def create_model(self):
allkeys = set(self.allinfo_local) | set(self.allinfo_remote)
allkeys = sorted(allkeys)
model = QStandardItemModel(self)
model.setHorizontalHeaderLabels(self._header_labels)
current_index = -1
for i, file_path in enumerate(allkeys):
datainfo = self._parse_info(file_path)
item1 = QStandardItem()
item1.setData(" " if datainfo.islocal else "", Qt.DisplayRole)
item1.setData(self.IndicatorBrushes[0], Qt.ForegroundRole)
item1.setData(datainfo, Qt.UserRole)
item2 = QStandardItem(datainfo.title)
item3 = QStandardItem()
item3.setData(datainfo.size, Qt.DisplayRole)
item4 = QStandardItem()
item4.setData(datainfo.instances, Qt.DisplayRole)
item5 = QStandardItem()
item5.setData(datainfo.variables, Qt.DisplayRole)
item6 = QStandardItem()
item6.setData(datainfo.target, Qt.DisplayRole)
if datainfo.target:
item6.setIcon(variable_icon(datainfo.target))
item7 = QStandardItem()
item7.setData(", ".join(datainfo.tags) if datainfo.tags else "",
Qt.DisplayRole)
row = [item1, item2, item3, item4, item5, item6, item7]
model.appendRow(row)
if os.path.join(*file_path) == self.selected_id:
current_index = i
return model, current_index
def create_model(self):
allkeys = set(self.allinfo_local) | set(self.allinfo_remote)
allkeys = sorted(allkeys)
model = QStandardItemModel(self)
model.setHorizontalHeaderLabels(self._header_labels)
current_index = -1
for i, file_path in enumerate(allkeys):
datainfo = self._parse_info(file_path)
item1 = QStandardItem()
item1.setData(" " if datainfo.islocal else "", Qt.DisplayRole)
item1.setData(datainfo, Qt.UserRole)
item2 = QStandardItem(datainfo.title)
item3 = QStandardItem()
item3.setData(datainfo.size, Qt.DisplayRole)
item4 = QStandardItem()
item4.setData(datainfo.instances, Qt.DisplayRole)
item5 = QStandardItem()
item5.setData(datainfo.variables, Qt.DisplayRole)
item6 = QStandardItem()
item6.setData(datainfo.target, Qt.DisplayRole)
if datainfo.target:
item6.setIcon(variable_icon(datainfo.target))
item7 = QStandardItem()
item7.setData(", ".join(datainfo.tags) if datainfo.tags else "",
Qt.DisplayRole)
row = [item1, item2, item3, item4, item5, item6, item7]
model.appendRow(row)
if os.path.join(*file_path) == self.selected_id:
current_index = i
return model, current_index
def on_done(self, result: Results):
model = QStandardItemModel()
for item in result.items:
model.appendRow(item)
model.setSortRole(Qt.UserRole)
model.setHorizontalHeaderLabels(Header.labels())
self.filter_proxy_model.setSourceModel(model)
self.tree_view.selectionModel().selectionChanged.connect(self.commit)
self.filter_view()
self.update_info_box()
def create_model(self):
allkeys = set(self.allinfo_local)
if self.allinfo_remote is not None:
allkeys = allkeys | set(self.allinfo_remote)
allkeys = sorted(allkeys)
model = QStandardItemModel(self)
model.setHorizontalHeaderLabels(self._header_labels)
current_index = -1
for i, file_path in enumerate(allkeys):
data_info = self._parse_info(file_path)
row = []
for info_tag, header_setting in self.HEADER_SCHEMA:
item = QStandardItem()
try:
data = data_info.__getattribute__(info_tag)
except AttributeError:
# unknown tag in JSON
data = ''
# first column indicating cached data sets
if info_tag == 'islocal':
item.setData(' ' if data else '', Qt.DisplayRole)
item.setData(data_info, Qt.UserRole)
else:
# parse taxid to common name
if info_tag == 'taxid' and data in common_taxids():
data = shortname(data)[0].title()
if info_tag == 'tags':
data = ', '.join(data) if data else ''
item.setData(data, Qt.DisplayRole)
# set icon to Target column
if info_tag == 'target' and data:
item.setIcon(
Orange.widgets.data.owdatasets.variable_icon(data))
row.append(item)
model.appendRow(row)
if os.path.join(*file_path) == self.selected_id:
current_index = i
return model, current_index
class OWTestLearners(OWWidget):
name = "Test & Score"
description = "Cross-validation accuracy estimation."
icon = "icons/TestLearners1.svg"
priority = 100
inputs = [("Learner", Learner, "set_learner", widget.Multiple),
("Data", Table, "set_train_data", widget.Default),
("Test Data", Table, "set_test_data"),
("Preprocessor", Preprocess, "set_preprocessor")]
outputs = [("Predictions", Table),
("Evaluation Results", Results)]
settingsHandler = settings.ClassValuesContextHandler()
#: Resampling/testing types
KFold, ShuffleSplit, LeaveOneOut, TestOnTrain, TestOnTest = 0, 1, 2, 3, 4
#: Numbers of folds
NFolds = [2, 3, 5, 10, 20]
#: Number of repetitions
NRepeats = [2, 3, 5, 10, 20, 50, 100]
#: Sample sizes
SampleSizes = [5, 10, 20, 25, 30, 33, 40, 50, 60, 66, 70, 75, 80, 90, 95]
#: Selected resampling type
resampling = settings.Setting(0)
#: Number of folds for K-fold cross validation
n_folds = settings.Setting(3)
#: Stratified sampling for K-fold
cv_stratified = settings.Setting(True)
#: Number of repeats for ShuffleSplit sampling
n_repeats = settings.Setting(3)
#: ShuffleSplit sample size
sample_size = settings.Setting(9)
#: Stratified sampling for Random Sampling
shuffle_stratified = settings.Setting(True)
TARGET_AVERAGE = "(Average over classes)"
class_selection = settings.ContextSetting(TARGET_AVERAGE)
class Error(OWWidget.Error):
train_data_empty = Msg("Train data set is empty.")
test_data_empty = Msg("Test data set is empty.")
class_required = Msg("Train data input requires a target variable.")
too_many_classes = Msg("Too many target variables.")
class_required_test = Msg("Test data input requires a target variable.")
too_many_folds = Msg("Number of folds exceeds the data size")
class_inconsistent = Msg("Test and train data sets "
"have different target variables.")
class Warning(OWWidget.Warning):
missing_data = \
Msg("Instances with unknown target values were removed from{}data.")
test_data_missing = Msg("Missing separate test data input.")
scores_not_computed = Msg("Some scores could not be computed.")
test_data_unused = Msg("Test data is present but unused. "
"Select 'Test on test data' to use it.")
class Information(OWWidget.Information):
data_sampled = Msg("Train data has been sampled")
test_data_sampled = Msg("Test data has been sampled")
def __init__(self):
super().__init__()
self.data = None
self.test_data = None
self.preprocessor = None
self.train_data_missing_vals = False
self.test_data_missing_vals = False
#: An Ordered dictionary with current inputs and their testing results.
self.learners = OrderedDict()
sbox = gui.vBox(self.controlArea, "Sampling")
rbox = gui.radioButtons(
sbox, self, "resampling", callback=self._param_changed)
gui.appendRadioButton(rbox, "Cross validation")
ibox = gui.indentedBox(rbox)
gui.comboBox(
ibox, self, "n_folds", label="Number of folds: ",
items=[str(x) for x in self.NFolds], maximumContentsLength=3,
orientation=Qt.Horizontal, callback=self.kfold_changed)
gui.checkBox(
ibox, self, "cv_stratified", "Stratified",
callback=self.kfold_changed)
gui.appendRadioButton(rbox, "Random sampling")
ibox = gui.indentedBox(rbox)
gui.comboBox(
ibox, self, "n_repeats", label="Repeat train/test: ",
items=[str(x) for x in self.NRepeats], maximumContentsLength=3,
orientation=Qt.Horizontal, callback=self.shuffle_split_changed)
gui.comboBox(
ibox, self, "sample_size", label="Training set size: ",
items=["{} %".format(x) for x in self.SampleSizes],
maximumContentsLength=5, orientation=Qt.Horizontal,
callback=self.shuffle_split_changed)
gui.checkBox(
ibox, self, "shuffle_stratified", "Stratified",
callback=self.shuffle_split_changed)
gui.appendRadioButton(rbox, "Leave one out")
gui.appendRadioButton(rbox, "Test on train data")
gui.appendRadioButton(rbox, "Test on test data")
self.cbox = gui.vBox(self.controlArea, "Target Class")
self.class_selection_combo = gui.comboBox(
self.cbox, self, "class_selection", items=[],
sendSelectedValue=True, valueType=str,
callback=self._on_target_class_changed,
contentsLength=8)
gui.rubber(self.controlArea)
self.view = gui.TableView(
wordWrap=True,
)
header = self.view.horizontalHeader()
header.setSectionResizeMode(QHeaderView.ResizeToContents)
header.setDefaultAlignment(Qt.AlignCenter)
header.setStretchLastSection(False)
self.result_model = QStandardItemModel(self)
self.result_model.setHorizontalHeaderLabels(["Method"])
self.view.setModel(self.result_model)
self.view.setItemDelegate(ItemDelegate())
box = gui.vBox(self.mainArea, "Evaluation Results")
box.layout().addWidget(self.view)
def sizeHint(self):
return QSize(780, 1)
def set_learner(self, learner, key):
"""
Set the input `learner` for `key`.
"""
if key in self.learners and learner is None:
# Removed
del self.learners[key]
else:
self.learners[key] = Input(learner, None, None)
self._invalidate([key])
def set_train_data(self, data):
"""
Set the input training dataset.
"""
self.Information.data_sampled.clear()
self.Error.train_data_empty.clear()
if data is not None and not len(data):
self.Error.train_data_empty()
data = None
if data and not data.domain.class_vars:
self.Error.class_required()
data = None
elif data and len(data.domain.class_vars) > 1:
self.Error.too_many_classes()
data = None
else:
self.Error.class_required.clear()
self.Error.too_many_classes.clear()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.train_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = RemoveNaNClasses(data)
else:
self.Warning.missing_data.clear()
self.data = data
self.closeContext()
if data is not None:
self._update_class_selection()
self.openContext(data.domain.class_var)
self._invalidate()
def set_test_data(self, data):
"""
Set the input separate testing dataset.
"""
self.Information.test_data_sampled.clear()
self.Error.test_data_empty.clear()
if data is not None and not len(data):
self.Error.test_data_empty()
data = None
if data and not data.domain.class_var:
self.Error.class_required()
data = None
else:
self.Error.class_required_test.clear()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.test_data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.test_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = RemoveNaNClasses(data)
else:
self.Warning.missing_data.clear()
self.test_data = data
if self.resampling == OWTestLearners.TestOnTest:
self._invalidate()
def _which_missing_data(self):
return {(True, True): " ", # both, don't specify
(True, False): " train ",
(False, True): " test "}[(self.train_data_missing_vals,
self.test_data_missing_vals)]
def set_preprocessor(self, preproc):
"""
Set the input preprocessor to apply on the training data.
"""
self.preprocessor = preproc
self._invalidate()
def handleNewSignals(self):
"""Reimplemented from OWWidget.handleNewSignals."""
self._update_class_selection()
self.commit()
def kfold_changed(self):
self.resampling = OWTestLearners.KFold
self._param_changed()
def shuffle_split_changed(self):
self.resampling = OWTestLearners.ShuffleSplit
self._param_changed()
def _param_changed(self):
self._invalidate()
def _update_results(self):
"""
Run/evaluate the learners.
"""
self.Warning.test_data_unused.clear()
self.Warning.test_data_missing.clear()
self.warning()
self.Error.class_inconsistent.clear()
self.Error.too_many_folds.clear()
self.error()
if self.data is None:
return
class_var = self.data.domain.class_var
if self.resampling == OWTestLearners.TestOnTest:
if self.test_data is None:
if not self.Error.test_data_empty.is_shown():
self.Warning.test_data_missing()
return
elif self.test_data.domain.class_var != class_var:
self.Error.class_inconsistent()
return
# items in need of an update
items = [(key, slot) for key, slot in self.learners.items()
if slot.results is None]
learners = [slot.learner for _, slot in items]
if len(items) == 0:
return
if self.test_data is not None and \
self.resampling != OWTestLearners.TestOnTest:
self.Warning.test_data_unused()
rstate = 42
def update_progress(finished):
self.progressBarSet(100 * finished)
common_args = dict(
store_data=True,
preprocessor=self.preprocessor,
callback=update_progress,
n_jobs=-1,
)
self.setStatusMessage("Running")
with self.progressBar():
try:
folds = self.NFolds[self.n_folds]
if self.resampling == OWTestLearners.KFold:
if len(self.data) < folds:
self.Error.too_many_folds()
return
warnings = []
results = Orange.evaluation.CrossValidation(
self.data, learners, k=folds,
random_state=rstate, warnings=warnings, **common_args)
if warnings:
self.warning(warnings[0])
elif self.resampling == OWTestLearners.LeaveOneOut:
results = Orange.evaluation.LeaveOneOut(
self.data, learners, **common_args)
elif self.resampling == OWTestLearners.ShuffleSplit:
train_size = self.SampleSizes[self.sample_size] / 100
results = Orange.evaluation.ShuffleSplit(
self.data, learners,
n_resamples=self.NRepeats[self.n_repeats],
train_size=train_size, test_size=None,
stratified=self.shuffle_stratified,
random_state=rstate, **common_args)
elif self.resampling == OWTestLearners.TestOnTrain:
results = Orange.evaluation.TestOnTrainingData(
self.data, learners, **common_args)
elif self.resampling == OWTestLearners.TestOnTest:
results = Orange.evaluation.TestOnTestData(
self.data, self.test_data, learners, **common_args)
else:
assert False
except (RuntimeError, ValueError) as e:
self.error(str(e))
self.setStatusMessage("")
return
else:
self.error()
learner_key = {slot.learner: key for key, slot in self.learners.items()}
for learner, result in zip(learners, results.split_by_model()):
stats = None
if class_var.is_discrete:
scorers = classification_stats.scores
elif class_var.is_continuous:
scorers = regression_stats.scores
else:
scorers = None
if scorers:
ex = result.failed[0]
if ex:
stats = [Try.Fail(ex)] * len(scorers)
result = Try.Fail(ex)
else:
stats = [Try(lambda: score(result)) for score in scorers]
result = Try.Success(result)
key = learner_key[learner]
self.learners[key] = \
self.learners[key]._replace(results=result, stats=stats)
self.setStatusMessage("")
def _update_header(self):
# Set the correct horizontal header labels on the results_model.
headers = ["Method"]
if self.data is not None:
if self.data.domain.has_discrete_class:
headers.extend(classification_stats.headers)
else:
headers.extend(regression_stats.headers)
# remove possible extra columns from the model.
for i in reversed(range(len(headers),
self.result_model.columnCount())):
self.result_model.takeColumn(i)
self.result_model.setHorizontalHeaderLabels(headers)
def _update_stats_model(self):
# Update the results_model with up to date scores.
# Note: The target class specific scores (if requested) are
# computed as needed in this method.
model = self.view.model()
# clear the table model, but preserving the header labels
for r in reversed(range(model.rowCount())):
model.takeRow(r)
target_index = None
if self.data is not None:
class_var = self.data.domain.class_var
if self.data.domain.has_discrete_class and \
self.class_selection != self.TARGET_AVERAGE:
target_index = class_var.values.index(self.class_selection)
else:
class_var = None
errors = []
has_missing_scores = False
for key, slot in self.learners.items():
name = learner_name(slot.learner)
head = QStandardItem(name)
head.setData(key, Qt.UserRole)
if isinstance(slot.results, Try.Fail):
head.setToolTip(str(slot.results.exception))
head.setText("{} (error)".format(name))
head.setForeground(QtGui.QBrush(Qt.red))
errors.append("{name} failed with error:\n"
"{exc.__class__.__name__}: {exc!s}"
.format(name=name, exc=slot.results.exception))
row = [head]
if class_var is not None and class_var.is_discrete and \
target_index is not None:
if slot.results is not None and slot.results.success:
ovr_results = results_one_vs_rest(
slot.results.value, target_index)
stats = [Try(lambda: score(ovr_results))
for score in classification_stats.scores]
else:
stats = None
else:
stats = slot.stats
if stats is not None:
for stat in stats:
item = QStandardItem()
if stat.success:
item.setText("{:.3f}".format(stat.value[0]))
else:
item.setToolTip(str(stat.exception))
has_missing_scores = True
row.append(item)
model.appendRow(row)
self.error("\n".join(errors), shown=bool(errors))
self.Warning.scores_not_computed(shown=has_missing_scores)
def _update_class_selection(self):
self.class_selection_combo.setCurrentIndex(-1)
self.class_selection_combo.clear()
if not self.data:
return
if self.data.domain.has_discrete_class:
self.cbox.setVisible(True)
class_var = self.data.domain.class_var
items = [self.TARGET_AVERAGE] + class_var.values
self.class_selection_combo.addItems(items)
class_index = 0
if self.class_selection in class_var.values:
class_index = class_var.values.index(self.class_selection) + 1
self.class_selection_combo.setCurrentIndex(class_index)
self.class_selection = items[class_index]
else:
self.cbox.setVisible(False)
def _on_target_class_changed(self):
self._update_stats_model()
def _invalidate(self, which=None):
# Invalidate learner results for `which` input keys
# (if None then all learner results are invalidated)
if which is None:
which = self.learners.keys()
model = self.view.model()
statmodelkeys = [model.item(row, 0).data(Qt.UserRole)
for row in range(model.rowCount())]
for key in which:
self.learners[key] = \
self.learners[key]._replace(results=None, stats=None)
if key in statmodelkeys:
row = statmodelkeys.index(key)
for c in range(1, model.columnCount()):
item = model.item(row, c)
if item is not None:
item.setData(None, Qt.DisplayRole)
item.setData(None, Qt.ToolTipRole)
self.commit()
def commit(self):
"""Recompute and output the results"""
self._update_header()
# Update the view to display the model names
self._update_stats_model()
self._update_results()
self._update_stats_model()
valid = [slot for slot in self.learners.values()
if slot.results is not None and slot.results.success]
if valid:
# Evaluation results
combined = results_merge([slot.results.value for slot in valid])
combined.learner_names = [learner_name(slot.learner)
for slot in valid]
# Predictions & Probabilities
predictions = combined.get_augmented_data(combined.learner_names)
else:
combined = None
predictions = None
self.send("Evaluation Results", combined)
self.send("Predictions", predictions)
def send_report(self):
"""Report on the testing schema and results"""
if not self.data or not self.learners:
return
if self.resampling == self.KFold:
stratified = 'Stratified ' if self.cv_stratified else ''
items = [("Sampling type", "{}{}-fold Cross validation".
format(stratified, self.NFolds[self.n_folds]))]
elif self.resampling == self.LeaveOneOut:
items = [("Sampling type", "Leave one out")]
elif self.resampling == self.ShuffleSplit:
stratified = 'Stratified ' if self.shuffle_stratified else ''
items = [("Sampling type",
"{}Shuffle split, {} random samples with {}% data "
.format(stratified, self.NRepeats[self.n_repeats],
self.SampleSizes[self.sample_size]))]
elif self.resampling == self.TestOnTrain:
items = [("Sampling type", "No sampling, test on training data")]
elif self.resampling == self.TestOnTest:
items = [("Sampling type", "No sampling, test on testing data")]
else:
items = []
if self.data.domain.has_discrete_class:
items += [("Target class", self.class_selection.strip("()"))]
if items:
self.report_items("Settings", items)
self.report_table("Scores", self.view)
def get_gds_model(progress=lambda val: None):
"""
Initialize and return a GDS datasets model.
:param progress: A progress callback.
:rval tuple:
A tuple of (QStandardItemModel, geo.GDSInfo, [geo.GDS])
.. note::
The returned QStandardItemModel's thread affinity is set to
the GUI thread.
