_model = DomainModel(dmod.MIXED, valid_types=dmod.PRIMITIVE)
self.cb_attr_x = gui.comboBox(
box, self, "attr_x", label="Axis x:", callback=self.update_attr,
model=self.xy_model, **common_options)
self.cb_attr_y = gui.comboBox(
box, self, "attr_y", label="Axis y:", callback=self.update_attr,
model=self.xy_model, **common_options)
vizrank_box = gui.hBox(box)
gui.separator(vizrank_box, width=common_options["labelWidth"])
self.vizrank, self.vizrank_button = ScatterPlotVizRank.add_vizrank(
vizrank_box, self, "Find Informative Projections", self.set_attr)
gui.separator(box)
g = self.graph.gui
g.add_widgets([g.JitterSizeSlider,
g.JitterNumericValues], box)
self.sampling = gui.auto_commit(
self.controlArea, self, "auto_sample", "Sample", box="Sampling",
callback=self.switch_sampling, commit=lambda: self.add_data(1))
self.sampling.setVisible(False)
g.point_properties_box(self.controlArea)
self.models = [self.xy_model] + g.points_models
box_plot_prop = gui.vBox(self.controlArea, "Plot Properties")
g.add_widgets([g.ShowLegend,
g.ShowGridLines,
g.ToolTipShowsAll,
g.ClassDensity,
g.RegressionLine,
g.LabelOnlySelected], box_plot_prop)
self.graph.box_zoom_select(self.controlArea)
self.controlArea.layout().addStretch(100)
self.icons = gui.attributeIconDict
p = self.graph.plot_widget.palette()
self.graph.set_palette(p)
gui.auto_commit(self.controlArea, self, "auto_send_selection",
"Send Selection", "Send Automatically")
self.graph.zoom_actions(self)
def keyPressEvent(self, event):
super().keyPressEvent(event)
self.graph.update_tooltip(event.modifiers())
def keyReleaseEvent(self, event):
super().keyReleaseEvent(event)
self.graph.update_tooltip(event.modifiers())
def reset_graph_data(self, *_):
if self.data is not None:
self.graph.rescale_data()
self.update_graph()
def _vizrank_color_change(self):
self.vizrank.initialize()
is_enabled = self.data is not None and not self.data.is_sparse() and \
len([v for v in chain(self.data.domain.variables, self.data.domain.metas)
if v.is_primitive]) > 2\
and len(self.data) > 1
self.vizrank_button.setEnabled(
is_enabled and self.graph.attr_color is not None and
not np.isnan(self.data.get_column_view(self.graph.attr_color)[0].astype(float)).all())
if is_enabled and self.graph.attr_color is None:
self.vizrank_button.setToolTip("Color variable has to be selected.")
else:
self.vizrank_button.setToolTip("")
@Inputs.data
def set_data(self, data):
self.clear_messages()
self.Information.sampled_sql.clear()
self.__timer.stop()
self.sampling.setVisible(False)
self.sql_data = None
if isinstance(data, SqlTable):
if data.approx_len() < 4000:
data = Table(data)
else:
self.Information.sampled_sql()
self.sql_data = data
data_sample = data.sample_time(0.8, no_cache=True)
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.sampling.setVisible(True)
if self.auto_sample:
self.__timer.start()
if data is not None and (len(data) == 0 or len(data.domain) == 0):
data = None
if self.data and data and self.data.checksum() == data.checksum():
return
self.closeContext()
same_domain = (self.data and data and
data.domain.checksum() == self.data.domain.checksum())
self.data = data
if not same_domain:
self.init_attr_values()
self.openContext(self.data)
self._vizrank_color_change()
def findvar(name, iterable):
"""Find a Orange.data.Variable in `iterable` by name"""
for el in iterable:
if isinstance(el, Orange.data.Variable) and el.name == name:
return el
return None
# handle restored settings from < 3.3.9 when attr_* were stored
# by name
if isinstance(self.attr_x, str):
self.attr_x = findvar(self.attr_x, self.xy_model)
if isinstance(self.attr_y, str):
self.attr_y = findvar(self.attr_y, self.xy_model)
if isinstance(self.graph.attr_label, str):
self.graph.attr_label = findvar(
self.graph.attr_label, self.graph.gui.label_model)
if isinstance(self.graph.attr_color, str):
self.graph.attr_color = findvar(
self.graph.attr_color, self.graph.gui.color_model)
if isinstance(self.graph.attr_shape, str):
self.graph.attr_shape = findvar(
self.graph.attr_shape, self.graph.gui.shape_model)
if isinstance(self.graph.attr_size, str):
self.graph.attr_size = findvar(
self.graph.attr_size, self.graph.gui.size_model)
def add_data(self, time=0.4):
if self.data and len(self.data) > 2000:
return self.__timer.stop()
data_sample = self.sql_data.sample_time(time, no_cache=True)
if data_sample:
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.data = Table.concatenate((self.data, data), axis=0)
self.handleNewSignals()
def switch_sampling(self):
self.__timer.stop()
if self.auto_sample and self.sql_data:
self.add_data()
self.__timer.start()
@Inputs.data_subset
def set_subset_data(self, subset_data):
self.warning()
if isinstance(subset_data, SqlTable):
if subset_data.approx_len() < AUTO_DL_LIMIT:
subset_data = Table(subset_data)
else:
