def get_control_by_type(param):
if isinstance(param, tuple):
options = param[1:]
param = param[0]
else:
options = ()
if param == 'str':
result = QLineEdit()
if 'pwd' in options:
result.setEchoMode(QLineEdit.Password)
if 'optional' in options:
result.setPlaceholderText('(Optional)')
elif param == 'big_str':
result = QPlainTextEdit()
elif param == 'label':
result = QLabel()
if 'url' in options:
result.setTextFormat(Qt.RichText)
result.setTextInteractionFlags(Qt.TextBrowserInteraction)
result.setOpenExternalLinks(True)
elif param == 'checkbox':
result = QCheckBox()
else:
raise RuntimeError()
return result
class DocumentEditor(QGroupBox):
text_changed = Signal(str, str)
remove_clicked = Signal()
def __init__(self, title, text, parent=None):
super().__init__(parent)
self.setLayout(QVBoxLayout())
self.title_le = QLineEdit()
self.title_le.setPlaceholderText("Document title")
self.title_le.setText(title)
self.title_le.editingFinished.connect(self._on_text_changed)
self.text_area = CustomQPlainTextEdit()
self.text_area.setPlaceholderText("Document text")
self.text_area.setPlainText(text)
self.text_area.editingFinished.connect(self._on_text_changed)
remove_button = QPushButton("x")
remove_button.setFixedWidth(35)
remove_button.setFocusPolicy(Qt.NoFocus)
remove_button.clicked.connect(self._on_remove_clicked)
box = gui.hBox(self)
box.layout().addWidget(self.title_le)
box.layout().addWidget(remove_button)
self.layout().addWidget(self.text_area)
def _on_text_changed(self):
self.text_changed.emit(self.title_le.text(),
self.text_area.toPlainText())
def _on_remove_clicked(self):
self.remove_clicked.emit()
class OWPCA(widget.OWWidget):
name = "PCA"
description = "Principal component analysis with a scree-diagram."
icon = "icons/PCA.svg"
priority = 3050
keywords = ["principal component analysis", "linear transformation"]
class Inputs:
data = Input("Data", Table)
class Outputs:
transformed_data = Output("Transformed data", Table)
components = Output("Components", Table)
pca = Output("PCA", PCA, dynamic=False)
preprocessor = Output("Preprocessor", Preprocess)
settingsHandler = settings.DomainContextHandler()
ncomponents = settings.Setting(2)
variance_covered = settings.Setting(100)
batch_size = settings.Setting(100)
address = settings.Setting('')
auto_update = settings.Setting(True)
auto_commit = settings.Setting(True)
normalize = settings.ContextSetting(True)
decomposition_idx = settings.ContextSetting(0)
maxp = settings.Setting(20)
axis_labels = settings.Setting(10)
graph_name = "plot.plotItem"
class Warning(widget.OWWidget.Warning):
trivial_components = widget.Msg(
"All components of the PCA are trivial (explain 0 variance). "
"Input data is constant (or near constant).")
class Error(widget.OWWidget.Error):
no_features = widget.Msg("At least 1 feature is required")
no_instances = widget.Msg("At least 1 data instance is required")
sparse_data = widget.Msg("Sparse data is not supported")
def __init__(self):
super().__init__()
self.data = None
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = False
self._init_projector()
# Components Selection
box = gui.vBox(self.controlArea, "Components Selection")
form = QFormLayout()
box.layout().addLayout(form)
self.components_spin = gui.spin(
box, self, "ncomponents", 1, MAX_COMPONENTS,
callback=self._update_selection_component_spin,
keyboardTracking=False
)
self.components_spin.setSpecialValueText("All")
self.variance_spin = gui.spin(
box, self, "variance_covered", 1, 100,
callback=self._update_selection_variance_spin,
keyboardTracking=False
)
self.variance_spin.setSuffix("%")
form.addRow("Components:", self.components_spin)
form.addRow("Variance covered:", self.variance_spin)
# Incremental learning
self.sampling_box = gui.vBox(self.controlArea, "Incremental learning")
self.addresstext = QLineEdit(box)
self.addresstext.setPlaceholderText('Remote server')
if self.address:
self.addresstext.setText(self.address)
self.sampling_box.layout().addWidget(self.addresstext)
form = QFormLayout()
self.sampling_box.layout().addLayout(form)
self.batch_spin = gui.spin(
self.sampling_box, self, "batch_size", 50, 100000, step=50,
keyboardTracking=False)
form.addRow("Batch size ~ ", self.batch_spin)
self.start_button = gui.button(
self.sampling_box, self, "Start remote computation",
callback=self.start, autoDefault=False,
tooltip="Start/abort computation on the server")
self.start_button.setEnabled(False)
gui.checkBox(self.sampling_box, self, "auto_update",
"Periodically fetch model", callback=self.update_model)
self.__timer = QTimer(self, interval=2000)
self.__timer.timeout.connect(self.get_model)
self.sampling_box.setVisible(remotely)
# Decomposition
self.decomposition_box = gui.radioButtons(
self.controlArea, self,
"decomposition_idx", [d.name for d in DECOMPOSITIONS],
box="Decomposition", callback=self._update_decomposition
)
# Options
self.options_box = gui.vBox(self.controlArea, "Options")
self.normalize_box = gui.checkBox(
self.options_box, self, "normalize",
"Normalize data", callback=self._update_normalize
)
self.maxp_spin = gui.spin(
self.options_box, self, "maxp", 1, MAX_COMPONENTS,
label="Show only first", callback=self._setup_plot,
keyboardTracking=False
)
self.controlArea.layout().addStretch()
gui.auto_commit(self.controlArea, self, "auto_commit", "Apply",
checkbox_label="Apply automatically")
self.plot = pg.PlotWidget(background="w")
axis = self.plot.getAxis("bottom")
axis.setLabel("Principal Components")
axis = self.plot.getAxis("left")
axis.setLabel("Proportion of variance")
self.plot_horlabels = []
self.plot_horlines = []
self.plot.getViewBox().setMenuEnabled(False)
self.plot.getViewBox().setMouseEnabled(False, False)
self.plot.showGrid(True, True, alpha=0.5)
self.plot.setRange(xRange=(0.0, 1.0), yRange=(0.0, 1.0))
self.mainArea.layout().addWidget(self.plot)
self._update_normalize()
def update_model(self):
self.get_model()
if self.auto_update and self.rpca and not self.rpca.ready():
self.__timer.start(2000)
else:
self.__timer.stop()
def update_buttons(self, sparse_data=False):
if sparse_data:
self.normalize = False
buttons = self.decomposition_box.buttons
for cls, button in zip(DECOMPOSITIONS, buttons):
button.setDisabled(sparse_data and not cls.supports_sparse)
if not buttons[self.decomposition_idx].isEnabled():
# Set decomposition index to first sparse-enabled decomposition
for i, cls in enumerate(DECOMPOSITIONS):
if cls.supports_sparse:
self.decomposition_idx = i
break
self._init_projector()
def start(self):
if 'Abort' in self.start_button.text():
self.rpca.abort()
self.__timer.stop()
self.start_button.setText("Start remote computation")
else:
self.address = self.addresstext.text()
with remote.server(self.address):
from Orange.projection.pca import RemotePCA
maxiter = (1e5 + self.data.approx_len()) / self.batch_size * 3
self.rpca = RemotePCA(self.data, self.batch_size, int(maxiter))
self.update_model()
self.start_button.setText("Abort remote computation")
@Inputs.data
def set_data(self, data):
self.closeContext()
self.clear_messages()
self.clear()
self.start_button.setEnabled(False)
self.information()
self.data = None
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
elif not remotely:
self.information("Data has been sampled")
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
else: # data was big and remote available
self.sampling_box.setVisible(True)
self.start_button.setText("Start remote computation")
self.start_button.setEnabled(True)
if not isinstance(data, SqlTable):
self.sampling_box.setVisible(False)
if isinstance(data, Table):
if len(data.domain.attributes) == 0:
self.Error.no_features()
self.clear_outputs()
return
if len(data) == 0:
self.Error.no_instances()
self.clear_outputs()
return
self.openContext(data)
sparse_data = data is not None and data.is_sparse()
self.normalize_box.setDisabled(sparse_data)
self.update_buttons(sparse_data=sparse_data)
self.data = data
self.fit()
def fit(self):
self.clear()
self.Warning.trivial_components.clear()
if self.data is None:
return
data = self.data
self._pca_projector.preprocessors = \
self._pca_preprocessors + ([Normalize()] if self.normalize else [])
if not isinstance(data, SqlTable):
pca = self._pca_projector(data)
variance_ratio = pca.explained_variance_ratio_
cumulative = numpy.cumsum(variance_ratio)
if numpy.isfinite(cumulative[-1]):
self.components_spin.setRange(0, len(cumulative))
self._pca = pca
self._variance_ratio = variance_ratio
self._cumulative = cumulative
self._setup_plot()
else:
self.Warning.trivial_components()
self.unconditional_commit()
def clear(self):
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = None
self.plot_horlabels = []
self.plot_horlines = []
self.plot.clear()
def clear_outputs(self):
self.Outputs.transformed_data.send(None)
self.Outputs.components.send(None)
self.Outputs.pca.send(self._pca_projector)
self.Outputs.preprocessor.send(None)
def get_model(self):
if self.rpca is None:
return
if self.rpca.ready():
self.__timer.stop()
self.start_button.setText("Restart (finished)")
self._pca = self.rpca.get_state()
if self._pca is None:
return
self._variance_ratio = self._pca.explained_variance_ratio_
self._cumulative = numpy.cumsum(self._variance_ratio)
self._setup_plot()
self._transformed = None
self.commit()
def _setup_plot(self):
self.plot.clear()
if self._pca is None:
return
explained_ratio = self._variance_ratio
explained = self._cumulative
p = min(len(self._variance_ratio), self.maxp)
self.plot.plot(numpy.arange(p), explained_ratio[:p],
pen=pg.mkPen(QColor(Qt.red), width=2),
antialias=True,
name="Variance")
self.plot.plot(numpy.arange(p), explained[:p],
pen=pg.mkPen(QColor(Qt.darkYellow), width=2),
antialias=True,
name="Cumulative Variance")
cutpos = self._nselected_components() - 1
self._line = pg.InfiniteLine(
angle=90, pos=cutpos, movable=True, bounds=(0, p - 1))
self._line.setCursor(Qt.SizeHorCursor)
self._line.setPen(pg.mkPen(QColor(Qt.black), width=2))
self._line.sigPositionChanged.connect(self._on_cut_changed)
self.plot.addItem(self._line)
self.plot_horlines = (
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)),
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)))
self.plot_horlabels = (
pg.TextItem(color=QColor(Qt.black), anchor=(1, 0)),
pg.TextItem(color=QColor(Qt.black), anchor=(1, 1)))
for item in self.plot_horlabels + self.plot_horlines:
self.plot.addItem(item)
self._set_horline_pos()
self.plot.setRange(xRange=(0.0, p - 1), yRange=(0.0, 1.0))
self._update_axis()
def _set_horline_pos(self):
cutidx = self.ncomponents - 1
for line, label, curve in zip(self.plot_horlines, self.plot_horlabels,
(self._variance_ratio, self._cumulative)):
y = curve[cutidx]
line.setData([-1, cutidx], 2 * [y])
label.setPos(cutidx, y)
label.setPlainText("{:.3f}".format(y))
def _on_cut_changed(self, line):
# cut changed by means of a cut line over the scree plot.
value = int(round(line.value()))
self._line.setValue(value)
current = self._nselected_components()
components = value + 1
if not (self.ncomponents == 0 and
components == len(self._variance_ratio)):
self.ncomponents = components
self._set_horline_pos()
if self._pca is not None:
var = self._cumulative[components - 1]
if numpy.isfinite(var):
self.variance_covered = int(var * 100)
if current != self._nselected_components():
self._invalidate_selection()
def _update_selection_component_spin(self):
# cut changed by "ncomponents" spin.
if self._pca is None:
self._invalidate_selection()
return
if self.ncomponents == 0:
# Special "All" value
cut = len(self._variance_ratio)
else:
cut = self.ncomponents
var = self._cumulative[cut - 1]
if numpy.isfinite(var):
self.variance_covered = int(var * 100)
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_selection_variance_spin(self):
# cut changed by "max variance" spin.
if self._pca is None:
return
cut = numpy.searchsorted(self._cumulative,
self.variance_covered / 100.0) + 1
cut = min(cut, len(self._cumulative))
self.ncomponents = cut
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_normalize(self):
self.fit()
if self.data is None:
self._invalidate_selection()
def _init_projector(self):
cls = DECOMPOSITIONS[self.decomposition_idx]
self._pca_projector = cls(n_components=MAX_COMPONENTS)
self._pca_projector.component = self.ncomponents
self._pca_preprocessors = cls.preprocessors
def _update_decomposition(self):
self._init_projector()
self._update_normalize()
def _nselected_components(self):
"""Return the number of selected components."""
if self._pca is None:
return 0
if self.ncomponents == 0:
# Special "All" value
max_comp = len(self._variance_ratio)
else:
max_comp = self.ncomponents
var_max = self._cumulative[max_comp - 1]
if var_max != numpy.floor(self.variance_covered / 100.0):
cut = max_comp
assert numpy.isfinite(var_max)
self.variance_covered = int(var_max * 100)
else:
self.ncomponents = cut = numpy.searchsorted(
self._cumulative, self.variance_covered / 100.0) + 1
return cut
def _invalidate_selection(self):
self.commit()
def _update_axis(self):
p = min(len(self._variance_ratio), self.maxp)
axis = self.plot.getAxis("bottom")
d = max((p-1)//(self.axis_labels-1), 1)
axis.setTicks([[(i, str(i+1)) for i in range(0, p, d)]])
def commit(self):
transformed = components = pp = None
if self._pca is not None:
if self._transformed is None:
# Compute the full transform (MAX_COMPONENTS components) only once.
self._transformed = self._pca(self.data)
transformed = self._transformed
domain = Domain(
transformed.domain.attributes[:self.ncomponents],
self.data.domain.class_vars,
self.data.domain.metas
)
transformed = transformed.from_table(domain, transformed)
# prevent caching new features by defining compute_value
dom = Domain([ContinuousVariable(a.name, compute_value=lambda _: None)
for a in self._pca.orig_domain.attributes],
metas=[StringVariable(name='component')])
metas = numpy.array([['PC{}'.format(i + 1)
for i in range(self.ncomponents)]],
dtype=object).T
components = Table(dom, self._pca.components_[:self.ncomponents],
metas=metas)
components.name = 'components'
pp = ApplyDomain(domain, "PCA")
self._pca_projector.component = self.ncomponents
self.Outputs.transformed_data.send(transformed)
self.Outputs.components.send(components)
self.Outputs.pca.send(self._pca_projector)
self.Outputs.preprocessor.send(pp)
def send_report(self):
if self.data is None:
return
self.report_items((
("Decomposition", DECOMPOSITIONS[self.decomposition_idx].name),
("Normalize data", str(self.normalize)),
("Selected components", self.ncomponents),
("Explained variance", "{:.3f} %".format(self.variance_covered))
))
self.report_plot()
@classmethod
def migrate_settings(cls, settings, version):
if "variance_covered" in settings:
# Due to the error in gh-1896 the variance_covered was persisted
# as a NaN value, causing a TypeError in the widgets `__init__`.
vc = settings["variance_covered"]
if isinstance(vc, numbers.Real):
if numpy.isfinite(vc):
vc = int(vc)
else:
vc = 100
settings["variance_covered"] = vc
if settings.get("ncomponents", 0) > MAX_COMPONENTS:
settings["ncomponents"] = MAX_COMPONENTS
def variables_filter(model, parent=None):
"""
GUI components: ListView with a lineedit which works as a filter. One can write
a variable name in a edit box and possible matches are then shown in a listview.
"""
def update_completer_model():
""" This gets called when the model for available attributes changes
through either drag/drop or the left/right button actions.
"""
nonlocal original_completer_items
items = [var.name for var in model]
items += [
"%s=%s" % item for v in model for item in v.attributes.items()
]
new = sorted(set(items))
if new != original_completer_items:
original_completer_items = new
completer_model.setStringList(original_completer_items)
def update_completer_prefix():
""" Prefixes all items in the completer model with the current
already done completion to enable the completion of multiple keywords.
"""
nonlocal original_completer_items
prefix = str(completer.completionPrefix())
if not prefix.endswith(" ") and " " in prefix:
prefix, _ = prefix.rsplit(" ", 1)
items = [prefix + " " + item for item in original_completer_items]
else:
items = original_completer_items
old = list(map(str, completer_model.stringList()))
if set(old) != set(items):
completer_model.setStringList(items)
original_completer_items = []
filter_edit = QLineEdit()
filter_edit.setToolTip("Filter the list of available variables.")
filter_edit.setPlaceholderText("Filter")
completer_model = QStringListModel()
completer = QCompleter(completer_model)
completer.setCompletionMode(QCompleter.InlineCompletion)
completer.setModelSorting(QCompleter.CaseSensitivelySortedModel)
filter_edit.setCompleter(completer)
completer_navigator = CompleterNavigator(parent)
filter_edit.installEventFilter(completer_navigator)
proxy = VariableFilterProxyModel()
proxy.setSourceModel(model)
view = VariablesListItemView(acceptedType=Orange.data.Variable)
view.setModel(proxy)
model.dataChanged.connect(update_completer_model)
model.rowsInserted.connect(update_completer_model)
model.rowsRemoved.connect(update_completer_model)
filter_edit.textChanged.connect(update_completer_prefix)
filter_edit.textChanged.connect(proxy.set_filter_string)
return filter_edit, view
class SignInForm(QDialog):
def __init__(self, flags, *args, **kwargs):
super().__init__(flags, *args, **kwargs)
self.cm: CredentialManager = CredentialManager(CREDENTIAL_MANAGER_SERVICE)
self.setWindowTitle('Sign in')
self.setFixedSize(400, 250)
self.server_cb_label = QLabel('Server *')
self.server_cb = QComboBox(self)
self.server_cb.addItems(RESOLWE_URLS)
self.server_cb.setEditable(True)
self.username_label = QLabel('Username *')
self.username_line_edit = QLineEdit(self)
self.username_line_edit.setPlaceholderText('Enter correct username')
self.username_line_edit.returnPressed.connect(self.sign_in)
self.username_line_edit.textChanged.connect(self.handle_sign_in_btn)
self.password_label = QLabel('Password *')
self.password_line_edit = QLineEdit(self)
self.password_line_edit.setPlaceholderText('Enter correct password')
self.password_line_edit.returnPressed.connect(self.sign_in)
self.password_line_edit.textChanged.connect(self.handle_sign_in_btn)
self.password_line_edit.setEchoMode(QLineEdit.Password)
self.sign_in_btn = QPushButton('Sign in', self)
self.sign_in_btn.setDisabled(True)
self.sign_in_btn.clicked.connect(self.sign_in)
self.error_msg = QLabel('Unable to log in with provided credentials.')
self.error_msg.setStyleSheet('color:red')
self.error_msg.hide()
layout = QVBoxLayout(self)
layout.addWidget(self.server_cb_label)
layout.addWidget(self.server_cb)
layout.addWidget(self.username_label)
layout.addWidget(self.username_line_edit)
layout.addWidget(self.password_label)
layout.addWidget(self.password_line_edit)
layout.addWidget(self.error_msg)
layout.addStretch()
layout.addWidget(self.sign_in_btn)
self.resolwe_instance = None
def handle_sign_in_btn(self):
self.sign_in_btn.setEnabled(
True if self.username_line_edit.text() and self.password_line_edit.text() else False
)
def sign_in(self):
self.server_cb_label.setStyleSheet(None)
self.username_label.setStyleSheet(None)
self.password_label.setStyleSheet(None)
self.error_msg.hide()
server = self.server_cb.currentText()
username = self.cm.username if self.cm.username else self.username_line_edit.text()
password = self.cm.password if self.cm.password else self.password_line_edit.text()
if not server:
self.server_cb_label.setStyleSheet('color:red')
return
if not username:
self.username_label.setStyleSheet('color:red')
return
if not password:
self.password_label.setStyleSheet('color:red')
return
try:
self.resolwe_instance = connect(username, password, url=server)
except ResolweAuthException:
self.error_msg.show()
return
self.cm.username = username
self.cm.password = password
self.accept()
class OWSql(OWWidget):
name = "SQL Table"
id = "orange.widgets.data.sql"
description = "Load data set from SQL."
icon = "icons/SQLTable.svg"
priority = 10
category = "Data"
keywords = ["data", "file", "load", "read"]
class Outputs:
data = Output(
"Data",
Table,
doc="Attribute-valued data set read from the input file.")