"""
progress(1)
info = geo.GDSInfo()
search_keys = ["dataset_id", "title", "platform_organism", "description"]
cache_dir = serverfiles.localpath(geo.DOMAIN)
gds_link = "http://www.ncbi.nlm.nih.gov/sites/GDSbrowser?acc={0}"
pm_link = "http://www.ncbi.nlm.nih.gov/pubmed/{0}"
gds_list = []
def is_cached(gds):
return os.path.exists(
os.path.join(cache_dir, gds["dataset_id"]) + ".soft.gz")
def item(displayvalue, item_values={}):
item = QStandardItem()
item.setData(displayvalue, Qt.DisplayRole)
for role, value in item_values.items():
item.setData(value, role)
return item
def gds_to_row(gds):
#: Text for easier full search.
search_text = " | ".join(
[gds.get(key, "").lower() for key in search_keys])
row = [
item(" " if is_cached(gds) else "", {TextFilterRole: search_text}),
item(gds["dataset_id"],
{LinkRole: gds_link.format(gds["dataset_id"])}),
item(gds["title"]),
item(gds["platform_organism"]),
item(len(gds["samples"])),
item(gds["feature_count"]),
item(gds["gene_count"]),
item(len(gds["subsets"])),
item(
gds.get("pubmed_id", ""), {
LinkRole:
pm_link.format(gds["pubmed_id"])
if gds.get("pubmed_id") else None
})
]
return row
model = QStandardItemModel()
model.setHorizontalHeaderLabels([
"", "ID", "Title", "Organism", "Samples", "Features", "Genes",
"Subsets", "PubMedID"
])
progress(20)
for gds in info.values():
model.appendRow(gds_to_row(gds))
gds_list.append(gds)
progress(50)
if QThread.currentThread() is not QCoreApplication.instance().thread():
model.moveToThread(QCoreApplication.instance().thread())
return model, info, gds_list
class OWKMeans(widget.OWWidget):
name = "k-Means"
description = "k-Means clustering algorithm with silhouette-based " \
"quality estimation."
icon = "icons/KMeans.svg"
priority = 2100
inputs = [("Data", Table, "set_data")]
outputs = [("Annotated Data", Table, widget.Default), ("Centroids", Table)]
class Error(widget.OWWidget.Error):
failed = widget.Msg("Clustering failed\nError: {}")
INIT_KMEANS, INIT_RANDOM = range(2)
INIT_METHODS = "Initialize with KMeans++", "Random initialization"
SILHOUETTE, INTERCLUSTER, DISTANCES = range(3)
SCORING_METHODS = [
("Silhouette", lambda km: km.silhouette, False, True),
("Inter-cluster distance", lambda km: km.inter_cluster, True, False),
("Distance to centroids", lambda km: km.inertia, True, False)
]
OUTPUT_CLASS, OUTPUT_ATTRIBUTE, OUTPUT_META = range(3)
OUTPUT_METHODS = ("Class", "Feature", "Meta")
resizing_enabled = False
k = Setting(3)
k_from = Setting(2)
k_to = Setting(8)
optimize_k = Setting(False)
max_iterations = Setting(300)
n_init = Setting(10)
smart_init = Setting(INIT_KMEANS)
scoring = Setting(SILHOUETTE)
append_cluster_ids = Setting(True)
place_cluster_ids = Setting(OUTPUT_CLASS)
output_name = Setting("Cluster")
auto_run = Setting(True)
def __init__(self):
super().__init__()
self.data = None
self.km = None
self.optimization_runs = []
box = gui.vBox(self.controlArea, "Number of Clusters")
layout = QGridLayout()
self.n_clusters = bg = gui.radioButtonsInBox(box,
self,
"optimize_k", [],
orientation=layout,
callback=self.update)
layout.addWidget(
gui.appendRadioButton(bg, "Fixed:", addToLayout=False), 1, 1)
sb = gui.hBox(None, margin=0)
self.fixedSpinBox = gui.spin(sb,
self,
"k",
minv=2,
maxv=30,
controlWidth=60,
alignment=Qt.AlignRight,
callback=self.update_k)
gui.rubber(sb)
layout.addWidget(sb, 1, 2)
layout.addWidget(
gui.appendRadioButton(bg, "Optimized from", addToLayout=False), 2,
1)
ftobox = gui.hBox(None)
ftobox.layout().setContentsMargins(0, 0, 0, 0)
layout.addWidget(ftobox)
gui.spin(ftobox,
self,
"k_from",
minv=2,
maxv=29,
controlWidth=60,
alignment=Qt.AlignRight,
callback=self.update_from)
gui.widgetLabel(ftobox, "to")
self.fixedSpinBox = gui.spin(ftobox,
self,
"k_to",
minv=3,
maxv=30,
controlWidth=60,
alignment=Qt.AlignRight,
callback=self.update_to)
gui.rubber(ftobox)
layout.addWidget(gui.widgetLabel(None, "Scoring: "), 5, 1,
Qt.AlignRight)
layout.addWidget(
gui.comboBox(None,
self,
"scoring",
label="Scoring",
items=list(zip(*self.SCORING_METHODS))[0],
callback=self.update), 5, 2)
box = gui.vBox(self.controlArea, "Initialization")
gui.comboBox(box,
self,
"smart_init",
items=self.INIT_METHODS,
callback=self.update)
layout = QGridLayout()
box2 = gui.widgetBox(box, orientation=layout)
box2.setSizePolicy(QSizePolicy.Maximum, QSizePolicy.Maximum)
layout.addWidget(gui.widgetLabel(None, "Re-runs: "), 0, 0,
Qt.AlignLeft)
sb = gui.hBox(None, margin=0)
layout.addWidget(sb, 0, 1)
gui.lineEdit(sb,
self,
"n_init",
controlWidth=60,
valueType=int,
validator=QIntValidator(),
callback=self.update)
layout.addWidget(gui.widgetLabel(None, "Maximal iterations: "), 1, 0,
Qt.AlignLeft)
sb = gui.hBox(None, margin=0)
layout.addWidget(sb, 1, 1)
gui.lineEdit(sb,
self,
"max_iterations",
controlWidth=60,
valueType=int,
validator=QIntValidator(),
callback=self.update)
box = gui.vBox(self.controlArea, "Output")
gui.comboBox(box,
self,
"place_cluster_ids",
label="Append cluster ID as:",
orientation=Qt.Horizontal,
callback=self.send_data,
items=self.OUTPUT_METHODS)
gui.lineEdit(box,
self,
"output_name",
label="Name:",
orientation=Qt.Horizontal,
callback=self.send_data)
gui.separator(self.buttonsArea, 30)
self.apply_button = gui.auto_commit(self.buttonsArea,
self,
"auto_run",
"Apply",
box=None,
commit=self.commit)
gui.rubber(self.controlArea)
self.table_model = QStandardItemModel(self)
self.table_model.setHorizontalHeaderLabels(["k", "Score"])
self.table_model.setColumnCount(2)
self.table_box = gui.vBox(self.mainArea,
"Optimization Report",
addSpace=0)
table = self.table_view = QTableView(self.table_box)
table.setHorizontalScrollMode(QTableView.ScrollPerPixel)
table.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
table.setSelectionMode(QTableView.SingleSelection)
table.setSelectionBehavior(QTableView.SelectRows)
table.verticalHeader().hide()
self.bar_delegate = gui.ColoredBarItemDelegate(self, color=Qt.cyan)
table.setItemDelegateForColumn(1, self.bar_delegate)
table.setModel(self.table_model)
table.selectionModel().selectionChanged.connect(
self.table_item_selected)
table.setColumnWidth(0, 40)
table.setColumnWidth(1, 120)
table.horizontalHeader().setStretchLastSection(True)
self.setSizePolicy(QSizePolicy.Preferred, QSizePolicy.Preferred)
self.mainArea.setSizePolicy(QSizePolicy.Maximum, QSizePolicy.Preferred)
self.table_box.setSizePolicy(QSizePolicy.Fixed,
QSizePolicy.MinimumExpanding)
self.table_view.setSizePolicy(QSizePolicy.Preferred,
QSizePolicy.MinimumExpanding)
self.table_box.layout().addWidget(self.table_view)
self.hide_show_opt_results()
def adjustSize(self):
self.ensurePolished()
s = self.sizeHint()
self.resize(s)
def hide_show_opt_results(self):
[self.mainArea.hide, self.mainArea.show][self.optimize_k]()
QTimer.singleShot(100, self.adjustSize)
def sizeHint(self):
s = self.controlArea.sizeHint()
if self.optimize_k and not self.mainArea.isHidden():
s.setWidth(s.width() + self.mainArea.sizeHint().width() +
4 * self.childrenRect().x())
return s
def update_k(self):
self.optimize_k = False
self.update()
def update_from(self):
self.k_to = max(self.k_from + 1, self.k_to)
self.optimize_k = True
self.update()
def update_to(self):
self.k_from = min(self.k_from, self.k_to - 1)
self.optimize_k = True
self.update()
def set_optimization(self):
self.updateOptimizationGui()
self.update()
def check_data_size(self, n, msg_group):
msg_group.add_message(
"not_enough_data",
"Too few ({}) unique data instances for {} clusters")
if n > len(self.data):
msg_group.not_enough_data(len(self.data), n)
return False
else:
msg_group.not_enough_data.clear()
return True
def run_optimization(self):
# Disabling is needed since this function is not reentrant
# Fast clicking on, say, "To: " causes multiple calls
try:
self.controlArea.setDisabled(True)
self.optimization_runs = []
error = ""
if not self.check_data_size(self.k_from, self.Error):
return
self.check_data_size(self.k_to, self.Warning)
k_to = min(self.k_to, len(self.data))
kmeans = KMeans(
init=['random', 'k-means++'][self.smart_init],
n_init=self.n_init,
max_iter=self.max_iterations,
compute_silhouette_score=self.scoring == self.SILHOUETTE)
with self.progressBar(k_to - self.k_from + 1) as progress:
for k in range(self.k_from, k_to + 1):
progress.advance()
kmeans.params["n_clusters"] = k
try:
self.optimization_runs.append((k, kmeans(self.data)))
except BaseException as exc:
error = str(exc)
self.optimization_runs.append((k, error))
if all(
isinstance(score, str)
for _, score in self.optimization_runs):
self.Error.failed(error) # Report just the last error
self.optimization_runs = []
finally:
self.controlArea.setDisabled(False)
self.show_results()
self.send_data()
def cluster(self):
if not self.check_data_size(self.k, self.Error):
return
try:
self.km = KMeans(n_clusters=self.k,
init=['random', 'k-means++'][self.smart_init],
n_init=self.n_init,
max_iter=self.max_iterations)(self.data)
except BaseException as exc:
self.Error.failed(str(exc))
self.km = None
self.send_data()
def run(self):
self.clear_messages()
if not self.data:
return
if self.optimize_k:
self.run_optimization()
else:
self.cluster()
def commit(self):
self.run()
def show_results(self):
_, scoring_method, minimize, normal = self.SCORING_METHODS[
self.scoring]
k_scores = [(k,
scoring_method(run) if not isinstance(run, str) else run)
for k, run in self.optimization_runs]
scores = [score for _, score in k_scores if not isinstance(score, str)]
min_score, max_score = min(scores, default=0), max(scores, default=1)
best_score = min_score if minimize else max_score
if normal:
min_score, max_score = 0, 1
nplaces = 3
else:
nplaces = min(5,
np.floor(abs(math.log(max(max_score, 1e-10)))) + 2)
score_span = (max_score - min_score) or 1
self.bar_delegate.scale = (min_score, max_score)
self.bar_delegate.float_fmt = "%%.%if" % int(nplaces)
model = self.table_model
model.setRowCount(len(k_scores))
no_selection = True
for i, (k, score) in enumerate(k_scores):
item0 = model.item(i, 0) or QStandardItem()
item0.setData(k, Qt.DisplayRole)
item0.setTextAlignment(Qt.AlignCenter)
model.setItem(i, 0, item0)
item = model.item(i, 1) or QStandardItem()
if not isinstance(score, str):
item.setData(score, Qt.DisplayRole)
item.setData(None, Qt.ToolTipRole)
bar_ratio = 0.95 * (score - min_score) / score_span
item.setData(bar_ratio, gui.BarRatioRole)
if no_selection and score == best_score:
self.table_view.selectRow(i)
no_selection = False
color = Qt.black
flags = Qt.ItemIsEnabled | Qt.ItemIsSelectable
else:
item.setData("clustering failed", Qt.DisplayRole)
item.setData(score, Qt.ToolTipRole)
item.setData(None, gui.BarRatioRole)
color = Qt.gray
flags = Qt.NoItemFlags
item0.setData(QBrush(color), Qt.ForegroundRole)
item0.setFlags(flags)
item.setData(QBrush(color), Qt.ForegroundRole)
item.setFlags(flags)
model.setItem(i, 1, item)
self.table_view.resizeRowsToContents()
self.table_view.show()
if minimize:
self.table_box.setTitle("Scoring (smaller is better)")
else:
self.table_box.setTitle("Scoring (bigger is better)")
QTimer.singleShot(0, self.adjustSize)
def update(self):
self.hide_show_opt_results()
self.run()
def selected_row(self):
indices = self.table_view.selectedIndexes()
rows = {ind.row() for ind in indices}
if len(rows) == 1:
return rows.pop()
def table_item_selected(self):
row = self.selected_row()
if row is not None:
self.send_data()
def send_data(self):
if self.optimize_k:
row = self.selected_row() if self.optimization_runs else None
km = self.optimization_runs[row][1] if row is not None else None
else:
km = self.km
if not self.data or not km:
self.send("Annotated Data", None)
self.send("Centroids", None)
return
clust_var = DiscreteVariable(
self.output_name, values=["C%d" % (x + 1) for x in range(km.k)])
clust_ids = km(self.data)
domain = self.data.domain
attributes, classes = domain.attributes, domain.class_vars
meta_attrs = domain.metas
if self.place_cluster_ids == self.OUTPUT_CLASS:
if classes:
meta_attrs += classes
classes = [clust_var]
elif self.place_cluster_ids == self.OUTPUT_ATTRIBUTE:
attributes += (clust_var, )
else:
meta_attrs += (clust_var, )
domain = Domain(attributes, classes, meta_attrs)
new_table = Table.from_table(domain, self.data)
new_table.get_column_view(clust_var)[0][:] = clust_ids.X.ravel()
centroids = Table(Domain(km.pre_domain.attributes), km.centroids)
self.send("Annotated Data", new_table)
self.send("Centroids", centroids)
@check_sql_input
def set_data(self, data):
self.data = data
if data is None:
self.Error.clear()
self.Warning.clear()
self.table_model.setRowCount(0)
self.send("Annotated Data", None)
self.send("Centroids", None)
else:
self.data = data
self.run()
def send_report(self):
k_clusters = self.k
if self.optimize_k and self.optimization_runs and self.selected_row(
) is not None:
k_clusters = self.optimization_runs[self.selected_row()][1].k
self.report_items(
(("Number of clusters",
k_clusters), ("Optimization", self.optimize_k != 0
and "{}, {} re-runs limited to {} steps".format(
self.INIT_METHODS[self.smart_init].lower(),
self.n_init, self.max_iterations)),
("Cluster ID in output", self.append_cluster_ids
and "'{}' (as {})".format(
self.output_name,
self.OUTPUT_METHODS[self.place_cluster_ids].lower()))))
if self.data:
self.report_data("Data", self.data)
if self.optimize_k:
self.report_table(
"Scoring by {}".format(
self.SCORING_METHODS[self.scoring][0]),
self.table_view)
def __on_enrichment_finished(self, results):
assert QThread.currentThread() is self.thread()
self.__state &= ~OWSetEnrichment.RunningEnrichment
query, reference, results = results
if self.annotationsChartView.model():
self.annotationsChartView.model().clear()
nquery = len(query)
nref = len(reference)
maxcount = max((len(e.query_mapped) for _, e in results), default=1)
maxrefcount = max((len(e.reference_mapped) for _, e in results),
default=1)
nspaces = int(math.ceil(math.log10(maxcount or 1)))
refspaces = int(math.ceil(math.log(maxrefcount or 1)))
query_fmt = "%" + str(nspaces) + "s (%.2f%%)"
ref_fmt = "%" + str(refspaces) + "s (%.2f%%)"
def fmt_count(fmt, count, total):
return fmt % (count, 100.0 * count / (total or 1))
fmt_query_count = partial(fmt_count, query_fmt)
fmt_ref_count = partial(fmt_count, ref_fmt)
linkFont = QFont(self.annotationsChartView.viewOptions().font)
linkFont.setUnderline(True)
def item(value=None, tooltip=None, user=None):
si = QStandardItem()
if value is not None:
si.setData(value, Qt.DisplayRole)
if tooltip is not None:
si.setData(tooltip, Qt.ToolTipRole)
if user is not None:
si.setData(user, Qt.UserRole)
else:
si.setData(value, Qt.UserRole)
return si
model = QStandardItemModel()
model.setSortRole(Qt.UserRole)
model.setHorizontalHeaderLabels([
"Category", "Term", "Count", "Reference count", "p-value", "FDR",
"Enrichment"
])
for i, (gset, enrich) in enumerate(results):
if len(enrich.query_mapped) == 0:
continue
nquery_mapped = len(enrich.query_mapped)
nref_mapped = len(enrich.reference_mapped)
row = [
item(", ".join(gset.hierarchy)),
item(gsname(gset), tooltip=gset.link),
item(fmt_query_count(nquery_mapped, nquery),
tooltip=nquery_mapped,
user=nquery_mapped),
item(fmt_ref_count(nref_mapped, nref),
tooltip=nref_mapped,
user=nref_mapped),
item(fmtp(enrich.p_value), user=enrich.p_value),
item(
), # column 5, FDR, is computed in filterAnnotationsChartView
item(enrich.enrichment_score,
tooltip="%.3f" % enrich.enrichment_score,
user=enrich.enrichment_score)
]
row[0].geneset = gset
row[0].enrichment = enrich
row[1].setData(gset.link, gui.LinkRole)
row[1].setFont(linkFont)
row[1].setForeground(QColor(Qt.blue))
model.appendRow(row)
self.annotationsChartView.setModel(model)
self.annotationsChartView.selectionModel().selectionChanged.connect(
self.commit)
if not model.rowCount():
self.warning(0, "No enriched sets found.")
else:
self.warning(0)
allnames = set(
gsname(geneset) for geneset, (count, _, _, _) in results if count)
allnames |= reduce(operator.ior,
(set(word_split(name)) for name in allnames), set())
self.filterCompleter.setModel(None)
self.completerModel = QStringListModel(sorted(allnames))
self.filterCompleter.setModel(self.completerModel)
if results:
max_score = max(
(e.enrichment_score
for _, e in results if np.isfinite(e.enrichment_score)),
default=1)
self.annotationsChartView.setItemDelegateForColumn(
6, BarItemDelegate(self, scale=(0.0, max_score)))
self.annotationsChartView.setItemDelegateForColumn(
1, gui.LinkStyledItemDelegate(self.annotationsChartView))
header = self.annotationsChartView.header()
for i in range(model.columnCount()):
sh = self.annotationsChartView.sizeHintForColumn(i)
sh = max(sh, header.sectionSizeHint(i))
self.annotationsChartView.setColumnWidth(i, max(min(sh, 300), 30))
# self.annotationsChartView.resizeColumnToContents(i)
self.filterAnnotationsChartView()
self.progressBarFinished()
self.setStatusMessage("")
class OWGeneSets(OWWidget):
name = "Gene Sets"
description = ""
icon = "icons/OWGeneSets.svg"
priority = 9
want_main_area = True
settingsHandler = OrganismContextHandler()
# settings
auto_commit = Setting(True)
stored_selections = ContextSetting([])
organism = ContextSetting(None)
class Inputs:
genes = Input("Genes", Table)
class Outputs:
matched_genes = Output("Matched Genes", Table)
class Information(OWWidget.Information):
pass
class Error(OWWidget.Error):
missing_annotation = Msg(ERROR_ON_MISSING_ANNOTATION)
missing_gene_id = Msg(ERROR_ON_MISSING_GENE_ID)
missing_tax_id = Msg(ERROR_ON_MISSING_TAX_ID)
cant_reach_host = Msg("Host orange.biolab.si is unreachable.")
cant_load_organisms = Msg(
"No available organisms, please check your connection.")
def __init__(self):
super().__init__()
# commit
self.commit_button = None
# progress bar
self.progress_bar = None
self.progress_bar_iterations = None
# data
self.input_data = None
self.input_genes = None
self.tax_id = None
self.use_attr_names = None
self.gene_id_attribute = None
self.gene_id_column = None
self.input_info = None
self.num_of_sel_genes = 0
# filter
self.lineEdit_filter = None
self.search_pattern = ''
self.organism_select_combobox = None
# data model view
self.data_view = None
self.data_model = None
# gene matcher NCBI
self.gene_matcher = None
# filter proxy model
self.filter_proxy_model = None
# hierarchy widget
self.hierarchy_widget = None
self.hierarchy_state = None
# threads
self.threadpool = QThreadPool(self)
self.workers = None
# gui
self.setup_gui()
def _progress_advance(self):
# GUI should be updated in main thread. That's why we are calling advance method here
if self.progress_bar:
self.progress_bar.advance()
def __get_genes(self):
self.input_genes = []
if self.use_attr_names:
for variable in self.input_data.domain.attributes:
self.input_genes.append(
str(variable.attributes.get(self.gene_id_attribute, '?')))
else:
genes, _ = self.input_data.get_column_view(self.gene_id_column)
self.input_genes = [str(g) for g in genes]
@Inputs.genes
def handle_input(self, data):
self.closeContext()
self.Error.clear()
if data:
self.input_data = data
self.tax_id = str(self.input_data.attributes.get(TAX_ID, None))
self.use_attr_names = self.input_data.attributes.get(
GENE_AS_ATTRIBUTE_NAME, None)
self.gene_id_attribute = self.input_data.attributes.get(
GENE_ID_ATTRIBUTE, None)
self.gene_id_column = self.input_data.attributes.get(
GENE_ID_COLUMN, None)
if not (self.use_attr_names is not None and
((self.gene_id_attribute is None) ^
(self.gene_id_column is None))):
if self.tax_id is None:
self.Error.missing_annotation()
return
self.Error.missing_gene_id()
return
elif self.tax_id is None:
self.Error.missing_tax_id()
return
self.openContext(self.tax_id)
self.__get_genes()
self.download_gene_sets()
def update_info_box(self):
info_string = ''
if self.input_genes:
info_string += '{} unique gene names on input.\n'.format(
len(self.input_genes))
info_string += '{} genes on output.\n'.format(
self.num_of_sel_genes)
else:
info_string += 'No genes on input.\n'
self.input_info.setText(info_string)
def on_gene_sets_download(self, result):