self.warning("Data subset does not support large Sql tables")
subset_data = None
self.subset_data = subset_data
self.controls.graph.alpha_value.setEnabled(subset_data is None)
# called when all signals are received, so the graph is updated only once
def handleNewSignals(self):
self.graph.new_data(self.data, self.subset_data)
if self.attribute_selection_list and self.graph.domain and \
all(attr in self.graph.domain
for attr in self.attribute_selection_list):
self.attr_x = self.attribute_selection_list[0]
self.attr_y = self.attribute_selection_list[1]
self.attribute_selection_list = None
self.update_graph()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
self.cb_reg_line.setEnabled(self.graph.can_draw_regresssion_line())
if self.data is not None and self.__pending_selection_restore is not None:
self.apply_selection(self.__pending_selection_restore)
self.__pending_selection_restore = None
self.unconditional_commit()
def apply_selection(self, selection):
"""Apply `selection` to the current plot."""
if self.data is not None:
self.graph.selection = np.zeros(len(self.data), dtype=np.uint8)
self.selection_group = [x for x in selection if x[0] < len(self.data)]
selection_array = np.array(self.selection_group).T
self.graph.selection[selection_array[0]] = selection_array[1]
self.graph.update_colors(keep_colors=True)
@Inputs.features
def set_shown_attributes(self, attributes):
if attributes and len(attributes) >= 2:
self.attribute_selection_list = attributes[:2]
else:
self.attribute_selection_list = None
def init_attr_values(self):
domain = self.data and self.data.domain
for model in self.models:
model.set_domain(domain)
self.attr_x = self.xy_model[0] if self.xy_model else None
self.attr_y = self.xy_model[1] if len(self.xy_model) >= 2 \
else self.attr_x
self.graph.attr_color = self.data.domain.class_var if domain else None
self.graph.attr_shape = None
self.graph.attr_size = None
self.graph.attr_label = None
def set_attr(self, attr_x, attr_y):
self.attr_x, self.attr_y = attr_x, attr_y
self.update_attr()
def update_attr(self):
self.update_graph()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
self.cb_reg_line.setEnabled(self.graph.can_draw_regresssion_line())
self.send_features()
def update_colors(self):
self._vizrank_color_change()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
def update_density(self):
self.update_graph(reset_view=False)
def update_regression_line(self):
self.update_graph(reset_view=False)
def update_graph(self, reset_view=True, **_):
self.graph.zoomStack = []
if self.graph.data is None:
return
self.graph.update_data(self.attr_x, self.attr_y, reset_view)
def selection_changed(self):
# Store current selection in a setting that is stored in workflow
if isinstance(self.data, SqlTable):
selection = None
elif self.data is not None:
selection = self.graph.get_selection()
else:
selection = None
if selection is not None and len(selection):
self.selection_group = list(zip(selection, self.graph.selection[selection]))
else:
self.selection_group = None
self.commit()
def send_data(self):
# TODO: Implement selection for sql data
def _get_selected():
if not len(selection):
return None
return create_groups_table(data, graph.selection, False, "Group")
def _get_annotated():
if graph.selection is not None and np.max(graph.selection) > 1:
return create_groups_table(data, graph.selection)
else:
return create_annotated_table(data, selection)
graph = self.graph
data = self.data
selection = graph.get_selection()
self.Outputs.annotated_data.send(_get_annotated())
self.Outputs.selected_data.send(_get_selected())
def send_features(self):
features = [attr for attr in [self.attr_x, self.attr_y] if attr]
self.Outputs.features.send(features or None)
def commit(self):
self.send_data()
self.send_features()
def get_widget_name_extension(self):
if self.data is not None:
return "{} vs {}".format(self.attr_x.name, self.attr_y.name)
def send_report(self):
if self.data is None:
return
def name(var):
return var and var.name
caption = report.render_items_vert((
("Color", name(self.graph.attr_color)),
("Label", name(self.graph.attr_label)),
("Shape", name(self.graph.attr_shape)),
("Size", name(self.graph.attr_size)),
("Jittering", (self.attr_x.is_discrete or
self.attr_y.is_discrete or
self.graph.jitter_continuous) and
self.graph.jitter_size)))
self.report_plot()
if caption:
self.report_caption(caption)
def onDeleteWidget(self):
super().onDeleteWidget()
self.graph.plot_widget.getViewBox().deleteLater()
self.graph.plot_widget.clear()
@classmethod
def migrate_settings(cls, settings, version):
if version < 2 and "selection" in settings and settings["selection"]:
settings["selection_group"] = [(a, 1) for a in settings["selection"]]
class OWScatterPlot(OWWidget):
"""Scatterplot visualization with explorative analysis and intelligent
data visualization enhancements."""