settings_version = 2
want_main_area = False
resizing_enabled = False
host = Setting(None)
port = Setting(None)
database = Setting(None)
schema = Setting(None)
username = ""
password = ""
table = Setting(None)
sql = Setting("")
guess_values = Setting(True)
download = Setting(False)
materialize = Setting(False)
materialize_table_name = Setting("")
class Information(OWWidget.Information):
data_sampled = Msg("Data description was generated from a sample.")
class Error(OWWidget.Error):
connection = Msg("{}")
no_backends = Msg("Please install a backend to use this widget")
missing_extension = Msg("Database is missing extension{}: {}")
def __init__(self):
super().__init__()
self.backend = None
self.data_desc_table = None
self.database_desc = None
vbox = gui.vBox(self.controlArea, "Server", addSpace=True)
box = gui.vBox(vbox)
self.backendmodel = BackendModel(Backend.available_backends())
self.backendcombo = QComboBox(box)
if len(self.backendmodel):
self.backendcombo.setModel(self.backendmodel)
else:
self.Error.no_backends()
box.setEnabled(False)
box.layout().addWidget(self.backendcombo)
self.servertext = QLineEdit(box)
self.servertext.setPlaceholderText('Server')
self.servertext.setToolTip('Server')
self.servertext.editingFinished.connect(self._load_credentials)
if self.host:
self.servertext.setText(self.host if not self.port else '{}:{}'.
format(self.host, self.port))
box.layout().addWidget(self.servertext)
self.databasetext = QLineEdit(box)
self.databasetext.setPlaceholderText('Database[/Schema]')
self.databasetext.setToolTip('Database or optionally Database/Schema')
if self.database:
self.databasetext.setText(
self.database if not self.schema else '{}/{}'.
format(self.database, self.schema))
box.layout().addWidget(self.databasetext)
self.usernametext = QLineEdit(box)
self.usernametext.setPlaceholderText('Username')
self.usernametext.setToolTip('Username')
box.layout().addWidget(self.usernametext)
self.passwordtext = QLineEdit(box)
self.passwordtext.setPlaceholderText('Password')
self.passwordtext.setToolTip('Password')
self.passwordtext.setEchoMode(QLineEdit.Password)
box.layout().addWidget(self.passwordtext)
self._load_credentials()
tables = gui.hBox(box)
self.tablemodel = TableModel()
self.tablecombo = QComboBox(
minimumContentsLength=35,
sizeAdjustPolicy=QComboBox.AdjustToMinimumContentsLength)
self.tablecombo.setModel(self.tablemodel)
self.tablecombo.setToolTip('table')
tables.layout().addWidget(self.tablecombo)
self.tablecombo.activated[int].connect(self.select_table)
self.connectbutton = gui.button(tables,
self,
'↻',
callback=self.connect)
self.connectbutton.setSizePolicy(QSizePolicy.Fixed, QSizePolicy.Fixed)
tables.layout().addWidget(self.connectbutton)
self.custom_sql = gui.vBox(box)
self.custom_sql.setVisible(False)
self.sqltext = QTextEdit(self.custom_sql)
self.sqltext.setPlainText(self.sql)
self.custom_sql.layout().addWidget(self.sqltext)
mt = gui.hBox(self.custom_sql)
cb = gui.checkBox(mt, self, 'materialize', 'Materialize to table ')
cb.setToolTip('Save results of the query in a table')
le = gui.lineEdit(mt, self, 'materialize_table_name')
le.setToolTip('Save results of the query in a table')
self.executebtn = gui.button(self.custom_sql,
self,
'Execute',
callback=self.open_table)
box.layout().addWidget(self.custom_sql)
gui.checkBox(box,
self,
"guess_values",
"Auto-discover discrete variables",
callback=self.open_table)
gui.checkBox(box,
self,
"download",
"Download data to local memory",
callback=self.open_table)
gui.rubber(self.buttonsArea)
QTimer.singleShot(0, self.connect)
def _load_credentials(self):
self._parse_host_port()
cm = self._credential_manager(self.host, self.port)
self.username = cm.username
self.password = cm.password
if self.username:
self.usernametext.setText(self.username)
if self.password:
self.passwordtext.setText(self.password)
def _save_credentials(self):
cm = self._credential_manager(self.host, self.port)
cm.username = self.username
cm.password = self.password
def _credential_manager(self, host, port):
return CredentialManager("SQL Table: {}:{}".format(host, port))
def error(self, id=0, text=""):
super().error(id, text)
err_style = 'QLineEdit {border: 2px solid red;}'
if 'server' in text or 'host' in text:
self.servertext.setStyleSheet(err_style)
else:
self.servertext.setStyleSheet('')
if 'role' in text:
self.usernametext.setStyleSheet(err_style)
else:
self.usernametext.setStyleSheet('')
if 'database' in text:
self.databasetext.setStyleSheet(err_style)
else:
self.databasetext.setStyleSheet('')
def _parse_host_port(self):
hostport = self.servertext.text().split(':')
self.host = hostport[0]
self.port = hostport[1] if len(hostport) == 2 else None
def connect(self):
self._parse_host_port()
self.database, _, self.schema = self.databasetext.text().partition('/')
self.username = self.usernametext.text() or None
self.password = self.passwordtext.text() or None
try:
if self.backendcombo.currentIndex() < 0:
return
backend = self.backendmodel[self.backendcombo.currentIndex()]
self.backend = backend(
dict(host=self.host,
port=self.port,
database=self.database,
user=self.username,
password=self.password))
self.Error.connection.clear()
self._save_credentials()
self.database_desc = OrderedDict(
(("Host", self.host), ("Port", self.port),
("Database", self.database), ("User name", self.username)))
self.refresh_tables()
self.select_table()
except BackendError as err:
error = str(err).split('\n')[0]
self.Error.connection(error)
self.database_desc = self.data_desc_table = None
self.tablecombo.clear()
def refresh_tables(self):
self.tablemodel.clear()
self.Error.missing_extension.clear()
if self.backend is None:
self.data_desc_table = None
return
self.tablemodel.append("Select a table")
self.tablemodel.extend(self.backend.list_tables(self.schema))
self.tablemodel.append("Custom SQL")
def select_table(self):
curIdx = self.tablecombo.currentIndex()
if self.tablecombo.itemText(curIdx) != "Custom SQL":
self.custom_sql.setVisible(False)
return self.open_table()
else:
self.custom_sql.setVisible(True)
self.data_desc_table = None
self.database_desc["Table"] = "(None)"
self.table = None
#self.Error.missing_extension(
# 's' if len(missing) > 1 else '',
# ', '.join(missing),
# shown=missing)
def open_table(self):
table = self.get_table()
self.data_desc_table = table
self.Outputs.data.send(table)
def get_table(self):
if self.tablecombo.currentIndex() <= 0:
if self.database_desc:
self.database_desc["Table"] = "(None)"
self.data_desc_table = None
return
if self.tablecombo.currentIndex() < self.tablecombo.count() - 1:
self.table = self.tablemodel[self.tablecombo.currentIndex()]
self.database_desc["Table"] = self.table
if "Query" in self.database_desc:
del self.database_desc["Query"]
else:
self.sql = self.table = self.sqltext.toPlainText()
if self.materialize:
import psycopg2
if not self.materialize_table_name:
self.Error.connection(
"Specify a table name to materialize the query")
return
try:
with self.backend.execute_sql_query(
"DROP TABLE IF EXISTS " +
self.materialize_table_name):
pass
with self.backend.execute_sql_query(
"CREATE TABLE " + self.materialize_table_name +
" AS " + self.table):
pass
with self.backend.execute_sql_query(
"ANALYZE " + self.materialize_table_name):
pass
self.table = self.materialize_table_name
except (psycopg2.ProgrammingError, BackendError) as ex:
self.Error.connection(str(ex))
return
try:
table = SqlTable(dict(host=self.host,
port=self.port,
database=self.database,
user=self.username,
password=self.password),
self.table,
backend=type(self.backend),
inspect_values=False)
except BackendError as ex:
self.Error.connection(str(ex))
return
self.Error.connection.clear()
sample = False
if table.approx_len() > LARGE_TABLE and self.guess_values:
confirm = QMessageBox(self)
confirm.setIcon(QMessageBox.Warning)
confirm.setText("Attribute discovery might take "
"a long time on large tables.\n"
"Do you want to auto discover attributes?")
confirm.addButton("Yes", QMessageBox.YesRole)
no_button = confirm.addButton("No", QMessageBox.NoRole)
sample_button = confirm.addButton("Yes, on a sample",
QMessageBox.YesRole)
confirm.exec()
if confirm.clickedButton() == no_button:
self.guess_values = False
elif confirm.clickedButton() == sample_button:
sample = True
self.Information.clear()
if self.guess_values:
QApplication.setOverrideCursor(QCursor(Qt.WaitCursor))
if sample:
s = table.sample_time(1)
domain = s.get_domain(inspect_values=True)
self.Information.data_sampled()
else:
domain = table.get_domain(inspect_values=True)
QApplication.restoreOverrideCursor()
table.domain = domain
if self.download:
if table.approx_len() > MAX_DL_LIMIT:
QMessageBox.warning(
self, 'Warning', "Data is too big to download.\n"
"Consider using the Data Sampler widget to download "
"a sample instead.")
self.download = False
elif table.approx_len() > AUTO_DL_LIMIT:
confirm = QMessageBox.question(
self, 'Question', "Data appears to be big. Do you really "
"want to download it to local memory?",
QMessageBox.Yes | QMessageBox.No, QMessageBox.No)
if confirm == QMessageBox.No:
self.download = False
if self.download:
table.download_data(MAX_DL_LIMIT)
table = Table(table)
return table
def send_report(self):
if not self.database_desc:
self.report_paragraph("No database connection.")
return
self.report_items("Database", self.database_desc)
if self.data_desc_table:
self.report_items("Data",
report.describe_data(self.data_desc_table))
@classmethod
def migrate_settings(cls, settings, version):
if version < 2:
# Until Orange version 3.4.4 username and password had been stored
# in Settings.
cm = cls._credential_manager(settings["host"], settings["port"])
cm.username = settings["username"]
cm.password = settings["password"]
class VizRankDialog(QDialog, ProgressBarMixin, WidgetMessagesMixin):
"""
Base class for VizRank dialogs, providing a GUI with a table and a button,
and the skeleton for managing the evaluation of visualizations.
Derived classes must provide methods
- `iterate_states` for generating combinations (e.g. pairs of attritutes),
- `compute_score(state)` for computing the score of a combination,
- `row_for_state(state)` that returns a list of items inserted into the
table for the given state.
and, optionally,
- `state_count` that returns the number of combinations (used for progress
bar)
- `on_selection_changed` that handles event triggered when the user selects
a table row. The method should emit signal
`VizRankDialog.selectionChanged(object)`.
- `bar_length` returns the length of the bar corresponding to the score.
The class provides a table and a button. A widget constructs a single
instance of this dialog in its `__init__`, like (in Sieve) by using a
convenience method :obj:`add_vizrank`::
self.vizrank, self.vizrank_button = SieveRank.add_vizrank(
box, self, "Score Combinations", self.set_attr)
When the widget receives new data, it must call the VizRankDialog's
method :obj:`VizRankDialog.initialize()` to clear the GUI and reset the
state.
Clicking the Start button calls method `run` (and renames the button to
Pause). Run sets up a progress bar by getting the number of combinations
from :obj:`VizRankDialog.state_count()`. It restores the paused state
(if any) and calls generator :obj:`VizRankDialog.iterate_states()`. For
each generated state, it calls :obj:`VizRankDialog.score(state)`, which
must return the score (lower is better) for this state. If the returned
state is not `None`, the data returned by `row_for_state` is inserted at
the appropriate place in the table.
Args:
master (Orange.widget.OWWidget): widget to which the dialog belongs
Attributes:
master (Orange.widget.OWWidget): widget to which the dialog belongs
captionTitle (str): the caption for the dialog. This can be a class
attribute. `captionTitle` is used by the `ProgressBarMixin`.
"""
captionTitle = ""
NEGATIVE_COLOR = QColor(70, 190, 250)
POSITIVE_COLOR = QColor(170, 242, 43)
processingStateChanged = Signal(int)
progressBarValueChanged = Signal(float)
messageActivated = Signal(Msg)
messageDeactivated = Signal(Msg)
selectionChanged = Signal(object)
class Information(WidgetMessagesMixin.Information):
nothing_to_rank = Msg("There is nothing to rank.")
def __init__(self, master):
"""Initialize the attributes and set up the interface"""
QDialog.__init__(self, master, windowTitle=self.captionTitle)
WidgetMessagesMixin.__init__(self)
self.setLayout(QVBoxLayout())
self.insert_message_bar()
self.layout().insertWidget(0, self.message_bar)
self.master = master
self.keep_running = False
self.scheduled_call = None
self.saved_state = None
self.saved_progress = 0
self.scores = []
self.add_to_model = queue.Queue()
self.update_timer = QTimer(self)
self.update_timer.timeout.connect(self._update)
self.update_timer.setInterval(200)
self._thread = None
self._worker = None
self.filter = QLineEdit()
self.filter.setPlaceholderText("Filter ...")
self.filter.textChanged.connect(self.filter_changed)
self.layout().addWidget(self.filter)
# Remove focus from line edit
self.setFocus(Qt.ActiveWindowFocusReason)
self.rank_model = QStandardItemModel(self)
self.model_proxy = QSortFilterProxyModel(self,
filterCaseSensitivity=False)
self.model_proxy.setSourceModel(self.rank_model)
self.rank_table = view = QTableView(
selectionBehavior=QTableView.SelectRows,
selectionMode=QTableView.SingleSelection,
showGrid=False,
editTriggers=gui.TableView.NoEditTriggers)
if self._has_bars:
view.setItemDelegate(TableBarItem())
else:
view.setItemDelegate(HorizontalGridDelegate())
view.setModel(self.model_proxy)
view.selectionModel().selectionChanged.connect(
self.on_selection_changed)
view.horizontalHeader().setStretchLastSection(True)
view.horizontalHeader().hide()
self.layout().addWidget(view)
self.button = gui.button(self,
self,
"Start",
callback=self.toggle,
default=True)
@property
def _has_bars(self):
return type(self).bar_length is not VizRankDialog.bar_length
@classmethod
def add_vizrank(cls, widget, master, button_label, set_attr_callback):
"""
Equip the widget with VizRank button and dialog, and monkey patch the
widget's `closeEvent` and `hideEvent` to close/hide the vizrank, too.
Args:
widget (QWidget): the widget into whose layout to insert the button
master (Orange.widgets.widget.OWWidget): the master widget
button_label: the label for the button
set_attr_callback: the callback for setting the projection chosen
in the vizrank
Returns:
tuple with Vizrank dialog instance and push button
"""
# Monkey patching could be avoided by mixing-in the class (not
# necessarily a good idea since we can make a mess of multiple
# defined/derived closeEvent and hideEvent methods). Furthermore,
# per-class patching would be better than per-instance, but we don't
# want to mess with meta-classes either.
vizrank = cls(master)
button = gui.button(widget,
master,
button_label,
callback=vizrank.reshow,
enabled=False)
vizrank.selectionChanged.connect(lambda args: set_attr_callback(*args))
master_close_event = master.closeEvent
master_hide_event = master.hideEvent
master_delete_event = master.onDeleteWidget
def closeEvent(event):
vizrank.close()
master_close_event(event)
def hideEvent(event):
vizrank.hide()
master_hide_event(event)
def deleteEvent():
vizrank.keep_running = False
if vizrank._thread is not None and vizrank._thread.isRunning():
vizrank._thread.quit()
vizrank._thread.wait()
master_delete_event()
master.closeEvent = closeEvent
master.hideEvent = hideEvent
master.onDeleteWidget = deleteEvent
return vizrank, button
def reshow(self):
"""Put the widget on top of all windows
"""
self.show()
self.raise_()
self.activateWindow()
def initialize(self):
"""
Clear and initialize the dialog.
This method must be called by the widget when the data is reset,
e.g. from `set_data` handler.
"""
if self._thread is not None and self._thread.isRunning():
self.keep_running = False
self._thread.quit()
self._thread.wait()
self.keep_running = False
self.scheduled_call = None
self.saved_state = None
self.saved_progress = 0
self.update_timer.stop()
self.progressBarFinished()
self.scores = []
self._update_model() # empty queue
self.rank_model.clear()
self.button.setText("Start")
self.button.setEnabled(self.check_preconditions())
self._thread = QThread(self)
self._worker = Worker(self)
self._worker.moveToThread(self._thread)
self._worker.stopped.connect(self._thread.quit)
self._worker.stopped.connect(self._select_first_if_none)
self._worker.stopped.connect(self._stopped)
self._worker.done.connect(self._done)
self._thread.started.connect(self._worker.do_work)
def filter_changed(self, text):
self.model_proxy.setFilterFixedString(text)
def stop_and_reset(self, reset_method=None):
if self.keep_running:
self.scheduled_call = reset_method or self.initialize
self.keep_running = False
else:
self.initialize()
def check_preconditions(self):
"""Check whether there is sufficient data for ranking."""
return True
def on_selection_changed(self, selected, deselected):
"""
Set the new visualization in the widget when the user select a
row in the table.
If derived class does not reimplement this, the table gives the
information but the user can't click it to select the visualization.
Args:
selected: the index of the selected item
deselected: the index of the previously selected item
"""
pass
def iterate_states(self, initial_state):
"""
Generate all possible states (e.g. attribute combinations) for the
given data. The content of the generated states is specific to the
visualization.
This method must be defined in the derived classes.
Args:
initial_state: initial state; None if this is the first call
"""
raise NotImplementedError
def state_count(self):
"""
Return the number of states for the progress bar.
Derived classes should implement this to ensure the proper behaviour of
the progress bar"""
return 0
def compute_score(self, state):
"""
Abstract method for computing the score for the given state. Smaller
scores are better.
Args:
state: the state, e.g. the combination of attributes as generated
by :obj:`state_count`.
"""
raise NotImplementedError
def bar_length(self, score):
"""Compute the bar length (between 0 and 1) corresponding to the score.
Return `None` if the score cannot be normalized.
"""
return None
def row_for_state(self, score, state):
"""
Abstract method that return the items that are inserted into the table.