# make sure this happens in the main thread.
# Qt insists that widgets be created within the GUI(main) thread.
assert threading.current_thread() == threading.main_thread()
self.progress_bar.finish()
self.setStatusMessage('')
tax_id, sets = result
self.set_hierarchy_model(self.hierarchy_widget,
*hierarchy_tree(tax_id, sets))
self.set_selected_hierarchies()
self.update_info_box()
self.workers = defaultdict(list)
self.progress_bar_iterations = dict()
for selected_hierarchy in self.get_hierarchies():
gene_sets = geneset.load_gene_sets(selected_hierarchy)
worker = Worker(get_collections,
gene_sets,
set(self.input_genes),
progress_callback=True,
partial_result=True)
worker.signals.error.connect(self.handle_error)
worker.signals.finished.connect(self.handle_worker_finished)
worker.signals.progress.connect(self._progress_advance)
worker.signals.partial_result.connect(self.populate_data_model)
worker.setAutoDelete(False)
self.workers[selected_hierarchy] = worker
self.progress_bar_iterations[selected_hierarchy] = len(gene_sets)
self.display_gene_sets()
def handle_worker_finished(self):
# We check if all workers have completed. If not, continue
# dirty hax, is this ok?
if self.progress_bar and self.progress_bar.widget.progressBarValue == 100:
self.progress_bar.finish()
self.setStatusMessage('')
self.hierarchy_widget.setDisabled(False)
# adjust column width
for i in range(len(DATA_HEADER_LABELS) - 1):
self.data_view.resizeColumnToContents(i)
self.filter_proxy_model.setSourceModel(self.data_model)
def populate_data_model(self, partial_result):
assert threading.current_thread() == threading.main_thread()
if partial_result:
self.data_model.appendRow(partial_result)
def set_hierarchy_model(self, model, tax_id, sets):
def beautify_displayed_text(text):
if '_' in text:
return text.replace('_', ' ').title()
else:
return text
# TODO: maybe optimize this code?
for key, value in sets.items():
item = QTreeWidgetItem(model, [beautify_displayed_text(key)])
item.setFlags(item.flags()
& (Qt.ItemIsUserCheckable | ~Qt.ItemIsSelectable
| Qt.ItemIsEnabled))
item.setExpanded(True)
item.tax_id = tax_id
item.hierarchy = key
if value:
item.setFlags(item.flags() | Qt.ItemIsTristate)
self.set_hierarchy_model(item, tax_id, value)
else:
if item.parent():
item.hierarchy = ((item.parent().hierarchy, key), tax_id)
if not item.childCount() and not item.parent():
item.hierarchy = ((key, ), tax_id)
def download_gene_sets(self):
self.Error.clear()
# reset hierarchy widget state
self.hierarchy_widget.clear()
# clear data view
self.init_item_model()
# get all gene sets for selected organism
gene_sets = geneset.list_all(organism=self.tax_id)
# init progress bar
self.progress_bar = ProgressBar(self, iterations=len(gene_sets) * 100)
# status message
self.setStatusMessage('downloading sets')
worker = Worker(download_gene_sets, gene_sets, progress_callback=True)
worker.signals.progress.connect(self._progress_advance)
worker.signals.result.connect(self.on_gene_sets_download)
worker.signals.error.connect(self.handle_error)
# move download process to worker thread
self.threadpool.start(worker)
def display_gene_sets(self):
self.init_item_model()
self.hierarchy_widget.setDisabled(True)
only_selected_hier = self.get_hierarchies(only_selected=True)
# init progress bar
iterations = sum([
self.progress_bar_iterations[hier] for hier in only_selected_hier
])
self.progress_bar = ProgressBar(self, iterations=iterations)
self.setStatusMessage('displaying gene sets')
if not only_selected_hier:
self.progress_bar.finish()
self.setStatusMessage('')
self.hierarchy_widget.setDisabled(False)
return
# save setting on selected hierarchies
self.stored_selections = only_selected_hier
# save context
self.closeContext()
for selected_hierarchy in only_selected_hier:
self.threadpool.start(self.workers[selected_hierarchy])
self.openContext(self.tax_id)
def handle_error(self, ex):
self.progress_bar.finish()
self.setStatusMessage('')
if isinstance(ex, ConnectionError):
self.Error.cant_reach_host()
print(ex)
def set_selected_hierarchies(self):
iterator = QTreeWidgetItemIterator(self.hierarchy_widget,
QTreeWidgetItemIterator.All)
while iterator.value():
# note: if hierarchy value is not a tuple, then this is just top level qTreeWidgetItem that
# holds subcategories. We don't want to display all sets from category
if type(iterator.value().hierarchy) is not str:
if iterator.value().hierarchy in self.stored_selections:
iterator.value().setCheckState(0, Qt.Checked)
else:
iterator.value().setCheckState(0, Qt.Unchecked)
iterator += 1
# if no items are checked, we check first one at random
if len(self.get_hierarchies(only_selected=True)) == 0:
iterator = QTreeWidgetItemIterator(
self.hierarchy_widget, QTreeWidgetItemIterator.NotChecked)
while iterator.value():
if type(iterator.value().hierarchy) is not str:
iterator.value().setCheckState(0, Qt.Checked)
return
iterator += 1
def get_hierarchies(self, **kwargs):
""" return selected hierarchy
"""
only_selected = kwargs.get('only_selected', None)
sets_to_display = list()
if only_selected:
iterator = QTreeWidgetItemIterator(self.hierarchy_widget,
QTreeWidgetItemIterator.Checked)
else:
iterator = QTreeWidgetItemIterator(self.hierarchy_widget)
while iterator.value():
# note: if hierarchy value is not a tuple, then this is just top level qTreeWidgetItem that
# holds subcategories. We don't want to display all sets from category
if type(iterator.value().hierarchy) is not str:
if not only_selected:
sets_to_display.append(iterator.value().hierarchy)
else:
if not iterator.value().isDisabled():
sets_to_display.append(iterator.value().hierarchy)
iterator += 1
return sets_to_display
def commit(self):
selection_model = self.data_view.selectionModel()
if selection_model:
# genes_from_set = selection_model.selectedRows(GENES)
matched_genes = selection_model.selectedRows(MATCHED)
if matched_genes and self.input_genes:
genes = [
model_index.data(Qt.UserRole)
for model_index in matched_genes
]
output_genes = [
gene_name for gene_name in list(set.union(*genes))
]
self.num_of_sel_genes = len(output_genes)
self.update_info_box()
if self.use_attr_names:
selected = [
column for column in self.input_data.domain.attributes
if self.gene_id_attribute in column.attributes
and str(column.attributes[
self.gene_id_attribute]) in output_genes
]
domain = Domain(selected,
self.input_data.domain.class_vars,
self.input_data.domain.metas)
new_data = self.input_data.from_table(
domain, self.input_data)
self.Outputs.matched_genes.send(new_data)
else:
selected_rows = []
for row_index, row in enumerate(self.input_data):
gene_in_row = str(row[self.gene_id_column])
if gene_in_row in self.input_genes and gene_in_row in output_genes:
selected_rows.append(row_index)
if selected_rows:
selected = self.input_data[selected_rows]
else:
selected = None
self.Outputs.matched_genes.send(selected)
def setup_gui(self):
# control area
info_box = vBox(self.controlArea, 'Input info')
self.input_info = widgetLabel(info_box)
hierarchy_box = widgetBox(self.controlArea, "Entity Sets")
self.hierarchy_widget = QTreeWidget(self)
self.hierarchy_widget.setEditTriggers(QTreeView.NoEditTriggers)
self.hierarchy_widget.setHeaderLabels(HIERARCHY_HEADER_LABELS)
self.hierarchy_widget.itemClicked.connect(self.display_gene_sets)
hierarchy_box.layout().addWidget(self.hierarchy_widget)
self.commit_button = auto_commit(self.controlArea,
self,
"auto_commit",
"&Commit",
box=False)
# rubber(self.controlArea)
# main area
self.filter_proxy_model = QSortFilterProxyModel(self.data_view)
self.filter_proxy_model.setFilterKeyColumn(3)
self.data_view = QTreeView()
self.data_view.setModel(self.filter_proxy_model)
self.data_view.setAlternatingRowColors(True)
self.data_view.sortByColumn(2, Qt.DescendingOrder)
self.data_view.setSortingEnabled(True)
self.data_view.setSelectionMode(QTreeView.ExtendedSelection)
self.data_view.setEditTriggers(QTreeView.NoEditTriggers)
self.data_view.viewport().setMouseTracking(True)
self.data_view.setItemDelegateForColumn(
TERM, LinkStyledItemDelegate(self.data_view))
self.data_view.selectionModel().selectionChanged.connect(self.commit)
self.lineEdit_filter = lineEdit(self.mainArea, self, 'search_pattern',
'Filter gene sets:')
self.lineEdit_filter.setPlaceholderText('search pattern ...')
self.lineEdit_filter.textChanged.connect(
self.filter_proxy_model.setFilterRegExp)
self.mainArea.layout().addWidget(self.data_view)
def init_item_model(self):
if self.data_model:
self.data_model.clear()
self.filter_proxy_model.setSourceModel(None)
else:
self.data_model = QStandardItemModel()
self.data_model.setSortRole(Qt.UserRole)
self.data_model.setHorizontalHeaderLabels(DATA_HEADER_LABELS)
def sizeHint(self):
return QSize(1280, 960)
def get_gds_model(progress=lambda val: None):
"""
Initialize and return a GDS datasets model.
:param progress: A progress callback.
:rval tuple:
A tuple of (QStandardItemModel, geo.GDSInfo, [geo.GDS])
.. note::
The returned QStandardItemModel's thread affinity is set to
the GUI thread.
"""
progress(1)
info = geo.GDSInfo()
search_keys = ["dataset_id", "title", "platform_organism", "description"]
cache_dir = serverfiles.localpath(geo.DOMAIN)
gds_link = "http://www.ncbi.nlm.nih.gov/sites/GDSbrowser?acc={0}"
pm_link = "http://www.ncbi.nlm.nih.gov/pubmed/{0}"
gds_list = []
def is_cached(gds):
return os.path.exists(os.path.join(cache_dir, gds["dataset_id"]) +
".soft.gz")
def item(displayvalue, item_values={}):
item = QStandardItem()
item.setData(displayvalue, Qt.DisplayRole)
for role, value in item_values.items():
item.setData(value, role)
return item
def gds_to_row(gds):
#: Text for easier full search.
search_text = " | ".join([gds.get(key, "").lower()
for key in search_keys])
row = [
item(" " if is_cached(gds) else "",
{TextFilterRole: search_text}),
item(gds["dataset_id"],
{LinkRole: gds_link.format(gds["dataset_id"])}),
item(gds["title"]),
item(gds["platform_organism"]),
item(len(gds["samples"])),
item(gds["feature_count"]),
item(gds["gene_count"]),
item(len(gds["subsets"])),
item(gds.get("pubmed_id", ""),
{LinkRole: pm_link.format(gds["pubmed_id"])
if gds.get("pubmed_id")
else None})
]
return row
model = QStandardItemModel()
model.setHorizontalHeaderLabels(
["", "ID", "Title", "Organism", "Samples", "Features",
"Genes", "Subsets", "PubMedID"]
)
progress(20)
for gds in info.values():
model.appendRow(gds_to_row(gds))
gds_list.append(gds)
progress(50)
if QThread.currentThread() is not QCoreApplication.instance().thread():
model.moveToThread(QCoreApplication.instance().thread())
return model, info, gds_list
def __set_index(self, f):
# type: (Future) -> None
# set results from `list_remote` query.
assert QThread.currentThread() is self.thread()
assert f.done()
self.setBlocking(False)
self.setStatusMessage("")
allinfolocal = self.list_local()
try:
res = f.result()
except Exception:
log.exception("Error while fetching updated index")
if not allinfolocal:
self.Error.no_remote_datasets()
else:
self.Warning.only_local_datasets()
res = {}
allinforemote = res # type: Dict[Tuple[str, str], dict]
allkeys = set(allinfolocal)
if allinforemote is not None:
allkeys = allkeys | set(allinforemote)
allkeys = sorted(allkeys)
def info(file_path):
if file_path in allinforemote:
info = allinforemote[file_path]
else:
info = allinfolocal[file_path]
islocal = file_path in allinfolocal
isremote = file_path in allinforemote
outdated = islocal and isremote and (
allinforemote[file_path].get('version', '') !=
allinfolocal[file_path].get('version', ''))
islocal &= not outdated
prefix = os.path.join('', *file_path[:-1])
filename = file_path[-1]
return namespace(
prefix=prefix, filename=filename,
title=info.get("title", filename),
datetime=info.get("datetime", None),
description=info.get("description", None),
references=info.get("references", []),
seealso=info.get("seealso", []),
source=info.get("source", None),
year=info.get("year", None),
instances=info.get("instances", None),
variables=info.get("variables", None),
target=info.get("target", None),
missing=info.get("missing", None),
tags=info.get("tags", []),
size=info.get("size", None),
islocal=islocal,
outdated=outdated
)
model = QStandardItemModel(self)
model.setHorizontalHeaderLabels(HEADER)
current_index = -1
for i, file_path in enumerate(allkeys):
datainfo = info(file_path)
item1 = QStandardItem()
item1.setData(" " if datainfo.islocal else "", Qt.DisplayRole)
item1.setData(datainfo, Qt.UserRole)
item2 = QStandardItem(datainfo.title)
item3 = QStandardItem()
item3.setData(datainfo.size, Qt.DisplayRole)
item4 = QStandardItem()
item4.setData(datainfo.instances, Qt.DisplayRole)
item5 = QStandardItem()
item5.setData(datainfo.variables, Qt.DisplayRole)
item6 = QStandardItem()
item6.setData(datainfo.target, Qt.DisplayRole)
if datainfo.target:
item6.setIcon(variable_icon(datainfo.target))
item7 = QStandardItem()
item7.setData(", ".join(datainfo.tags) if datainfo.tags else "",
Qt.DisplayRole)
row = [item1, item2, item3, item4, item5, item6, item7]
model.appendRow(row)
if os.path.join(*file_path) == self.selected_id:
current_index = i
hs = self.view.header().saveState()
model_ = self.view.model().sourceModel()
self.view.model().setSourceModel(model)
self.view.header().restoreState(hs)
model_.deleteLater()
model_.setParent(None)
self.view.selectionModel().selectionChanged.connect(
self.__on_selection
)
# Update the info text
self.infolabel.setText(format_info(model.rowCount(), len(allinfolocal)))
if current_index != -1:
selmodel = self.view.selectionModel()
selmodel.select(
self.view.model().mapFromSource(model.index(current_index, 0)),
QItemSelectionModel.ClearAndSelect | QItemSelectionModel.Rows)
class OWRank(OWWidget):
name = "Rank"
description = "Rank and filter data features by their relevance."
icon = "icons/Rank.svg"
priority = 1102
buttons_area_orientation = Qt.Vertical
inputs = [("Data", Table, "setData"),
("Scorer", score.Scorer, "set_learner", widget.Multiple)]
outputs = [("Reduced Data", Table, widget.Default), ("Scores", Table)]
SelectNone, SelectAll, SelectManual, SelectNBest = range(4)
cls_default_selected = Setting({"Gain Ratio", "Gini Decrease"})
reg_default_selected = Setting({"Univariate Linear Regression", "RReliefF"})
selectMethod = Setting(SelectNBest)
nSelected = Setting(5)
auto_apply = Setting(True)
# Header state for discrete/continuous/no_class scores
headerState = Setting([None, None, None])
settings_version = 1
settingsHandler = DomainContextHandler()
selected_rows = ContextSetting([])
gain = inf_gain = gini = anova = chi2 = ulr = relief = rrelief = fcbc = True
_score_vars = ["gain", "inf_gain", "gini", "anova", "chi2", "relief",
"fcbc", "ulr", "rrelief"]
class Warning(OWWidget.Warning):
no_target_var = Msg("Data does not have a target variable")
class Error(OWWidget.Error):
invalid_type = Msg("Cannot handle target variable type {}")
inadequate_learner = Msg("{}")
def __init__(self):
super().__init__()
self.measure_scores = None
self.update_scores = True
self.usefulAttributes = []
self.learners = {}
self.labels = []
self.out_domain_desc = None
self.all_measures = SCORES
self.selectedMeasures = dict([(m.name, True) for m
in self.all_measures])
# Discrete (0) or continuous (1) class mode
self.rankMode = 0
self.data = None
self.discMeasures = [m for m in self.all_measures if
issubclass(DiscreteVariable, m.score.class_type)]
self.contMeasures = [m for m in self.all_measures if
issubclass(ContinuousVariable, m.score.class_type)]
self.score_checks = []
self.cls_scoring_box = gui.vBox(None, "Scoring for Classification")
self.reg_scoring_box = gui.vBox(None, "Scoring for Regression")
boxes = [self.cls_scoring_box] * 7 + [self.reg_scoring_box] * 2
for _score, var, box in zip(SCORES, self._score_vars, boxes):
check = gui.checkBox(
box, self, var, label=_score.name,
callback=lambda val=_score: self.measuresSelectionChanged(val))
self.score_checks.append(check)
self.score_stack = QStackedWidget(self)
self.score_stack.addWidget(self.cls_scoring_box)
self.score_stack.addWidget(self.reg_scoring_box)
self.score_stack.addWidget(QWidget())
self.controlArea.layout().addWidget(self.score_stack)
gui.rubber(self.controlArea)
selMethBox = gui.vBox(
self.controlArea, "Select Attributes", addSpace=True)
grid = QGridLayout()
grid.setContentsMargins(6, 0, 6, 0)
self.selectButtons = QButtonGroup()
self.selectButtons.buttonClicked[int].connect(self.setSelectMethod)
def button(text, buttonid, toolTip=None):
b = QRadioButton(text)
self.selectButtons.addButton(b, buttonid)
if toolTip is not None:
b.setToolTip(toolTip)
return b
b1 = button(self.tr("None"), OWRank.SelectNone)
b2 = button(self.tr("All"), OWRank.SelectAll)
b3 = button(self.tr("Manual"), OWRank.SelectManual)
b4 = button(self.tr("Best ranked:"), OWRank.SelectNBest)
s = gui.spin(selMethBox, self, "nSelected", 1, 100,
callback=self.nSelectedChanged)
grid.addWidget(b1, 0, 0)
grid.addWidget(b2, 1, 0)
grid.addWidget(b3, 2, 0)
grid.addWidget(b4, 3, 0)
grid.addWidget(s, 3, 1)
self.selectButtons.button(self.selectMethod).setChecked(True)
selMethBox.layout().addLayout(grid)
gui.auto_commit(selMethBox, self, "auto_apply", "Send", box=False)
# Discrete, continuous and no_class table views are stacked
self.ranksViewStack = QStackedLayout()
self.mainArea.layout().addLayout(self.ranksViewStack)
self.discRanksView = QTableView()
self.ranksViewStack.addWidget(self.discRanksView)
self.discRanksView.setSelectionBehavior(QTableView.SelectRows)
self.discRanksView.setSelectionMode(QTableView.MultiSelection)
self.discRanksView.setSortingEnabled(True)
self.discRanksLabels = ["#"] + [m.shortname for m in self.discMeasures]
self.discRanksModel = QStandardItemModel(self)
self.discRanksModel.setHorizontalHeaderLabels(self.discRanksLabels)
self.discRanksProxyModel = MySortProxyModel(self)
self.discRanksProxyModel.setSourceModel(self.discRanksModel)
self.discRanksView.setModel(self.discRanksProxyModel)
self.discRanksView.setColumnWidth(0, 20)
self.discRanksView.selectionModel().selectionChanged.connect(
self.commit
)
self.discRanksView.pressed.connect(self.onSelectItem)
self.discRanksView.horizontalHeader().sectionClicked.connect(
self.headerClick
)
self.discRanksView.verticalHeader().sectionClicked.connect(
self.onSelectItem
)
if self.headerState[0] is not None:
self.discRanksView.horizontalHeader().restoreState(
self.headerState[0])
self.contRanksView = QTableView()
self.ranksViewStack.addWidget(self.contRanksView)
self.contRanksView.setSelectionBehavior(QTableView.SelectRows)
self.contRanksView.setSelectionMode(QTableView.MultiSelection)
self.contRanksView.setSortingEnabled(True)
self.contRanksLabels = ["#"] + [m.shortname for m in self.contMeasures]
self.contRanksModel = QStandardItemModel(self)
self.contRanksModel.setHorizontalHeaderLabels(self.contRanksLabels)
self.contRanksProxyModel = MySortProxyModel(self)
self.contRanksProxyModel.setSourceModel(self.contRanksModel)
self.contRanksView.setModel(self.contRanksProxyModel)
self.contRanksView.setColumnWidth(0, 20)
self.contRanksView.selectionModel().selectionChanged.connect(
self.commit
)
self.contRanksView.pressed.connect(self.onSelectItem)
self.contRanksView.horizontalHeader().sectionClicked.connect(
self.headerClick
)
self.contRanksView.verticalHeader().sectionClicked.connect(
self.onSelectItem
)
if self.headerState[1] is not None:
self.contRanksView.horizontalHeader().restoreState(
self.headerState[1])
self.noClassRanksView = QTableView()
self.ranksViewStack.addWidget(self.noClassRanksView)
self.noClassRanksView.setSelectionBehavior(QTableView.SelectRows)
self.noClassRanksView.setSelectionMode(QTableView.MultiSelection)
self.noClassRanksView.setSortingEnabled(True)
self.noClassRanksLabels = ["#"]
self.noClassRanksModel = QStandardItemModel(self)
self.noClassRanksModel.setHorizontalHeaderLabels(self.noClassRanksLabels)
self.noClassRanksProxyModel = MySortProxyModel(self)
self.noClassRanksProxyModel.setSourceModel(self.noClassRanksModel)
self.noClassRanksView.setModel(self.noClassRanksProxyModel)
self.noClassRanksView.setColumnWidth(0, 20)
self.noClassRanksView.selectionModel().selectionChanged.connect(
self.commit
)
self.noClassRanksView.pressed.connect(self.onSelectItem)
self.noClassRanksView.horizontalHeader().sectionClicked.connect(
self.headerClick
)
self.noClassRanksView.verticalHeader().sectionClicked.connect(
self.onSelectItem
)
if self.headerState[2] is not None:
self.noClassRanksView.horizontalHeader().restoreState(
self.headerState[2])
# Switch the current view to Discrete
self.switchRanksMode(0)
self.resetInternals()
self.updateDelegates()
self.updateVisibleScoreColumns()
self.resize(690, 500)
self.measure_scores = table((len(self.measures), 0), None)
def switchRanksMode(self, index):
"""
Switch between discrete/continuous/no_class mode
"""
self.rankMode = index
self.ranksViewStack.setCurrentIndex(index)
if index == 0:
self.ranksView = self.discRanksView
self.ranksModel = self.discRanksModel
self.ranksProxyModel = self.discRanksProxyModel
self.measures = self.discMeasures
self.selected_checks = self.cls_default_selected
self.reg_scoring_box.setSizePolicy(QSizePolicy.Ignored,
QSizePolicy.Ignored)
self.cls_scoring_box.setSizePolicy(QSizePolicy.Expanding,
QSizePolicy.Expanding)
elif index == 1:
self.ranksView = self.contRanksView
self.ranksModel = self.contRanksModel
self.ranksProxyModel = self.contRanksProxyModel
self.measures = self.contMeasures
self.selected_checks = self.reg_default_selected
self.cls_scoring_box.setSizePolicy(QSizePolicy.Ignored,
QSizePolicy.Ignored)
self.reg_scoring_box.setSizePolicy(QSizePolicy.Expanding,
QSizePolicy.Expanding)
else:
self.ranksView = self.noClassRanksView
self.ranksModel = self.noClassRanksModel
self.ranksProxyModel = self.noClassRanksProxyModel
self.measures = []
self.selected_checks = set()
self.reg_scoring_box.setSizePolicy(QSizePolicy.Ignored,
QSizePolicy.Ignored)
self.cls_scoring_box.setSizePolicy(QSizePolicy.Ignored,
QSizePolicy.Ignored)
shape = (len(self.measures) + len(self.learners), 0)
self.measure_scores = table(shape, None)
self.update_scores = False
for check, score in zip(self.score_checks, SCORES):
check.setChecked(score.name in self.selected_checks)
self.update_scores = True
self.score_stack.setCurrentIndex(index)
self.updateVisibleScoreColumns()
@check_sql_input
def setData(self, data):
self.closeContext()
self.clear_messages()
self.resetInternals()
self.data = data
self.switchRanksMode(0)
if self.data is not None:
domain = self.data.domain
attrs = domain.attributes
self.usefulAttributes = [attr for attr in attrs
if attr.is_discrete or attr.is_continuous]
if domain.has_continuous_class:
self.switchRanksMode(1)
elif not domain.class_var:
self.Warning.no_target_var()
self.switchRanksMode(2)
elif not domain.has_discrete_class:
self.Error.invalid_type(type(domain.class_var).__name__)
if issparse(self.data.X): # keep only measures supporting sparse data
self.measures = [m for m in self.measures
if m.score.supports_sparse_data]
self.ranksModel.setRowCount(len(attrs))
for i, a in enumerate(attrs):
if a.is_discrete:
v = len(a.values)
else:
v = "C"
item = ScoreValueItem()
item.setData(v, Qt.DisplayRole)
self.ranksModel.setItem(i, 0, item)
item = QStandardItem(a.name)
item.setData(gui.attributeIconDict[a], Qt.DecorationRole)
self.ranksModel.setVerticalHeaderItem(i, item)
shape = (len(self.measures) + len(self.learners), len(attrs))
self.measure_scores = table(shape, None)
self.updateScores()
else:
self.send("Scores", None)
self.selected_rows = []
self.openContext(data)
self.selectMethodChanged()
self.commit()
def get_selection(self):
selection = self.ranksView.selectionModel().selection()
return list(set(ind.row() for ind in selection.indexes()))
def set_learner(self, learner, lid=None):
if learner is None and lid is not None:
del self.learners[lid]
elif learner is not None:
self.learners[lid] = score_meta(
learner.name,
learner.name,
learner
)
attrs_len = 0 if not self.data else len(self.data.domain.attributes)
shape = (len(self.learners), attrs_len)
self.measure_scores = self.measure_scores[:len(self.measures)]
self.measure_scores += table(shape, None)
self.contRanksModel.setHorizontalHeaderLabels(self.contRanksLabels)
self.discRanksModel.setHorizontalHeaderLabels(self.discRanksLabels)
self.noClassRanksModel.setHorizontalHeaderLabels(
self.noClassRanksLabels)
measures_mask = [False] * len(self.measures)
measures_mask += [True for _ in self.learners]
self.updateScores(measures_mask)
self.commit()
def updateScores(self, measuresMask=None):
"""
Update the current computed scores.