name = 'Scatter Plot'
description = "Interactive scatter plot visualization with " \
"intelligent data visualization enhancements."
icon = "icons/ScatterPlot.svg"
priority = 140
class Inputs:
data = Input("Data", Table, default=True)
data_subset = Input("Data Subset", Table)
features = Input("Features", AttributeList)
class Outputs:
selected_data = Output("Selected Data", Table, default=True)
annotated_data = Output(ANNOTATED_DATA_SIGNAL_NAME, Table)
features = Output("Features", AttributeList, dynamic=False)
settings_version = 2
settingsHandler = DomainContextHandler()
auto_send_selection = Setting(True)
auto_sample = Setting(True)
toolbar_selection = Setting(0)
attr_x = ContextSetting(None)
attr_y = ContextSetting(None)
selection_group = Setting(None, schema_only=True)
graph = SettingProvider(OWScatterPlotGraph)
jitter_sizes = [0, 0.1, 0.5, 1, 2, 3, 4, 5, 7, 10]
graph_name = "graph.plot_widget.plotItem"
class Information(OWWidget.Information):
sampled_sql = Msg("Large SQL table; showing a sample.")
def __init__(self):
super().__init__()
box = gui.vBox(self.mainArea, True, margin=0)
self.graph = OWScatterPlotGraph(self, box, "ScatterPlot")
box.layout().addWidget(self.graph.plot_widget)
plot = self.graph.plot_widget
axispen = QPen(self.palette().color(QPalette.Text))
axis = plot.getAxis("bottom")
axis.setPen(axispen)
axis = plot.getAxis("left")
axis.setPen(axispen)
self.data = None # Orange.data.Table
self.subset_data = None # Orange.data.Table
self.data_metas_X = None # self.data, where primitive metas are moved to X
self.sql_data = None # Orange.data.sql.table.SqlTable
self.attribute_selection_list = None # list of Orange.data.Variable
self.__timer = QTimer(self, interval=1200)
self.__timer.timeout.connect(self.add_data)
common_options = dict(labelWidth=50,
orientation=Qt.Horizontal,
sendSelectedValue=True,
valueType=str)
box = gui.vBox(self.controlArea, "Axis Data")
dmod = DomainModel
self.xy_model = DomainModel(dmod.MIXED, valid_types=dmod.PRIMITIVE)
self.cb_attr_x = gui.comboBox(box,
self,
"attr_x",
label="Axis x:",
callback=self.update_attr,
model=self.xy_model,
**common_options)
self.cb_attr_y = gui.comboBox(box,
self,
"attr_y",
label="Axis y:",
callback=self.update_attr,
model=self.xy_model,
**common_options)
vizrank_box = gui.hBox(box)
gui.separator(vizrank_box, width=common_options["labelWidth"])
self.vizrank, self.vizrank_button = ScatterPlotVizRank.add_vizrank(
vizrank_box, self, "Find Informative Projections", self.set_attr)
gui.separator(box)
g = self.graph.gui
g.add_widgets([g.JitterSizeSlider, g.JitterNumericValues], box)
self.sampling = gui.auto_commit(self.controlArea,
self,
"auto_sample",
"Sample",
box="Sampling",
callback=self.switch_sampling,
commit=lambda: self.add_data(1))
self.sampling.setVisible(False)
g.point_properties_box(self.controlArea)
self.models = [self.xy_model] + g.points_models
box_plot_prop = gui.vBox(self.controlArea, "Plot Properties")
g.add_widgets([
g.ShowLegend, g.ShowGridLines, g.ToolTipShowsAll, g.ClassDensity,
g.RegressionLine, g.LabelOnlySelected
], box_plot_prop)
self.graph.box_zoom_select(self.controlArea)
self.controlArea.layout().addStretch(100)
self.icons = gui.attributeIconDict
p = self.graph.plot_widget.palette()
self.graph.set_palette(p)
gui.auto_commit(self.controlArea, self, "auto_send_selection",
"Send Selection", "Send Automatically")
self.graph.zoom_actions(self)
def keyPressEvent(self, event):
super().keyPressEvent(event)
self.graph.update_tooltip(event.modifiers())
def keyReleaseEvent(self, event):
super().keyReleaseEvent(event)
self.graph.update_tooltip(event.modifiers())
def reset_graph_data(self, *_):
if self.data is not None:
self.graph.rescale_data()
self.update_graph()
@Inputs.data
def set_data(self, data):
self.clear_messages()
self.Information.sampled_sql.clear()
self.__timer.stop()
self.sampling.setVisible(False)
self.sql_data = None
if isinstance(data, SqlTable):
if data.approx_len() < 4000:
data = Table(data)
else:
self.Information.sampled_sql()