Args:
score: score, computed by :obj:`compute_score`
state: the state, e.g. combination of attributes
"""
raise NotImplementedError
def _select_first_if_none(self):
if not self.rank_table.selectedIndexes():
self.rank_table.selectRow(0)
def _done(self):
self.button.setText("Finished")
self.button.setEnabled(False)
self.keep_running = False
self.saved_state = None
def _stopped(self):
self.update_timer.stop()
self.progressBarFinished()
self._update_model()
self.stopped()
if self.scheduled_call:
self.scheduled_call()
def _update(self):
self._update_model()
self._update_progress()
def _update_progress(self):
self.progressBarSet(
int(self.saved_progress * 100 / max(1, self.state_count())))
def _update_model(self):
try:
while True:
pos, row_items = self.add_to_model.get_nowait()
self.rank_model.insertRow(pos, row_items)
except queue.Empty:
pass
def toggle(self):
"""Start or pause the computation."""
self.keep_running = not self.keep_running
if self.keep_running:
self.button.setText("Pause")
self.progressBarInit()
self.update_timer.start()
self.before_running()
self._thread.start()
else:
self.button.setText("Continue")
self._thread.quit()
# Need to sync state (the worker must read the keep_running
# state and stop) for reliable restart.
self._thread.wait()
def before_running(self):
"""Code that is run before running vizrank in its own thread"""
pass
def stopped(self):
"""Code that is run after stopping the vizrank thread"""
pass
class OWSelectAttributes(widget.OWWidget):
name = "Select Columns"
description = "Select columns from the data table and assign them to " \
"data features, classes or meta variables."
icon = "icons/SelectColumns.svg"
priority = 100
class Inputs:
data = Input("Data", Table)
class Outputs:
data = Output("Data", Table)
features = Output("Features", widget.AttributeList, dynamic=False)
want_main_area = False
want_control_area = True
settingsHandler = SelectAttributesDomainContextHandler()
domain_role_hints = ContextSetting({})
auto_commit = Setting(True)
def __init__(self):
super().__init__()
self.controlArea = QWidget(self.controlArea)
self.layout().addWidget(self.controlArea)
layout = QGridLayout()
self.controlArea.setLayout(layout)
layout.setContentsMargins(4, 4, 4, 4)
box = gui.vBox(self.controlArea,
"Available Variables",
addToLayout=False)
self.filter_edit = QLineEdit()
self.filter_edit.setToolTip("Filter the list of available variables.")
box.layout().addWidget(self.filter_edit)
if hasattr(self.filter_edit, "setPlaceholderText"):
self.filter_edit.setPlaceholderText("Filter")
self.completer = QCompleter()
self.completer.setCompletionMode(QCompleter.InlineCompletion)
self.completer_model = QStringListModel()
self.completer.setModel(self.completer_model)
self.completer.setModelSorting(QCompleter.CaseSensitivelySortedModel)
self.filter_edit.setCompleter(self.completer)
self.completer_navigator = CompleterNavigator(self)
self.filter_edit.installEventFilter(self.completer_navigator)
def dropcompleted(action):
if action == Qt.MoveAction:
self.commit()
self.available_attrs = VariableListModel(enable_dnd=True)
self.available_attrs_proxy = VariableFilterProxyModel()
self.available_attrs_proxy.setSourceModel(self.available_attrs)
self.available_attrs_view = VariablesListItemView(
acceptedType=Orange.data.Variable)
self.available_attrs_view.setModel(self.available_attrs_proxy)
aa = self.available_attrs
aa.dataChanged.connect(self.update_completer_model)
aa.rowsInserted.connect(self.update_completer_model)
aa.rowsRemoved.connect(self.update_completer_model)
self.available_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.available_attrs_view))
self.available_attrs_view.dragDropActionDidComplete.connect(
dropcompleted)
self.filter_edit.textChanged.connect(self.update_completer_prefix)
self.filter_edit.textChanged.connect(
self.available_attrs_proxy.set_filter_string)
box.layout().addWidget(self.available_attrs_view)
layout.addWidget(box, 0, 0, 3, 1)
box = gui.vBox(self.controlArea, "Features", addToLayout=False)
self.used_attrs = VariableListModel(enable_dnd=True)
self.used_attrs_view = VariablesListItemView(
acceptedType=(Orange.data.DiscreteVariable,
Orange.data.ContinuousVariable))
self.used_attrs_view.setModel(self.used_attrs)
self.used_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.used_attrs_view))
self.used_attrs_view.dragDropActionDidComplete.connect(dropcompleted)
box.layout().addWidget(self.used_attrs_view)
layout.addWidget(box, 0, 2, 1, 1)
box = gui.vBox(self.controlArea, "Target Variable", addToLayout=False)
self.class_attrs = ClassVarListItemModel(enable_dnd=True)
self.class_attrs_view = ClassVariableItemView(
acceptedType=(Orange.data.DiscreteVariable,
Orange.data.ContinuousVariable))
self.class_attrs_view.setModel(self.class_attrs)
self.class_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.class_attrs_view))
self.class_attrs_view.dragDropActionDidComplete.connect(dropcompleted)
self.class_attrs_view.setMaximumHeight(72)
box.layout().addWidget(self.class_attrs_view)
layout.addWidget(box, 1, 2, 1, 1)
box = gui.vBox(self.controlArea, "Meta Attributes", addToLayout=False)
self.meta_attrs = VariableListModel(enable_dnd=True)
self.meta_attrs_view = VariablesListItemView(
acceptedType=Orange.data.Variable)
self.meta_attrs_view.setModel(self.meta_attrs)
self.meta_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.meta_attrs_view))
self.meta_attrs_view.dragDropActionDidComplete.connect(dropcompleted)
box.layout().addWidget(self.meta_attrs_view)
layout.addWidget(box, 2, 2, 1, 1)
bbox = gui.vBox(self.controlArea, addToLayout=False, margin=0)
layout.addWidget(bbox, 0, 1, 1, 1)
self.up_attr_button = gui.button(bbox,
self,
"Up",
callback=partial(
self.move_up,
self.used_attrs_view))
self.move_attr_button = gui.button(bbox,
self,
">",
callback=partial(
self.move_selected,
self.used_attrs_view))
self.down_attr_button = gui.button(bbox,
self,
"Down",
callback=partial(
self.move_down,
self.used_attrs_view))
bbox = gui.vBox(self.controlArea, addToLayout=False, margin=0)
layout.addWidget(bbox, 1, 1, 1, 1)
self.up_class_button = gui.button(bbox,
self,
"Up",
callback=partial(
self.move_up,
self.class_attrs_view))
self.move_class_button = gui.button(bbox,
self,
">",
callback=partial(
self.move_selected,
self.class_attrs_view,
exclusive=False))
self.down_class_button = gui.button(bbox,
self,
"Down",
callback=partial(
self.move_down,
self.class_attrs_view))
bbox = gui.vBox(self.controlArea, addToLayout=False, margin=0)
layout.addWidget(bbox, 2, 1, 1, 1)
self.up_meta_button = gui.button(bbox,
self,
"Up",
callback=partial(
self.move_up,
self.meta_attrs_view))
self.move_meta_button = gui.button(bbox,
self,
">",
callback=partial(
self.move_selected,
self.meta_attrs_view))
self.down_meta_button = gui.button(bbox,
self,
"Down",
callback=partial(
self.move_down,
self.meta_attrs_view))
autobox = gui.auto_commit(None, self, "auto_commit", "Send")
layout.addWidget(autobox, 3, 0, 1, 3)
reset = gui.button(None, self, "Reset", callback=self.reset)
autobox.layout().insertWidget(0, self.report_button)
autobox.layout().insertWidget(1, reset)
autobox.layout().insertSpacing(2, 10)
reset.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Preferred)
self.report_button.setSizePolicy(QSizePolicy.Expanding,
QSizePolicy.Preferred)
layout.setRowStretch(0, 4)
layout.setRowStretch(1, 0)
layout.setRowStretch(2, 2)
layout.setHorizontalSpacing(0)
self.controlArea.setLayout(layout)
self.data = None
self.output_data = None
self.original_completer_items = []
self.resize(500, 600)
@Inputs.data
def set_data(self, data=None):
self.update_domain_role_hints()
self.closeContext()
self.data = data
if data is not None:
self.openContext(data)
all_vars = data.domain.variables + data.domain.metas
var_sig = lambda attr: (attr.name, vartype(attr))
domain_hints = {
var_sig(attr): ("attribute", i)
for i, attr in enumerate(data.domain.attributes)
}
domain_hints.update({
var_sig(attr): ("meta", i)
for i, attr in enumerate(data.domain.metas)
})
if data.domain.class_vars:
domain_hints.update({
var_sig(attr): ("class", i)
for i, attr in enumerate(data.domain.class_vars)
})
# update the hints from context settings
domain_hints.update(self.domain_role_hints)
attrs_for_role = lambda role: [
(domain_hints[var_sig(attr)][1], attr) for attr in all_vars
if domain_hints[var_sig(attr)][0] == role
]
attributes = [
attr for place, attr in sorted(attrs_for_role("attribute"),
key=lambda a: a[0])
]
classes = [
attr for place, attr in sorted(attrs_for_role("class"),
key=lambda a: a[0])
]
metas = [
attr for place, attr in sorted(attrs_for_role("meta"),
key=lambda a: a[0])
]
available = [
attr for place, attr in sorted(attrs_for_role("available"),
key=lambda a: a[0])
]
self.used_attrs[:] = attributes
self.class_attrs[:] = classes
self.meta_attrs[:] = metas
self.available_attrs[:] = available
else:
self.used_attrs[:] = []
self.class_attrs[:] = []
self.meta_attrs[:] = []
self.available_attrs[:] = []
self.unconditional_commit()
def update_domain_role_hints(self):
""" Update the domain hints to be stored in the widgets settings.
"""
hints_from_model = lambda role, model: [(
(attr.name, vartype(attr)),
(role, i)) for i, attr in enumerate(model)]
hints = dict(hints_from_model("available", self.available_attrs))
hints.update(hints_from_model("attribute", self.used_attrs))
hints.update(hints_from_model("class", self.class_attrs))
hints.update(hints_from_model("meta", self.meta_attrs))
self.domain_role_hints = hints
def selected_rows(self, view):
""" Return the selected rows in the view.
"""
rows = view.selectionModel().selectedRows()
model = view.model()
if isinstance(model, QSortFilterProxyModel):
rows = [model.mapToSource(r) for r in rows]
return [r.row() for r in rows]
def move_rows(self, view, rows, offset):
model = view.model()
newrows = [min(max(0, row + offset), len(model) - 1) for row in rows]
for row, newrow in sorted(zip(rows, newrows), reverse=offset > 0):
model[row], model[newrow] = model[newrow], model[row]
selection = QItemSelection()
for nrow in newrows:
index = model.index(nrow, 0)
selection.select(index, index)
view.selectionModel().select(selection,
QItemSelectionModel.ClearAndSelect)
self.commit()
def move_up(self, view):
selected = self.selected_rows(view)
self.move_rows(view, selected, -1)
def move_down(self, view):
selected = self.selected_rows(view)
self.move_rows(view, selected, 1)
def move_selected(self, view, exclusive=False):
if self.selected_rows(view):
self.move_selected_from_to(view, self.available_attrs_view)
elif self.selected_rows(self.available_attrs_view):
self.move_selected_from_to(self.available_attrs_view, view,
exclusive)
def move_selected_from_to(self, src, dst, exclusive=False):
self.move_from_to(src, dst, self.selected_rows(src), exclusive)
def move_from_to(self, src, dst, rows, exclusive=False):
src_model = source_model(src)
attrs = [src_model[r] for r in rows]
for s1, s2 in reversed(list(slices(rows))):
del src_model[s1:s2]
dst_model = source_model(dst)
dst_model.extend(attrs)
self.commit()
def update_interface_state(self,
focus=None,
selected=None,
deselected=None):
for view in [
self.available_attrs_view, self.used_attrs_view,
self.class_attrs_view, self.meta_attrs_view
]:
if view is not focus and not view.hasFocus(
) and self.selected_rows(view):
view.selectionModel().clear()
def selected_vars(view):
model = source_model(view)
return [model[i] for i in self.selected_rows(view)]
available_selected = selected_vars(self.available_attrs_view)
attrs_selected = selected_vars(self.used_attrs_view)
class_selected = selected_vars(self.class_attrs_view)
meta_selected = selected_vars(self.meta_attrs_view)
available_types = set(map(type, available_selected))
all_primitive = all(var.is_primitive() for var in available_types)
move_attr_enabled = (available_selected and all_primitive) or \
attrs_selected
self.move_attr_button.setEnabled(bool(move_attr_enabled))
if move_attr_enabled:
self.move_attr_button.setText(">" if available_selected else "<")
move_class_enabled = (all_primitive
and available_selected) or class_selected
self.move_class_button.setEnabled(bool(move_class_enabled))
if move_class_enabled:
self.move_class_button.setText(">" if available_selected else "<")
move_meta_enabled = available_selected or meta_selected
self.move_meta_button.setEnabled(bool(move_meta_enabled))
if move_meta_enabled:
self.move_meta_button.setText(">" if available_selected else "<")
def update_completer_model(self, *_):
""" This gets called when the model for available attributes changes
through either drag/drop or the left/right button actions.
"""
vars = list(self.available_attrs)
items = [var.name for var in vars]
items += [
"%s=%s" % item for v in vars for item in v.attributes.items()
]
new = sorted(set(items))
if new != self.original_completer_items:
self.original_completer_items = new
self.completer_model.setStringList(self.original_completer_items)
def update_completer_prefix(self, filter):
""" Prefixes all items in the completer model with the current
already done completion to enable the completion of multiple keywords.
"""
prefix = str(self.completer.completionPrefix())
if not prefix.endswith(" ") and " " in prefix:
prefix, _ = prefix.rsplit(" ", 1)
items = [
prefix + " " + item for item in self.original_completer_items
]
else:
items = self.original_completer_items
old = list(map(str, self.completer_model.stringList()))
if set(old) != set(items):
self.completer_model.setStringList(items)
def commit(self):
self.update_domain_role_hints()
if self.data is not None:
attributes = list(self.used_attrs)
class_var = list(self.class_attrs)
metas = list(self.meta_attrs)
domain = Orange.data.Domain(attributes, class_var, metas)
newdata = self.data.transform(domain)
self.output_data = newdata
self.Outputs.data.send(newdata)
self.Outputs.features.send(widget.AttributeList(attributes))
else:
self.output_data = None
self.Outputs.data.send(None)
self.Outputs.features.send(None)
def reset(self):
if self.data is not None:
self.available_attrs[:] = []
self.used_attrs[:] = self.data.domain.attributes
self.class_attrs[:] = self.data.domain.class_vars
self.meta_attrs[:] = self.data.domain.metas
self.update_domain_role_hints()
self.commit()
def send_report(self):
if not self.data or not self.output_data:
return
in_domain, out_domain = self.data.domain, self.output_data.domain
self.report_domain("Input data", self.data.domain)
if (in_domain.attributes, in_domain.class_vars,
in_domain.metas) == (out_domain.attributes,
out_domain.class_vars, out_domain.metas):
self.report_paragraph("Output data", "No changes.")
else:
self.report_domain("Output data", self.output_data.domain)
diff = list(
set(in_domain.variables + in_domain.metas) -
set(out_domain.variables + out_domain.metas))
if diff:
text = "%i (%s)" % (len(diff), ", ".join(x.name for x in diff))
self.report_items((("Removed", text), ))
class VizRankDialog(QDialog, ProgressBarMixin, WidgetMessagesMixin):
"""
Base class for VizRank dialogs, providing a GUI with a table and a button,
and the skeleton for managing the evaluation of visualizations.
Derived classes must provide methods
- `iterate_states` for generating combinations (e.g. pairs of attritutes),
- `compute_score(state)` for computing the score of a combination,
- `row_for_state(state)` that returns a list of items inserted into the
table for the given state.
and, optionally,
- `state_count` that returns the number of combinations (used for progress
bar)
- `on_selection_changed` that handles event triggered when the user selects
a table row. The method should emit signal
`VizRankDialog.selectionChanged(object)`.
- `bar_length` returns the length of the bar corresponding to the score.
The class provides a table and a button. A widget constructs a single
instance of this dialog in its `__init__`, like (in Sieve) by using a
convenience method :obj:`add_vizrank`::
self.vizrank, self.vizrank_button = SieveRank.add_vizrank(
box, self, "Score Combinations", self.set_attr)
When the widget receives new data, it must call the VizRankDialog's
method :obj:`VizRankDialog.initialize()` to clear the GUI and reset the
state.
Clicking the Start button calls method `run` (and renames the button to
Pause). Run sets up a progress bar by getting the number of combinations
from :obj:`VizRankDialog.state_count()`. It restores the paused state
(if any) and calls generator :obj:`VizRankDialog.iterate_states()`. For
each generated state, it calls :obj:`VizRankDialog.score(state)`, which
must return the score (lower is better) for this state. If the returned
state is not `None`, the data returned by `row_for_state` is inserted at
the appropriate place in the table.
Args:
master (Orange.widget.OWWidget): widget to which the dialog belongs
Attributes:
master (Orange.widget.OWWidget): widget to which the dialog belongs
captionTitle (str): the caption for the dialog. This can be a class
attribute. `captionTitle` is used by the `ProgressBarMixin`.
"""
captionTitle = ""
processingStateChanged = Signal(int)
progressBarValueChanged = Signal(float)
messageActivated = Signal(Msg)
messageDeactivated = Signal(Msg)
selectionChanged = Signal(object)
class Information(WidgetMessagesMixin.Information):
nothing_to_rank = Msg("There is nothing to rank.")
def __init__(self, master):
"""Initialize the attributes and set up the interface"""
QDialog.__init__(self, master, windowTitle=self.captionTitle)
WidgetMessagesMixin.__init__(self)
self.setLayout(QVBoxLayout())
self.insert_message_bar()
self.layout().insertWidget(0, self.message_bar)
self.master = master
self.keep_running = False
self.scheduled_call = None
self.saved_state = None
self.saved_progress = 0
self.scores = []
self.add_to_model = queue.Queue()
self.update_timer = QTimer(self)
self.update_timer.timeout.connect(self._update)
self.update_timer.setInterval(200)
self._thread = None
self._worker = None
self.filter = QLineEdit()
self.filter.setPlaceholderText("Filter ...")
self.filter.textChanged.connect(self.filter_changed)
self.layout().addWidget(self.filter)
# Remove focus from line edit
self.setFocus(Qt.ActiveWindowFocusReason)
self.rank_model = QStandardItemModel(self)
self.model_proxy = QSortFilterProxyModel(
self, filterCaseSensitivity=False)
self.model_proxy.setSourceModel(self.rank_model)
self.rank_table = view = QTableView(
selectionBehavior=QTableView.SelectRows,
selectionMode=QTableView.SingleSelection,
showGrid=False,
editTriggers=gui.TableView.NoEditTriggers)
if self._has_bars:
view.setItemDelegate(TableBarItem())
else:
view.setItemDelegate(HorizontalGridDelegate())
view.setModel(self.model_proxy)
view.selectionModel().selectionChanged.connect(
self.on_selection_changed)
view.horizontalHeader().setStretchLastSection(True)
view.horizontalHeader().hide()
self.layout().addWidget(view)
self.button = gui.button(
self, self, "Start", callback=self.toggle, default=True)
@property
def _has_bars(self):
return type(self).bar_length is not VizRankDialog.bar_length
@classmethod
def add_vizrank(cls, widget, master, button_label, set_attr_callback):
"""
Equip the widget with VizRank button and dialog, and monkey patch the
widget's `closeEvent` and `hideEvent` to close/hide the vizrank, too.
Args:
widget (QWidget): the widget into whose layout to insert the button
master (Orange.widgets.widget.OWWidget): the master widget
button_label: the label for the button
set_attr_callback: the callback for setting the projection chosen
in the vizrank
Returns:
tuple with Vizrank dialog instance and push button
"""
# Monkey patching could be avoided by mixing-in the class (not
# necessarily a good idea since we can make a mess of multiple
# defined/derived closeEvent and hideEvent methods). Furthermore,
# per-class patching would be better than per-instance, but we don't
# want to mess with meta-classes either.
vizrank = cls(master)
button = gui.button(
widget, master, button_label, callback=vizrank.reshow,
enabled=False)
vizrank.selectionChanged.connect(lambda args: set_attr_callback(*args))
master_close_event = master.closeEvent
master_hide_event = master.hideEvent
master_delete_event = master.onDeleteWidget
def closeEvent(event):
vizrank.close()
master_close_event(event)
def hideEvent(event):
vizrank.hide()
master_hide_event(event)
def deleteEvent():
vizrank.keep_running = False
if vizrank._thread is not None and vizrank._thread.isRunning():
vizrank._thread.quit()
vizrank._thread.wait()
master_delete_event()
master.closeEvent = closeEvent
master.hideEvent = hideEvent
master.onDeleteWidget = deleteEvent
return vizrank, button
def reshow(self):
"""Put the widget on top of all windows
"""
self.show()
self.raise_()
self.activateWindow()
def initialize(self):
"""
Clear and initialize the dialog.
This method must be called by the widget when the data is reset,
e.g. from `set_data` handler.
"""
if self._thread is not None and self._thread.isRunning():
self.keep_running = False
self._thread.quit()
self._thread.wait()
self.keep_running = False
self.scheduled_call = None
self.saved_state = None
self.saved_progress = 0
self.update_timer.stop()
self.progressBarFinished()
self.scores = []
self._update_model() # empty queue
self.rank_model.clear()
self.button.setText("Start")
self.button.setEnabled(self.check_preconditions())
self._thread = QThread(self)
self._worker = Worker(self)
self._worker.moveToThread(self._thread)
self._worker.stopped.connect(self._thread.quit)
self._worker.stopped.connect(self._select_first_if_none)
self._worker.stopped.connect(self._stopped)
self._worker.done.connect(self._done)
self._thread.started.connect(self._worker.do_work)
def filter_changed(self, text):
self.model_proxy.setFilterFixedString(text)
def stop_and_reset(self, reset_method=None):
if self.keep_running:
self.scheduled_call = reset_method or self.initialize
self.keep_running = False
else:
self.initialize()
def check_preconditions(self):
"""Check whether there is sufficient data for ranking."""
return True
def on_selection_changed(self, selected, deselected):
"""
Set the new visualization in the widget when the user select a
row in the table.
If derived class does not reimplement this, the table gives the
information but the user can't click it to select the visualization.
Args:
selected: the index of the selected item
deselected: the index of the previously selected item
"""
pass
def iterate_states(self, initial_state):
"""
Generate all possible states (e.g. attribute combinations) for the
given data. The content of the generated states is specific to the
visualization.
This method must be defined in the derived classes.
Args:
initial_state: initial state; None if this is the first call
"""
raise NotImplementedError
def state_count(self):
"""
Return the number of states for the progress bar.
Derived classes should implement this to ensure the proper behaviour of
the progress bar"""
return 0
def compute_score(self, state):
"""
Abstract method for computing the score for the given state. Smaller
scores are better.
Args:
state: the state, e.g. the combination of attributes as generated
by :obj:`state_count`.
"""
raise NotImplementedError
def bar_length(self, score):
"""Compute the bar length (between 0 and 1) corresponding to the score.
Return `None` if the score cannot be normalized.