If `measuresMask` is given it must be an list of bool values
indicating what measures should be recomputed.
"""
if not self.data:
return
if self.data.has_missing():
self.information("Missing values have been imputed.")
measures = self.measures + [v for k, v in self.learners.items()]
if measuresMask is None:
# Update all selected measures
measuresMask = [self.selectedMeasures.get(m.name)
for m in self.measures]
measuresMask = measuresMask + [v.name for k, v in
self.learners.items()]
data = self.data
learner_col = len(self.measures)
if len(measuresMask) <= len(self.measures) or \
measuresMask[len(self.measures)]:
self.labels = []
self.Error.inadequate_learner.clear()
self.setStatusMessage("Running")
with self.progressBar():
n_measure_update = len([x for x in measuresMask if x is not False])
count = 0
for index, (meas, mask) in enumerate(zip(measures, measuresMask)):
if not mask:
continue
self.progressBarSet(90 * count / n_measure_update)
count += 1
if index < len(self.measures):
estimator = meas.score()
try:
self.measure_scores[index] = estimator(data)
except ValueError:
self.measure_scores[index] = []
for attr in data.domain.attributes:
try:
self.measure_scores[index].append(
estimator(data, attr))
except ValueError:
self.measure_scores[index].append(None)
else:
learner = meas.score
if isinstance(learner, Learner) and \
not learner.check_learner_adequacy(self.data.domain):
self.Error.inadequate_learner(
learner.learner_adequacy_err_msg)
scores = table((1, len(data.domain.attributes)))
else:
scores = meas.score.score_data(data)
for i, row in enumerate(scores):
self.labels.append(meas.shortname + str(i + 1))
if len(self.measure_scores) > learner_col:
self.measure_scores[learner_col] = row
else:
self.measure_scores.append(row)
learner_col += 1
self.progressBarSet(90)
self.contRanksModel.setHorizontalHeaderLabels(
self.contRanksLabels + self.labels
)
self.discRanksModel.setHorizontalHeaderLabels(
self.discRanksLabels + self.labels
)
self.noClassRanksModel.setHorizontalHeaderLabels(
self.noClassRanksLabels + self.labels
)
self.updateRankModel(measuresMask)
self.ranksProxyModel.invalidate()
self.selectMethodChanged()
self.send("Scores", self.create_scores_table(self.labels))
self.setStatusMessage("")
def updateRankModel(self, measuresMask):
"""
Update the rankModel.
"""
values = []
diff = len(self.measure_scores) - len(measuresMask)
if len(measuresMask):
measuresMask += [measuresMask[-1]] * diff
for i in range(self.ranksModel.columnCount() - 1,
len(self.measure_scores), -1):
self.ranksModel.removeColumn(i)
for i, (scores, m) in enumerate(zip(self.measure_scores, measuresMask)):
if not m and self.ranksModel.item(0, i + 1):
values.append([])
continue
values_one = []
for j, _score in enumerate(scores):
values_one.append(_score)
item = self.ranksModel.item(j, i + 1)
if not item:
item = ScoreValueItem()
self.ranksModel.setItem(j, i + 1, item)
item.setData(_score, Qt.DisplayRole)
values.append(values_one)
for i, (vals, m) in enumerate(zip(values, measuresMask)):
if not m:
continue
valid_vals = [v for v in vals if v is not None]
if valid_vals:
vmin, vmax = min(valid_vals), max(valid_vals)
for j, v in enumerate(vals):
if v is not None:
# Set the bar ratio role for i-th measure.
ratio = float((v - vmin) / ((vmax - vmin) or 1))
item = self.ranksModel.item(j, i + 1)
item.setData(ratio, gui.BarRatioRole)
self.ranksView.setColumnWidth(0, 20)
self.ranksView.resizeRowsToContents()
def resetInternals(self):
self.data = None
self.usefulAttributes = []
self.ranksModel.setRowCount(0)
def onSelectItem(self, index):
"""
Called when the user selects/unselects an item in the table view.
"""
self.selectMethod = OWRank.SelectManual # Manual
self.selectButtons.button(self.selectMethod).setChecked(True)
self.commit()
def setSelectMethod(self, method):
if self.selectMethod != method:
self.selectMethod = method
self.selectButtons.button(method).setChecked(True)
self.selectMethodChanged()
def selectMethodChanged(self):
self.autoSelection()
self.ranksView.setFocus()
def nSelectedChanged(self):
self.selectMethod = OWRank.SelectNBest
self.selectButtons.button(self.selectMethod).setChecked(True)
self.selectMethodChanged()
def autoSelection(self):
selModel = self.ranksView.selectionModel()
rowCount = self.ranksModel.rowCount()
columnCount = self.ranksModel.columnCount()
model = self.ranksProxyModel
if self.selectMethod == OWRank.SelectNone:
selection = QItemSelection()
elif self.selectMethod == OWRank.SelectAll:
selection = QItemSelection(
model.index(0, 0),
model.index(rowCount - 1, columnCount - 1)
)
elif self.selectMethod == OWRank.SelectNBest:
nSelected = min(self.nSelected, rowCount)
selection = QItemSelection(
model.index(0, 0),
model.index(nSelected - 1, columnCount - 1)
)
else:
selection = QItemSelection()
if len(self.selected_rows):
selection = QItemSelection()
for row in self.selected_rows:
selection.append(QItemSelectionRange(
model.index(row, 0), model.index(row, columnCount - 1)))
selModel.select(selection, QItemSelectionModel.ClearAndSelect)
def headerClick(self, index):
if index >= 1 and self.selectMethod == OWRank.SelectNBest:
# Reselect the top ranked attributes
self.autoSelection()
# Store the header states
disc = bytes(self.discRanksView.horizontalHeader().saveState())
cont = bytes(self.contRanksView.horizontalHeader().saveState())
no_class = bytes(self.noClassRanksView.horizontalHeader().saveState())
self.headerState = [disc, cont, no_class]
def measuresSelectionChanged(self, measure):
"""Measure selection has changed. Update column visibility.
"""
checked = self.selectedMeasures[measure.name]
self.selectedMeasures[measure.name] = not checked
if not checked:
self.selected_checks.add(measure.name)
elif measure.name in self.selected_checks:
self.selected_checks.remove(measure.name)
measures_mask = [False] * len(self.measures)
measures_mask += [False for _ in self.learners]
# Update scores for shown column if they are not yet computed.
if measure in self.measures and self.measure_scores:
index = self.measures.index(measure)
if all(s is None for s in self.measure_scores[index]):
measures_mask[index] = True
if self.update_scores:
self.updateScores(measures_mask)
self.updateVisibleScoreColumns()
def updateVisibleScoreColumns(self):
"""
Update the visible columns of the scores view.
"""
for i, measure in enumerate(self.measures):
shown = self.selectedMeasures.get(measure.name)
self.ranksView.setColumnHidden(i + 1, not shown)
self.ranksView.setColumnWidth(i + 1, 100)
index = self.ranksView.horizontalHeader().sortIndicatorSection()
if self.ranksView.isColumnHidden(index):
self.headerState[self.rankMode] = None
if self.headerState[self.rankMode] is None:
def get_sort_by_col(measures, selected_measures):
cols = [i + 1 for i, m in enumerate(measures) if
m.name in selected_measures]
return cols[0] if cols else len(measures) + 1
col = get_sort_by_col(self.measures, self.selected_checks)
self.ranksView.sortByColumn(col, Qt.DescendingOrder)
self.autoSelection()
def updateDelegates(self):
self.contRanksView.setItemDelegate(gui.ColoredBarItemDelegate(self))
self.discRanksView.setItemDelegate(gui.ColoredBarItemDelegate(self))
self.noClassRanksView.setItemDelegate(gui.ColoredBarItemDelegate(self))
def send_report(self):
if not self.data:
return
self.report_domain("Input", self.data.domain)
self.report_table("Ranks", self.ranksView, num_format="{:.3f}")
if self.out_domain_desc is not None:
self.report_items("Output", self.out_domain_desc)
def commit(self):
self.selected_rows = self.get_selection()
if self.data and len(self.data.domain.attributes) == len(
self.selected_rows):
self.selectMethod = OWRank.SelectAll
self.selectButtons.button(self.selectMethod).setChecked(True)
selected = self.selectedAttrs()
if not self.data or not selected:
self.send("Reduced Data", None)
self.out_domain_desc = None
else:
data = Table(Domain(selected, self.data.domain.class_var,
self.data.domain.metas), self.data)
self.send("Reduced Data", data)
self.out_domain_desc = report.describe_domain(data.domain)
def selectedAttrs(self):
if self.data:
inds = self.ranksView.selectionModel().selectedRows(0)
source = self.ranksProxyModel.mapToSource
inds = map(source, inds)
inds = [ind.row() for ind in inds]
return [self.data.domain.attributes[i] for i in inds]
else:
return []
def create_scores_table(self, labels):
indices = [i for i, m in enumerate(self.measures)
if self.selectedMeasures.get(m.name, False)]
measures = [s.name for s in self.measures if
self.selectedMeasures.get(s.name, False)]
measures += [label for label in labels]
if not measures:
return None
features = [ContinuousVariable(s) for s in measures]
metas = [StringVariable("Feature name")]
domain = Domain(features, metas=metas)
scores = np.array([row for i, row in enumerate(self.measure_scores)
if i in indices or i >= len(self.measures)]).T
feature_names = np.array([a.name for a in self.data.domain.attributes])
# Reshape to 2d array as Table does not like 1d arrays
feature_names = feature_names[:, None]
new_table = Table(domain, scores, metas=feature_names)
new_table.name = "Feature Scores"
return new_table
@classmethod
def migrate_settings(cls, settings, version):
if not version:
# Before fc5caa1e1d716607f1f5c4e0b0be265c23280fa0
# headerState had length 2
headerState = settings.get("headerState", None)
if headerState is not None and \
isinstance(headerState, tuple) and \
len(headerState) < 3:
headerState = (list(headerState) + [None] * 3)[:3]
settings["headerState"] = headerState
class OWGeneSets(OWWidget):
name = "Gene Sets"
description = ""
icon = "icons/OWGeneSets.svg"
priority = 9
want_main_area = True
COUNT, GENES, CATEGORY, TERM = range(4)
DATA_HEADER_LABELS = ["Count", 'Genes In Set', 'Category', 'Term']
organism = Setting(None, schema_only=True)
stored_gene_sets_selection = Setting([], schema_only=True)
selected_rows = Setting([], schema_only=True)
custom_gene_set_indicator = Setting(None, schema_only=True)
min_count = Setting(5)
use_min_count = Setting(True)
auto_commit = Setting(True)
class Inputs:
genes = Input("Data", Table)
custom_sets = Input('Custom Gene Sets', Table)
class Outputs:
matched_genes = Output("Matched Genes", Table)
class Information(OWWidget.Information):
pass
class Warning(OWWidget.Warning):
all_sets_filtered = Msg('All sets were filtered out.')
class Error(OWWidget.Error):
organism_mismatch = Msg('Organism in input data and custom gene sets does not match')
missing_annotation = Msg(ERROR_ON_MISSING_ANNOTATION)
missing_gene_id = Msg(ERROR_ON_MISSING_GENE_ID)
missing_tax_id = Msg(ERROR_ON_MISSING_TAX_ID)
cant_reach_host = Msg("Host orange.biolab.si is unreachable.")
cant_load_organisms = Msg("No available organisms, please check your connection.")
def __init__(self):
super().__init__()
# commit
self.commit_button = None
# progress bar
self.progress_bar = None
self.progress_bar_iterations = None
# data
self.input_data = None
self.input_genes = []
self.tax_id = None
self.use_attr_names = None
self.gene_id_attribute = None
self.gene_id_column = None
# custom gene sets
self.custom_data = None
self.feature_model = DomainModel(valid_types=(DiscreteVariable, StringVariable))
self.custom_gs_col_box = None
self.gs_label_combobox = None
self.custom_tax_id = None
self.custom_use_attr_names = None
self.custom_gene_id_attribute = None
self.custom_gene_id_column = None
self.num_of_custom_sets = None
# Gene Sets widget
self.gs_widget = None
# info box
self.input_info = None
self.num_of_sel_genes = 0
# filter
self.line_edit_filter = None
self.search_pattern = ''
self.organism_select_combobox = None
# data model view
self.data_view = None
self.data_model = None
# gene matcher NCBI
self.gene_matcher = None
# filter proxy model
self.filter_proxy_model = None
# hierarchy widget
self.hierarchy_widget = None
self.hierarchy_state = None
# spinbox
self.spin_widget = None
# threads
self.threadpool = QThreadPool(self)
self.workers = None
self._task = None # type: Optional[Task]
self._executor = ThreadExecutor()
# gui
self.setup_gui()
def __reset_widget_state(self):
self.update_info_box()
# clear data view
self.init_item_model()
# reset filters
self.setup_filter_model()
def cancel(self):
"""
Cancel the current task (if any).
"""
if self._task is not None:
self._task.cancel()
assert self._task.future.done()
# disconnect the `_task_finished` slot
self._task.watcher.done.disconnect(self._init_gene_sets_finished)
self._task = None
@Slot()
def progress_advance(self):
# GUI should be updated in main thread. That's why we are calling advance method here
if self.progress_bar:
self.progress_bar.advance()
def __get_input_genes(self):
self.input_genes = []
if self.use_attr_names:
for variable in self.input_data.domain.attributes:
self.input_genes.append(str(variable.attributes.get(self.gene_id_attribute, '?')))
else:
genes, _ = self.input_data.get_column_view(self.gene_id_column)
self.input_genes = [str(g) for g in genes]
def handle_custom_gene_sets(self, select_customs_flag=False):
if self.custom_gene_set_indicator:
if self.custom_data is not None and self.custom_gene_id_column is not None:
if self.__check_organism_mismatch():
# self.gs_label_combobox.setDisabled(True)
self.Error.organism_mismatch()
self.gs_widget.update_gs_hierarchy()
return
if isinstance(self.custom_gene_set_indicator, DiscreteVariable):
labels = self.custom_gene_set_indicator.values
gene_sets_names = [
labels[int(idx)] for idx in self.custom_data.get_column_view(self.custom_gene_set_indicator)[0]
]
else:
gene_sets_names, _ = self.custom_data.get_column_view(self.custom_gene_set_indicator)
self.num_of_custom_sets = len(set(gene_sets_names))
gene_names, _ = self.custom_data.get_column_view(self.custom_gene_id_column)
hierarchy_title = (self.custom_data.name if self.custom_data.name else 'Custom sets',)
try:
self.gs_widget.add_custom_sets(
gene_sets_names,
gene_names,
hierarchy_title=hierarchy_title,
select_customs_flag=select_customs_flag,
)
except geneset.GeneSetException:
pass
# self.gs_label_combobox.setDisabled(False)
else:
self.gs_widget.update_gs_hierarchy()
self.update_info_box()
def update_tree_view(self):
self.init_gene_sets()
def invalidate(self):
# clear
self.__reset_widget_state()
self.update_info_box()
if self.input_data is not None:
# setup
self.__get_input_genes()
self.update_tree_view()
def __check_organism_mismatch(self):
""" Check if organisms from different inputs match.
:return: True if there is a mismatch
"""
if self.tax_id is not None and self.custom_tax_id is not None:
return self.tax_id != self.custom_tax_id
return False
def __get_reference_genes(self):
self.reference_genes = []
if self.reference_attr_names:
for variable in self.reference_data.domain.attributes:
self.reference_genes.append(str(variable.attributes.get(self.reference_gene_id_attribute, '?')))
else:
genes, _ = self.reference_data.get_column_view(self.reference_gene_id_column)
self.reference_genes = [str(g) for g in genes]
@Inputs.custom_sets
def handle_custom_input(self, data):
self.Error.clear()
self.__reset_widget_state()
self.custom_data = None
self.custom_tax_id = None
self.custom_use_attr_names = None
self.custom_gene_id_attribute = None
self.custom_gene_id_column = None
self.feature_model.set_domain(None)
if data:
self.custom_data = data
self.feature_model.set_domain(self.custom_data.domain)
self.custom_tax_id = str(self.custom_data.attributes.get(TAX_ID, None))
self.custom_use_attr_names = self.custom_data.attributes.get(GENE_AS_ATTRIBUTE_NAME, None)
self.custom_gene_id_attribute = self.custom_data.attributes.get(GENE_ID_ATTRIBUTE, None)
self.custom_gene_id_column = self.custom_data.attributes.get(GENE_ID_COLUMN, None)
if self.gs_label_combobox is None:
self.gs_label_combobox = comboBox(
self.custom_gs_col_box,
self,
"custom_gene_set_indicator",
sendSelectedValue=True,
model=self.feature_model,
callback=self.on_gene_set_indicator_changed,
)
self.custom_gs_col_box.show()
if self.custom_gene_set_indicator in self.feature_model:
index = self.feature_model.indexOf(self.custom_gene_set_indicator)
self.custom_gene_set_indicator = self.feature_model[index]
else:
self.custom_gene_set_indicator = self.feature_model[0]
else:
self.custom_gs_col_box.hide()
self.gs_widget.clear_custom_sets()
self.handle_custom_gene_sets(select_customs_flag=self.custom_gene_set_indicator is not None)
self.invalidate()
@Inputs.genes
def handle_genes_input(self, data):
self.Error.clear()
self.__reset_widget_state()
# clear output
self.Outputs.matched_genes.send(None)
# clear input values
self.input_genes = []
self.input_data = None
self.tax_id = None
self.use_attr_names = None
self.gene_id_attribute = None
self.gs_widget.clear()
self.gs_widget.clear_gene_sets()
self.update_info_box()
if data:
self.input_data = data
self.tax_id = str(self.input_data.attributes.get(TAX_ID, None))
self.use_attr_names = self.input_data.attributes.get(GENE_AS_ATTRIBUTE_NAME, None)
self.gene_id_attribute = self.input_data.attributes.get(GENE_ID_ATTRIBUTE, None)
self.gene_id_column = self.input_data.attributes.get(GENE_ID_COLUMN, None)
self.update_info_box()
if not (
self.use_attr_names is not None and ((self.gene_id_attribute is None) ^ (self.gene_id_column is None))
):
if self.tax_id is None:
self.Error.missing_annotation()
return
self.Error.missing_gene_id()
return
elif self.tax_id is None:
self.Error.missing_tax_id()
return
if self.__check_organism_mismatch():
self.Error.organism_mismatch()
return
self.gs_widget.load_gene_sets(self.tax_id)
# if input data change, we need to refresh custom sets
if self.custom_data:
self.gs_widget.clear_custom_sets()
self.handle_custom_gene_sets()
self.invalidate()
def update_info_box(self):
info_string = ''
if self.input_genes:
info_string += '{} unique gene names on input.\n'.format(len(self.input_genes))
info_string += '{} genes on output.\n'.format(self.num_of_sel_genes)
else:
if self.input_data:
if not any([self.gene_id_column, self.gene_id_attribute]):
info_string += 'Input data with incorrect meta data.\nUse Gene Name Matcher widget.'
else:
info_string += 'No data on input.\n'
if self.custom_data:
info_string += '{} marker genes in {} sets\n'.format(self.custom_data.X.shape[0], self.num_of_custom_sets)
self.input_info.setText(info_string)
def create_partial(self):
return partial(
self.set_items,
self.gs_widget.gs_object,
self.stored_gene_sets_selection,
set(self.input_genes),
self.callback,
)
def callback(self):
if self._task.cancelled:
raise KeyboardInterrupt()
if self.progress_bar:
methodinvoke(self, "progress_advance")()
def init_gene_sets(self):
if self._task is not None:
self.cancel()
assert self._task is None
self._task = Task()
self.init_item_model()
# save setting on selected hierarchies
self.stored_gene_sets_selection = self.gs_widget.get_hierarchies(only_selected=True)
f = self.create_partial()
progress_iterations = sum(
(
len(g_set)
for hier, g_set in self.gs_widget.gs_object.map_hierarchy_to_sets().items()
if hier in self.stored_gene_sets_selection
)
)
self.progress_bar = ProgressBar(self, iterations=progress_iterations)
self._task.future = self._executor.submit(f)
self._task.watcher = FutureWatcher(self._task.future)
self._task.watcher.done.connect(self._init_gene_sets_finished)
@Slot(concurrent.futures.Future)
def _init_gene_sets_finished(self, f):
assert self.thread() is QThread.currentThread()
assert threading.current_thread() == threading.main_thread()
assert self._task is not None
assert self._task.future is f
assert f.done()
self._task = None
self.progress_bar.finish()
self.setStatusMessage('')
try:
results = f.result() # type: list
[self.data_model.appendRow(model_item) for model_item in results]
self.filter_proxy_model.setSourceModel(self.data_model)
self.data_view.selectionModel().selectionChanged.connect(self.commit)
self.filter_data_view()
self.set_selection()
self.update_info_box()
except Exception as ex:
print(ex)
def create_filters(self):
search_term = self.search_pattern.lower().strip().split()
filters = [
FilterProxyModel.Filter(
self.TERM, Qt.DisplayRole, lambda value: all(fs in value.lower() for fs in search_term)
)
]
if self.use_min_count:
filters.append(FilterProxyModel.Filter(self.COUNT, Qt.DisplayRole, lambda value: value >= self.min_count))
return filters
def filter_data_view(self):
filter_proxy = self.filter_proxy_model # type: FilterProxyModel
model = filter_proxy.sourceModel() # type: QStandardItemModel
if isinstance(model, QStandardItemModel):
# apply filtering rules
filter_proxy.set_filters(self.create_filters())
if model.rowCount() and not filter_proxy.rowCount():
self.Warning.all_sets_filtered()
else:
self.Warning.clear()
def set_selection(self):
if len(self.selected_rows):
view = self.data_view
model = self.data_model
row_model_indexes = [model.indexFromItem(model.item(i)) for i in self.selected_rows]
proxy_rows = [self.filter_proxy_model.mapFromSource(i).row() for i in row_model_indexes]
if model.rowCount() <= self.selected_rows[-1]:
return
header_count = view.header().count() - 1
selection = QItemSelection()
for row_index in proxy_rows:
selection.append(
QItemSelectionRange(
self.filter_proxy_model.index(row_index, 0),
self.filter_proxy_model.index(row_index, header_count),
)
)
view.selectionModel().select(selection, QItemSelectionModel.ClearAndSelect)
def commit(self):
selection_model = self.data_view.selectionModel()
if selection_model:
selection = selection_model.selectedRows(self.COUNT)
self.selected_rows = [self.filter_proxy_model.mapToSource(sel).row() for sel in selection]
if selection and self.input_genes:
genes = [model_index.data(Qt.UserRole) for model_index in selection]
output_genes = [gene_name for gene_name in list(set.union(*genes))]
self.num_of_sel_genes = len(output_genes)
self.update_info_box()
if self.use_attr_names:
selected = [
column
for column in self.input_data.domain.attributes
if self.gene_id_attribute in column.attributes
and str(column.attributes[self.gene_id_attribute]) in output_genes
]
domain = Domain(selected, self.input_data.domain.class_vars, self.input_data.domain.metas)
new_data = self.input_data.from_table(domain, self.input_data)
self.Outputs.matched_genes.send(new_data)
else:
# create filter from selected column for genes
only_known = table_filter.FilterStringList(self.gene_id_column, output_genes)
# apply filter to the data
data_table = table_filter.Values([only_known])(self.input_data)
self.Outputs.matched_genes.send(data_table)
def assign_delegates(self):
self.data_view.setItemDelegateForColumn(self.GENES, NumericalColumnDelegate(self))
self.data_view.setItemDelegateForColumn(self.COUNT, NumericalColumnDelegate(self))
def setup_filter_model(self):
self.filter_proxy_model = FilterProxyModel()
self.filter_proxy_model.setFilterKeyColumn(self.TERM)
self.data_view.setModel(self.filter_proxy_model)
def setup_filter_area(self):
h_layout = QHBoxLayout()
h_layout.setSpacing(100)
h_widget = widgetBox(self.mainArea, orientation=h_layout)
spin(
h_widget,
self,
'min_count',
0,
1000,
label='Count',
tooltip='Minimum genes count',
checked='use_min_count',
callback=self.filter_data_view,
callbackOnReturn=True,
checkCallback=self.filter_data_view,
)
self.line_edit_filter = lineEdit(h_widget, self, 'search_pattern')
self.line_edit_filter.setPlaceholderText('Filter gene sets ...')