self.sql_data = data
data_sample = data.sample_time(0.8, no_cache=True)
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.sampling.setVisible(True)
if self.auto_sample:
self.__timer.start()
if data is not None and (len(data) == 0 or len(data.domain) == 0):
data = None
if self.data and data and self.data.checksum() == data.checksum():
return
self.closeContext()
same_domain = (self.data and data and data.domain.checksum()
== self.data.domain.checksum())
self.data = data
self.data_metas_X = self.move_primitive_metas_to_X(data)
if not same_domain:
self.init_attr_values()
self.vizrank.initialize()
self.vizrank.attrs = self.data.domain.attributes if self.data is not None else []
self.vizrank_button.setEnabled(
self.data is not None and not self.data.is_sparse()
and self.data.domain.class_var is not None
and len(self.data.domain.attributes) > 1 and len(self.data) > 1)
if self.data is not None and self.data.domain.class_var is None \
and len(self.data.domain.attributes) > 1 and len(self.data) > 1:
self.vizrank_button.setToolTip(
"Data with a class variable is required.")
else:
self.vizrank_button.setToolTip("")
self.openContext(self.data)
def findvar(name, iterable):
"""Find a Orange.data.Variable in `iterable` by name"""
for el in iterable:
if isinstance(el, Orange.data.Variable) and el.name == name:
return el
return None
# handle restored settings from < 3.3.9 when attr_* were stored
# by name
if isinstance(self.attr_x, str):
self.attr_x = findvar(self.attr_x, self.xy_model)
if isinstance(self.attr_y, str):
self.attr_y = findvar(self.attr_y, self.xy_model)
if isinstance(self.graph.attr_label, str):
self.graph.attr_label = findvar(self.graph.attr_label,
self.graph.gui.label_model)
if isinstance(self.graph.attr_color, str):
self.graph.attr_color = findvar(self.graph.attr_color,
self.graph.gui.color_model)
if isinstance(self.graph.attr_shape, str):
self.graph.attr_shape = findvar(self.graph.attr_shape,
self.graph.gui.shape_model)
if isinstance(self.graph.attr_size, str):
self.graph.attr_size = findvar(self.graph.attr_size,
self.graph.gui.size_model)
def add_data(self, time=0.4):
if self.data and len(self.data) > 2000:
return self.__timer.stop()
data_sample = self.sql_data.sample_time(time, no_cache=True)
if data_sample:
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.data = Table.concatenate((self.data, data), axis=0)
self.data_metas_X = self.move_primitive_metas_to_X(self.data)
self.handleNewSignals()
def switch_sampling(self):
self.__timer.stop()
if self.auto_sample and self.sql_data:
self.add_data()
self.__timer.start()
def move_primitive_metas_to_X(self, data):
if data is not None:
new_attrs = [
a for a in data.domain.attributes + data.domain.metas
if a.is_primitive()
]
new_metas = [m for m in data.domain.metas if not m.is_primitive()]
new_domain = Domain(new_attrs, data.domain.class_vars, new_metas)
data = data.transform(new_domain)
return data
@Inputs.data_subset
def set_subset_data(self, subset_data):
self.warning()
if isinstance(subset_data, SqlTable):
if subset_data.approx_len() < AUTO_DL_LIMIT:
subset_data = Table(subset_data)
else:
self.warning("Data subset does not support large Sql tables")
subset_data = None
self.subset_data = self.move_primitive_metas_to_X(subset_data)
self.controls.graph.alpha_value.setEnabled(subset_data is None)
# called when all signals are received, so the graph is updated only once
def handleNewSignals(self):
self.graph.new_data(self.data_metas_X, self.subset_data)
if self.attribute_selection_list and self.graph.domain and \
all(attr in self.graph.domain
for attr in self.attribute_selection_list):
self.attr_x = self.attribute_selection_list[0]
self.attr_y = self.attribute_selection_list[1]
self.attribute_selection_list = None
self.update_graph()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
self.cb_reg_line.setEnabled(self.graph.can_draw_regresssion_line())
self.apply_selection()
self.unconditional_commit()
def apply_selection(self):
"""Apply selection saved in workflow."""