"""
return None
def row_for_state(self, score, state):
"""
Abstract method that return the items that are inserted into the table.
Args:
score: score, computed by :obj:`compute_score`
state: the state, e.g. combination of attributes
"""
raise NotImplementedError
def _select_first_if_none(self):
if not self.rank_table.selectedIndexes():
self.rank_table.selectRow(0)
def _done(self):
self.button.setText("Finished")
self.button.setEnabled(False)
self.keep_running = False
self.saved_state = None
def _stopped(self):
self.update_timer.stop()
self.progressBarFinished()
self._update_model()
self.stopped()
if self.scheduled_call:
self.scheduled_call()
def _update(self):
self._update_model()
self._update_progress()
def _update_progress(self):
self.progressBarSet(int(self.saved_progress * 100 / max(1, self.state_count())))
def _update_model(self):
try:
while True:
pos, row_items = self.add_to_model.get_nowait()
self.rank_model.insertRow(pos, row_items)
except queue.Empty:
pass
def toggle(self):
"""Start or pause the computation."""
self.keep_running = not self.keep_running
if self.keep_running:
self.button.setText("Pause")
self.progressBarInit()
self.update_timer.start()
self.before_running()
self._thread.start()
else:
self.button.setText("Continue")
self._thread.quit()
# Need to sync state (the worker must read the keep_running
# state and stop) for reliable restart.
self._thread.wait()
def before_running(self):
"""Code that is run before running vizrank in its own thread"""
pass
def stopped(self):
"""Code that is run after stopping the vizrank thread"""
pass
class OWPCA(widget.OWWidget):
name = "PCA"
description = "Principal component analysis with a scree-diagram."
icon = "icons/PCA.svg"
priority = 3050
inputs = [("Data", Table, "set_data")]
outputs = [("Transformed data", Table), ("Components", Table),
("PCA", PCA)]
ncomponents = settings.Setting(2)
variance_covered = settings.Setting(100)
batch_size = settings.Setting(100)
address = settings.Setting('')
auto_update = settings.Setting(True)
auto_commit = settings.Setting(True)
normalize = settings.Setting(True)
maxp = settings.Setting(20)
axis_labels = settings.Setting(10)
graph_name = "plot.plotItem"
def __init__(self):
super().__init__()
self.data = None
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = False
self._pca_projector = PCA()
self._pca_projector.component = self.ncomponents
self._pca_preprocessors = PCA.preprocessors
# Components Selection
box = gui.vBox(self.controlArea, "Components Selection")
form = QFormLayout()
box.layout().addLayout(form)
self.components_spin = gui.spin(
box,
self,
"ncomponents",
0,
1000,
callback=self._update_selection_component_spin,
keyboardTracking=False)
self.components_spin.setSpecialValueText("All")
self.variance_spin = gui.spin(
box,
self,
"variance_covered",
1,
100,
callback=self._update_selection_variance_spin,
keyboardTracking=False)
self.variance_spin.setSuffix("%")
form.addRow("Components:", self.components_spin)
form.addRow("Variance covered:", self.variance_spin)
# Incremental learning
self.sampling_box = gui.vBox(self.controlArea, "Incremental learning")
self.addresstext = QLineEdit(box)
self.addresstext.setPlaceholderText('Remote server')
if self.address:
self.addresstext.setText(self.address)
self.sampling_box.layout().addWidget(self.addresstext)
form = QFormLayout()
self.sampling_box.layout().addLayout(form)
self.batch_spin = gui.spin(self.sampling_box,
self,
"batch_size",
50,
100000,
step=50,
keyboardTracking=False)
form.addRow("Batch size ~ ", self.batch_spin)
self.start_button = gui.button(
self.sampling_box,
self,
"Start remote computation",
callback=self.start,
autoDefault=False,
tooltip="Start/abort computation on the server")
self.start_button.setEnabled(False)
gui.checkBox(self.sampling_box,
self,
"auto_update",
"Periodically fetch model",
callback=self.update_model)
self.__timer = QTimer(self, interval=2000)
self.__timer.timeout.connect(self.get_model)
self.sampling_box.setVisible(remotely)
# Options
self.options_box = gui.vBox(self.controlArea, "Options")
gui.checkBox(self.options_box,
self,
"normalize",
"Normalize data",
callback=self._update_normalize)
self.maxp_spin = gui.spin(self.options_box,
self,
"maxp",
1,
100,
label="Show only first",
callback=self._setup_plot,
keyboardTracking=False)
self.controlArea.layout().addStretch()
gui.auto_commit(self.controlArea,
self,
"auto_commit",
"Apply",
checkbox_label="Apply automatically")
self.plot = pg.PlotWidget(background="w")
axis = self.plot.getAxis("bottom")
axis.setLabel("Principal Components")
axis = self.plot.getAxis("left")
axis.setLabel("Proportion of variance")
self.plot_horlabels = []
self.plot_horlines = []
self.plot.getViewBox().setMenuEnabled(False)
self.plot.getViewBox().setMouseEnabled(False, False)
self.plot.showGrid(True, True, alpha=0.5)
self.plot.setRange(xRange=(0.0, 1.0), yRange=(0.0, 1.0))
self.mainArea.layout().addWidget(self.plot)
self._update_normalize()
def update_model(self):
self.get_model()
if self.auto_update and self.rpca and not self.rpca.ready():
self.__timer.start(2000)
else:
self.__timer.stop()
def start(self):
if 'Abort' in self.start_button.text():
self.rpca.abort()
self.__timer.stop()
self.start_button.setText("Start remote computation")
else:
self.address = self.addresstext.text()
with remote.server(self.address):
from Orange.projection.pca import RemotePCA
maxiter = (1e5 + self.data.approx_len()) / self.batch_size * 3
self.rpca = RemotePCA(self.data, self.batch_size, int(maxiter))
self.update_model()
self.start_button.setText("Abort remote computation")
def set_data(self, data):
self.information()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
elif not remotely:
self.information("Data has been sampled")
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.data = data
self.fit()
def fit(self):
self.clear()
self.start_button.setEnabled(False)
if self.data is None:
return
data = self.data
self._transformed = None
if isinstance(data, SqlTable): # data was big and remote available
self.sampling_box.setVisible(True)
self.start_button.setText("Start remote computation")
self.start_button.setEnabled(True)
else:
self.sampling_box.setVisible(False)
pca = self._pca_projector(data)
variance_ratio = pca.explained_variance_ratio_
cumulative = numpy.cumsum(variance_ratio)
self.components_spin.setRange(0, len(cumulative))
self._pca = pca
self._variance_ratio = variance_ratio
self._cumulative = cumulative
self._setup_plot()
self.unconditional_commit()
def clear(self):
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = None
self.plot_horlabels = []
self.plot_horlines = []
self.plot.clear()
def get_model(self):
if self.rpca is None:
return
if self.rpca.ready():
self.__timer.stop()
self.start_button.setText("Restart (finished)")
self._pca = self.rpca.get_state()
if self._pca is None:
return
self._variance_ratio = self._pca.explained_variance_ratio_
self._cumulative = numpy.cumsum(self._variance_ratio)
self._setup_plot()
self._transformed = None
self.commit()
def _setup_plot(self):
self.plot.clear()
explained_ratio = self._variance_ratio
explained = self._cumulative
p = min(len(self._variance_ratio), self.maxp)
self.plot.plot(numpy.arange(p),
explained_ratio[:p],
pen=pg.mkPen(QColor(Qt.red), width=2),
antialias=True,
name="Variance")
self.plot.plot(numpy.arange(p),
explained[:p],
pen=pg.mkPen(QColor(Qt.darkYellow), width=2),
antialias=True,
name="Cumulative Variance")
cutpos = self._nselected_components() - 1
self._line = pg.InfiniteLine(angle=90,
pos=cutpos,
movable=True,
bounds=(0, p - 1))
self._line.setCursor(Qt.SizeHorCursor)
self._line.setPen(pg.mkPen(QColor(Qt.black), width=2))
self._line.sigPositionChanged.connect(self._on_cut_changed)
self.plot.addItem(self._line)
self.plot_horlines = (
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)),
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)))
self.plot_horlabels = (pg.TextItem(color=QColor(Qt.black),
anchor=(1, 0)),
pg.TextItem(color=QColor(Qt.black),
anchor=(1, 1)))
for item in self.plot_horlabels + self.plot_horlines:
self.plot.addItem(item)
self._set_horline_pos()
self.plot.setRange(xRange=(0.0, p - 1), yRange=(0.0, 1.0))
self._update_axis()
def _set_horline_pos(self):
cutidx = self.ncomponents - 1
for line, label, curve in zip(
self.plot_horlines, self.plot_horlabels,
(self._variance_ratio, self._cumulative)):
y = curve[cutidx]
line.setData([-1, cutidx], 2 * [y])
label.setPos(cutidx, y)
label.setPlainText("{:.3f}".format(y))
def _on_cut_changed(self, line):
# cut changed by means of a cut line over the scree plot.
value = int(round(line.value()))
self._line.setValue(value)
current = self._nselected_components()
components = value + 1
if not (self.ncomponents == 0
and components == len(self._variance_ratio)):
self.ncomponents = components
self._set_horline_pos()
if self._pca is not None:
self.variance_covered = self._cumulative[components - 1] * 100
if current != self._nselected_components():
self._invalidate_selection()
def _update_selection_component_spin(self):
# cut changed by "ncomponents" spin.
if self._pca is None:
self._invalidate_selection()
return
if self.ncomponents == 0:
# Special "All" value
cut = len(self._variance_ratio)
else:
cut = self.ncomponents
self.variance_covered = self._cumulative[cut - 1] * 100
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_selection_variance_spin(self):
# cut changed by "max variance" spin.
if self._pca is None:
return
cut = numpy.searchsorted(self._cumulative,
self.variance_covered / 100.0) + 1
self.ncomponents = cut
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_normalize(self):
if self.normalize:
pp = self._pca_preprocessors + [Normalize()]
else:
pp = self._pca_preprocessors
self._pca_projector.preprocessors = pp
self.fit()
if self.data is None:
self._invalidate_selection()
def _nselected_components(self):
"""Return the number of selected components."""
if self._pca is None:
return 0
if self.ncomponents == 0:
# Special "All" value
max_comp = len(self._variance_ratio)
else:
max_comp = self.ncomponents
var_max = self._cumulative[max_comp - 1]
if var_max != numpy.floor(self.variance_covered / 100.0):
cut = max_comp
self.variance_covered = var_max * 100
else:
self.ncomponents = cut = numpy.searchsorted(
self._cumulative, self.variance_covered / 100.0) + 1
return cut
def _invalidate_selection(self):
self.commit()
def _update_axis(self):
p = min(len(self._variance_ratio), self.maxp)
axis = self.plot.getAxis("bottom")
d = max((p - 1) // (self.axis_labels - 1), 1)
axis.setTicks([[(i, str(i + 1)) for i in range(0, p, d)]])
def commit(self):
transformed = components = None
if self._pca is not None:
if self._transformed is None:
# Compute the full transform (all components) only once.
self._transformed = self._pca(self.data)
transformed = self._transformed
domain = Domain(transformed.domain.attributes[:self.ncomponents],
self.data.domain.class_vars,
self.data.domain.metas)
transformed = transformed.from_table(domain, transformed)
dom = Domain(self._pca.orig_domain.attributes,
metas=[StringVariable(name='component')])
metas = numpy.array(
[['PC{}'.format(i + 1) for i in range(self.ncomponents)]],
dtype=object).T
components = Table(dom,
self._pca.components_[:self.ncomponents],
metas=metas)
components.name = 'components'
self._pca_projector.component = self.ncomponents
self.send("Transformed data", transformed)
self.send("Components", components)
self.send("PCA", self._pca_projector)
def send_report(self):
if self.data is None:
return
self.report_items(
(("Selected components", self.ncomponents),
("Explained variance", "{:.3f} %".format(self.variance_covered))))
self.report_plot()
class OWMarkerGenes(widget.OWWidget):
name = "Marker Genes"
icon = 'icons/OWMarkerGenes.svg'
priority = 130
replaces = ['orangecontrib.single_cell.widgets.owmarkergenes.OWMarkerGenes']
class Warning(widget.OWWidget.Warning):
using_local_files = widget.Msg("Can't connect to serverfiles. Using cached files.")
class Outputs:
genes = widget.Output("Genes", Table)
want_main_area = True
want_control_area = True
auto_commit = Setting(True)
selected_source = Setting("")
selected_organism = Setting("")
selected_root_attribute = Setting(0)
settingsHandler = MarkerGroupContextHandler() # noqa: N815
selected_genes = settings.ContextSetting([])
settings_version = 2
_data = None
_available_sources = None
def __init__(self) -> None:
super().__init__()
# define the layout
main_area = QWidget(self.mainArea)
self.mainArea.layout().addWidget(main_area)
layout = QGridLayout()
main_area.setLayout(layout)
layout.setContentsMargins(4, 4, 4, 4)
# filter line edit
self.filter_line_edit = QLineEdit()
self.filter_line_edit.setPlaceholderText("Filter marker genes")
layout.addWidget(self.filter_line_edit, 0, 0, 1, 3)
# define available markers view
self.available_markers_view = TreeView()
box = gui.vBox(self.mainArea, "Available markers", addToLayout=False)
box.layout().addWidget(self.available_markers_view)
layout.addWidget(box, 1, 0, 2, 1)
# create selected markers view
self.selected_markers_view = TreeView()
box = gui.vBox(self.mainArea, "Selected markers", addToLayout=False)
box.layout().addWidget(self.selected_markers_view)
layout.addWidget(box, 1, 2, 2, 1)
self.available_markers_view.otherView = self.selected_markers_view
self.selected_markers_view.otherView = self.available_markers_view
# buttons
box = gui.vBox(self.mainArea, addToLayout=False, margin=0)
layout.addWidget(box, 1, 1, 1, 1)
self.move_button = gui.button(box, self, ">", callback=self._move_selected)
self._init_description_area(layout)
self._init_control_area()
self._load_data()
def _init_description_area(self, layout: QLayout) -> None:
"""
Function define an info area with description of the genes and add it to the layout.
"""
box = gui.widgetBox(self.mainArea, "Description", addToLayout=False)
self.descriptionlabel = QTextBrowser(
openExternalLinks=True, textInteractionFlags=(Qt.TextSelectableByMouse | Qt.LinksAccessibleByMouse)
)
box.setMaximumHeight(self.descriptionlabel.fontMetrics().height() * (NUM_LINES_TEXT + 3))
# description filed
self.descriptionlabel.setText("Select a gene to see information.")
self.descriptionlabel.setFrameStyle(QTextBrowser.NoFrame)
# no (white) text background
self.descriptionlabel.viewport().setAutoFillBackground(False)
box.layout().addWidget(self.descriptionlabel)
layout.addWidget(box, 3, 0, 1, 3)
def _init_control_area(self) -> None:
"""
Function defines dropdowns and the button in the control area.
"""
box = gui.widgetBox(self.controlArea, 'Database', margin=0)
self.source_index = -1
self.db_source_cb = gui.comboBox(box, self, 'source_index')
self.db_source_cb.activated[int].connect(self._set_db_source_index)
box = gui.widgetBox(self.controlArea, 'Organism', margin=0)
self.organism_index = -1
self.group_cb = gui.comboBox(box, self, 'organism_index')
self.group_cb.activated[int].connect(self._set_group_index)
box = gui.widgetBox(self.controlArea, 'Group by', margin=0)
self.group_by_cb = gui.comboBox(
box, self, 'selected_root_attribute', items=GROUP_BY_ITEMS, callback=self._setup
)
gui.rubber(self.controlArea)
gui.auto_commit(self.controlArea, self, "auto_commit", "Commit")
def sizeHint(self):
return super().sizeHint().expandedTo(QSize(900, 500))
@property
def available_sources(self) -> dict:
return self._available_sources
@available_sources.setter
def available_sources(self, value: dict) -> None:
"""
Set _available_sources variable, add them to dropdown, and select the source that was previously.
"""
self._available_sources = value
items = sorted(list(value.keys()), reverse=True) # panglao first
try:
idx = items.index(self.selected_source)
except ValueError:
idx = -1
self.db_source_cb.clear()
self.db_source_cb.addItems(items)
if idx != -1:
self.source_index = idx
self.selected_source = items[idx]
elif items:
self.source_index = min(max(self.source_index, 0), len(items) - 1)
self._set_db_source_index(self.source_index)
@property
def data(self) -> Table:
return self._data
@data.setter
def data(self, value: Table):
"""
Set the source data. The data is then filtered on the first meta column (group).
Select set dropdown with the groups and select the one that was selected previously.
"""
self._data = value
domain = value.domain
if domain.metas:
group = domain.metas[0]
groupcol, _ = value.get_column_view(group)
if group.is_string:
group_values = list(set(groupcol))
elif group.is_discrete:
group_values = group.values
else:
raise TypeError("Invalid column type")
group_values = sorted(group_values) # human first
try:
idx = group_values.index(self.selected_organism)
except ValueError:
idx = -1
self.group_cb.clear()
self.group_cb.addItems(group_values)
if idx != -1:
self.organism_index = idx
self.selected_organism = group_values[idx]
elif group_values:
self.organism_index = min(max(self.organism_index, 0), len(group_values) - 1)
self._set_group_index(self.organism_index)
def _load_data(self) -> None:
"""
Collect available data sources (marker genes data sets).
"""
self.Warning.using_local_files.clear()
found_sources = {}
try:
found_sources.update(serverfiles.ServerFiles().allinfo(SERVER_FILES_DOMAIN))
except requests.exceptions.ConnectionError:
found_sources.update(serverfiles.allinfo(SERVER_FILES_DOMAIN))
self.Warning.using_local_files()
self.available_sources = {item.get('title').split(': ')[-1]: item for item in found_sources.values()}
def _source_changed(self) -> None:
"""
Respond on change of the source and download the data.
"""
if self.available_sources:
file_name = self.available_sources[self.selected_source]['filename']
try:
serverfiles.update(SERVER_FILES_DOMAIN, file_name)
except requests.exceptions.ConnectionError:
# try to update file. Ignore network errors.
pass
try:
file_path = serverfiles.localpath_download(SERVER_FILES_DOMAIN, file_name)
except requests.exceptions.ConnectionError as err:
# Unexpected error.
raise err
self.data = Table.from_file(file_path)
def _setup(self) -> None:
"""
Setup the views with data.
"""
self.closeContext()
self.selected_genes = []
self.openContext((self.selected_organism, self.selected_source))
data_not_selected, data_selected = self._filter_data_group(self.data)
# add model to available markers view
group_by = GROUP_BY_ITEMS[self.selected_root_attribute]
tree_model = TreeModel(data_not_selected, group_by)
proxy_model = FilterProxyModel(self.filter_line_edit)
proxy_model.setSourceModel(tree_model)
self.available_markers_view.setModel(proxy_model)
self.available_markers_view.selectionModel().selectionChanged.connect(
partial(self._on_selection_changed, self.available_markers_view)
)
tree_model = TreeModel(data_selected, group_by)
proxy_model = FilterProxyModel(self.filter_line_edit)
proxy_model.setSourceModel(tree_model)
self.selected_markers_view.setModel(proxy_model)
self.selected_markers_view.selectionModel().selectionChanged.connect(
partial(self._on_selection_changed, self.selected_markers_view)
)
self.selected_markers_view.model().sourceModel().data_added.connect(self._selected_markers_changed)
self.selected_markers_view.model().sourceModel().data_removed.connect(self._selected_markers_changed)
# update output and messages
self._selected_markers_changed()
def _filter_data_group(self, data: Table) -> Tuple[Table, Tuple]:
"""
Function filter the table based on the selected group (Mouse, Human) and divide them in two groups based on
selected_data variable.
Parameters
----------
data
Table to be filtered
Returns
-------
data_not_selected
Data that will initially be in available markers view.
data_selected
Data that will initially be in selected markers view.
"""
group = data.domain.metas[0]
gvec = data.get_column_view(group)[0]
if group.is_string:
mask = gvec == self.selected_organism
else:
mask = gvec == self.organism_index
data = data[mask]
# divide data based on selected_genes variable (context)
unique_gene_names = np.core.defchararray.add(
data.get_column_view("Entrez ID")[0].astype(str), data.get_column_view("Cell Type")[0].astype(str)
)
mask = np.isin(unique_gene_names, self.selected_genes)
data_not_selected = data[~mask]
data_selected = data[mask]
return data_not_selected, data_selected
def commit(self) -> None:
rows = self.selected_markers_view.model().sourceModel().rootItem.get_data_rows()
if len(rows) > 0:
metas = [r.metas for r in rows]
data = Table.from_numpy(self.data.domain, np.empty((len(metas), 0)), metas=np.array(metas))
# always false for marker genes data tables in single cell
data.attributes[GENE_AS_ATTRIBUTE_NAME] = False
# set taxonomy id in data.attributes
data.attributes[TAX_ID] = MAP_GROUP_TO_TAX_ID.get(self.selected_organism, '')
# set column id flag
data.attributes[GENE_ID_COLUMN] = "Entrez ID"
data.name = 'Marker Genes'
else:
data = None
self.Outputs.genes.send(data)
def _update_description(self, view: TreeView) -> None:
"""
Upate the description about the gene. Only in case when one gene is selected.