self.line_edit_filter.textChanged.connect(self.filter_data_view)
def on_gene_set_indicator_changed(self):
# self._handle_future_model()
self.gs_widget.clear_custom_sets()
self.handle_custom_gene_sets()
self.invalidate()
def setup_control_area(self):
# Control area
self.input_info = widgetLabel(widgetBox(self.controlArea, "Info", addSpace=True), 'No data on input.\n')
self.custom_gs_col_box = box = vBox(self.controlArea, 'Custom Gene Set Term Column')
box.hide()
gene_sets_box = widgetBox(self.controlArea, "Gene Sets")
self.gs_widget = GeneSetsSelection(gene_sets_box, self, 'stored_gene_sets_selection')
self.gs_widget.hierarchy_tree_widget.itemClicked.connect(self.update_tree_view)
self.commit_button = auto_commit(self.controlArea, self, "auto_commit", "&Commit", box=False)
def setup_gui(self):
# control area
self.setup_control_area()
# main area
self.data_view = QTreeView()
self.setup_filter_model()
self.setup_filter_area()
self.data_view.setAlternatingRowColors(True)
self.data_view.sortByColumn(self.COUNT, Qt.DescendingOrder)
self.data_view.setSortingEnabled(True)
self.data_view.setSelectionMode(QTreeView.ExtendedSelection)
self.data_view.setEditTriggers(QTreeView.NoEditTriggers)
self.data_view.viewport().setMouseTracking(False)
self.data_view.setItemDelegateForColumn(self.TERM, LinkStyledItemDelegate(self.data_view))
self.mainArea.layout().addWidget(self.data_view)
self.data_view.header().setSectionResizeMode(QHeaderView.ResizeToContents)
self.assign_delegates()
@staticmethod
def set_items(gene_sets, sets_to_display, genes, callback):
model_items = []
if not genes:
return
for gene_set in sorted(gene_sets):
if gene_set.hierarchy not in sets_to_display:
continue
callback()
matched_set = gene_set.genes & genes
if len(matched_set) > 0:
category_column = QStandardItem()
term_column = QStandardItem()
count_column = QStandardItem()
genes_column = QStandardItem()
category_column.setData(", ".join(gene_set.hierarchy), Qt.DisplayRole)
term_column.setData(gene_set.name, Qt.DisplayRole)
term_column.setData(gene_set.name, Qt.ToolTipRole)
term_column.setData(gene_set.link, LinkRole)
term_column.setForeground(QColor(Qt.blue))
count_column.setData(matched_set, Qt.UserRole)
count_column.setData(len(matched_set), Qt.DisplayRole)
genes_column.setData(len(gene_set.genes), Qt.DisplayRole)
genes_column.setData(
set(gene_set.genes), Qt.UserRole
) # store genes to get then on output on selection
model_items.append([count_column, genes_column, category_column, term_column])
return model_items
def init_item_model(self):
if self.data_model:
self.data_model.clear()
self.setup_filter_model()
else:
self.data_model = QStandardItemModel()
self.data_model.setSortRole(Qt.UserRole)
self.data_model.setHorizontalHeaderLabels(self.DATA_HEADER_LABELS)
def sizeHint(self):
return QSize(1280, 960)
class OWTestLearners(OWWidget):
name = "Test & Score"
description = "Cross-validation accuracy estimation."
icon = "icons/TestLearners1.svg"
priority = 100
inputs = [("Learner", Learner, "set_learner", widget.Multiple),
("Data", Table, "set_train_data", widget.Default),
("Test Data", Table, "set_test_data"),
("Preprocessor", Preprocess, "set_preprocessor")]
outputs = [("Predictions", Table),
("Evaluation Results", Results)]
settingsHandler = settings.ClassValuesContextHandler()
#: Resampling/testing types
KFold, ShuffleSplit, LeaveOneOut, TestOnTrain, TestOnTest = 0, 1, 2, 3, 4
#: Numbers of folds
NFolds = [2, 3, 5, 10, 20]
#: Number of repetitions
NRepeats = [2, 3, 5, 10, 20, 50, 100]
#: Sample sizes
SampleSizes = [5, 10, 20, 25, 30, 33, 40, 50, 60, 66, 70, 75, 80, 90, 95]
#: Selected resampling type
resampling = settings.Setting(0)
#: Number of folds for K-fold cross validation
n_folds = settings.Setting(3)
#: Stratified sampling for K-fold
cv_stratified = settings.Setting(True)
#: Number of repeats for ShuffleSplit sampling
n_repeats = settings.Setting(3)
#: ShuffleSplit sample size
sample_size = settings.Setting(9)
#: Stratified sampling for Random Sampling
shuffle_stratified = settings.Setting(True)
TARGET_AVERAGE = "(Average over classes)"
class_selection = settings.ContextSetting(TARGET_AVERAGE)
class Error(OWWidget.Error):
class_required = Msg("Train data input requires a target variable.")
too_many_classes = Msg("Too many target variables.")
class_required_test = Msg("Test data input requires a target variable.")
too_many_folds = Msg("Number of folds exceeds the data size")
class_inconsistent = Msg("Test and train data sets "
"have different target variables.")
class Warning(OWWidget.Warning):
missing_data = \
Msg("Instances with unknown target values were removed from{}data.")
test_data_missing = Msg("Missing separate test data input.")
scores_not_computed = Msg("Some scores could not be computed.")
test_data_unused = Msg("Test data is present but unused. "
"Select 'Test on test data' to use it.")
class Information(OWWidget.Information):
data_sampled = Msg("Train data has been sampled")
test_data_sampled = Msg("Test data has been sampled")
def __init__(self):
super().__init__()
self.data = None
self.test_data = None
self.preprocessor = None
self.train_data_missing_vals = False
self.test_data_missing_vals = False
#: An Ordered dictionary with current inputs and their testing results.
self.learners = OrderedDict()
sbox = gui.vBox(self.controlArea, "Sampling")
rbox = gui.radioButtons(
sbox, self, "resampling", callback=self._param_changed)
gui.appendRadioButton(rbox, "Cross validation")
ibox = gui.indentedBox(rbox)
gui.comboBox(
ibox, self, "n_folds", label="Number of folds: ",
items=[str(x) for x in self.NFolds], maximumContentsLength=3,
orientation=Qt.Horizontal, callback=self.kfold_changed)
gui.checkBox(
ibox, self, "cv_stratified", "Stratified",
callback=self.kfold_changed)
gui.appendRadioButton(rbox, "Random sampling")
ibox = gui.indentedBox(rbox)
gui.comboBox(
ibox, self, "n_repeats", label="Repeat train/test: ",
items=[str(x) for x in self.NRepeats], maximumContentsLength=3,
orientation=Qt.Horizontal, callback=self.shuffle_split_changed)
gui.comboBox(
ibox, self, "sample_size", label="Training set size: ",
items=["{} %".format(x) for x in self.SampleSizes],
maximumContentsLength=5, orientation=Qt.Horizontal,
callback=self.shuffle_split_changed)
gui.checkBox(
ibox, self, "shuffle_stratified", "Stratified",
callback=self.shuffle_split_changed)
gui.appendRadioButton(rbox, "Leave one out")
gui.appendRadioButton(rbox, "Test on train data")
gui.appendRadioButton(rbox, "Test on test data")
self.cbox = gui.vBox(self.controlArea, "Target Class")
self.class_selection_combo = gui.comboBox(
self.cbox, self, "class_selection", items=[],
sendSelectedValue=True, valueType=str,
callback=self._on_target_class_changed,
contentsLength=8)
gui.rubber(self.controlArea)
self.view = gui.TableView(
wordWrap=True,
)
header = self.view.horizontalHeader()
header.setSectionResizeMode(QHeaderView.ResizeToContents)
header.setDefaultAlignment(Qt.AlignCenter)
header.setStretchLastSection(False)
self.result_model = QStandardItemModel(self)
self.result_model.setHorizontalHeaderLabels(["Method"])
self.view.setModel(self.result_model)
self.view.setItemDelegate(ItemDelegate())
box = gui.vBox(self.mainArea, "Evaluation Results")
box.layout().addWidget(self.view)
def sizeHint(self):
return QSize(780, 1)
def set_learner(self, learner, key):
"""
Set the input `learner` for `key`.
"""
if key in self.learners and learner is None:
# Removed
del self.learners[key]
else:
self.learners[key] = Input(learner, None, None)
self._invalidate([key])
def set_train_data(self, data):
"""
Set the input training dataset.
"""
self.Information.data_sampled.clear()
if data and not data.domain.class_vars:
self.Error.class_required()
data = None
elif data and len(data.domain.class_vars) > 1:
self.Error.too_many_classes()
data = None
else:
self.Error.class_required.clear()
self.Error.too_many_classes.clear()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.train_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = RemoveNaNClasses(data)
else:
self.Warning.missing_data.clear()
self.data = data
self.closeContext()
if data is not None:
self._update_class_selection()
self.openContext(data.domain.class_var)
self._invalidate()
def set_test_data(self, data):
"""
Set the input separate testing dataset.
"""
self.Information.test_data_sampled.clear()
if data and not data.domain.class_var:
self.Error.class_required()
data = None
else:
self.Error.class_required_test.clear()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.test_data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.test_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = RemoveNaNClasses(data)
else:
self.Warning.missing_data.clear()
self.test_data = data
if self.resampling == OWTestLearners.TestOnTest:
self._invalidate()
def _which_missing_data(self):
return {(True, True): " ", # both, don't specify
(True, False): " train ",
(False, True): " test "}[(self.train_data_missing_vals,
self.test_data_missing_vals)]
def set_preprocessor(self, preproc):
"""
Set the input preprocessor to apply on the training data.
"""
self.preprocessor = preproc
self._invalidate()
def handleNewSignals(self):
"""Reimplemented from OWWidget.handleNewSignals."""
self._update_class_selection()
self.commit()
def kfold_changed(self):
self.resampling = OWTestLearners.KFold
self._param_changed()
def shuffle_split_changed(self):
self.resampling = OWTestLearners.ShuffleSplit
self._param_changed()
def _param_changed(self):
self._invalidate()
def _update_results(self):
"""
Run/evaluate the learners.
"""
self.Warning.test_data_unused.clear()
self.Warning.test_data_missing.clear()
self.warning()
self.Error.class_inconsistent.clear()
self.Error.too_many_folds.clear()
self.error()
if self.data is None:
return
class_var = self.data.domain.class_var
if self.resampling == OWTestLearners.TestOnTest:
if self.test_data is None:
self.Warning.test_data_missing()
return
elif self.test_data.domain.class_var != class_var:
self.Error.class_inconsistent()
return
# items in need of an update
items = [(key, slot) for key, slot in self.learners.items()
if slot.results is None]
learners = [slot.learner for _, slot in items]
if len(items) == 0:
return
if self.test_data is not None and \
self.resampling != OWTestLearners.TestOnTest:
self.Warning.test_data_unused()
rstate = 42
def update_progress(finished):
self.progressBarSet(100 * finished)
common_args = dict(
store_data=True,
preprocessor=self.preprocessor,
callback=update_progress,
n_jobs=-1,
)
self.setStatusMessage("Running")
with self.progressBar():
try:
folds = self.NFolds[self.n_folds]
if self.resampling == OWTestLearners.KFold:
if len(self.data) < folds:
self.Error.too_many_folds()
return
warnings = []
results = Orange.evaluation.CrossValidation(
self.data, learners, k=folds,
random_state=rstate, warnings=warnings, **common_args)
if warnings:
self.warning(warnings[0])
elif self.resampling == OWTestLearners.LeaveOneOut:
results = Orange.evaluation.LeaveOneOut(
self.data, learners, **common_args)
elif self.resampling == OWTestLearners.ShuffleSplit:
train_size = self.SampleSizes[self.sample_size] / 100
results = Orange.evaluation.ShuffleSplit(
self.data, learners,
n_resamples=self.NRepeats[self.n_repeats],
train_size=train_size, test_size=None,
stratified=self.shuffle_stratified,
random_state=rstate, **common_args)
elif self.resampling == OWTestLearners.TestOnTrain:
results = Orange.evaluation.TestOnTrainingData(
self.data, learners, **common_args)
elif self.resampling == OWTestLearners.TestOnTest:
results = Orange.evaluation.TestOnTestData(
self.data, self.test_data, learners, **common_args)
else:
assert False
except (RuntimeError, ValueError) as e:
self.error(str(e))
self.setStatusMessage("")
return
else:
self.error()
learner_key = {slot.learner: key for key, slot in self.learners.items()}
for learner, result in zip(learners, results.split_by_model()):
stats = None
if class_var.is_discrete:
scorers = classification_stats.scores
elif class_var.is_continuous:
scorers = regression_stats.scores
else:
scorers = None
if scorers:
ex = result.failed[0]
if ex:
stats = [Try.Fail(ex)] * len(scorers)
result = Try.Fail(ex)
else:
stats = [Try(lambda: score(result)) for score in scorers]
result = Try.Success(result)
key = learner_key[learner]
self.learners[key] = \
self.learners[key]._replace(results=result, stats=stats)
self.setStatusMessage("")
def _update_header(self):
# Set the correct horizontal header labels on the results_model.
headers = ["Method"]
if self.data is not None:
if self.data.domain.has_discrete_class:
headers.extend(classification_stats.headers)
else:
headers.extend(regression_stats.headers)
# remove possible extra columns from the model.
for i in reversed(range(len(headers),
self.result_model.columnCount())):
self.result_model.takeColumn(i)
self.result_model.setHorizontalHeaderLabels(headers)
def _update_stats_model(self):
# Update the results_model with up to date scores.
# Note: The target class specific scores (if requested) are
# computed as needed in this method.
model = self.view.model()
# clear the table model, but preserving the header labels
for r in reversed(range(model.rowCount())):
model.takeRow(r)
target_index = None
if self.data is not None:
class_var = self.data.domain.class_var
if self.data.domain.has_discrete_class and \
self.class_selection != self.TARGET_AVERAGE:
target_index = class_var.values.index(self.class_selection)
else:
class_var = None
errors = []
has_missing_scores = False
for key, slot in self.learners.items():
name = learner_name(slot.learner)
head = QStandardItem(name)
head.setData(key, Qt.UserRole)
if isinstance(slot.results, Try.Fail):
head.setToolTip(str(slot.results.exception))
head.setText("{} (error)".format(name))
head.setForeground(QtGui.QBrush(Qt.red))
errors.append("{name} failed with error:\n"
"{exc.__class__.__name__}: {exc!s}"
.format(name=name, exc=slot.results.exception))
row = [head]
if class_var is not None and class_var.is_discrete and \
target_index is not None:
if slot.results is not None and slot.results.success:
ovr_results = results_one_vs_rest(
slot.results.value, target_index)
stats = [Try(lambda: score(ovr_results))
for score in classification_stats.scores]
else:
stats = None
else:
stats = slot.stats
if stats is not None:
for stat in stats:
item = QStandardItem()
if stat.success:
item.setText("{:.3f}".format(stat.value[0]))
else:
item.setToolTip(str(stat.exception))
has_missing_scores = True
row.append(item)
model.appendRow(row)
self.error("\n".join(errors), shown=bool(errors))
self.Warning.scores_not_computed(shown=has_missing_scores)
def _update_class_selection(self):
self.class_selection_combo.setCurrentIndex(-1)
self.class_selection_combo.clear()
if not self.data:
return
if self.data.domain.has_discrete_class:
self.cbox.setVisible(True)
class_var = self.data.domain.class_var
items = [self.TARGET_AVERAGE] + class_var.values
self.class_selection_combo.addItems(items)
class_index = 0
if self.class_selection in class_var.values:
class_index = class_var.values.index(self.class_selection) + 1
self.class_selection_combo.setCurrentIndex(class_index)
self.class_selection = items[class_index]
else:
self.cbox.setVisible(False)
def _on_target_class_changed(self):
self._update_stats_model()
def _invalidate(self, which=None):
# Invalidate learner results for `which` input keys
# (if None then all learner results are invalidated)
if which is None:
which = self.learners.keys()
model = self.view.model()
statmodelkeys = [model.item(row, 0).data(Qt.UserRole)
for row in range(model.rowCount())]
for key in which:
self.learners[key] = \
self.learners[key]._replace(results=None, stats=None)
if key in statmodelkeys:
row = statmodelkeys.index(key)
for c in range(1, model.columnCount()):
item = model.item(row, c)
if item is not None:
item.setData(None, Qt.DisplayRole)
item.setData(None, Qt.ToolTipRole)
self.commit()
def commit(self):
"""Recompute and output the results"""
self._update_header()
# Update the view to display the model names
self._update_stats_model()
self._update_results()
self._update_stats_model()
valid = [slot for slot in self.learners.values()
if slot.results is not None and slot.results.success]
if valid:
# Evaluation results
combined = results_merge([slot.results.value for slot in valid])
combined.learner_names = [learner_name(slot.learner)
for slot in valid]
# Predictions & Probabilities
predictions = combined.get_augmented_data(combined.learner_names)
else:
combined = None
predictions = None
self.send("Evaluation Results", combined)
self.send("Predictions", predictions)
def send_report(self):
"""Report on the testing schema and results"""
if not self.data or not self.learners:
return
if self.resampling == self.KFold:
stratified = 'Stratified ' if self.cv_stratified else ''
items = [("Sampling type", "{}{}-fold Cross validation".
format(stratified, self.NFolds[self.n_folds]))]
elif self.resampling == self.LeaveOneOut:
items = [("Sampling type", "Leave one out")]
elif self.resampling == self.ShuffleSplit:
stratified = 'Stratified ' if self.shuffle_stratified else ''
items = [("Sampling type",
"{}Shuffle split, {} random samples with {}% data "
.format(stratified, self.NRepeats[self.n_repeats],
self.SampleSizes[self.sample_size]))]
elif self.resampling == self.TestOnTrain:
items = [("Sampling type", "No sampling, test on training data")]
elif self.resampling == self.TestOnTest:
items = [("Sampling type", "No sampling, test on testing data")]
else:
items = []
if self.data.domain.has_discrete_class:
items += [("Target class", self.class_selection.strip("()"))]
if items:
self.report_items("Settings", items)
self.report_table("Scores", self.view)
def __set_index(self, f):