if self.data is not None and self.selection_group is not None:
self.graph.selection = np.zeros(len(self.data), dtype=np.uint8)
self.selection_group = [
x for x in self.selection_group if x[0] < len(self.data)
]
selection_array = np.array(self.selection_group).T
self.graph.selection[selection_array[0]] = selection_array[1]
self.graph.update_colors(keep_colors=True)
@Inputs.features
def set_shown_attributes(self, attributes):
if attributes and len(attributes) >= 2:
self.attribute_selection_list = attributes[:2]
else:
self.attribute_selection_list = None
def init_attr_values(self):
domain = self.data and self.data.domain
for model in self.models:
model.set_domain(domain)
self.attr_x = self.xy_model[0] if self.xy_model else None
self.attr_y = self.xy_model[1] if len(self.xy_model) >= 2 \
else self.attr_x
self.graph.attr_color = self.data.domain.class_var if domain else None
self.graph.attr_shape = None
self.graph.attr_size = None
self.graph.attr_label = None
def set_attr(self, attr_x, attr_y):
self.attr_x, self.attr_y = attr_x, attr_y
self.update_attr()
def update_attr(self):
self.update_graph()
self.cb_class_density.setEnabled(self.graph.can_draw_density())
self.cb_reg_line.setEnabled(self.graph.can_draw_regresssion_line())
self.send_features()
def update_colors(self):
self.cb_class_density.setEnabled(self.graph.can_draw_density())
def update_density(self):
self.update_graph(reset_view=False)
def update_regression_line(self):
self.update_graph(reset_view=False)
def update_graph(self, reset_view=True, **_):
self.graph.zoomStack = []
if self.graph.data is None:
return
self.graph.update_data(self.attr_x, self.attr_y, reset_view)
def selection_changed(self):
self.commit()
def send_data(self):
# TODO: Implement selection for sql data
def _get_selected():
if not len(selection):
return None
return create_groups_table(data, graph.selection, False, "Group")
def _get_annotated():
if graph.selection is not None and np.max(graph.selection) > 1:
return create_groups_table(data, graph.selection)
else:
return create_annotated_table(data, selection)
graph = self.graph
data = self.data
selection = graph.get_selection()
self.Outputs.annotated_data.send(_get_annotated())
self.Outputs.selected_data.send(_get_selected())
# Store current selection in a setting that is stored in workflow
if len(selection):
self.selection_group = list(
zip(selection, graph.selection[selection]))
else:
self.selection_group = None
def send_features(self):
features = [attr for attr in [self.attr_x, self.attr_y] if attr]
self.Outputs.features.send(features or None)
def commit(self):
self.send_data()
self.send_features()
def get_widget_name_extension(self):
if self.data is not None:
return "{} vs {}".format(self.attr_x.name, self.attr_y.name)
def send_report(self):
if self.data is None:
return
def name(var):
return var and var.name
caption = report.render_items_vert(
(("Color", name(self.graph.attr_color)),
("Label", name(self.graph.attr_label)),
("Shape", name(self.graph.attr_shape)),
("Size", name(self.graph.attr_size)),
("Jittering", (self.attr_x.is_discrete or self.attr_y.is_discrete
or self.graph.jitter_continuous)
and self.graph.jitter_size)))
self.report_plot()
if caption:
self.report_caption(caption)
def onDeleteWidget(self):
super().onDeleteWidget()
self.graph.plot_widget.getViewBox().deleteLater()
self.graph.plot_widget.clear()
@classmethod
def migrate_settings(cls, settings, version):
if version < 2 and "selection" in settings and settings["selection"]:
settings["selection_group"] = [(a, 1)
for a in settings["selection"]]