"""
selection = self._selected_rows(view)
qmodel = view.model().sourceModel()
if len(selection) > 1 or len(selection) == 0 or qmodel.node_from_index(selection[0]).data_row is None:
self.descriptionlabel.setText("Select a gene to see information.")
else:
data_row = qmodel.node_from_index(selection[0]).data_row
self.descriptionlabel.setHtml(
f"<b>Gene name:</b> {data_row['Name']}<br/>"
f"<b>Entrez ID:</b> {data_row['Entrez ID']}<br/>"
f"<b>Cell Type:</b> {data_row['Cell Type']}<br/>"
f"<b>Function:</b> {data_row['Function']}<br/>"
f"<b>Reference:</b> <a href='{data_row['URL']}'>{data_row['Reference']}</a>"
)
def _update_data_info(self) -> None:
"""
Updates output info in the control area.
"""
sel_model = self.selected_markers_view.model().sourceModel()
self.info.set_output_summary(f"Selected: {str(len(sel_model))}")
# callback functions
def _selected_markers_changed(self) -> None:
"""
This function is called when markers in the selected view are added or removed.
"""
rows = self.selected_markers_view.model().sourceModel().rootItem.get_data_rows()
self.selected_genes = [row["Entrez ID"].value + row["Cell Type"].value for row in rows]
self._update_data_info()
self.commit()
def _on_selection_changed(self, view: TreeView) -> None:
"""
When selection in one of the view changes in a view button should change a sign in the correct direction and
other view should reset the selection. Also gene description is updated.
"""
self.move_button.setText(">" if view is self.available_markers_view else "<")
if view is self.available_markers_view:
self.selected_markers_view.clearSelection()
else:
self.available_markers_view.clearSelection()
self._update_description(view)
def _set_db_source_index(self, source_index: int) -> None:
"""
Set the index of selected database source - index in a dropdown.
"""
self.source_index = source_index
self.selected_source = self.db_source_cb.itemText(source_index)
self._source_changed()
def _set_group_index(self, group_index: int) -> None:
"""
Set the index of organism - index in a dropdown.
"""
self.organism_index = group_index
self.selected_organism = self.group_cb.itemText(group_index)
self._setup()
def _move_selected(self) -> None:
"""
Move selected genes when button clicked.
"""
if self._selected_rows(self.selected_markers_view):
self._move_selected_from_to(self.selected_markers_view, self.available_markers_view)
elif self._selected_rows(self.available_markers_view):
self._move_selected_from_to(self.available_markers_view, self.selected_markers_view)
# support functions for callbacks
def _move_selected_from_to(self, src: TreeView, dst: TreeView) -> None:
"""
Function moves items from src model to dst model.
"""
selected_items = self._selected_rows(src)
src_model = src.model().sourceModel()
dst_model = dst.model().sourceModel()
# move data as mimeData from source to destination tree view
mime_data = src_model.mimeData(selected_items)
# remove nodes from the source view
src_model.remove_node_list(selected_items)
dst_model.dropMimeData(mime_data, Qt.MoveAction, -1, -1)
@staticmethod
def _selected_rows(view: TreeView) -> List[QModelIndex]:
"""
Return the selected rows in the view.
"""
rows = view.selectionModel().selectedRows()
return list(map(view.model().mapToSource, rows))
@classmethod
def migrate_settings(cls, settings, version=0):
def migrate_to_version_2():
settings["selected_source"] = settings.pop("selected_db_source", "")
settings["selected_organism"] = settings.pop("selected_group", "")
if "context_settings" in settings:
for co in settings["context_settings"]:
co.values["selected_genes"] = [g[0] + g[1] for g in co.values["selected_genes"]]
if version < 2:
migrate_to_version_2()
class ComboBoxSearch(QComboBox):
"""
A drop down list combo box with filter/search.
The popup list view is filtered by text entered in the filter field.
Note
----
`popup`, `lineEdit` and `completer` from the base QComboBox class are
unused. Setting/modifying them will have no effect.
"""
# NOTE: Setting editable + QComboBox.NoInsert policy + ... did not achieve
# the same results.
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.__searchline = QLineEdit(self, visible=False, frame=False)
self.__searchline.setAttribute(Qt.WA_MacShowFocusRect, False)
self.__searchline.setFocusProxy(self)
self.__popup = None # type: Optional[QAbstractItemModel]
self.__proxy = None # type: Optional[QSortFilterProxyModel]
self.__popupTimer = QElapsedTimer()
self.setFocusPolicy(Qt.ClickFocus | Qt.TabFocus)
def showPopup(self):
# type: () -> None
"""
Reimplemented from QComboBox.showPopup
Popup up a customized view and filter edit line.
Note
----
The .popup(), .lineEdit(), .completer() of the base class are not used.
"""
if self.__popup is not None:
# We have user entered state that cannot be disturbed
# (entered filter text, scroll offset, ...)
return # pragma: no cover
if self.count() == 0:
return
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
popup = QListView(
uniformItemSizes=True,
horizontalScrollBarPolicy=Qt.ScrollBarAlwaysOff,
verticalScrollBarPolicy=Qt.ScrollBarAsNeeded,
iconSize=self.iconSize(),
)
popup.setFocusProxy(self.__searchline)
popup.setParent(self, Qt.Popup | Qt.FramelessWindowHint)
popup.setItemDelegate(_ComboBoxListDelegate(popup))
proxy = QSortFilterProxyModel(
popup, filterCaseSensitivity=Qt.CaseInsensitive
)
proxy.setFilterKeyColumn(self.modelColumn())
proxy.setSourceModel(self.model())
popup.setModel(proxy)
root = proxy.mapFromSource(self.rootModelIndex())
popup.setRootIndex(root)
self.__popup = popup
self.__proxy = proxy
self.__searchline.setText("")
self.__searchline.setPlaceholderText("Filter...")
self.__searchline.setVisible(True)
self.__searchline.textEdited.connect(proxy.setFilterFixedString)
style = self.style() # type: QStyle
popuprect_origin = style.subControlRect(
QStyle.CC_ComboBox, opt, QStyle.SC_ComboBoxListBoxPopup, self
) # type: QRect
popuprect_origin = QRect(
self.mapToGlobal(popuprect_origin.topLeft()),
popuprect_origin.size()
)
editrect = style.subControlRect(
QStyle.CC_ComboBox, opt, QStyle.SC_ComboBoxEditField, self
) # type: QRect
self.__searchline.setGeometry(editrect)
desktop = QApplication.desktop()
screenrect = desktop.availableGeometry(self) # type: QRect
# get the height for the view
listrect = QRect()
for i in range(min(proxy.rowCount(root), self.maxVisibleItems())):
index = proxy.index(i, self.modelColumn(), root)
if index.isValid():
listrect = listrect.united(popup.visualRect(index))
if listrect.height() >= screenrect.height():
break
window = popup.window() # type: QWidget
window.ensurePolished()
if window.layout() is not None:
window.layout().activate()
else:
QApplication.sendEvent(window, QEvent(QEvent.LayoutRequest))
margins = qwidget_margin_within(popup.viewport(), window)
height = (listrect.height() + 2 * popup.spacing() +
margins.top() + margins.bottom())
popup_size = (QSize(popuprect_origin.width(), height)
.expandedTo(window.minimumSize())
.boundedTo(window.maximumSize())
.boundedTo(screenrect.size()))
popuprect = QRect(popuprect_origin.bottomLeft(), popup_size)
popuprect = dropdown_popup_geometry(
popuprect, popuprect_origin, screenrect)
popup.setGeometry(popuprect)
current = proxy.mapFromSource(
self.model().index(self.currentIndex(), self.modelColumn(),
self.rootModelIndex()))
popup.setCurrentIndex(current)
popup.scrollTo(current, QAbstractItemView.EnsureVisible)
popup.show()
popup.setFocus(Qt.PopupFocusReason)
popup.installEventFilter(self)
popup.viewport().installEventFilter(self)
popup.viewport().setMouseTracking(True)
self.update()
self.__popupTimer.restart()
def hidePopup(self):
"""Reimplemented"""
if self.__popup is not None:
popup = self.__popup
self.__popup = self.__proxy = None
popup.setFocusProxy(None)
popup.hide()
popup.deleteLater()
popup.removeEventFilter(self)
popup.viewport().removeEventFilter(self)
# need to call base hidePopup even though the base showPopup was not
# called (update internal state wrt. 'pressed' arrow, ...)
super().hidePopup()
self.__searchline.hide()
self.update()
def initStyleOption(self, option):
# type: (QStyleOptionComboBox) -> None
super().initStyleOption(option)
option.editable = True
def __updateGeometries(self):
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
editarea = self.style().subControlRect(
QStyle.CC_ComboBox, opt, QStyle.SC_ComboBoxEditField, self)
self.__searchline.setGeometry(editarea)
def resizeEvent(self, event):
"""Reimplemented."""
super().resizeEvent(event)
self.__updateGeometries()
def paintEvent(self, event):
"""Reimplemented."""
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
painter = QStylePainter(self)
painter.drawComplexControl(QStyle.CC_ComboBox, opt)
if not self.__searchline.isVisibleTo(self):
opt.editable = False
painter.drawControl(QStyle.CE_ComboBoxLabel, opt)
def eventFilter(self, obj, event): # pylint: disable=too-many-branches
# type: (QObject, QEvent) -> bool
"""Reimplemented."""
etype = event.type()
if etype == QEvent.FocusOut and self.__popup is not None:
self.hidePopup()
return True
if etype == QEvent.Hide and self.__popup is not None:
self.hidePopup()
return False
if etype == QEvent.KeyPress or etype == QEvent.KeyRelease or \
etype == QEvent.ShortcutOverride and obj is self.__popup:
event = event # type: QKeyEvent
key, modifiers = event.key(), event.modifiers()
if key in (Qt.Key_Enter, Qt.Key_Return, Qt.Key_Select):
current = self.__popup.currentIndex()
if current.isValid():
self.__activateProxyIndex(current)
elif key in (Qt.Key_Up, Qt.Key_Down,
Qt.Key_PageUp, Qt.Key_PageDown):
return False #
elif key in (Qt.Key_Tab, Qt.Key_Backtab):
pass
elif key == Qt.Key_Escape or \
(key == Qt.Key_F4 and modifiers & Qt.AltModifier):
self.__popup.hide()
return True
else:
# pass the input events to the filter edit line (no propagation
# up the parent chain).
self.__searchline.event(event)
if event.isAccepted():
return True
if etype == QEvent.MouseButtonRelease and self.__popup is not None \
and obj is self.__popup.viewport() \
and self.__popupTimer.elapsed() >= \
QApplication.doubleClickInterval():
event = event # type: QMouseEvent
index = self.__popup.indexAt(event.pos())
if index.isValid():
self.__activateProxyIndex(index)
if etype == QEvent.MouseMove and self.__popup is not None \
and obj is self.__popup.viewport():
event = event # type: QMouseEvent
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
style = self.style() # type: QStyle
if style.styleHint(QStyle.SH_ComboBox_ListMouseTracking, opt, self):
index = self.__popup.indexAt(event.pos())
if index.isValid() and \
index.flags() & (Qt.ItemIsEnabled | Qt.ItemIsSelectable):
self.__popup.setCurrentIndex(index)
if etype == QEvent.MouseButtonPress and self.__popup is obj:
# Popup border or out of window mouse button press/release.
# At least on windows this needs to be handled.
style = self.style()
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
opt.subControls = QStyle.SC_All
opt.activeSubControls = QStyle.SC_ComboBoxArrow
pos = self.mapFromGlobal(event.globalPos())
sc = style.hitTestComplexControl(QStyle.CC_ComboBox, opt, pos, self)
if sc != QStyle.SC_None:
self.__popup.setAttribute(Qt.WA_NoMouseReplay)
self.hidePopup()
return super().eventFilter(obj, event)
def __activateProxyIndex(self, index):
# type: (QModelIndex) -> None
# Set current and activate the source index corresponding to the proxy
# index in the popup's model.
if self.__popup is not None and index.isValid():
proxy = self.__popup.model()
assert index.model() is proxy
index = proxy.mapToSource(index)
assert index.model() is self.model()
if index.isValid() and \
index.flags() & (Qt.ItemIsEnabled | Qt.ItemIsSelectable):
self.hidePopup()
text = self.itemText(index.row())
self.setCurrentIndex(index.row())
self.activated[int].emit(index.row())
self.activated[str].emit(text)
class OWBaseSql(OWWidget, openclass=True):
"""Base widget for connecting to a database.
Override `get_backend` when subclassing to get corresponding backend.
"""
class Outputs:
data = Output("Data", Table)
class Error(OWWidget.Error):
connection = Msg("{}")
want_main_area = False
resizing_enabled = False
host = Setting(None) # type: Optional[str]
port = Setting(None) # type: Optional[str]
database = Setting(None) # type: Optional[str]
schema = Setting(None) # type: Optional[str]
username = ""
password = ""
def __init__(self):
super().__init__()
self.backend = None # type: Optional[Backend]
self.data_desc_table = None # type: Optional[Table]
self.database_desc = None # type: Optional[OrderedDict]
self._setup_gui()
self.connect()
def _setup_gui(self):
self.controlArea.setMinimumWidth(360)
vbox = gui.vBox(self.controlArea, "Server", addSpace=True)
self.serverbox = gui.vBox(vbox)
self.servertext = QLineEdit(self.serverbox)
self.servertext.setPlaceholderText("Server")
self.servertext.setToolTip("Server")
self.servertext.editingFinished.connect(self._load_credentials)
if self.host:
self.servertext.setText(self.host if not self.port else
"{}:{}".format(self.host, self.port))
self.serverbox.layout().addWidget(self.servertext)
self.databasetext = QLineEdit(self.serverbox)
self.databasetext.setPlaceholderText("Database[/Schema]")
self.databasetext.setToolTip("Database or optionally Database/Schema")
if self.database:
self.databasetext.setText(
self.database if not self.schema else
"{}/{}".format(self.database, self.schema))
self.serverbox.layout().addWidget(self.databasetext)
self.usernametext = QLineEdit(self.serverbox)
self.usernametext.setPlaceholderText("Username")
self.usernametext.setToolTip("Username")
self.serverbox.layout().addWidget(self.usernametext)
self.passwordtext = QLineEdit(self.serverbox)
self.passwordtext.setPlaceholderText("Password")
self.passwordtext.setToolTip("Password")
self.passwordtext.setEchoMode(QLineEdit.Password)
self.serverbox.layout().addWidget(self.passwordtext)
self._load_credentials()
self.connectbutton = gui.button(self.serverbox, self, "Connect",
callback=self.connect)
self.connectbutton.setSizePolicy(QSizePolicy.Fixed, QSizePolicy.Fixed)
def _load_credentials(self):
self._parse_host_port()
cm = self._credential_manager(self.host, self.port)
self.username = cm.username
self.password = cm.password
if self.username:
self.usernametext.setText(self.username)
if self.password:
self.passwordtext.setText(self.password)
def _save_credentials(self):
cm = self._credential_manager(self.host, self.port)
cm.username = self.username or ""
cm.password = self.password or ""
@staticmethod
def _credential_manager(host, port):
return CredentialManager("SQL Table: {}:{}".format(host, port))
def _parse_host_port(self):
hostport = self.servertext.text().split(":")
self.host = hostport[0]
self.port = hostport[1] if len(hostport) == 2 else None
def _check_db_settings(self):
self._parse_host_port()
self.database, _, self.schema = self.databasetext.text().partition("/")
self.username = self.usernametext.text() or None
self.password = self.passwordtext.text() or None
def connect(self):
self.clear()
self._check_db_settings()
if not self.host or not self.database:
return
try:
backend = self.get_backend()
if backend is None:
return
self.backend = backend(dict(
host=self.host,
port=self.port,
database=self.database,
user=self.username,
password=self.password
))
self.on_connection_success()
except BackendError as err:
self.on_connection_error(err)
def get_backend(self) -> Type[Backend]:
"""
Derived widgets should override this to get corresponding backend.
Returns
-------
backend: Type[Backend]
"""
raise NotImplementedError
def on_connection_success(self):
self._save_credentials()
self.database_desc = OrderedDict((
("Host", self.host), ("Port", self.port),
("Database", self.database), ("User name", self.username)
))
def on_connection_error(self, err):
error = str(err).split("\n")[0]
self.Error.connection(error)
def open_table(self):
data = self.get_table()
self.data_desc_table = data
self.Outputs.data.send(data)
info = str(len(data)) if data else self.info.NoOutput
self.info.set_output_summary(info)
def get_table(self) -> Table:
"""
Derived widgets should override this to get corresponding table.
Returns
-------
table: Table
"""
raise NotImplementedError
def clear(self):
self.Error.connection.clear()
self.database_desc = None
self.data_desc_table = None
self.Outputs.data.send(None)
self.info.set_output_summary(self.info.NoOutput)
def send_report(self):
if not self.database_desc:
self.report_paragraph("No database connection.")
return
self.report_items("Database", self.database_desc)
if self.data_desc_table:
self.report_items("Data",
report.describe_data(self.data_desc_table))
class OWPCA(widget.OWWidget):
name = "PCA"
description = "Principal component analysis with a scree-diagram."