# type: (Future) -> None
# set results from `list_remote` query.
assert QThread.currentThread() is self.thread()
assert f.done()
self.setBlocking(False)
self.setStatusMessage("")
allinfolocal = list_local()
try:
res = f.result()
except Exception as er:
log = logging.getLogger(__name__)
log.exception("Error while fetching updated index")
if not allinfolocal:
self.error("Could not fetch data set list")
else:
self.warning("Could not fetch data sets list, only local "
"cached data sets are shown")
res = {}
allinforemote = res # type: Dict[Tuple[str, str], dict]
allkeys = set(allinfolocal)
if allinforemote is not None:
allkeys = allkeys | set(allinforemote)
allkeys = sorted(allkeys)
def info(prefix, filename):
if (prefix, filename) in allinforemote:
info = allinforemote[prefix, filename]
else:
info = allinfolocal[prefix, filename]
islocal = (prefix, filename) in allinfolocal
return namespace(prefix=prefix,
filename=filename,
title=info.get("title", filename),
datetime=info.get("datetime", None),
description=info.get("description", None),
reference=info.get("reference", None),
instances=info.get("instances", None),
variables=info.get("variables", None),
target=info.get("target", None),
missing=info.get("missing", None),
tags=info.get("tags", []),
size=info.get("size", None),
islocal=islocal)
model = QStandardItemModel(self)
model.setHorizontalHeaderLabels(HEADER)
current_index = -1
for i, (prefix, filename) in enumerate(allkeys):
datainfo = info(prefix, filename)
item1 = QStandardItem()
item1.setData(" " if datainfo.islocal else "", Qt.DisplayRole)
item1.setData(datainfo, Qt.UserRole)
item2 = QStandardItem(datainfo.title)
item3 = QStandardItem()
item3.setData(datainfo.size, Qt.DisplayRole)
item4 = QStandardItem()
item4.setData(datainfo.instances, Qt.DisplayRole)
item5 = QStandardItem()
item5.setData(datainfo.variables, Qt.DisplayRole)
item6 = QStandardItem()
item6.setData(datainfo.target, Qt.DisplayRole)
item6.setIcon(variable_icon(datainfo.target))
item7 = QStandardItem()
item7.setData(", ".join(datainfo.tags), Qt.DisplayRole)
row = [item1, item2, item3, item4, item5, item6, item7]
model.appendRow(row)
if (prefix, filename) == self.selected_id:
current_index = i
hs = self.view.header().saveState()
model_ = self.view.model()
self.view.setModel(model)
self.view.header().restoreState(hs)
model_.deleteLater()
model_.setParent(None)
self.view.selectionModel().selectionChanged.connect(
self.__on_selection)
# Update the info text
self.infolabel.setText("{} datasets \n{} datasets cached".format(
model.rowCount(), len(allinfolocal)))
if current_index != -1:
selmodel = self.view.selectionModel()
selmodel.select(
model.index(current_index, 0),
QItemSelectionModel.ClearAndSelect | QItemSelectionModel.Rows)
class OWSetEnrichment(widget.OWWidget):
name = "Set Enrichment"
description = ""
icon = "../widgets/icons/GeneSetEnrichment.svg"
priority = 5000
inputs = [("Data", Orange.data.Table, "setData", widget.Default),
("Reference", Orange.data.Table, "setReference")]
outputs = [("Data subset", Orange.data.Table)]
settingsHandler = settings.DomainContextHandler()
taxid = settings.ContextSetting(None)
speciesIndex = settings.ContextSetting(0)
genesinrows = settings.ContextSetting(False)
geneattr = settings.ContextSetting(0)
categoriesCheckState = settings.ContextSetting({})
useReferenceData = settings.Setting(False)
useMinCountFilter = settings.Setting(True)
useMaxPValFilter = settings.Setting(True)
useMaxFDRFilter = settings.Setting(True)
minClusterCount = settings.Setting(3)
maxPValue = settings.Setting(0.01)
maxFDR = settings.Setting(0.01)
autocommit = settings.Setting(False)
Ready, Initializing, Loading, RunningEnrichment = 0, 1, 2, 4
class Error(widget.OWWidget.Error):
no_gene_names = Msg("Input data contains no columns with gene names")
no_data_onInput = Msg("No data on input")
class Warning(widget.OWWidget.Warning):
no_sets_found = Msg("No enriched sets found")
def __init__(self, parent=None):
super().__init__(parent)
self.geneMatcherSettings = [False, False, True, False]
self.data = None
self.referenceData = None
self.taxid_list = []
self.__genematcher = (None, fulfill(gene.matcher([])))
self.__invalidated = False
self.currentAnnotatedCategories = []
self.state = None
self.__state = OWSetEnrichment.Initializing
box = gui.widgetBox(self.controlArea, "Info")
self.infoBox = gui.widgetLabel(box, "Info")
self.infoBox.setText("No data on input.\n")
self.speciesComboBox = gui.comboBox(
self.controlArea, self,
"speciesIndex", "Species",
callback=self.__on_speciesIndexChanged)
box = gui.widgetBox(self.controlArea, "Entity names")
self.geneAttrComboBox = gui.comboBox(
box, self, "geneattr", "Entity feature", sendSelectedValue=0,
callback=self.updateAnnotations)
cb = gui.checkBox(
box, self, "genesinrows", "Use feature names",
callback=self.updateAnnotations,
disables=[(-1, self.geneAttrComboBox)])
cb.makeConsistent()
# gui.button(box, self, "Gene matcher settings",
# callback=self.updateGeneMatcherSettings,
# tooltip="Open gene matching settings dialog")
self.referenceRadioBox = gui.radioButtonsInBox(
self.controlArea,
self, "useReferenceData",
["All entities", "Reference set (input)"],
tooltips=["Use entire genome (for gene set enrichment) or all " +
"available entities for reference",
"Use entities from Reference Examples input signal " +
"as reference"],
box="Reference", callback=self.updateAnnotations)
box = gui.widgetBox(self.controlArea, "Entity Sets")
self.groupsWidget = QTreeWidget(self)
self.groupsWidget.setHeaderLabels(["Category"])
box.layout().addWidget(self.groupsWidget)
hLayout = QHBoxLayout()
hLayout.setSpacing(10)
hWidget = gui.widgetBox(self.mainArea, orientation=hLayout)
gui.spin(hWidget, self, "minClusterCount",
0, 100, label="Entities",
tooltip="Minimum entity count",
callback=self.filterAnnotationsChartView,
callbackOnReturn=True,
checked="useMinCountFilter",
checkCallback=self.filterAnnotationsChartView)
pvalfilterbox = gui.widgetBox(hWidget, orientation="horizontal")
cb = gui.checkBox(
pvalfilterbox, self, "useMaxPValFilter", "p-value",
callback=self.filterAnnotationsChartView)
sp = gui.doubleSpin(
pvalfilterbox, self, "maxPValue", 0.0, 1.0, 0.0001,
tooltip="Maximum p-value",
callback=self.filterAnnotationsChartView,
callbackOnReturn=True,
)
sp.setEnabled(self.useMaxFDRFilter)
cb.toggled[bool].connect(sp.setEnabled)
pvalfilterbox.layout().setAlignment(cb, Qt.AlignRight)
pvalfilterbox.layout().setAlignment(sp, Qt.AlignLeft)
fdrfilterbox = gui.widgetBox(hWidget, orientation="horizontal")
cb = gui.checkBox(
fdrfilterbox, self, "useMaxFDRFilter", "FDR",
callback=self.filterAnnotationsChartView)
sp = gui.doubleSpin(
fdrfilterbox, self, "maxFDR", 0.0, 1.0, 0.0001,
tooltip="Maximum False discovery rate",
callback=self.filterAnnotationsChartView,
callbackOnReturn=True,
)
sp.setEnabled(self.useMaxFDRFilter)
cb.toggled[bool].connect(sp.setEnabled)
fdrfilterbox.layout().setAlignment(cb, Qt.AlignRight)
fdrfilterbox.layout().setAlignment(sp, Qt.AlignLeft)
self.filterLineEdit = QLineEdit(self, placeholderText="Search ...")
self.filterCompleter = QCompleter(self.filterLineEdit)
self.filterLineEdit.setCompleter(self.filterCompleter)
hLayout.addWidget(self.filterLineEdit)
self.mainArea.layout().addWidget(hWidget)
self.filterLineEdit.textChanged.connect(
self.filterAnnotationsChartView)
self.annotationsChartView = QTreeView(
alternatingRowColors=True,
sortingEnabled=True,
selectionMode=QTreeView.ExtendedSelection,
rootIsDecorated=False,
editTriggers=QTreeView.NoEditTriggers,
)
self.source_model = QStandardItemModel()
self.source_model.setSortRole(Qt.UserRole)
self.source_model.setHorizontalHeaderLabels(
["Category", "Term", "Count", "Reference count", "p-value",
"FDR", "Enrichment"])
self.proxy_model = CustomFilterModel(self.annotationsChartView)
self.proxy_model.setFilterKeyColumn(1) # filter only by name (second column).
self.annotationsChartView.setModel(self.proxy_model)
self.annotationsChartView.viewport().setMouseTracking(True)
self.mainArea.layout().addWidget(self.annotationsChartView)
contextEventFilter = gui.VisibleHeaderSectionContextEventFilter(
self.annotationsChartView)
self.annotationsChartView.header().installEventFilter(contextEventFilter)
self.groupsWidget.itemClicked.connect(self.subsetSelectionChanged)
gui.auto_commit(self.controlArea, self, "autocommit", "Commit")
self.setBlocking(True)
task = EnsureDownloaded(
[(taxonomy.Taxonomy.DOMAIN, taxonomy.Taxonomy.FILENAME),
(geneset.sfdomain, "index.pck")]
)
task.finished.connect(self.__initialize_finish)
self.setStatusMessage("Initializing")
self._executor = ThreadExecutor(
parent=self, threadPool=QThreadPool(self))
self._executor.submit(task)
def string_search(self):
self.annotationsChartView.model().setFilterFixedString(self.filterLineEdit.text())
def sizeHint(self):
return QSize(1024, 600)
def __initialize_finish(self):
# Finalize the the widget's initialization (preferably after
# ensuring all required databases have been downloaded.
sets = geneset.list_all()
taxids = set(taxonomy.common_taxids() +
list(filter(None, [tid for _, tid, _ in sets])))
organisms = [(tid, name_or_none(tid)) for tid in taxids]
organisms = [(tid, name) for tid, name in organisms
if name is not None]
organisms = [(None, "None")] + sorted(organisms)
taxids = [tid for tid, _ in organisms]
names = [name for _, name in organisms]
self.taxid_list = taxids
self.speciesComboBox.clear()
self.speciesComboBox.addItems(names)
self.genesets = sets
if self.taxid in self.taxid_list:
taxid = self.taxid
else:
taxid = self.taxid_list[0]
self.taxid = None
self.setCurrentOrganism(taxid)
self.setBlocking(False)
self.__state = OWSetEnrichment.Ready
self.setStatusMessage("")
def setCurrentOrganism(self, taxid):
"""Set the current organism `taxid`."""
if taxid not in self.taxid_list:
taxid = self.taxid_list[min(self.speciesIndex,
len(self.taxid_list) - 1)]
if self.taxid != taxid:
self.taxid = taxid
self.speciesIndex = self.taxid_list.index(taxid)
self.refreshHierarchy()
self._invalidateGeneMatcher()
self._invalidate()
def currentOrganism(self):
"""Return the current organism taxid"""
return self.taxid
def __on_speciesIndexChanged(self):
taxid = self.taxid_list[self.speciesIndex]
self.taxid = "< Do not look >"
self.setCurrentOrganism(taxid)
if self.__invalidated and self.data is not None:
self.updateAnnotations()
def clear(self):
"""Clear/reset the widget state."""
self._cancelPending()
self.state = None
self.__state = self.__state & ~OWSetEnrichment.RunningEnrichment
self._clearView()
if self.annotationsChartView.model() is not None:
self.annotationsChartView.model().clear()
self.geneAttrComboBox.clear()
self.geneAttrs = []
self._updatesummary()
def _cancelPending(self):
"""Cancel pending tasks."""
if self.state is not None:
self.state.results.cancel()
self.state.namematcher.cancel()
self.state.cancelled = True
def _clearView(self):
"""Clear the enrichment report view (main area)."""
if self.annotationsChartView.model() is not None:
self.annotationsChartView.model().clear()
def setData(self, data=None):
"""Set the input dataset with query gene names"""
if self.__state & OWSetEnrichment.Initializing:
self.__initialize_finish()
self.Error.clear()
self.closeContext()
self.clear()
self.groupsWidget.clear()
self.data = data
if data is not None:
varlist = [var for var in data.domain.variables + data.domain.metas
if isinstance(var, Orange.data.StringVariable)]
self.geneAttrs = varlist
for var in varlist:
self.geneAttrComboBox.addItem(*gui.attributeItem(var))
oldtaxid = self.taxid
self.geneattr = min(self.geneattr, len(self.geneAttrs) - 1)
taxid = data_hints.get_hint(data, "taxid", "")
if taxid in self.taxid_list:
self.speciesIndex = self.taxid_list.index(taxid)
self.taxid = taxid
self.genesinrows = data_hints.get_hint(
data, "genesinrows", self.genesinrows)
self.openContext(data)
if oldtaxid != self.taxid:
self.taxid = "< Do not look >"
self.setCurrentOrganism(taxid)
self.refreshHierarchy()
self._invalidate()
def setReference(self, data=None):
"""Set the (optional) input dataset with reference gene names."""
self.referenceData = data
self.referenceRadioBox.setEnabled(bool(data))
if self.useReferenceData:
self._invalidate()
def handleNewSignals(self):
if self.__invalidated:
self.updateAnnotations()
def _invalidateGeneMatcher(self):
_, f = self.__genematcher
f.cancel()
self.__genematcher = (None, fulfill(gene.matcher([])))
def _invalidate(self):
self.__invalidated = True
def genesFromTable(self, table):
if self.genesinrows:
genes = [attr.name for attr in table.domain.attributes]
else:
geneattr = self.geneAttrs[self.geneattr]
genes = [str(ex[geneattr]) for ex in table]
return genes
def getHierarchy(self, taxid):
def recursive_dict():
return defaultdict(recursive_dict)
collection = recursive_dict()
def collect(col, hier):
if hier:
collect(col[hier[0]], hier[1:])
for hierarchy, t_id, _ in self.genesets:
collect(collection[t_id], hierarchy)
return (taxid, collection[taxid]), (None, collection[None])
def setHierarchy(self, hierarchy, hierarchy_noorg):
self.groupsWidgetItems = {}
def fill(col, parent, full=(), org=""):
for key, value in sorted(col.items()):
full_cat = full + (key,)
item = QTreeWidgetItem(parent, [key])
item.setFlags(item.flags() | Qt.ItemIsUserCheckable |
Qt.ItemIsSelectable | Qt.ItemIsEnabled)
if value:
item.setFlags(item.flags() | Qt.ItemIsTristate)
checked = self.categoriesCheckState.get(
(full_cat, org), Qt.Checked)
item.setData(0, Qt.CheckStateRole, checked)
item.setExpanded(True)
item.category = full_cat
item.organism = org
self.groupsWidgetItems[full_cat] = item
fill(value, item, full_cat, org=org)
self.groupsWidget.clear()
fill(hierarchy[1], self.groupsWidget, org=hierarchy[0])
fill(hierarchy_noorg[1], self.groupsWidget, org=hierarchy_noorg[0])
def refreshHierarchy(self):
self.setHierarchy(*self.getHierarchy(taxid=self.taxid_list[self.speciesIndex]))
def selectedCategories(self):
"""
Return a list of currently selected hierarchy keys.
A key is a tuple of identifiers from the root to the leaf of
the hierarchy tree.
"""
return [key for key, check in self.getHierarchyCheckState().items()
if check == Qt.Checked]
def getHierarchyCheckState(self):
def collect(item, full=()):
checked = item.checkState(0)
name = str(item.data(0, Qt.DisplayRole))
full_cat = full + (name,)
result = [((full_cat, item.organism), checked)]
for i in range(item.childCount()):
result.extend(collect(item.child(i), full_cat))
return result
items = [self.groupsWidget.topLevelItem(i)
for i in range(self.groupsWidget.topLevelItemCount())]
states = itertools.chain(*(collect(item) for item in items))
return dict(states)
def subsetSelectionChanged(self, item, column):
# The selected geneset (hierarchy) subset has been changed by the
# user. Update the displayed results.
# Update the stored state (persistent settings)
self.categoriesCheckState = self.getHierarchyCheckState()
categories = self.selectedCategories()
if self.data is not None:
if self._nogenematching() or \
not set(categories) <= set(self.currentAnnotatedCategories):
self.updateAnnotations()
else:
# compute FDR according the selected categories
self.compute_fdr()
self.filterAnnotationsChartView()
def updateGeneMatcherSettings(self):
raise NotImplementedError
from .OWGOEnrichmentAnalysis import GeneMatcherDialog
dialog = GeneMatcherDialog(self, defaults=self.geneMatcherSettings, enabled=[True] * 4, modal=True)
if dialog.exec_():
self.geneMatcherSettings = [getattr(dialog, item[0]) for item in dialog.items]
self._invalidateGeneMatcher()
if self.data is not None:
self.updateAnnotations()
def _genematcher(self):
"""
Return a Future[gene.SequenceMatcher]
"""
taxid = self.taxid_list[self.speciesIndex]
current, matcher_f = self.__genematcher
if taxid == current and \
not matcher_f.cancelled():
return matcher_f
self._invalidateGeneMatcher()
if taxid is None:
self.__genematcher = (None, fulfill(gene.matcher([])))
return self.__genematcher[1]
matchers = [gene.GMGO, gene.GMKEGG, gene.GMNCBI, gene.GMAffy]
matchers = [m for m, use in zip(matchers, self.geneMatcherSettings)
if use]
def create():
return gene.matcher([m(taxid) for m in matchers])
matcher_f = self._executor.submit(create)
self.__genematcher = (taxid, matcher_f)
return self.__genematcher[1]
def _nogenematching(self):
return self.taxid is None or not any(self.geneMatcherSettings)
def updateAnnotations(self):
if self.data is None:
return
assert not self.__state & OWSetEnrichment.Initializing
self._cancelPending()
self._clearView()
self.Warning.clear()
self.Error.clear()
if not self.genesinrows and len(self.geneAttrs) == 0:
self.Error.no_gene_names()
return
self.__state = OWSetEnrichment.RunningEnrichment
taxid = self.taxid_list[self.speciesIndex]
self.taxid = taxid
categories = self.selectedCategories()
clusterGenes = self.genesFromTable(self.data)
if self.referenceData is not None and self.useReferenceData:
referenceGenes = self.genesFromTable(self.referenceData)
else:
referenceGenes = None
self.currentAnnotatedCategories = categories
genematcher = self._genematcher()
self.progressBarInit()
## Load collections in a worker thread
# TODO: Use cached collections if already loaded and
# use ensure_genesetsdownloaded with progress report (OWSelectGenes)
collections = self._executor.submit(geneset.collections, *categories)
def refset_null():
"""Return the default background reference set"""
col = collections.result()
return reduce(operator.ior, (set(g.genes) for g in col), set())
def refset_ncbi():
"""Return all NCBI gene names"""
geneinfo = gene.NCBIGeneInfo(taxid)
return set(geneinfo.keys())
def namematcher():
matcher = genematcher.result()
match = matcher.set_targets(ref_set.result())
match.umatch = memoize(match.umatch)
return match
def map_unames():
matcher = namematcher.result()
query = list(filter(None, map(matcher.umatch, querynames)))
reference = list(filter(None, map(matcher.umatch, ref_set.result())))
return query, reference
if self._nogenematching():
if referenceGenes is None:
ref_set = self._executor.submit(refset_null)
else:
ref_set = fulfill(referenceGenes)
else:
if referenceGenes == None:
ref_set = self._executor.submit(refset_ncbi)
else:
ref_set = fulfill(referenceGenes)
namematcher = self._executor.submit(namematcher)
querynames = clusterGenes
state = types.SimpleNamespace()
state.query_set = clusterGenes
state.reference_set = referenceGenes
state.namematcher = namematcher
state.query_count = len(set(clusterGenes))
state.reference_count = (len(set(referenceGenes))
if referenceGenes is not None else None)
state.cancelled = False
progress = methodinvoke(self, "_setProgress", (float,))
info = methodinvoke(self, "_setRunInfo", (str,))
@withtraceback
def run():
info("Loading data")
match = namematcher.result()
query, reference = map_unames()
gscollections = collections.result()
results = []
info("Running enrichment")
p = 0
for i, gset in enumerate(gscollections):
genes = set(filter(None, map(match.umatch, gset.genes)))
enr = set_enrichment(genes, reference, query)
results.append((gset, enr))
if state.cancelled:
raise UserInteruptException
pnew = int(100 * i / len(gscollections))
if pnew != p:
progress(pnew)
p = pnew
progress(100)
info("")
return query, reference, results
task = Task(function=run)
task.resultReady.connect(self.__on_enrichment_finished)
task.exceptionReady.connect(self.__on_enrichment_failed)
result = self._executor.submit(task)
state.results = result
self.state = state
self._updatesummary()
def __on_enrichment_failed(self, exception):
if not isinstance(exception, UserInteruptException):
print("ERROR:", exception, file=sys.stderr)
print(exception._traceback, file=sys.stderr)
self.progressBarFinished()
self.setStatusMessage("")
self.__state &= ~OWSetEnrichment.RunningEnrichment
def __on_enrichment_finished(self, results):
assert QThread.currentThread() is self.thread()
self.__state &= ~OWSetEnrichment.RunningEnrichment
query, reference, results = results
if self.annotationsChartView.model():
self.annotationsChartView.model().clear()
nquery = len(query)
nref = len(reference)
maxcount = max((len(e.query_mapped) for _, e in results),
default=1)
maxrefcount = max((len(e.reference_mapped) for _, e in results),
default=1)
nspaces = int(math.ceil(math.log10(maxcount or 1)))
refspaces = int(math.ceil(math.log(maxrefcount or 1)))
query_fmt = "%" + str(nspaces) + "s (%.2f%%)"
ref_fmt = "%" + str(refspaces) + "s (%.2f%%)"
def fmt_count(fmt, count, total):
return fmt % (count, 100.0 * count / (total or 1))
fmt_query_count = partial(fmt_count, query_fmt)
fmt_ref_count = partial(fmt_count, ref_fmt)
linkFont = QFont(self.annotationsChartView.viewOptions().font)
linkFont.setUnderline(True)
def item(value=None, tooltip=None, user=None):
si = QStandardItem()
if value is not None:
si.setData(value, Qt.DisplayRole)
if tooltip is not None:
si.setData(tooltip, Qt.ToolTipRole)
if user is not None:
si.setData(user, Qt.UserRole)
else:
si.setData(value, Qt.UserRole)
return si
model = self.source_model
for i, (gset, enrich) in enumerate(results):
if len(enrich.query_mapped) == 0:
continue
nquery_mapped = len(enrich.query_mapped)
nref_mapped = len(enrich.reference_mapped)
row = [
item(", ".join(gset.hierarchy)),
item(gsname(gset), tooltip=gset.link),
item(fmt_query_count(nquery_mapped, nquery),
tooltip=nquery_mapped, user=nquery_mapped),
item(fmt_ref_count(nref_mapped, nref),
tooltip=nref_mapped, user=nref_mapped),
item(fmtp(enrich.p_value), user=enrich.p_value),
item(), # column 5, FDR, is computed in filterAnnotationsChartView
item(enrich.enrichment_score,
tooltip="%.3f" % enrich.enrichment_score,
user=enrich.enrichment_score)
]
row[0].geneset = gset
row[0].enrichment = enrich
row[1].setData(gset.link, gui.LinkRole)
row[1].setFont(linkFont)
row[1].setForeground(QColor(Qt.blue))
model.appendRow(row)
self.annotationsChartView.selectionModel().selectionChanged.connect(
self.commit
)
if not model.rowCount():
self.Warning.no_sets_found()
else:
# compute FDR according the selected categories
self.compute_fdr()
# set source model if there are sets found
self.filterAnnotationsChartView()
self.proxy_model.setSourceModel(model)
self.Warning.clear()
self._updatesummary()
allnames = set(gsname(geneset)
for geneset, (count, _, _, _) in results if count)
allnames |= reduce(operator.ior,
(set(word_split(name)) for name in allnames),
set())
self.filterCompleter.setModel(None)
self.completerModel = QStringListModel(sorted(allnames))
self.filterCompleter.setModel(self.completerModel)
if results:
max_score = max((e.enrichment_score for _, e in results
if np.isfinite(e.enrichment_score)),
default=1)
self.annotationsChartView.setItemDelegateForColumn(
6, BarItemDelegate(self, scale=(0.0, max_score))
)
self.annotationsChartView.setItemDelegateForColumn(
1, gui.LinkStyledItemDelegate(self.annotationsChartView)
)
header = self.annotationsChartView.header()
for i in range(model.columnCount()):
sh = self.annotationsChartView.sizeHintForColumn(i)
sh = max(sh, header.sectionSizeHint(i))
self.annotationsChartView.setColumnWidth(i, max(min(sh, 300), 30))
# self.annotationsChartView.resizeColumnToContents(i)
self.progressBarFinished()
self.setStatusMessage("")
def _updatesummary(self):
state = self.state
if state is None:
self.Error.no_data_onInput()
self.Warning.clear()
self.infoBox.setText("No data on input.\n")
return
text = "{.query_count} unique names on input\n".format(state)
if state.results.done() and not state.results.exception():
mapped, _, _ = state.results.result()
ratio_mapped = (len(mapped) / state.query_count
if state.query_count else 0)
text += ("%i (%.1f%%) gene names matched" %
(len(mapped), 100.0 * ratio_mapped))
elif not state.results.done():
text += "..."
else:
text += "<Error {}>".format(str(state.results.exception()))
self.infoBox.setText(text)
# TODO: warn on no enriched sets found (i.e no query genes
# mapped to any set)
def return_filter_values(self):
categories = set(", ".join(category) for category, _ in self.selectedCategories())
return {
"useMinCountFilter": self.useMinCountFilter,
"useMaxPValFilter": self.useMaxPValFilter,
"minClusterCount": self.minClusterCount,
"useMaxFDRFilter": self.useMaxFDRFilter,
"maxPValue": self.maxPValue,
"categories": categories,
"maxFDR": self.maxFDR
}
def compute_fdr(self):
selected_categories = self.return_filter_values().get('categories')
data_model = self.source_model
# get pvalues of rows that match selected category
pvals = [(row_index, data_model.index(row_index, 4).data(role=Qt.UserRole))
for row_index in range(data_model.rowCount())
if data_model.index(row_index, 0).data(role=Qt.DisplayRole) in selected_categories]
fdrs = utils.stats.FDR([pval for _, pval in pvals])
# set calculated values to the data model
for (row_index, pval), fdr in zip(pvals, fdrs):
fdr_item = data_model.item(row_index, 5)
fdr_item.setData(fmtpdet(fdr), Qt.ToolTipRole)
fdr_item.setData(fmtp(fdr), Qt.DisplayRole)
fdr_item.setData(fdr, Qt.UserRole)
def filterAnnotationsChartView(self):
if self.__state & OWSetEnrichment.RunningEnrichment:
return
# set string pattern
self.proxy_model._pattern = str(self.filterLineEdit.text())
# set filter values to proxy model
self.proxy_model.set_filter_values(self.return_filter_values())
# filter model
self.proxy_model.invalidateFilter()
self._updatesummary()
@Slot(float)
def _setProgress(self, value):
assert QThread.currentThread() is self.thread()
self.progressBarSet(value, processEvents=None)
@Slot(str)
def _setRunInfo(self, text):
self.setStatusMessage(text)
def commit(self):
if self.data is None or \
self.__state & OWSetEnrichment.RunningEnrichment:
return
model = self.source_model
rows = self.annotationsChartView.selectionModel().selectedRows(0)
selected = [model.item(index.row(), 0) for index in rows]
mapped = reduce(operator.ior,
(set(item.enrichment.query_mapped)
for item in selected),
set())
assert self.state.namematcher.done()
matcher = self.state.namematcher.result()
axis = 1 if self.genesinrows else 0
if axis == 1:
mapped = [attr for attr in self.data.domain.attributes
if matcher.umatch(attr.name) in mapped]
newdomain = Orange.data.Domain(
mapped, self.data.domain.class_vars, self.data.domain.metas)
data = self.data.from_table(newdomain, self.data)
else:
geneattr = self.geneAttrs[self.geneattr]
selected = [i for i, ex in enumerate(self.data)
if matcher.umatch(str(ex[geneattr])) in mapped]
data = self.data[selected]
self.send("Data subset", data)
def onDeleteWidget(self):
if self.state is not None:
self._cancelPending()
self.state = None
self._executor.shutdown(wait=False)
class OWKMeans(widget.OWWidget):
name = "k-Means"
description = "k-Means clustering algorithm with silhouette-based " \
"quality estimation."
icon = "icons/KMeans.svg"
priority = 2100
inputs = [("Data", Table, "set_data")]
outputs = [("Annotated Data", Table, widget.Default),
("Centroids", Table)]
INIT_KMEANS, INIT_RANDOM = range(2)
INIT_METHODS = "Initialize with KMeans++", "Random initialization"
SILHOUETTE, INTERCLUSTER, DISTANCES = range(3)
SCORING_METHODS = [("Silhouette", lambda km: km.silhouette, False),
("Inter-cluster distance",
lambda km: km.inter_cluster, True),
("Distance to centroids",
lambda km: km.inertia, True)]
OUTPUT_CLASS, OUTPUT_ATTRIBUTE, OUTPUT_META = range(3)
OUTPUT_METHODS = ("Class", "Feature", "Meta")
resizing_enabled = False
k = Setting(3)
k_from = Setting(2)
k_to = Setting(8)
optimize_k = Setting(False)
max_iterations = Setting(300)
n_init = Setting(10)
smart_init = Setting(INIT_KMEANS)
scoring = Setting(SILHOUETTE)
append_cluster_ids = Setting(True)
place_cluster_ids = Setting(OUTPUT_CLASS)
output_name = Setting("Cluster")
auto_run = Setting(True)
def __init__(self):
super().__init__()
self.data = None
self.km = None
self.optimization_runs = []
box = gui.vBox(self.controlArea, "Number of Clusters")
layout = QGridLayout()
self.n_clusters = bg = gui.radioButtonsInBox(
box, self, "optimize_k", [], orientation=layout,
callback=self.update)
layout.addWidget(
gui.appendRadioButton(bg, "Fixed:", addToLayout=False),
1, 1)
sb = gui.hBox(None, margin=0)
self.fixedSpinBox = gui.spin(
sb, self, "k", minv=2, maxv=30,
controlWidth=60, alignment=Qt.AlignRight, callback=self.update_k)
gui.rubber(sb)
layout.addWidget(sb, 1, 2)
layout.addWidget(
gui.appendRadioButton(bg, "Optimized from", addToLayout=False), 2, 1)
ftobox = gui.hBox(None)
ftobox.layout().setContentsMargins(0, 0, 0, 0)
layout.addWidget(ftobox)
gui.spin(
ftobox, self, "k_from", minv=2, maxv=29,
controlWidth=60, alignment=Qt.AlignRight,
callback=self.update_from)
gui.widgetLabel(ftobox, "to")
self.fixedSpinBox = gui.spin(
ftobox, self, "k_to", minv=3, maxv=30,
controlWidth=60, alignment=Qt.AlignRight,
callback=self.update_to)
gui.rubber(ftobox)
layout.addWidget(gui.widgetLabel(None, "Scoring: "),
5, 1, Qt.AlignRight)
layout.addWidget(
gui.comboBox(
None, self, "scoring", label="Scoring",
items=list(zip(*self.SCORING_METHODS))[0],
callback=self.update), 5, 2)
box = gui.vBox(self.controlArea, "Initialization")
gui.comboBox(
box, self, "smart_init", items=self.INIT_METHODS,
callback=self.update)
layout = QGridLayout()
box2 = gui.widgetBox(box, orientation=layout)
box2.setSizePolicy(QSizePolicy.Maximum, QSizePolicy.Maximum)
layout.addWidget(gui.widgetLabel(None, "Re-runs: "),
0, 0, Qt.AlignLeft)
sb = gui.hBox(None, margin=0)
layout.addWidget(sb, 0, 1)
gui.lineEdit(
sb, self, "n_init", controlWidth=60,
valueType=int, validator=QIntValidator(),
callback=self.update)
layout.addWidget(gui.widgetLabel(None, "Maximal iterations: "),
1, 0, Qt.AlignLeft)
sb = gui.hBox(None, margin=0)
layout.addWidget(sb, 1, 1)
gui.lineEdit(sb, self, "max_iterations",
controlWidth=60, valueType=int,
validator=QIntValidator(),
callback=self.update)
box = gui.vBox(self.controlArea, "Output")
gui.comboBox(box, self, "place_cluster_ids",
label="Append cluster ID as:", orientation=Qt.Horizontal,
callback=self.send_data, items=self.OUTPUT_METHODS)
gui.lineEdit(box, self, "output_name",
label="Name:", orientation=Qt.Horizontal,
callback=self.send_data)
gui.separator(self.buttonsArea, 30)
self.apply_button = gui.auto_commit(
self.buttonsArea, self, "auto_run", "Apply", box=None,
commit=self.commit
)
gui.rubber(self.controlArea)
self.table_model = QStandardItemModel(self)
self.table_model.setHorizontalHeaderLabels(["k", "Score"])
self.table_model.setColumnCount(2)
self.table_box = gui.vBox(
self.mainArea, "Optimization Report", addSpace=0)
table = self.table_view = QTableView(self.table_box)
table.setHorizontalScrollMode(QTableView.ScrollPerPixel)
table.setHorizontalScrollBarPolicy(Qt.ScrollBarAlwaysOff)
table.setSelectionMode(QTableView.SingleSelection)
table.setSelectionBehavior(QTableView.SelectRows)
table.verticalHeader().hide()
table.setItemDelegateForColumn(1, gui.TableBarItem(self))
table.setModel(self.table_model)
table.selectionModel().selectionChanged.connect(
self.table_item_selected)
table.setColumnWidth(0, 40)
table.setColumnWidth(1, 120)
table.horizontalHeader().setStretchLastSection(True)
self.setSizePolicy(QSizePolicy.Preferred, QSizePolicy.Preferred)
self.mainArea.setSizePolicy(QSizePolicy.Maximum,
QSizePolicy.Preferred)
self.table_box.setSizePolicy(QSizePolicy.Fixed,
QSizePolicy.MinimumExpanding)
self.table_view.setSizePolicy(QSizePolicy.Preferred,
QSizePolicy.MinimumExpanding)
self.table_box.layout().addWidget(self.table_view)
self.hide_show_opt_results()
def adjustSize(self):
self.ensurePolished()
s = self.sizeHint()
self.resize(s)
def hide_show_opt_results(self):
[self.mainArea.hide, self.mainArea.show][self.optimize_k]()
QTimer.singleShot(100, self.adjustSize)
def sizeHint(self):
s = self.controlArea.sizeHint()
if self.optimize_k and not self.mainArea.isHidden():
s.setWidth(s.width() + self.mainArea.sizeHint().width() +
4 * self.childrenRect().x())
return s
def update_k(self):
self.optimize_k = False
self.update()
def update_from(self):
self.k_to = max(self.k_from + 1, self.k_to)
self.optimize_k = True
self.update()
def update_to(self):
self.k_from = min(self.k_from, self.k_to - 1)
self.optimize_k = True
self.update()
def set_optimization(self):
self.updateOptimizationGui()
self.update()
def check_data_size(self, n, msg_group):
msg_group.add_message(
"not_enough_data",
"Too few ({}) unique data instances for {} clusters")
if n > len(self.data):
msg_group.not_enough_data(len(self.data), n)
return False
else:
msg_group.not_enough_data.clear()
return True
def run_optimization(self):
# Disabling is needed since this function is not reentrant
# Fast clicking on, say, "To: " causes multiple calls
try:
self.controlArea.setDisabled(True)
self.optimization_runs = []
if not self.check_data_size(self.k_from, self.Error):
return
self.check_data_size(self.k_to, self.Warning)
k_to = min(self.k_to, len(self.data))
kmeans = KMeans(
init=['random', 'k-means++'][self.smart_init],
n_init=self.n_init, max_iter=self.max_iterations,
compute_silhouette_score=self.scoring == self.SILHOUETTE)
with self.progressBar(k_to - self.k_from + 1) as progress:
for k in range(self.k_from, k_to + 1):
progress.advance()
kmeans.params["n_clusters"] = k
self.optimization_runs.append((k, kmeans(self.data)))
finally:
self.controlArea.setDisabled(False)
self.show_results()
self.send_data()
def cluster(self):
if not self.check_data_size(self.k, self.Error):
return
self.km = KMeans(
n_clusters=self.k,
init=['random', 'k-means++'][self.smart_init],
n_init=self.n_init,
max_iter=self.max_iterations)(self.data)
self.send_data()
def run(self):
self.clear_messages()
if not self.data:
return
if self.optimize_k:
self.run_optimization()
else:
self.cluster()
def commit(self):
self.run()
def show_results(self):
minimize = self.SCORING_METHODS[self.scoring][2]
k_scores = [(k, self.SCORING_METHODS[self.scoring][1](run)) for
k, run in self.optimization_runs]
scores = list(zip(*k_scores))[1]
if minimize:
best_score, worst_score = min(scores), max(scores)
else:
best_score, worst_score = max(scores), min(scores)
best_run = scores.index(best_score)
score_span = (best_score - worst_score) or 1
max_score = max(scores)
nplaces = min(5, np.floor(abs(math.log(max(max_score, 1e-10)))) + 2)
fmt = "{{:.{}f}}".format(int(nplaces))
model = self.table_model
model.setRowCount(len(k_scores))
for i, (k, score) in enumerate(k_scores):
item = model.item(i, 0)
if item is None:
item = QStandardItem()
item.setData(k, Qt.DisplayRole)
item.setTextAlignment(Qt.AlignCenter)
model.setItem(i, 0, item)
item = model.item(i, 1)
if item is None:
item = QStandardItem()
item.setData(fmt.format(score) if not np.isnan(score) else 'out-of-memory error',
Qt.DisplayRole)
bar_ratio = 0.95 * (score - worst_score) / score_span
item.setData(bar_ratio, gui.TableBarItem.BarRole)
model.setItem(i, 1, item)
self.table_view.resizeRowsToContents()
self.table_view.selectRow(best_run)
self.table_view.show()
if minimize:
self.table_box.setTitle("Scoring (smaller is better)")
else:
self.table_box.setTitle("Scoring (bigger is better)")
QTimer.singleShot(0, self.adjustSize)
def update(self):
self.hide_show_opt_results()
self.run()
def selected_row(self):
indices = self.table_view.selectedIndexes()
rows = {ind.row() for ind in indices}
if len(rows) == 1:
return rows.pop()
def table_item_selected(self):
row = self.selected_row()
if row is not None:
self.send_data(row)
def send_data(self, row=None):
if self.optimize_k:
if row is None:
row = self.selected_row()
km = self.optimization_runs[row][1]
else:
km = self.km
if not self.data or not km:
self.send("Annotated Data", None)
self.send("Centroids", None)
return
clust_var = DiscreteVariable(
self.output_name, values=["C%d" % (x + 1) for x in range(km.k)])
clust_ids = km(self.data)
domain = self.data.domain
attributes, classes = domain.attributes, domain.class_vars
meta_attrs = domain.metas
if self.place_cluster_ids == self.OUTPUT_CLASS:
if classes:
meta_attrs += classes
classes = [clust_var]
elif self.place_cluster_ids == self.OUTPUT_ATTRIBUTE:
attributes += (clust_var, )
else:
meta_attrs += (clust_var, )
domain = Domain(attributes, classes, meta_attrs)
new_table = Table.from_table(domain, self.data)
new_table.get_column_view(clust_var)[0][:] = clust_ids.X.ravel()
centroids = Table(Domain(km.pre_domain.attributes), km.centroids)
self.send("Annotated Data", new_table)
self.send("Centroids", centroids)
@check_sql_input
def set_data(self, data):
self.data = data
if data is None:
self.table_model.setRowCount(0)
self.send("Annotated Data", None)
self.send("Centroids", None)
else:
self.data = data
self.run()
def send_report(self):
k_clusters = self.k
if self.optimize_k and self.optimization_runs and self.selected_row() is not None:
k_clusters = self.optimization_runs[self.selected_row()][1].k
self.report_items((
("Number of clusters", k_clusters),
("Optimization",
self.optimize_k != 0 and
"{}, {} re-runs limited to {} steps".format(
self.INIT_METHODS[self.smart_init].lower(),
self.n_init, self.max_iterations)),
("Cluster ID in output",
self.append_cluster_ids and
"'{}' (as {})".format(
self.output_name,
self.OUTPUT_METHODS[self.place_cluster_ids].lower()))
))
if self.data:
self.report_data("Data", self.data)
if self.optimize_k:
self.report_table(
"Scoring by {}".format(self.SCORING_METHODS[self.scoring][0]
),
self.table_view)
class ResolweDataWidget(QWidget):
def __init__(self, data_objects, descriptor_schema, *args, **kwargs):
super().__init__(*args, **kwargs)
self.ow = kwargs.get('parent', None)
self._data_objects = data_objects
self.descriptor_schema = descriptor_schema
self.header_schema = None
self.header = None
# set layout
layout = QVBoxLayout()
layout.setContentsMargins(0, 0, 0, 0)
self.setLayout(layout)
self.view = QTreeView()
self.view.setSortingEnabled(False)
self.view.setAlternatingRowColors(True)
self.view.setEditTriggers(QTreeView.NoEditTriggers)
self.view.setSelectionMode(QTreeView.SingleSelection)
self.model = QStandardItemModel()
self.display_data_objects()
self.layout().addWidget(self.view)
def __set_header_values(self):
if self.header_schema:
labels = [val.get('label', '?') for val in self.header_schema]
self.model.setHorizontalHeaderLabels(labels)
def __create_row(self, obj):
row_items = []
tabular_data = obj.descriptor.get('tabular', None)
output_data = obj.output.get('table', None)
# TODO: refactor this. Use file_name and size from obj.output instead of desc. schema
for schema_value in self.header_schema:
item = QStandardItem()
schema_key = schema_value['name']
data_info = tabular_data.get(schema_key,
'?') if tabular_data else '?'
if schema_key == 'file_name' and data_info == '?':
data_info = output_data.get('file',
'?') if output_data else '?'
elif schema_key == 'file_size' and data_info == '?':
data_info = output_data.get('size',
'?') if output_data else '?'
item.setData(data_info, Qt.DisplayRole)
row_items.append(item)
return row_items
def __populate_data_model(self):
if self.model:
self.model.clear()
for data_object in self.data_objects:
self.model.appendRow(self.__create_row(data_object))
def __parse_description_schema(self):
self.header_schema = []
if self.descriptor_schema:
for schema_value in self.descriptor_schema.schema:
if schema_value['name'] == 'tabular':
[
self.header_schema.append(value)
for value in schema_value['group']
]
if self.header_schema:
keys = [val.get('name', '?') for val in self.header_schema]
header_index = namedtuple('header_index',
[label for label in keys])
self.header = header_index(
*[index for index, _ in enumerate(keys)])
@property
def data_objects(self):
return self._data_objects
@data_objects.setter
def data_objects(self, data_objects):
self._data_objects = data_objects
self.display_data_objects()
def display_data_objects(self):
self.__parse_description_schema()
self.__populate_data_model()
self.__set_header_values()
self.view.setModel(self.model)
def set_target_column(self, target_column):
# type: (int) -> None
for row in range(self.model.rowCount()):
item = self.model.item(row, target_column)
item_data = item.data(role=Qt.DisplayRole)
if item_data:
item.setIcon(variable_icon(item_data))
def selected_data_object(self):
# type: () -> Data
rows = self.view.selectionModel().selectedRows()
assert 0 <= len(rows) <= 1
sel_row_index = rows[0].row() if rows else None
obj_range = range(len(self._data_objects))
assert sel_row_index in obj_range
try:
return self._data_objects[sel_row_index]
except IndexError:
# can this happen? self._data_objects can't
# be empty if model is constructed
pass
class OWGeneSets(OWWidget):
name = "Gene Set Enrichment"
description = ""
icon = "icons/OWGeneSets.svg"
priority = 9
want_main_area = True
settingsHandler = OrganismContextHandler()
# settings
auto_commit = Setting(True)
stored_selections = ContextSetting([])
organism = ContextSetting(None)
min_count = Setting(5)
use_min_count = Setting(True)
max_p_value = Setting(0.0001)
use_p_value = Setting(False)
max_fdr = Setting(0.01)
use_max_fdr = Setting(True)
use_reference_data = Setting(True)
COUNT, REFERENCE, P_VAL, FDR, ENRICHMENT, GENES, CATEGORY, TERM = range(8)
DATA_HEADER_LABELS = [
"Count", 'Reference', 'p-Value', 'FDR', 'Enrichment', 'Genes In Set',
'Category', 'Term'
]
class Inputs:
genes = Input("Genes", Table)
custom_sets = Input('Custom Gene Sets', Table)
reference = Input("Reference Genes", Table)
class Outputs:
matched_genes = Output("Matched Genes", Table)
class Information(OWWidget.Information):
pass
class Warning(OWWidget.Warning):
all_sets_filtered = Msg('All sets were filtered out.')
class Error(OWWidget.Error):
organism_mismatch = Msg(
'Organism in input data and custom gene sets does not match')
missing_annotation = Msg(ERROR_ON_MISSING_ANNOTATION)
missing_gene_id = Msg(ERROR_ON_MISSING_GENE_ID)
missing_tax_id = Msg(ERROR_ON_MISSING_TAX_ID)
cant_reach_host = Msg("Host orange.biolab.si is unreachable.")
cant_load_organisms = Msg(
"No available organisms, please check your connection.")
def __init__(self):
super().__init__()
# commit
self.commit_button = None
# gene sets object
self.gene_sets_obj = geneset.GeneSets()
# progress bar
self.progress_bar = None
self.progress_bar_iterations = None
# data
self.input_data = None
self.input_genes = []
self.tax_id = None
self.use_attr_names = None
self.gene_id_attribute = None
self.gene_id_column = None
# custom gene sets
self.custom_data = None
self.feature_model = DomainModel(valid_types=(DiscreteVariable,
StringVariable))
self.gene_set_label = None
self.gs_label_combobox = None
self.custom_tax_id = None
self.custom_use_attr_names = None
self.custom_gene_id_attribute = None
self.custom_gene_id_column = None
# reference genes
self.reference_radio_box = None
self.reference_data = None
self.reference_genes = None
self.reference_tax_id = None
self.reference_attr_names = None
self.reference_gene_id_attribute = None
self.reference_gene_id_column = None
# info box
self.input_info = None
self.num_of_sel_genes = 0
# filter
self.line_edit_filter = None
self.search_pattern = ''
self.organism_select_combobox = None
# data model view
self.data_view = None
self.data_model = None
# gene matcher NCBI
self.gene_matcher = None
# filter proxy model
self.filter_proxy_model = None
# hierarchy widget
self.hierarchy_widget = None
self.hierarchy_state = None
# spinbox
self.spin_widget = None
# threads
self.threadpool = QThreadPool(self)
self.workers = None
self._task = None # type: Optional[Task]
self._executor = ThreadExecutor()
# gui
self.setup_gui()
def __reset_widget_state(self):
# reset hierarchy widget state
self.hierarchy_widget.clear()
# clear data view
self.init_item_model()
# reset filters
self.setup_filter_model()
def cancel(self):
"""
Cancel the current task (if any).
"""
if self._task is not None:
self._task.cancel()
assert self._task.future.done()
# disconnect the `_task_finished` slot
self._task.watcher.done.disconnect(self._init_gene_sets_finished)
self._task = None
@Slot()
def progress_advance(self):
# GUI should be updated in main thread. That's why we are calling advance method here
if self.progress_bar:
self.progress_bar.advance()
def __get_input_genes(self):
self.input_genes = []
if self.use_attr_names:
for variable in self.input_data.domain.attributes:
self.input_genes.append(
str(variable.attributes.get(self.gene_id_attribute, '?')))
else:
genes, _ = self.input_data.get_column_view(self.gene_id_column)
self.input_genes = [str(g) for g in genes]
def __construct_custom_gene_sets(self):
custom_set_hier = ('Custom sets', )
# delete any custom sets if they exists
self.gene_sets_obj.delete_sets_by_hierarchy(custom_set_hier)
if self.custom_data and self.custom_gene_id_column:
gene_sets_names, _ = self.custom_data.get_column_view(
self.gene_set_label)
gene_names, _ = self.custom_data.get_column_view(
self.custom_gene_id_column)
temp_dict = defaultdict(list)
for set_name, gene_name in zip(gene_sets_names, gene_names):
temp_dict[set_name].append(gene_name)
g_sets = []
for key, value in temp_dict.items():
g_sets.append(
geneset.GeneSet(gs_id=key,
hierarchy=custom_set_hier,
organism=self.custom_tax_id,
name=key,
genes=set(value)))
self.gene_sets_obj.update(g_sets)
def __update_hierarchy(self):
self.set_hierarchy_model(
self.hierarchy_widget,
hierarchy_tree(self.gene_sets_obj.hierarchies()))
self.set_selected_hierarchies()
def update_tree_view(self):
if self.use_reference_data and self.reference_data:
self.init_gene_sets(reference_genes=self.reference_genes)
else:
self.init_gene_sets()
def invalidate(self):
# clear
self.__reset_widget_state()
self.update_info_box()
if self.input_data is not None:
# setup
self.__construct_custom_gene_sets()
self.__get_input_genes()
self.__update_hierarchy()
self.update_tree_view()
def __check_organism_mismatch(self):
""" Check if organisms from different inputs match.