icon = "icons/PCA.svg"
priority = 3050
inputs = [("Data", Table, "set_data")]
outputs = [("Transformed data", Table),
("Components", Table),
("PCA", PCA)]
ncomponents = settings.Setting(2)
variance_covered = settings.Setting(100)
batch_size = settings.Setting(100)
address = settings.Setting('')
auto_update = settings.Setting(True)
auto_commit = settings.Setting(True)
normalize = settings.Setting(True)
maxp = settings.Setting(20)
axis_labels = settings.Setting(10)
graph_name = "plot.plotItem"
def __init__(self):
super().__init__()
self.data = None
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = False
self._pca_projector = PCA()
self._pca_projector.component = self.ncomponents
self._pca_preprocessors = PCA.preprocessors
# Components Selection
box = gui.vBox(self.controlArea, "Components Selection")
form = QFormLayout()
box.layout().addLayout(form)
self.components_spin = gui.spin(
box, self, "ncomponents", 0, 1000,
callback=self._update_selection_component_spin,
keyboardTracking=False
)
self.components_spin.setSpecialValueText("All")
self.variance_spin = gui.spin(
box, self, "variance_covered", 1, 100,
callback=self._update_selection_variance_spin,
keyboardTracking=False
)
self.variance_spin.setSuffix("%")
form.addRow("Components:", self.components_spin)
form.addRow("Variance covered:", self.variance_spin)
# Incremental learning
self.sampling_box = gui.vBox(self.controlArea, "Incremental learning")
self.addresstext = QLineEdit(box)
self.addresstext.setPlaceholderText('Remote server')
if self.address:
self.addresstext.setText(self.address)
self.sampling_box.layout().addWidget(self.addresstext)
form = QFormLayout()
self.sampling_box.layout().addLayout(form)
self.batch_spin = gui.spin(
self.sampling_box, self, "batch_size", 50, 100000, step=50,
keyboardTracking=False)
form.addRow("Batch size ~ ", self.batch_spin)
self.start_button = gui.button(
self.sampling_box, self, "Start remote computation",
callback=self.start, autoDefault=False,
tooltip="Start/abort computation on the server")
self.start_button.setEnabled(False)
gui.checkBox(self.sampling_box, self, "auto_update",
"Periodically fetch model", callback=self.update_model)
self.__timer = QTimer(self, interval=2000)
self.__timer.timeout.connect(self.get_model)
self.sampling_box.setVisible(remotely)
# Options
self.options_box = gui.vBox(self.controlArea, "Options")
gui.checkBox(self.options_box, self, "normalize", "Normalize data",
callback=self._update_normalize)
self.maxp_spin = gui.spin(
self.options_box, self, "maxp", 1, 100,
label="Show only first", callback=self._setup_plot,
keyboardTracking=False
)
self.controlArea.layout().addStretch()
gui.auto_commit(self.controlArea, self, "auto_commit", "Apply",
checkbox_label="Apply automatically")
self.plot = pg.PlotWidget(background="w")
axis = self.plot.getAxis("bottom")
axis.setLabel("Principal Components")
axis = self.plot.getAxis("left")
axis.setLabel("Proportion of variance")
self.plot_horlabels = []
self.plot_horlines = []
self.plot.getViewBox().setMenuEnabled(False)
self.plot.getViewBox().setMouseEnabled(False, False)
self.plot.showGrid(True, True, alpha=0.5)
self.plot.setRange(xRange=(0.0, 1.0), yRange=(0.0, 1.0))
self.mainArea.layout().addWidget(self.plot)
self._update_normalize()
def update_model(self):
self.get_model()
if self.auto_update and self.rpca and not self.rpca.ready():
self.__timer.start(2000)
else:
self.__timer.stop()
def start(self):
if 'Abort' in self.start_button.text():
self.rpca.abort()
self.__timer.stop()
self.start_button.setText("Start remote computation")
else:
self.address = self.addresstext.text()
with remote.server(self.address):
from Orange.projection.pca import RemotePCA
maxiter = (1e5 + self.data.approx_len()) / self.batch_size * 3
self.rpca = RemotePCA(self.data, self.batch_size, int(maxiter))
self.update_model()
self.start_button.setText("Abort remote computation")
def set_data(self, data):
self.information()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
elif not remotely:
self.information("Data has been sampled")
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
self.data = data
self.fit()
def fit(self):
self.clear()
self.start_button.setEnabled(False)
if self.data is None:
return
data = self.data
self._transformed = None
if isinstance(data, SqlTable): # data was big and remote available
self.sampling_box.setVisible(True)
self.start_button.setText("Start remote computation")
self.start_button.setEnabled(True)
else:
self.sampling_box.setVisible(False)
pca = self._pca_projector(data)
variance_ratio = pca.explained_variance_ratio_
cumulative = numpy.cumsum(variance_ratio)
self.components_spin.setRange(0, len(cumulative))
self._pca = pca
self._variance_ratio = variance_ratio
self._cumulative = cumulative
self._setup_plot()
self.unconditional_commit()
def clear(self):
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = None
self.plot_horlabels = []
self.plot_horlines = []
self.plot.clear()
def get_model(self):
if self.rpca is None:
return
if self.rpca.ready():
self.__timer.stop()
self.start_button.setText("Restart (finished)")
self._pca = self.rpca.get_state()
if self._pca is None:
return
self._variance_ratio = self._pca.explained_variance_ratio_
self._cumulative = numpy.cumsum(self._variance_ratio)
self._setup_plot()
self._transformed = None
self.commit()
def _setup_plot(self):
self.plot.clear()
explained_ratio = self._variance_ratio
explained = self._cumulative
p = min(len(self._variance_ratio), self.maxp)
self.plot.plot(numpy.arange(p), explained_ratio[:p],
pen=pg.mkPen(QColor(Qt.red), width=2),
antialias=True,
name="Variance")
self.plot.plot(numpy.arange(p), explained[:p],
pen=pg.mkPen(QColor(Qt.darkYellow), width=2),
antialias=True,
name="Cumulative Variance")
cutpos = self._nselected_components() - 1
self._line = pg.InfiniteLine(
angle=90, pos=cutpos, movable=True, bounds=(0, p - 1))
self._line.setCursor(Qt.SizeHorCursor)
self._line.setPen(pg.mkPen(QColor(Qt.black), width=2))
self._line.sigPositionChanged.connect(self._on_cut_changed)
self.plot.addItem(self._line)
self.plot_horlines = (
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)),
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)))
self.plot_horlabels = (
pg.TextItem(color=QColor(Qt.black), anchor=(1, 0)),
pg.TextItem(color=QColor(Qt.black), anchor=(1, 1)))
for item in self.plot_horlabels + self.plot_horlines:
self.plot.addItem(item)
self._set_horline_pos()
self.plot.setRange(xRange=(0.0, p - 1), yRange=(0.0, 1.0))
self._update_axis()
def _set_horline_pos(self):
cutidx = self.ncomponents - 1
for line, label, curve in zip(self.plot_horlines, self.plot_horlabels,
(self._variance_ratio, self._cumulative)):
y = curve[cutidx]
line.setData([-1, cutidx], 2 * [y])
label.setPos(cutidx, y)
label.setPlainText("{:.3f}".format(y))
def _on_cut_changed(self, line):
# cut changed by means of a cut line over the scree plot.
value = int(round(line.value()))
self._line.setValue(value)
current = self._nselected_components()
components = value + 1
if not (self.ncomponents == 0 and
components == len(self._variance_ratio)):
self.ncomponents = components
self._set_horline_pos()
if self._pca is not None:
self.variance_covered = self._cumulative[components - 1] * 100
if current != self._nselected_components():
self._invalidate_selection()
def _update_selection_component_spin(self):
# cut changed by "ncomponents" spin.
if self._pca is None:
self._invalidate_selection()
return
if self.ncomponents == 0:
# Special "All" value
cut = len(self._variance_ratio)
else:
cut = self.ncomponents
self.variance_covered = self._cumulative[cut - 1] * 100
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_selection_variance_spin(self):
# cut changed by "max variance" spin.
if self._pca is None:
return
cut = numpy.searchsorted(self._cumulative,
self.variance_covered / 100.0) + 1
cut = min(cut, len(self._cumulative))
self.ncomponents = cut
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_normalize(self):
if self.normalize:
pp = self._pca_preprocessors + [Normalize()]
else:
pp = self._pca_preprocessors
self._pca_projector.preprocessors = pp
self.fit()
if self.data is None:
self._invalidate_selection()
def _nselected_components(self):
"""Return the number of selected components."""
if self._pca is None:
return 0
if self.ncomponents == 0:
# Special "All" value
max_comp = len(self._variance_ratio)
else:
max_comp = self.ncomponents
var_max = self._cumulative[max_comp - 1]
if var_max != numpy.floor(self.variance_covered / 100.0):
cut = max_comp
self.variance_covered = var_max * 100
else:
self.ncomponents = cut = numpy.searchsorted(
self._cumulative, self.variance_covered / 100.0) + 1
return cut
def _invalidate_selection(self):
self.commit()
def _update_axis(self):
p = min(len(self._variance_ratio), self.maxp)
axis = self.plot.getAxis("bottom")
d = max((p-1)//(self.axis_labels-1), 1)
axis.setTicks([[(i, str(i+1)) for i in range(0, p, d)]])
def commit(self):
transformed = components = None
if self._pca is not None:
if self._transformed is None:
# Compute the full transform (all components) only once.
self._transformed = self._pca(self.data)
transformed = self._transformed
domain = Domain(
transformed.domain.attributes[:self.ncomponents],
self.data.domain.class_vars,
self.data.domain.metas
)
transformed = transformed.from_table(domain, transformed)
dom = Domain(self._pca.orig_domain.attributes,
metas=[StringVariable(name='component')])
metas = numpy.array([['PC{}'.format(i + 1)
for i in range(self.ncomponents)]],
dtype=object).T
components = Table(dom, self._pca.components_[:self.ncomponents],
metas=metas)
components.name = 'components'
self._pca_projector.component = self.ncomponents
self.send("Transformed data", transformed)
self.send("Components", components)
self.send("PCA", self._pca_projector)
def send_report(self):
if self.data is None:
return
self.report_items((
("Selected components", self.ncomponents),
("Explained variance", "{:.3f} %".format(self.variance_covered))
))
self.report_plot()
def variables_filter(model, parent=None, accepted_type=Orange.data.Variable,
view_type=VariablesListItemView):
"""
GUI components: ListView with a lineedit which works as a filter. One can write
a variable name in a edit box and possible matches are then shown in a listview.
"""
def update_completer_model():
""" This gets called when the model for available attributes changes
through either drag/drop or the left/right button actions.
"""
nonlocal original_completer_items
items = ["%s=%s" % item for v in model for item in v.attributes.items()]
new = sorted(set(items))
if new != original_completer_items:
original_completer_items = new
completer_model.setStringList(original_completer_items)
def update_completer_prefix():
""" Prefixes all items in the completer model with the current
already done completion to enable the completion of multiple keywords.
"""
nonlocal original_completer_items
prefix = str(completer.completionPrefix())
if not prefix.endswith(" ") and " " in prefix:
prefix, _ = prefix.rsplit(" ", 1)
items = [prefix + " " + item
for item in original_completer_items]
else:
items = original_completer_items
old = list(map(str, completer_model.stringList()))
if set(old) != set(items):
completer_model.setStringList(items)
original_completer_items = []
filter_edit = QLineEdit()
filter_edit.setToolTip("Filter the list of available variables.")
filter_edit.setPlaceholderText("Filter")
completer_model = QStringListModel()
completer = QCompleter(completer_model, filter_edit)
completer.setCompletionMode(QCompleter.InlineCompletion)
completer.setModelSorting(QCompleter.CaseSensitivelySortedModel)
filter_edit.setCompleter(completer)
completer_navigator = CompleterNavigator(parent)
filter_edit.installEventFilter(completer_navigator)
proxy = VariableFilterProxyModel()
proxy.setSourceModel(model)
view = view_type(acceptedType=accepted_type)
view.setModel(proxy)
model.dataChanged.connect(update_completer_model)
model.rowsInserted.connect(update_completer_model)
model.rowsRemoved.connect(update_completer_model)
filter_edit.textChanged.connect(update_completer_prefix)
filter_edit.textChanged.connect(proxy.set_filter_string)
return filter_edit, view
class OWPCA(widget.OWWidget):
name = "PCA"
description = "Principal component analysis with a scree-diagram."
icon = "icons/PCA.svg"
priority = 3050
class Inputs:
data = Input("Data", Table)
class Outputs:
transformed_data = Output("Transformed data", Table)
components = Output("Components", Table)
pca = Output("PCA", PCA, dynamic=False)
settingsHandler = settings.DomainContextHandler()
ncomponents = settings.Setting(2)
variance_covered = settings.Setting(100)
batch_size = settings.Setting(100)
address = settings.Setting('')
auto_update = settings.Setting(True)
auto_commit = settings.Setting(True)
normalize = settings.ContextSetting(True)
decomposition_idx = settings.ContextSetting(0)
maxp = settings.Setting(20)
axis_labels = settings.Setting(10)
graph_name = "plot.plotItem"
class Warning(widget.OWWidget.Warning):
trivial_components = widget.Msg(
"All components of the PCA are trivial (explain 0 variance). "
"Input data is constant (or near constant).")
class Error(widget.OWWidget.Error):
no_features = widget.Msg("At least 1 feature is required")
no_instances = widget.Msg("At least 1 data instance is required")
sparse_data = widget.Msg("Sparse data is not supported")
def __init__(self):
super().__init__()
self.data = None
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = False
self._init_projector()
# Components Selection
box = gui.vBox(self.controlArea, "Components Selection")
form = QFormLayout()
box.layout().addLayout(form)
self.components_spin = gui.spin(
box,
self,
"ncomponents",
1,
MAX_COMPONENTS,
callback=self._update_selection_component_spin,
keyboardTracking=False)
self.components_spin.setSpecialValueText("All")
self.variance_spin = gui.spin(
box,
self,
"variance_covered",
1,
100,
callback=self._update_selection_variance_spin,
keyboardTracking=False)
self.variance_spin.setSuffix("%")
form.addRow("Components:", self.components_spin)
form.addRow("Variance covered:", self.variance_spin)
# Incremental learning
self.sampling_box = gui.vBox(self.controlArea, "Incremental learning")
self.addresstext = QLineEdit(box)
self.addresstext.setPlaceholderText('Remote server')
if self.address:
self.addresstext.setText(self.address)
self.sampling_box.layout().addWidget(self.addresstext)
form = QFormLayout()
self.sampling_box.layout().addLayout(form)
self.batch_spin = gui.spin(self.sampling_box,
self,
"batch_size",
50,
100000,
step=50,
keyboardTracking=False)
form.addRow("Batch size ~ ", self.batch_spin)
self.start_button = gui.button(
self.sampling_box,
self,
"Start remote computation",
callback=self.start,
autoDefault=False,
tooltip="Start/abort computation on the server")
self.start_button.setEnabled(False)
gui.checkBox(self.sampling_box,
self,
"auto_update",
"Periodically fetch model",
callback=self.update_model)
self.__timer = QTimer(self, interval=2000)
self.__timer.timeout.connect(self.get_model)
self.sampling_box.setVisible(remotely)
# Decomposition
self.decomposition_box = gui.radioButtons(
self.controlArea,
self,
"decomposition_idx", [d.name for d in DECOMPOSITIONS],
box="Decomposition",
callback=self._update_decomposition)
# Options
self.options_box = gui.vBox(self.controlArea, "Options")
self.normalize_box = gui.checkBox(self.options_box,
self,
"normalize",
"Normalize data",
callback=self._update_normalize)
self.maxp_spin = gui.spin(self.options_box,
self,
"maxp",
1,
MAX_COMPONENTS,
label="Show only first",
callback=self._setup_plot,
keyboardTracking=False)
self.controlArea.layout().addStretch()
gui.auto_commit(self.controlArea,
self,
"auto_commit",
"Apply",
checkbox_label="Apply automatically")
self.plot = pg.PlotWidget(background="w")
axis = self.plot.getAxis("bottom")
axis.setLabel("Principal Components")
axis = self.plot.getAxis("left")
axis.setLabel("Proportion of variance")
self.plot_horlabels = []
self.plot_horlines = []
self.plot.getViewBox().setMenuEnabled(False)
self.plot.getViewBox().setMouseEnabled(False, False)
self.plot.showGrid(True, True, alpha=0.5)
self.plot.setRange(xRange=(0.0, 1.0), yRange=(0.0, 1.0))
self.mainArea.layout().addWidget(self.plot)
self._update_normalize()
def update_model(self):
self.get_model()
if self.auto_update and self.rpca and not self.rpca.ready():
self.__timer.start(2000)
else:
self.__timer.stop()
def update_buttons(self, sparse_data=False):
if sparse_data:
self.normalize = False
buttons = self.decomposition_box.buttons
for cls, button in zip(DECOMPOSITIONS, buttons):
button.setDisabled(sparse_data and not cls.supports_sparse)
if not buttons[self.decomposition_idx].isEnabled():
# Set decomposition index to first sparse-enabled decomposition
for i, cls in enumerate(DECOMPOSITIONS):
if cls.supports_sparse:
self.decomposition_idx = i
break
self._init_projector()
def start(self):
if 'Abort' in self.start_button.text():
self.rpca.abort()
self.__timer.stop()
self.start_button.setText("Start remote computation")
else:
self.address = self.addresstext.text()
with remote.server(self.address):
from Orange.projection.pca import RemotePCA
maxiter = (1e5 + self.data.approx_len()) / self.batch_size * 3
self.rpca = RemotePCA(self.data, self.batch_size, int(maxiter))
self.update_model()
self.start_button.setText("Abort remote computation")
@Inputs.data
def set_data(self, data):
self.closeContext()
self.clear_messages()
self.clear()
self.start_button.setEnabled(False)
self.information()
self.data = None
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
elif not remotely:
self.information("Data has been sampled")
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(2000, partial=True)
data = Table(data_sample)
else: # data was big and remote available
self.sampling_box.setVisible(True)
self.start_button.setText("Start remote computation")
self.start_button.setEnabled(True)
if not isinstance(data, SqlTable):
self.sampling_box.setVisible(False)
if isinstance(data, Table):
if len(data.domain.attributes) == 0:
self.Error.no_features()
self.clear_outputs()
return
if len(data) == 0:
self.Error.no_instances()
self.clear_outputs()
return
self.openContext(data)
sparse_data = data is not None and data.is_sparse()
self.normalize_box.setDisabled(sparse_data)
self.update_buttons(sparse_data=sparse_data)
self.data = data
self.fit()
def fit(self):
self.clear()
self.Warning.trivial_components.clear()
if self.data is None:
return
data = self.data
if not isinstance(data, SqlTable):
pca = self._pca_projector(data)
variance_ratio = pca.explained_variance_ratio_
cumulative = numpy.cumsum(variance_ratio)
if numpy.isfinite(cumulative[-1]):
self.components_spin.setRange(0, len(cumulative))
self._pca = pca
self._variance_ratio = variance_ratio
self._cumulative = cumulative
self._setup_plot()
else:
self.Warning.trivial_components()
self.unconditional_commit()
def clear(self):
self._pca = None
self._transformed = None
self._variance_ratio = None
self._cumulative = None
self._line = None
self.plot_horlabels = []
self.plot_horlines = []
self.plot.clear()
def clear_outputs(self):
self.Outputs.transformed_data.send(None)
self.Outputs.components.send(None)
self.Outputs.pca.send(self._pca_projector)
def get_model(self):
if self.rpca is None:
return
if self.rpca.ready():
self.__timer.stop()
self.start_button.setText("Restart (finished)")
self._pca = self.rpca.get_state()
if self._pca is None:
return
self._variance_ratio = self._pca.explained_variance_ratio_
self._cumulative = numpy.cumsum(self._variance_ratio)
self._setup_plot()
self._transformed = None
self.commit()
def _setup_plot(self):
self.plot.clear()
if self._pca is None:
return
explained_ratio = self._variance_ratio
explained = self._cumulative
p = min(len(self._variance_ratio), self.maxp)
self.plot.plot(numpy.arange(p),
explained_ratio[:p],
pen=pg.mkPen(QColor(Qt.red), width=2),
antialias=True,
name="Variance")
self.plot.plot(numpy.arange(p),
explained[:p],
pen=pg.mkPen(QColor(Qt.darkYellow), width=2),
antialias=True,
name="Cumulative Variance")
cutpos = self._nselected_components() - 1
self._line = pg.InfiniteLine(angle=90,
pos=cutpos,
movable=True,
bounds=(0, p - 1))
self._line.setCursor(Qt.SizeHorCursor)
self._line.setPen(pg.mkPen(QColor(Qt.black), width=2))
self._line.sigPositionChanged.connect(self._on_cut_changed)
self.plot.addItem(self._line)
self.plot_horlines = (
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)),
pg.PlotCurveItem(pen=pg.mkPen(QColor(Qt.blue), style=Qt.DashLine)))
self.plot_horlabels = (pg.TextItem(color=QColor(Qt.black),
anchor=(1, 0)),
pg.TextItem(color=QColor(Qt.black),
anchor=(1, 1)))
for item in self.plot_horlabels + self.plot_horlines:
self.plot.addItem(item)
self._set_horline_pos()
self.plot.setRange(xRange=(0.0, p - 1), yRange=(0.0, 1.0))
self._update_axis()
def _set_horline_pos(self):
cutidx = self.ncomponents - 1
for line, label, curve in zip(
self.plot_horlines, self.plot_horlabels,
(self._variance_ratio, self._cumulative)):
y = curve[cutidx]
line.setData([-1, cutidx], 2 * [y])
label.setPos(cutidx, y)
label.setPlainText("{:.3f}".format(y))
def _on_cut_changed(self, line):
# cut changed by means of a cut line over the scree plot.
value = int(round(line.value()))
self._line.setValue(value)
current = self._nselected_components()
components = value + 1
if not (self.ncomponents == 0
and components == len(self._variance_ratio)):
self.ncomponents = components
self._set_horline_pos()
if self._pca is not None:
var = self._cumulative[components - 1]
if numpy.isfinite(var):
self.variance_covered = int(var * 100)
if current != self._nselected_components():
self._invalidate_selection()
def _update_selection_component_spin(self):
# cut changed by "ncomponents" spin.
if self._pca is None:
self._invalidate_selection()
return
if self.ncomponents == 0:
# Special "All" value
cut = len(self._variance_ratio)
else:
cut = self.ncomponents
var = self._cumulative[cut - 1]
if numpy.isfinite(var):
self.variance_covered = int(var * 100)
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_selection_variance_spin(self):
# cut changed by "max variance" spin.
if self._pca is None:
return
cut = numpy.searchsorted(self._cumulative,
self.variance_covered / 100.0) + 1
cut = min(cut, len(self._cumulative))
self.ncomponents = cut
if numpy.floor(self._line.value()) + 1 != cut:
self._line.setValue(cut - 1)
self._invalidate_selection()
def _update_normalize(self):
if self.normalize:
pp = self._pca_preprocessors + [Normalize()]
else:
pp = self._pca_preprocessors
self._pca_projector.preprocessors = pp
self.fit()
if self.data is None:
self._invalidate_selection()
def _init_projector(self):
cls = DECOMPOSITIONS[self.decomposition_idx]
self._pca_projector = cls(n_components=MAX_COMPONENTS)
self._pca_projector.component = self.ncomponents
self._pca_preprocessors = cls.preprocessors
def _update_decomposition(self):
self._init_projector()
self._update_normalize()
def _nselected_components(self):
"""Return the number of selected components."""
if self._pca is None:
return 0
if self.ncomponents == 0:
# Special "All" value
max_comp = len(self._variance_ratio)
else:
max_comp = self.ncomponents
var_max = self._cumulative[max_comp - 1]
if var_max != numpy.floor(self.variance_covered / 100.0):
cut = max_comp
assert numpy.isfinite(var_max)
self.variance_covered = int(var_max * 100)
else:
self.ncomponents = cut = numpy.searchsorted(
self._cumulative, self.variance_covered / 100.0) + 1
return cut
def _invalidate_selection(self):
self.commit()
def _update_axis(self):
p = min(len(self._variance_ratio), self.maxp)
axis = self.plot.getAxis("bottom")
d = max((p - 1) // (self.axis_labels - 1), 1)
axis.setTicks([[(i, str(i + 1)) for i in range(0, p, d)]])
def commit(self):
transformed = components = None
if self._pca is not None:
if self._transformed is None:
# Compute the full transform (MAX_COMPONENTS components) only once.
self._transformed = self._pca(self.data)
transformed = self._transformed
domain = Domain(transformed.domain.attributes[:self.ncomponents],
self.data.domain.class_vars,
self.data.domain.metas)
transformed = transformed.from_table(domain, transformed)
dom = Domain([
ContinuousVariable(a.name)
for a in self._pca.orig_domain.attributes
],
metas=[StringVariable(name='component')])
metas = numpy.array(
[['PC{}'.format(i + 1) for i in range(self.ncomponents)]],
dtype=object).T
components = Table(dom,
self._pca.components_[:self.ncomponents],
metas=metas)
components.name = 'components'
self._pca_projector.component = self.ncomponents
self.Outputs.transformed_data.send(transformed)
self.Outputs.components.send(components)
self.Outputs.pca.send(self._pca_projector)
def send_report(self):
if self.data is None:
return
self.report_items(
(("Decomposition", DECOMPOSITIONS[self.decomposition_idx].name),
("Normalize data", str(self.normalize)), ("Selected components",
self.ncomponents),
("Explained variance", "{:.3f} %".format(self.variance_covered))))
self.report_plot()
@classmethod
def migrate_settings(cls, settings, version):
if "variance_covered" in settings:
# Due to the error in gh-1896 the variance_covered was persisted
# as a NaN value, causing a TypeError in the widgets `__init__`.
vc = settings["variance_covered"]
if isinstance(vc, numbers.Real):
if numpy.isfinite(vc):
vc = int(vc)
else:
vc = 100
settings["variance_covered"] = vc
if settings["ncomponents"] > MAX_COMPONENTS:
settings["ncomponents"] = MAX_COMPONENTS
class ComboBoxSearch(QComboBox):
"""
A drop down list combo box with filter/search.
The popup list view is filtered by text entered in the filter field.
Note
----
`popup`, `lineEdit` and `completer` from the base QComboBox class are
unused. Setting/modifying them will have no effect.
"""
# NOTE: Setting editable + QComboBox.NoInsert policy + ... did not achieve
# the same results.
def __init__(self, parent=None, **kwargs):
self.__maximumContentsLength = MAXIMUM_CONTENTS_LENGTH
self.__searchline = QLineEdit(visible=False, frame=False)
self.__searchline.setAttribute(Qt.WA_MacShowFocusRect, False)
self.__popup = None # type: Optional[QAbstractItemModel]
self.__proxy = None # type: Optional[QSortFilterProxyModel]
self.__popupTimer = QElapsedTimer()
super().__init__(parent, **kwargs)
self.__searchline.setParent(self)
self.__searchline.setFocusProxy(self)
self.setFocusPolicy(Qt.ClickFocus | Qt.TabFocus)
def setMaximumContentsLength(self, length): # type: (int) -> None
"""
Set the maximum contents length hint.
The hint specifies the upper bound on the `sizeHint` and
`minimumSizeHint` width specified in character length.
Set to 0 or negative value to disable.
Note
----
This property does not affect the widget's `maximumSize`.
The widget can still grow depending on its `sizePolicy`.
Parameters
----------
length : int
Maximum contents length hint.
"""
if self.__maximumContentsLength != length:
self.__maximumContentsLength = length
self.updateGeometry()
def _get_size_hint(self):
sh = super().sizeHint()
if self.__maximumContentsLength > 0:
width = (
self.fontMetrics().width("X") * self.__maximumContentsLength
+ self.iconSize().width() + 4
)
sh = sh.boundedTo(QSize(width, sh.height()))
return sh
def sizeHint(self): # type: () -> QSize
# reimplemented
return self._get_size_hint()
def minimumSizeHint(self): # type: () -> QSize
# reimplemented
return self._get_size_hint()
def showPopup(self):
# type: () -> None
"""
Reimplemented from QComboBox.showPopup
Popup up a customized view and filter edit line.
Note
----
The .popup(), .lineEdit(), .completer() of the base class are not used.
"""
if self.__popup is not None:
# We have user entered state that cannot be disturbed
# (entered filter text, scroll offset, ...)
return # pragma: no cover
if self.count() == 0:
return
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
popup = QListView(
uniformItemSizes=True,
horizontalScrollBarPolicy=Qt.ScrollBarAlwaysOff,
verticalScrollBarPolicy=Qt.ScrollBarAsNeeded,
iconSize=self.iconSize(),
)
popup.setFocusProxy(self.__searchline)
popup.setParent(self, Qt.Popup | Qt.FramelessWindowHint)
popup.setItemDelegate(_ComboBoxListDelegate(popup))
proxy = QSortFilterProxyModel(
popup, filterCaseSensitivity=Qt.CaseInsensitive
)
proxy.setFilterKeyColumn(self.modelColumn())
proxy.setSourceModel(self.model())
popup.setModel(proxy)
root = proxy.mapFromSource(self.rootModelIndex())
popup.setRootIndex(root)
self.__popup = popup
self.__proxy = proxy
self.__searchline.setText("")
self.__searchline.setPlaceholderText("Filter...")
self.__searchline.setVisible(True)
self.__searchline.textEdited.connect(proxy.setFilterFixedString)
style = self.style() # type: QStyle
popuprect_origin = style.subControlRect(
QStyle.CC_ComboBox, opt, QStyle.SC_ComboBoxListBoxPopup, self
) # type: QRect
popuprect_origin = QRect(
self.mapToGlobal(popuprect_origin.topLeft()),
popuprect_origin.size()
)
editrect = style.subControlRect(
QStyle.CC_ComboBox, opt, QStyle.SC_ComboBoxEditField, self
) # type: QRect
self.__searchline.setGeometry(editrect)
desktop = QApplication.desktop()
screenrect = desktop.availableGeometry(self) # type: QRect
# get the height for the view
listrect = QRect()
for i in range(min(proxy.rowCount(root), self.maxVisibleItems())):
index = proxy.index(i, self.modelColumn(), root)
if index.isValid():
listrect = listrect.united(popup.visualRect(index))
if listrect.height() >= screenrect.height():
break
window = popup.window() # type: QWidget
window.ensurePolished()
if window.layout() is not None:
window.layout().activate()
else:
QApplication.sendEvent(window, QEvent(QEvent.LayoutRequest))
margins = qwidget_margin_within(popup.viewport(), window)
height = (listrect.height() + 2 * popup.spacing() +
margins.top() + margins.bottom())
popup_size = (QSize(popuprect_origin.width(), height)
.expandedTo(window.minimumSize())
.boundedTo(window.maximumSize())
.boundedTo(screenrect.size()))
popuprect = QRect(popuprect_origin.bottomLeft(), popup_size)
popuprect = dropdown_popup_geometry(
popuprect, popuprect_origin, screenrect)
popup.setGeometry(popuprect)
current = proxy.mapFromSource(
self.model().index(self.currentIndex(), self.modelColumn(),
self.rootModelIndex()))
popup.setCurrentIndex(current)
popup.scrollTo(current, QAbstractItemView.EnsureVisible)
popup.show()
popup.setFocus(Qt.PopupFocusReason)
popup.installEventFilter(self)
popup.viewport().installEventFilter(self)
popup.viewport().setMouseTracking(True)
self.update()
self.__popupTimer.restart()
def hidePopup(self):
"""Reimplemented"""
if self.__popup is not None:
popup = self.__popup
self.__popup = self.__proxy = None
popup.setFocusProxy(None)
popup.hide()
popup.deleteLater()
popup.removeEventFilter(self)
popup.viewport().removeEventFilter(self)
# need to call base hidePopup even though the base showPopup was not
# called (update internal state wrt. 'pressed' arrow, ...)
super().hidePopup()
self.__searchline.hide()
self.update()
def initStyleOption(self, option):
# type: (QStyleOptionComboBox) -> None
super().initStyleOption(option)
option.editable = True
def __updateGeometries(self):
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
editarea = self.style().subControlRect(
QStyle.CC_ComboBox, opt, QStyle.SC_ComboBoxEditField, self)
self.__searchline.setGeometry(editarea)
def resizeEvent(self, event):
"""Reimplemented."""
super().resizeEvent(event)
self.__updateGeometries()
def paintEvent(self, event):
"""Reimplemented."""
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
painter = QStylePainter(self)
painter.drawComplexControl(QStyle.CC_ComboBox, opt)
if not self.__searchline.isVisibleTo(self):
opt.editable = False
painter.drawControl(QStyle.CE_ComboBoxLabel, opt)
def eventFilter(self, obj, event): # pylint: disable=too-many-branches
# type: (QObject, QEvent) -> bool
"""Reimplemented."""
etype = event.type()
if etype == QEvent.FocusOut and self.__popup is not None:
self.hidePopup()
return True
if etype == QEvent.Hide and self.__popup is not None:
self.hidePopup()
return False
if etype == QEvent.KeyPress or etype == QEvent.KeyRelease or \
etype == QEvent.ShortcutOverride and obj is self.__popup:
event = event # type: QKeyEvent
key, modifiers = event.key(), event.modifiers()
if key in (Qt.Key_Enter, Qt.Key_Return, Qt.Key_Select):
current = self.__popup.currentIndex()
if current.isValid():
self.__activateProxyIndex(current)
elif key in (Qt.Key_Up, Qt.Key_Down,
Qt.Key_PageUp, Qt.Key_PageDown):
return False #
elif key in (Qt.Key_Tab, Qt.Key_Backtab):
pass
elif key == Qt.Key_Escape or \
(key == Qt.Key_F4 and modifiers & Qt.AltModifier):
self.__popup.hide()
return True
else:
# pass the input events to the filter edit line (no propagation
# up the parent chain).
self.__searchline.event(event)
if event.isAccepted():
return True
if etype == QEvent.MouseButtonRelease and self.__popup is not None \
and obj is self.__popup.viewport() \
and self.__popupTimer.elapsed() >= \
QApplication.doubleClickInterval():
event = event # type: QMouseEvent
index = self.__popup.indexAt(event.pos())
if index.isValid():
self.__activateProxyIndex(index)
if etype == QEvent.MouseMove and self.__popup is not None \
and obj is self.__popup.viewport():
event = event # type: QMouseEvent
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
style = self.style() # type: QStyle
if style.styleHint(QStyle.SH_ComboBox_ListMouseTracking, opt, self):
index = self.__popup.indexAt(event.pos())
if index.isValid() and \
index.flags() & (Qt.ItemIsEnabled | Qt.ItemIsSelectable):
self.__popup.setCurrentIndex(index)
if etype == QEvent.MouseButtonPress and self.__popup is obj:
# Popup border or out of window mouse button press/release.
# At least on windows this needs to be handled.
style = self.style()
opt = QStyleOptionComboBox()
self.initStyleOption(opt)
opt.subControls = QStyle.SC_All
opt.activeSubControls = QStyle.SC_ComboBoxArrow
pos = self.mapFromGlobal(event.globalPos())
sc = style.hitTestComplexControl(QStyle.CC_ComboBox, opt, pos, self)
if sc != QStyle.SC_None:
self.__popup.setAttribute(Qt.WA_NoMouseReplay)
self.hidePopup()
return super().eventFilter(obj, event)
def __activateProxyIndex(self, index):
# type: (QModelIndex) -> None
# Set current and activate the source index corresponding to the proxy
# index in the popup's model.
if self.__popup is not None and index.isValid():
proxy = self.__popup.model()
assert index.model() is proxy
index = proxy.mapToSource(index)
assert index.model() is self.model()
if index.isValid() and \
index.flags() & (Qt.ItemIsEnabled | Qt.ItemIsSelectable):
self.hidePopup()
text = self.itemText(index.row())
self.setCurrentIndex(index.row())
self.activated[int].emit(index.row())
self.activated[str].emit(text)
class OWSql(OWWidget):
name = "SQL Table"
id = "orange.widgets.data.sql"
description = "Load data set from SQL."
icon = "icons/SQLTable.svg"
priority = 30
category = "Data"
keywords = ["data", "file", "load", "read", "SQL"]
class Outputs:
data = Output("Data", Table, doc="Attribute-valued data set read from the input file.")
settings_version = 2
want_main_area = False
resizing_enabled = False
host = Setting(None)
port = Setting(None)
database = Setting(None)
schema = Setting(None)
username = ""
password = ""
table = Setting(None)
sql = Setting("")
guess_values = Setting(True)
download = Setting(False)
materialize = Setting(False)
materialize_table_name = Setting("")
class Information(OWWidget.Information):
data_sampled = Msg("Data description was generated from a sample.")
class Error(OWWidget.Error):
connection = Msg("{}")
no_backends = Msg("Please install a backend to use this widget")
missing_extension = Msg("Database is missing extension{}: {}")
def __init__(self):
super().__init__()
self.backend = None
self.data_desc_table = None
self.database_desc = None
vbox = gui.vBox(self.controlArea, "Server", addSpace=True)
box = gui.vBox(vbox)
self.backends = BackendModel(Backend.available_backends())
self.backendcombo = QComboBox(box)
if len(self.backends):
self.backendcombo.setModel(self.backends)
else:
self.Error.no_backends()
box.setEnabled(False)
box.layout().addWidget(self.backendcombo)
self.servertext = QLineEdit(box)
self.servertext.setPlaceholderText('Server')
self.servertext.setToolTip('Server')
self.servertext.editingFinished.connect(self._load_credentials)
if self.host:
self.servertext.setText(self.host if not self.port else
'{}:{}'.format(self.host, self.port))
box.layout().addWidget(self.servertext)
self.databasetext = QLineEdit(box)
self.databasetext.setPlaceholderText('Database[/Schema]')
self.databasetext.setToolTip('Database or optionally Database/Schema')
if self.database:
self.databasetext.setText(
self.database if not self.schema else
'{}/{}'.format(self.database, self.schema))
box.layout().addWidget(self.databasetext)
self.usernametext = QLineEdit(box)
self.usernametext.setPlaceholderText('Username')
self.usernametext.setToolTip('Username')
box.layout().addWidget(self.usernametext)
self.passwordtext = QLineEdit(box)
self.passwordtext.setPlaceholderText('Password')
self.passwordtext.setToolTip('Password')
self.passwordtext.setEchoMode(QLineEdit.Password)
box.layout().addWidget(self.passwordtext)
self._load_credentials()
self.tables = TableModel()
tables = gui.hBox(box)
self.tablecombo = QComboBox(
minimumContentsLength=35,
sizeAdjustPolicy=QComboBox.AdjustToMinimumContentsLength
)
self.tablecombo.setModel(self.tables)
self.tablecombo.setToolTip('table')
tables.layout().addWidget(self.tablecombo)
self.connect()
index = self.tablecombo.findText(str(self.table))
if index != -1:
self.tablecombo.setCurrentIndex(index)
# set up the callback to select_table in case of selection change
self.tablecombo.activated[int].connect(self.select_table)
self.connectbutton = gui.button(
tables, self, '↻', callback=self.connect)
self.connectbutton.setSizePolicy(
QSizePolicy.Fixed, QSizePolicy.Fixed)
tables.layout().addWidget(self.connectbutton)
self.custom_sql = gui.vBox(box)
self.custom_sql.setVisible(False)
self.sqltext = QTextEdit(self.custom_sql)
self.sqltext.setPlainText(self.sql)
self.custom_sql.layout().addWidget(self.sqltext)
mt = gui.hBox(self.custom_sql)
cb = gui.checkBox(mt, self, 'materialize', 'Materialize to table ')
cb.setToolTip('Save results of the query in a table')
le = gui.lineEdit(mt, self, 'materialize_table_name')
le.setToolTip('Save results of the query in a table')
self.executebtn = gui.button(
self.custom_sql, self, 'Execute', callback=self.open_table)
box.layout().addWidget(self.custom_sql)
gui.checkBox(box, self, "guess_values",
"Auto-discover categorical variables",
callback=self.open_table)
gui.checkBox(box, self, "download",
"Download data to local memory",
callback=self.open_table)
gui.rubber(self.buttonsArea)
QTimer.singleShot(0, self.select_table)
def _load_credentials(self):
self._parse_host_port()
cm = self._credential_manager(self.host, self.port)
self.username = cm.username
self.password = cm.password
if self.username:
self.usernametext.setText(self.username)
if self.password:
self.passwordtext.setText(self.password)
def _save_credentials(self):
cm = self._credential_manager(self.host, self.port)
cm.username = self.username or ''
cm.password = self.password or ''
def _credential_manager(self, host, port):
return CredentialManager("SQL Table: {}:{}".format(host, port))
def error(self, id=0, text=""):
super().error(id, text)
err_style = 'QLineEdit {border: 2px solid red;}'
if 'server' in text or 'host' in text:
self.servertext.setStyleSheet(err_style)
else:
self.servertext.setStyleSheet('')
if 'role' in text:
self.usernametext.setStyleSheet(err_style)
else:
self.usernametext.setStyleSheet('')
if 'database' in text:
self.databasetext.setStyleSheet(err_style)
else:
self.databasetext.setStyleSheet('')
def _parse_host_port(self):
hostport = self.servertext.text().split(':')
self.host = hostport[0]
self.port = hostport[1] if len(hostport) == 2 else None
def connect(self):
self._parse_host_port()
self.database, _, self.schema = self.databasetext.text().partition('/')
self.username = self.usernametext.text() or None
self.password = self.passwordtext.text() or None
try:
if self.backendcombo.currentIndex() < 0:
return
backend = self.backends[self.backendcombo.currentIndex()]
self.backend = backend(dict(
host=self.host,
port=self.port,
database=self.database,
user=self.username,
password=self.password
))
self.Error.connection.clear()
self._save_credentials()
self.database_desc = OrderedDict((
("Host", self.host), ("Port", self.port),
("Database", self.database), ("User name", self.username)
))
self.refresh_tables()
except BackendError as err:
error = str(err).split('\n')[0]
self.Error.connection(error)
self.database_desc = self.data_desc_table = None
self.tablecombo.clear()
def refresh_tables(self):
self.tables.clear()
self.Error.missing_extension.clear()
if self.backend is None:
self.data_desc_table = None
return
self.tables.append("Select a table")
self.tables.append("Custom SQL")
self.tables.extend(self.backend.list_tables(self.schema))
# Called on tablecombo selection change:
def select_table(self):
curIdx = self.tablecombo.currentIndex()
if self.tablecombo.itemText(curIdx) != "Custom SQL":
self.custom_sql.setVisible(False)
return self.open_table()
else:
self.custom_sql.setVisible(True)
self.data_desc_table = None
self.database_desc["Table"] = "(None)"
self.table = None
if len(str(self.sql)) > 14:
return self.open_table()
#self.Error.missing_extension(
# 's' if len(missing) > 1 else '',
# ', '.join(missing),
# shown=missing)
def open_table(self):
table = self.get_table()
self.data_desc_table = table
self.Outputs.data.send(table)
def get_table(self):
curIdx = self.tablecombo.currentIndex()
if curIdx <= 0:
if self.database_desc:
self.database_desc["Table"] = "(None)"
self.data_desc_table = None
return
if self.tablecombo.itemText(curIdx) != "Custom SQL":
self.table = self.tables[self.tablecombo.currentIndex()]
self.database_desc["Table"] = self.table
if "Query" in self.database_desc:
del self.database_desc["Query"]
what = self.table
else:
what = self.sql = self.sqltext.toPlainText()
self.table = "Custom SQL"
if self.materialize:
import psycopg2
if not self.materialize_table_name:
self.Error.connection(
"Specify a table name to materialize the query")
return
try:
with self.backend.execute_sql_query("DROP TABLE IF EXISTS " +
self.materialize_table_name):
pass
with self.backend.execute_sql_query("CREATE TABLE " +
self.materialize_table_name +
" AS " + self.sql):
pass
with self.backend.execute_sql_query("ANALYZE " + self.materialize_table_name):
pass
except (psycopg2.ProgrammingError, BackendError) as ex:
self.Error.connection(str(ex))
return
try:
table = SqlTable(dict(host=self.host,
port=self.port,
database=self.database,
user=self.username,
password=self.password),
what,
backend=type(self.backend),
inspect_values=False)
except BackendError as ex:
self.Error.connection(str(ex))
return
self.Error.connection.clear()
sample = False
if table.approx_len() > LARGE_TABLE and self.guess_values:
confirm = QMessageBox(self)
confirm.setIcon(QMessageBox.Warning)
confirm.setText("Attribute discovery might take "
"a long time on large tables.\n"
"Do you want to auto discover attributes?")
confirm.addButton("Yes", QMessageBox.YesRole)
no_button = confirm.addButton("No", QMessageBox.NoRole)
sample_button = confirm.addButton("Yes, on a sample",
QMessageBox.YesRole)
confirm.exec()
if confirm.clickedButton() == no_button:
self.guess_values = False
elif confirm.clickedButton() == sample_button:
sample = True
self.Information.clear()
if self.guess_values:
QApplication.setOverrideCursor(QCursor(Qt.WaitCursor))
if sample:
s = table.sample_time(1)
domain = s.get_domain(inspect_values=True)
self.Information.data_sampled()
else:
domain = table.get_domain(inspect_values=True)
QApplication.restoreOverrideCursor()
table.domain = domain
if self.download:
if table.approx_len() > MAX_DL_LIMIT:
QMessageBox.warning(
self, 'Warning', "Data is too big to download.\n"
"Consider using the Data Sampler widget to download "
"a sample instead.")
self.download = False
elif table.approx_len() > AUTO_DL_LIMIT:
confirm = QMessageBox.question(
self, 'Question', "Data appears to be big. Do you really "
"want to download it to local memory?",
QMessageBox.Yes | QMessageBox.No, QMessageBox.No)
if confirm == QMessageBox.No:
self.download = False
if self.download:
table.download_data(MAX_DL_LIMIT)
table = Table(table)
return table
def send_report(self):
if not self.database_desc:
self.report_paragraph("No database connection.")
return
self.report_items("Database", self.database_desc)
if self.data_desc_table:
self.report_items("Data",
report.describe_data(self.data_desc_table))
@classmethod
def migrate_settings(cls, settings, version):
if version < 2:
# Until Orange version 3.4.4 username and password had been stored
# in Settings.
cm = cls._credential_manager(settings["host"], settings["port"])
cm.username = settings["username"]
cm.password = settings["password"]
class OWSelectAttributes(widget.OWWidget):
name = "Select Columns"
description = "Select columns from the data table and assign them to " \
"data features, classes or meta variables."
icon = "icons/SelectColumns.svg"
priority = 100
inputs = [("Data", Table, "set_data")]
outputs = [("Data", Table), ("Features", widget.AttributeList)]
want_main_area = False
want_control_area = True
settingsHandler = SelectAttributesDomainContextHandler()
domain_role_hints = ContextSetting({})
auto_commit = Setting(False)
def __init__(self):
super().__init__()
self.controlArea = QWidget(self.controlArea)
self.layout().addWidget(self.controlArea)
layout = QGridLayout()
self.controlArea.setLayout(layout)
layout.setContentsMargins(4, 4, 4, 4)
box = gui.vBox(self.controlArea, "Available Variables",
addToLayout=False)
self.filter_edit = QLineEdit()
self.filter_edit.setToolTip("Filter the list of available variables.")
box.layout().addWidget(self.filter_edit)
if hasattr(self.filter_edit, "setPlaceholderText"):
self.filter_edit.setPlaceholderText("Filter")
self.completer = QCompleter()
self.completer.setCompletionMode(QCompleter.InlineCompletion)
self.completer_model = QStringListModel()
self.completer.setModel(self.completer_model)
self.completer.setModelSorting(
QCompleter.CaseSensitivelySortedModel)
self.filter_edit.setCompleter(self.completer)
self.completer_navigator = CompleterNavigator(self)
self.filter_edit.installEventFilter(self.completer_navigator)
def dropcompleted(action):
if action == Qt.MoveAction:
self.commit()
self.available_attrs = VariableListModel(enable_dnd=True)
self.available_attrs_proxy = VariableFilterProxyModel()
self.available_attrs_proxy.setSourceModel(self.available_attrs)
self.available_attrs_view = VariablesListItemView(
acceptedType=Orange.data.Variable)
self.available_attrs_view.setModel(self.available_attrs_proxy)
aa = self.available_attrs
aa.dataChanged.connect(self.update_completer_model)
aa.rowsInserted.connect(self.update_completer_model)
aa.rowsRemoved.connect(self.update_completer_model)
self.available_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.available_attrs_view))
self.available_attrs_view.dragDropActionDidComplete.connect(dropcompleted)
self.filter_edit.textChanged.connect(self.update_completer_prefix)
self.filter_edit.textChanged.connect(
self.available_attrs_proxy.set_filter_string)
box.layout().addWidget(self.available_attrs_view)
layout.addWidget(box, 0, 0, 3, 1)
box = gui.vBox(self.controlArea, "Features", addToLayout=False)
self.used_attrs = VariableListModel(enable_dnd=True)
self.used_attrs_view = VariablesListItemView(
acceptedType=(Orange.data.DiscreteVariable,
Orange.data.ContinuousVariable))
self.used_attrs_view.setModel(self.used_attrs)
self.used_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.used_attrs_view))
self.used_attrs_view.dragDropActionDidComplete.connect(dropcompleted)
box.layout().addWidget(self.used_attrs_view)
layout.addWidget(box, 0, 2, 1, 1)
box = gui.vBox(self.controlArea, "Target Variable", addToLayout=False)
self.class_attrs = ClassVarListItemModel(enable_dnd=True)
self.class_attrs_view = ClassVariableItemView(
acceptedType=(Orange.data.DiscreteVariable,
Orange.data.ContinuousVariable))
self.class_attrs_view.setModel(self.class_attrs)
self.class_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.class_attrs_view))
self.class_attrs_view.dragDropActionDidComplete.connect(dropcompleted)
self.class_attrs_view.setMaximumHeight(24)
box.layout().addWidget(self.class_attrs_view)
layout.addWidget(box, 1, 2, 1, 1)
box = gui.vBox(self.controlArea, "Meta Attributes", addToLayout=False)
self.meta_attrs = VariableListModel(enable_dnd=True)
self.meta_attrs_view = VariablesListItemView(
acceptedType=Orange.data.Variable)
self.meta_attrs_view.setModel(self.meta_attrs)
self.meta_attrs_view.selectionModel().selectionChanged.connect(
partial(self.update_interface_state, self.meta_attrs_view))
self.meta_attrs_view.dragDropActionDidComplete.connect(dropcompleted)
box.layout().addWidget(self.meta_attrs_view)
layout.addWidget(box, 2, 2, 1, 1)
bbox = gui.vBox(self.controlArea, addToLayout=False, margin=0)
layout.addWidget(bbox, 0, 1, 1, 1)
self.up_attr_button = gui.button(bbox, self, "Up",
callback=partial(self.move_up, self.used_attrs_view))
self.move_attr_button = gui.button(bbox, self, ">",
callback=partial(self.move_selected, self.used_attrs_view))
self.down_attr_button = gui.button(bbox, self, "Down",
callback=partial(self.move_down, self.used_attrs_view))
bbox = gui.vBox(self.controlArea, addToLayout=False, margin=0)
layout.addWidget(bbox, 1, 1, 1, 1)
self.move_class_button = gui.button(bbox, self, ">",
callback=partial(self.move_selected,
self.class_attrs_view, exclusive=True))
bbox = gui.vBox(self.controlArea, addToLayout=False, margin=0)
layout.addWidget(bbox, 2, 1, 1, 1)
self.up_meta_button = gui.button(bbox, self, "Up",
callback=partial(self.move_up, self.meta_attrs_view))
self.move_meta_button = gui.button(bbox, self, ">",
callback=partial(self.move_selected, self.meta_attrs_view))
self.down_meta_button = gui.button(bbox, self, "Down",
callback=partial(self.move_down, self.meta_attrs_view))
autobox = gui.auto_commit(None, self, "auto_commit", "Send")
layout.addWidget(autobox, 3, 0, 1, 3)
reset = gui.button(None, self, "Reset", callback=self.reset)
autobox.layout().insertWidget(0, self.report_button)
autobox.layout().insertWidget(1, reset)
autobox.layout().insertSpacing(2, 10)
reset.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Preferred)
self.report_button.setSizePolicy(QSizePolicy.Expanding, QSizePolicy.Preferred)
layout.setRowStretch(0, 4)
layout.setRowStretch(1, 0)
layout.setRowStretch(2, 2)
layout.setHorizontalSpacing(0)
self.controlArea.setLayout(layout)
self.data = None
self.output_data = None
self.original_completer_items = []
self.resize(500, 600)
def set_data(self, data=None):
self.update_domain_role_hints()
self.closeContext()
self.data = data
if data is not None:
self.openContext(data)
all_vars = data.domain.variables + data.domain.metas
var_sig = lambda attr: (attr.name, vartype(attr))
domain_hints = {var_sig(attr): ("attribute", i)
for i, attr in enumerate(data.domain.attributes)}
domain_hints.update({var_sig(attr): ("meta", i)
for i, attr in enumerate(data.domain.metas)})
if data.domain.class_vars:
domain_hints.update(
{var_sig(attr): ("class", i)
for i, attr in enumerate(data.domain.class_vars)})
# update the hints from context settings
domain_hints.update(self.domain_role_hints)
attrs_for_role = lambda role: [
(domain_hints[var_sig(attr)][1], attr)
for attr in all_vars if domain_hints[var_sig(attr)][0] == role]
attributes = [
attr for place, attr in sorted(attrs_for_role("attribute"),
key=lambda a: a[0])]
classes = [
attr for place, attr in sorted(attrs_for_role("class"),
key=lambda a: a[0])]
metas = [
attr for place, attr in sorted(attrs_for_role("meta"),
key=lambda a: a[0])]
available = [
attr for place, attr in sorted(attrs_for_role("available"),
key=lambda a: a[0])]
self.used_attrs[:] = attributes
self.class_attrs[:] = classes
self.meta_attrs[:] = metas
self.available_attrs[:] = available
else:
self.used_attrs[:] = []
self.class_attrs[:] = []
self.meta_attrs[:] = []
self.available_attrs[:] = []
self.unconditional_commit()
def update_domain_role_hints(self):
""" Update the domain hints to be stored in the widgets settings.
"""
hints_from_model = lambda role, model: [
((attr.name, vartype(attr)), (role, i))
for i, attr in enumerate(model)]
hints = dict(hints_from_model("available", self.available_attrs))
hints.update(hints_from_model("attribute", self.used_attrs))
hints.update(hints_from_model("class", self.class_attrs))
hints.update(hints_from_model("meta", self.meta_attrs))
self.domain_role_hints = hints
def selected_rows(self, view):
""" Return the selected rows in the view.
"""
rows = view.selectionModel().selectedRows()
model = view.model()
if isinstance(model, QSortFilterProxyModel):
rows = [model.mapToSource(r) for r in rows]
return [r.row() for r in rows]
def move_rows(self, view, rows, offset):
model = view.model()
newrows = [min(max(0, row + offset), len(model) - 1) for row in rows]
for row, newrow in sorted(zip(rows, newrows), reverse=offset > 0):
model[row], model[newrow] = model[newrow], model[row]
selection = QItemSelection()
for nrow in newrows:
index = model.index(nrow, 0)
selection.select(index, index)
view.selectionModel().select(
selection, QItemSelectionModel.ClearAndSelect)
self.commit()
def move_up(self, view):
selected = self.selected_rows(view)
self.move_rows(view, selected, -1)
def move_down(self, view):
selected = self.selected_rows(view)
self.move_rows(view, selected, 1)
def move_selected(self, view, exclusive=False):
if self.selected_rows(view):
self.move_selected_from_to(view, self.available_attrs_view)
elif self.selected_rows(self.available_attrs_view):
self.move_selected_from_to(self.available_attrs_view, view,
exclusive)
def move_selected_from_to(self, src, dst, exclusive=False):
self.move_from_to(src, dst, self.selected_rows(src), exclusive)
def move_from_to(self, src, dst, rows, exclusive=False):
src_model = source_model(src)
attrs = [src_model[r] for r in rows]
if exclusive and len(attrs) != 1:
return
for s1, s2 in reversed(list(slices(rows))):
del src_model[s1:s2]
dst_model = source_model(dst)
if exclusive and len(dst_model) > 0:
src_model.append(dst_model[0])
del dst_model[0]
dst_model.extend(attrs)
self.commit()
def update_interface_state(self, focus=None, selected=None, deselected=None):
for view in [self.available_attrs_view, self.used_attrs_view,
self.class_attrs_view, self.meta_attrs_view]:
if view is not focus and not view.hasFocus() and self.selected_rows(view):
view.selectionModel().clear()
def selected_vars(view):
model = source_model(view)
return [model[i] for i in self.selected_rows(view)]
available_selected = selected_vars(self.available_attrs_view)
attrs_selected = selected_vars(self.used_attrs_view)
class_selected = selected_vars(self.class_attrs_view)
meta_selected = selected_vars(self.meta_attrs_view)
available_types = set(map(type, available_selected))
all_primitive = all(var.is_primitive()
for var in available_types)
move_attr_enabled = (available_selected and all_primitive) or \
attrs_selected
self.move_attr_button.setEnabled(bool(move_attr_enabled))
if move_attr_enabled:
self.move_attr_button.setText(">" if available_selected else "<")
move_class_enabled = (len(available_selected) == 1 and all_primitive) or \
class_selected
self.move_class_button.setEnabled(bool(move_class_enabled))
if move_class_enabled:
self.move_class_button.setText(">" if available_selected else "<")
move_meta_enabled = available_selected or meta_selected
self.move_meta_button.setEnabled(bool(move_meta_enabled))
if move_meta_enabled:
self.move_meta_button.setText(">" if available_selected else "<")
def update_completer_model(self, *_):
""" This gets called when the model for available attributes changes
through either drag/drop or the left/right button actions.
"""
vars = list(self.available_attrs)
items = [var.name for var in vars]
items += ["%s=%s" % item for v in vars for item in v.attributes.items()]
new = sorted(set(items))
if new != self.original_completer_items:
self.original_completer_items = new
self.completer_model.setStringList(self.original_completer_items)
def update_completer_prefix(self, filter):
""" Prefixes all items in the completer model with the current
already done completion to enable the completion of multiple keywords.
"""
prefix = str(self.completer.completionPrefix())
if not prefix.endswith(" ") and " " in prefix:
prefix, _ = prefix.rsplit(" ", 1)
items = [prefix + " " + item
for item in self.original_completer_items]
else:
items = self.original_completer_items
old = list(map(str, self.completer_model.stringList()))
if set(old) != set(items):
self.completer_model.setStringList(items)
def commit(self):
self.update_domain_role_hints()
if self.data is not None:
attributes = list(self.used_attrs)
class_var = list(self.class_attrs)
metas = list(self.meta_attrs)
domain = Orange.data.Domain(attributes, class_var, metas)
newdata = self.data.from_table(domain, self.data)
self.output_data = newdata
self.send("Data", newdata)
self.send("Features", widget.AttributeList(attributes))
else:
self.output_data = None
self.send("Data", None)
self.send("Features", None)
def reset(self):
if self.data is not None:
self.available_attrs[:] = []
self.used_attrs[:] = self.data.domain.attributes
self.class_attrs[:] = self.data.domain.class_vars
self.meta_attrs[:] = self.data.domain.metas
self.update_domain_role_hints()
self.commit()
def send_report(self):
if not self.data or not self.output_data:
return
in_domain, out_domain = self.data.domain, self.output_data.domain
self.report_domain("Input data", self.data.domain)
if (in_domain.attributes, in_domain.class_vars, in_domain.metas) == (
out_domain.attributes, out_domain.class_vars, out_domain.metas):
self.report_paragraph("Output data", "No changes.")
else:
self.report_domain("Output data", self.output_data.domain)
diff = list(set(in_domain.variables + in_domain.metas) -
set(out_domain.variables + out_domain.metas))
if diff:
text = "%i (%s)" % (len(diff), ", ".join(x.name for x in diff))
self.report_items((("Removed", text),))
class ListViewSearch(QListView):
"""
An QListView with an implicit and transparent row filtering.
"""
def __init__(self, *a, preferred_size=None, **ak):
super().__init__(*a, **ak)
self.__search = QLineEdit(self, placeholderText="Filter...")
self.__search.textEdited.connect(self.__setFilterString)
# Use an QSortFilterProxyModel for filtering. Note that this is
# never set on the view, only its rows insertes/removed signals are
# connected to observe an update row hidden state.
self.__pmodel = QSortFilterProxyModel(
self, filterCaseSensitivity=Qt.CaseInsensitive)
self.__pmodel.rowsAboutToBeRemoved.connect(
self.__filter_rowsAboutToBeRemoved)
self.__pmodel.rowsInserted.connect(self.__filter_rowsInserted)
self.__layout()
self.preferred_size = preferred_size
self.setMinimumHeight(100)
def setFilterPlaceholderText(self, text: str):
self.__search.setPlaceholderText(text)
def filterPlaceholderText(self) -> str:
return self.__search.placeholderText()
def setFilterProxyModel(self, proxy: QSortFilterProxyModel) -> None:
"""
Set an instance of QSortFilterProxyModel that will be used for filtering
the model. The `proxy` must be a filtering proxy only; it MUST not sort
the row of the model.
The FilterListView takes ownership of the proxy.
"""
self.__pmodel.rowsAboutToBeRemoved.disconnect(
self.__filter_rowsAboutToBeRemoved)
self.__pmodel.rowsInserted.disconnect(self.__filter_rowsInserted)
self.__pmodel = proxy
proxy.setParent(self)
self.__pmodel.rowsAboutToBeRemoved.connect(
self.__filter_rowsAboutToBeRemoved)
self.__pmodel.rowsInserted.connect(self.__filter_rowsInserted)
self.__pmodel.setSourceModel(self.model())
self.__filter_reset()
def filterProxyModel(self) -> QSortFilterProxyModel:
return self.__pmodel
def setModel(self, model: QAbstractItemModel) -> None:
super().setModel(model)
self.__pmodel.setSourceModel(model)
self.__filter_reset()
self.model().rowsInserted.connect(self.__model_rowInserted)
def setRootIndex(self, index: QModelIndex) -> None:
super().setRootIndex(index)
self.__filter_reset()
def __filter_reset(self):
root = self.rootIndex()
self.__filter(range(self.__pmodel.rowCount(root)))
def __setFilterString(self, string: str):
self.__pmodel.setFilterFixedString(string)
def setFilterString(self, string: str):
"""Set the filter string."""
self.__search.setText(string)
self.__pmodel.setFilterFixedString(string)
def filterString(self):
"""Return the filter string."""
return self.__search.text()
def __filter(self, rows: Iterable[int]) -> None:
"""Set hidden state for rows based on filter string"""
root = self.rootIndex()
pm = self.__pmodel
for r in rows:
self.setRowHidden(r, not pm.filterAcceptsRow(r, root))
def __filter_set(self, rows: Iterable[int], state: bool):
for r in rows:
self.setRowHidden(r, state)
def __filter_rowsAboutToBeRemoved(self, parent: QModelIndex, start: int,
end: int) -> None:
fmodel = self.__pmodel
mrange = QItemSelection(fmodel.index(start, 0, parent),
fmodel.index(end, 0, parent))
mranges = fmodel.mapSelectionToSource(mrange)
for mrange in mranges:
self.__filter_set(range(mrange.top(), mrange.bottom() + 1), True)
def __filter_rowsInserted(self, parent: QModelIndex, start: int,
end: int) -> None:
fmodel = self.__pmodel
mrange = QItemSelection(fmodel.index(start, 0, parent),
fmodel.index(end, 0, parent))
mranges = fmodel.mapSelectionToSource(mrange)
for mrange in mranges:
self.__filter_set(range(mrange.top(), mrange.bottom() + 1), False)
def __model_rowInserted(self, _, start: int, end: int) -> None:
"""
Filter elements when inserted in list - proxy model's rowsAboutToBeRemoved
is not called on elements that are hidden when inserting
"""
self.__filter(range(start, end + 1))
def resizeEvent(self, event: QResizeEvent) -> None:
super().resizeEvent(event)
def updateGeometries(self) -> None:
super().updateGeometries()
self.__layout()
def __layout(self):
margins = self.viewportMargins()
search = self.__search
sh = search.sizeHint()
size = self.size()
margins.setTop(sh.height())
vscroll = self.verticalScrollBar()
style = self.style()
transient = style.styleHint(QStyle.SH_ScrollBar_Transient, None,
vscroll)
w = size.width()
if vscroll.isVisibleTo(self) and not transient:
w = w - vscroll.width() - 1
search.setGeometry(0, 0, w, sh.height())
self.setViewportMargins(margins)
def sizeHint(self):
return (self.preferred_size
if self.preferred_size is not None else super().sizeHint())