:return: True if there is a mismatch
"""
if self.tax_id is not None and self.custom_tax_id is not None:
return self.tax_id != self.custom_tax_id
return False
def __get_reference_genes(self):
self.reference_genes = []
if self.reference_attr_names:
for variable in self.reference_data.domain.attributes:
self.reference_genes.append(
str(
variable.attributes.get(
self.reference_gene_id_attribute, '?')))
else:
genes, _ = self.reference_data.get_column_view(
self.reference_gene_id_column)
self.reference_genes = [str(g) for g in genes]
@Inputs.reference
def handle_reference_genes(self, data):
"""
Set the (optional) input dataset with reference gene names.
"""
if data:
self.reference_data = data
self.reference_tax_id = str(
self.reference_data.attributes.get(TAX_ID, None))
self.reference_attr_names = self.reference_data.attributes.get(
GENE_AS_ATTRIBUTE_NAME, None)
self.reference_gene_id_attribute = self.reference_data.attributes.get(
GENE_ID_ATTRIBUTE, None)
self.reference_gene_id_column = self.reference_data.attributes.get(
GENE_ID_COLUMN, None)
if not (self.reference_attr_names is not None and
((self.reference_gene_id_attribute is None) ^
(self.reference_gene_id_column is None))):
if self.reference_tax_id is None:
self.Error.missing_annotation()
return
self.Error.missing_gene_id()
return
elif self.reference_tax_id is None:
self.Error.missing_tax_id()
return
self.__get_reference_genes()
self.reference_radio_box.setEnabled(bool(self.reference_data))
self.invalidate()
@Inputs.custom_sets
def handle_custom_input(self, data):
self.Error.clear()
self.__reset_widget_state()
self.custom_data = None
self.custom_tax_id = None
self.custom_use_attr_names = None
self.custom_gene_id_attribute = None
self.custom_gene_id_column = None
self.gs_label_combobox.setDisabled(True)
self.feature_model.set_domain(None)
if data:
self.custom_data = data
self.custom_tax_id = str(
self.custom_data.attributes.get(TAX_ID, None))
self.custom_use_attr_names = self.custom_data.attributes.get(
GENE_AS_ATTRIBUTE_NAME, None)
self.custom_gene_id_attribute = self.custom_data.attributes.get(
GENE_ID_ATTRIBUTE, None)
self.custom_gene_id_column = self.custom_data.attributes.get(
GENE_ID_COLUMN, None)
if not (self.custom_use_attr_names is not None and
((self.custom_gene_id_attribute is None) ^
(self.custom_gene_id_column is None))):
if self.custom_tax_id is None:
self.Error.missing_annotation()
return
self.Error.missing_gene_id()
return
elif self.custom_tax_id is None:
self.Error.missing_tax_id()
return
if self.__check_organism_mismatch():
self.Error.organism_mismatch()
return
self.gs_label_combobox.setDisabled(False)
self.feature_model.set_domain(self.custom_data.domain)
if self.feature_model:
self.gene_set_label = self.feature_model[0]
self.invalidate()
@Inputs.genes
def handle_genes_input(self, data):
self.closeContext()
self.Error.clear()
self.__reset_widget_state()
# clear output
self.Outputs.matched_genes.send(None)
# clear input genes
self.input_genes = []
self.gs_label_combobox.setDisabled(True)
self.update_info_box()
if data:
self.input_data = data
self.tax_id = str(self.input_data.attributes.get(TAX_ID, None))
self.use_attr_names = self.input_data.attributes.get(
GENE_AS_ATTRIBUTE_NAME, None)
self.gene_id_attribute = self.input_data.attributes.get(
GENE_ID_ATTRIBUTE, None)
self.gene_id_column = self.input_data.attributes.get(
GENE_ID_COLUMN, None)
if not (self.use_attr_names is not None and
((self.gene_id_attribute is None) ^
(self.gene_id_column is None))):
if self.tax_id is None:
self.Error.missing_annotation()
return
self.Error.missing_gene_id()
return
elif self.tax_id is None:
self.Error.missing_tax_id()
return
if self.__check_organism_mismatch():
self.Error.organism_mismatch()
return
self.openContext(self.tax_id)
# if input data change, we need to set feature model again
if self.custom_data:
self.gs_label_combobox.setDisabled(False)
self.feature_model.set_domain(self.custom_data.domain)
if self.feature_model:
self.gene_set_label = self.feature_model[0]
self.download_gene_sets()
def update_info_box(self):
info_string = ''
if self.input_genes:
info_string += '{} unique gene names on input.\n'.format(
len(self.input_genes))
info_string += '{} genes on output.\n'.format(
self.num_of_sel_genes)
else:
info_string += 'No genes on input.\n'
self.input_info.setText(info_string)
def on_gene_sets_download(self, result):
# make sure this happens in the main thread.
# Qt insists that widgets be created within the GUI(main) thread.
assert threading.current_thread() == threading.main_thread()
self.setStatusMessage('')
if result:
for res in result:
g_sets = geneset.load_gene_sets(res, self.tax_id)
self.gene_sets_obj.update([g_set for g_set in g_sets])
# add custom sets if there are any
self.invalidate()
self.update_info_box()
def download_gene_sets(self):
self.Error.clear()
# reset hierarchy widget state
self.hierarchy_widget.clear()
# clear data view
self.init_item_model()
# get all gene sets for selected organism
gene_sets = geneset.list_all(organism=self.tax_id)
# status message
self.setStatusMessage('downloading sets')
worker = Worker(download_gene_sets, self.tax_id, gene_sets)
worker.signals.result.connect(self.on_gene_sets_download)
# move download process to worker thread
self.threadpool.start(worker)
def set_hierarchy_model(self, tree_widget, sets):
def beautify_displayed_text(text):
if '_' in text:
return text.replace('_', ' ').title()
else:
return text
# TODO: maybe optimize this code?
for key, value in sets.items():
item = QTreeWidgetItem(tree_widget, [beautify_displayed_text(key)])
item.setFlags(item.flags()
& (Qt.ItemIsUserCheckable | ~Qt.ItemIsSelectable
| Qt.ItemIsEnabled))
item.setExpanded(True)
item.hierarchy = key
if value:
item.setFlags(item.flags() | Qt.ItemIsTristate)
self.set_hierarchy_model(item, value)
else:
if item.parent():
item.hierarchy = (item.parent().hierarchy, key)
if not item.childCount() and not item.parent():
item.hierarchy = (key, )
def init_gene_sets(self, reference_genes=None):
if self._task is not None:
self.cancel()
assert self._task is None
self._task = Task()
progress_advance = methodinvoke(self, "progress_advance")
def callback():
if self._task.cancelled:
raise KeyboardInterrupt()
if self.progress_bar:
progress_advance()
if reference_genes is None:
reference_genes = self.gene_sets_obj.genes()
self.init_item_model()
sets_to_display = self.get_hierarchies(only_selected=True)
# save setting on selected hierarchies
self.stored_selections = sets_to_display
# save context
self.closeContext()
f = partial(self.set_items,
self.gene_sets_obj,
sets_to_display,
set(self.input_genes),
reference_genes,
self.min_count if self.use_min_count else 1,
callback=callback)
progress_iterations = sum([
len(g_set) for hier, g_set in
self.gene_sets_obj.map_hierarchy_to_sets().items()
if hier in sets_to_display
])
self.progress_bar = ProgressBar(self, iterations=progress_iterations)
self._task.future = self._executor.submit(f)
self._task.watcher = FutureWatcher(self._task.future)
self._task.watcher.done.connect(self._init_gene_sets_finished)
self.openContext(self.tax_id)
@Slot(concurrent.futures.Future)
def _init_gene_sets_finished(self, f):
assert self.thread() is QThread.currentThread()
assert threading.current_thread() == threading.main_thread()
assert self._task is not None
assert self._task.future is f
assert f.done()
self._task = None
self.progress_bar.finish()
self.setStatusMessage('')
try:
results = f.result() # type: list
[self.data_model.appendRow(model_item) for model_item in results]
self.filter_proxy_model.setSourceModel(self.data_model)
self._update_fdr()
self.filter_data_view()
except Exception as ex:
print(ex)
def set_selected_hierarchies(self):
iterator = QTreeWidgetItemIterator(self.hierarchy_widget,
QTreeWidgetItemIterator.All)
while iterator.value():
# note: if hierarchy value is not a tuple, then this is just top level qTreeWidgetItem that
# holds subcategories. We don't want to display all sets from category
if type(iterator.value().hierarchy) is not str:
if iterator.value().hierarchy in self.stored_selections:
iterator.value().setCheckState(0, Qt.Checked)
else:
iterator.value().setCheckState(0, Qt.Unchecked)
iterator += 1
# if no items are checked, we check first one at random
if len(self.get_hierarchies(only_selected=True)) == 0:
iterator = QTreeWidgetItemIterator(
self.hierarchy_widget, QTreeWidgetItemIterator.NotChecked)
while iterator.value():
if type(iterator.value().hierarchy) is not str:
iterator.value().setCheckState(0, Qt.Checked)
return
iterator += 1
def get_hierarchies(self, **kwargs):
""" return selected hierarchy
"""
only_selected = kwargs.get('only_selected', None)
sets_to_display = list()
if only_selected:
iterator = QTreeWidgetItemIterator(self.hierarchy_widget,
QTreeWidgetItemIterator.Checked)
else:
iterator = QTreeWidgetItemIterator(self.hierarchy_widget)
while iterator.value():
# note: if hierarchy value is not a tuple, then this is just top level qTreeWidgetItem that
# holds subcategories. We don't want to display all sets from category
if type(iterator.value().hierarchy) is not str:
if not only_selected:
sets_to_display.append(iterator.value().hierarchy)
else:
if not iterator.value().isDisabled():
sets_to_display.append(iterator.value().hierarchy)
iterator += 1
return sets_to_display
def filter_data_view(self):
filter_proxy = self.filter_proxy_model # type: FilterProxyModel
model = filter_proxy.sourceModel() # type: QStandardItemModel
assert isinstance(model, QStandardItemModel)
search_term = self.search_pattern.lower().strip().split()
# apply filtering rules
filters = [
FilterProxyModel.Filter(
self.TERM, Qt.DisplayRole,
lambda value: all(fs in value.lower() for fs in search_term))
]
# if self.use_min_count:
# filters.append(
# FilterProxyModel.Filter(
# self.COUNT, Qt.DisplayRole,
# lambda value: value >= self.min_count,
# )
# )
if self.use_p_value:
filters.append(
FilterProxyModel.Filter(
self.P_VAL, Qt.DisplayRole,
lambda value: value < self.max_p_value))
if self.use_max_fdr:
filters.append(
FilterProxyModel.Filter(self.FDR, Qt.DisplayRole,
lambda value: value < self.max_fdr))
filter_proxy.set_filters(filters)
if model.rowCount() and not filter_proxy.rowCount():
self.Warning.all_sets_filtered()
else:
self.Warning.clear()
def __get_source_data(self, proxy_row_index, column):
proxy_index = self.filter_proxy_model.index(proxy_row_index, column)
source_index = self.filter_proxy_model.mapToSource(proxy_index)
return source_index.data(role=Qt.DisplayRole)
def _update_fdr(self):
# Update the FDR in place due to a changed selected categories set and
# results for all of these categories are already available.
proxy = self.filter_proxy_model
model = self.filter_proxy_model.sourceModel()
if model is not None:
assert isinstance(model, QStandardItemModel)
p_values = [(i, self.__get_source_data(i, self.P_VAL))
for i in range(proxy.rowCount())]
fdr_values = FDR([p_val for _, p_val in p_values])
for i, fdr_val in zip([i for i, _ in p_values], fdr_values):
proxy_index = proxy.index(i, self.FDR)
source_index = self.filter_proxy_model.mapToSource(proxy_index)
source_item = model.item(source_index.row(), self.FDR)
source_item.setData(fdr_val, role=Qt.DisplayRole)
source_item.setData(fdr_val, role=Qt.ToolTipRole)
def commit(self):
selection_model = self.data_view.selectionModel()
if selection_model:
# genes_from_set = selection_model.selectedRows(GENES)
matched_genes = selection_model.selectedRows(self.COUNT)
if matched_genes and self.input_genes:
genes = [
model_index.data(Qt.UserRole)
for model_index in matched_genes
]
output_genes = [
gene_name for gene_name in list(set.union(*genes))
]
self.num_of_sel_genes = len(output_genes)
self.update_info_box()
if self.use_attr_names:
selected = [
column for column in self.input_data.domain.attributes
if self.gene_id_attribute in column.attributes
and str(column.attributes[
self.gene_id_attribute]) in output_genes
]
domain = Domain(selected,
self.input_data.domain.class_vars,
self.input_data.domain.metas)
new_data = self.input_data.from_table(
domain, self.input_data)
self.Outputs.matched_genes.send(new_data)
else:
selected_rows = []
for row_index, row in enumerate(self.input_data):
gene_in_row = str(row[self.gene_id_column])
if gene_in_row in self.input_genes and gene_in_row in output_genes:
selected_rows.append(row_index)
if selected_rows:
selected = self.input_data[selected_rows]
else:
selected = None
self.Outputs.matched_genes.send(selected)
def assign_delegates(self):
self.data_view.setItemDelegateForColumn(self.GENES,
NumericalColumnDelegate(self))
self.data_view.setItemDelegateForColumn(self.COUNT,
NumericalColumnDelegate(self))
self.data_view.setItemDelegateForColumn(self.REFERENCE,
NumericalColumnDelegate(self))
self.data_view.setItemDelegateForColumn(
self.P_VAL, NumericalColumnDelegate(self,
precision=2,
notation='e'))
self.data_view.setItemDelegateForColumn(
self.FDR, NumericalColumnDelegate(self, precision=2, notation='e'))
self.data_view.setItemDelegateForColumn(
self.ENRICHMENT, NumericalColumnDelegate(self, precision=1))
def setup_filter_model(self):
self.filter_proxy_model = FilterProxyModel()
self.filter_proxy_model.setFilterKeyColumn(self.TERM)
self.data_view.setModel(self.filter_proxy_model)
def setup_filter_area(self):
h_layout = QHBoxLayout()
h_layout.setSpacing(100)
h_widget = widgetBox(self.mainArea, orientation=h_layout)
spin(h_widget,
self,
'min_count',
0,
100,
label='Count',
tooltip='Minimum genes count',
checked='use_min_count',
callback=self.invalidate,
callbackOnReturn=True,
checkCallback=self.invalidate)
doubleSpin(h_widget,
self,
'max_p_value',
0.0,
1.0,
0.0001,
label='p-value',
tooltip='Maximum p-value of the enrichment score',
checked='use_p_value',
callback=self.filter_data_view,
callbackOnReturn=True,
checkCallback=self.filter_data_view)
doubleSpin(h_widget,
self,
'max_fdr',
0.0,
1.0,
0.0001,
label='FDR',
tooltip='Maximum false discovery rate',
checked='use_max_fdr',
callback=self.filter_data_view,
callbackOnReturn=True,
checkCallback=self.filter_data_view)
self.line_edit_filter = lineEdit(h_widget, self, 'search_pattern')
self.line_edit_filter.setPlaceholderText('Filter gene sets ...')
self.line_edit_filter.textChanged.connect(self.filter_data_view)
def setup_control_area(self):
info_box = vBox(self.controlArea, 'Info')
self.input_info = widgetLabel(info_box)
box = vBox(self.controlArea, "Custom Gene Sets")
self.gs_label_combobox = comboBox(box,
self,
"gene_set_label",
sendSelectedValue=True,
model=self.feature_model,
callback=self.invalidate)
self.gs_label_combobox.setDisabled(True)
self.reference_radio_box = radioButtonsInBox(
self.controlArea,
self,
"use_reference_data",
["Entire genome", "Reference gene set (input)"],
tooltips=[
"Use entire genome (for gene set enrichment)",
"Use reference set of genes"
],
box="Reference",
callback=self.invalidate)
self.reference_radio_box.setEnabled(False)
hierarchy_box = widgetBox(self.controlArea, "Gene Set Categories")
self.hierarchy_widget = QTreeWidget(self)
self.hierarchy_widget.setEditTriggers(QTreeView.NoEditTriggers)
self.hierarchy_widget.setHeaderLabels([' '])
self.hierarchy_widget.itemClicked.connect(self.update_tree_view)
hierarchy_box.layout().addWidget(self.hierarchy_widget)
self.commit_button = auto_commit(self.controlArea,
self,
"auto_commit",
"&Commit",
box=False)
def setup_gui(self):
# control area
self.setup_control_area()
# main area
self.data_view = QTreeView()
self.setup_filter_model()
self.setup_filter_area()
self.data_view.setAlternatingRowColors(True)
self.data_view.sortByColumn(self.COUNT, Qt.DescendingOrder)
self.data_view.setSortingEnabled(True)
self.data_view.setSelectionMode(QTreeView.ExtendedSelection)
self.data_view.setEditTriggers(QTreeView.NoEditTriggers)
self.data_view.viewport().setMouseTracking(False)
self.data_view.setItemDelegateForColumn(
self.TERM, LinkStyledItemDelegate(self.data_view))
self.data_view.selectionModel().selectionChanged.connect(self.commit)
self.mainArea.layout().addWidget(self.data_view)
self.data_view.header().setSectionResizeMode(
QHeaderView.ResizeToContents)
self.assign_delegates()
@staticmethod
def set_items(gene_sets, sets_to_display, genes, ref, count_treshold,
callback):
model_items = []
if not genes:
return
for gene_set in gene_sets:
if gene_set.hierarchy not in sets_to_display:
continue
enrichemnt_result = gene_set.set_enrichment(
ref, genes.intersection(ref))
callback()
if len(enrichemnt_result.query) >= count_treshold:
category_column = QStandardItem()
name_column = QStandardItem()
count_column = QStandardItem()
genes_column = QStandardItem()
ref_column = QStandardItem()
pval_column = QStandardItem()
fdr_column = QStandardItem()
enrichemnt_column = QStandardItem()
category_column.setData(", ".join(gene_set.hierarchy),
Qt.DisplayRole)
name_column.setData(gene_set.name, Qt.DisplayRole)
name_column.setData(gene_set.name, Qt.ToolTipRole)
name_column.setData(gene_set.link, LinkRole)
name_column.setForeground(QColor(Qt.blue))
count_column.setData(len(enrichemnt_result.query),
Qt.DisplayRole)
count_column.setData(set(enrichemnt_result.query), Qt.UserRole)
genes_column.setData(len(gene_set.genes), Qt.DisplayRole)
genes_column.setData(
set(gene_set.genes), Qt.UserRole
) # store genes to get then on output on selection
ref_column.setData(len(enrichemnt_result.reference),
Qt.DisplayRole)
pval_column.setData(enrichemnt_result.p_value, Qt.DisplayRole)
pval_column.setData(enrichemnt_result.p_value, Qt.ToolTipRole)
enrichemnt_column.setData(enrichemnt_result.enrichment_score,
Qt.DisplayRole)
enrichemnt_column.setData(enrichemnt_result.enrichment_score,
Qt.ToolTipRole)
model_items.append([
count_column, ref_column, pval_column, fdr_column,
enrichemnt_column, genes_column, category_column,
name_column
])
return model_items
def init_item_model(self):
if self.data_model:
self.data_model.clear()
self.setup_filter_model()
else:
self.data_model = QStandardItemModel()
self.data_model.setSortRole(Qt.UserRole)
self.data_model.setHorizontalHeaderLabels(self.DATA_HEADER_LABELS)
def sizeHint(self):
return QSize(1280, 960)
def __init__(self):
super().__init__()
self.local_cache_path = os.path.join(data_dir(), self.DATASET_DIR)
self.__awaiting_state = None # type: Optional[_FetchState]
box = gui.widgetBox(self.controlArea, "Info")
self.infolabel = QLabel(text="Initializing...\n\n")
box.layout().addWidget(self.infolabel)
gui.widgetLabel(self.mainArea, "Filter")
self.filterLineEdit = QLineEdit(
textChanged=self.filter
)
self.mainArea.layout().addWidget(self.filterLineEdit)
self.splitter = QSplitter(orientation=Qt.Vertical)
self.view = QTreeView(
sortingEnabled=True,
selectionMode=QTreeView.SingleSelection,
alternatingRowColors=True,
rootIsDecorated=False,
editTriggers=QTreeView.NoEditTriggers,
)
box = gui.widgetBox(self.splitter, "Description", addToLayout=False)
self.descriptionlabel = QLabel(
wordWrap=True,
textFormat=Qt.RichText,
)
self.descriptionlabel = QTextBrowser(
openExternalLinks=True,
textInteractionFlags=(Qt.TextSelectableByMouse |
Qt.LinksAccessibleByMouse)
)
self.descriptionlabel.setFrameStyle(QTextBrowser.NoFrame)
# no (white) text background
self.descriptionlabel.viewport().setAutoFillBackground(False)
box.layout().addWidget(self.descriptionlabel)
self.splitter.addWidget(self.view)
self.splitter.addWidget(box)
self.splitter.setSizes([300, 200])
self.splitter.splitterMoved.connect(
lambda:
setattr(self, "splitter_state", bytes(self.splitter.saveState()))
)
self.mainArea.layout().addWidget(self.splitter)
self.controlArea.layout().addStretch(10)
gui.auto_commit(self.controlArea, self, "auto_commit", "Send Data")
model = QStandardItemModel(self)
model.setHorizontalHeaderLabels(HEADER)
proxy = QSortFilterProxyModel()
proxy.setSourceModel(model)
proxy.setFilterKeyColumn(-1)
proxy.setFilterCaseSensitivity(False)
self.view.setModel(proxy)
if self.splitter_state:
self.splitter.restoreState(self.splitter_state)
self.view.setItemDelegateForColumn(
Header.Size, SizeDelegate(self))
self.view.setItemDelegateForColumn(
Header.Local, gui.IndicatorItemDelegate(self, role=Qt.DisplayRole))
self.view.setItemDelegateForColumn(
Header.Instances, NumericalDelegate(self))
self.view.setItemDelegateForColumn(
Header.Variables, NumericalDelegate(self))
self.view.resizeColumnToContents(Header.Local)
if self.header_state:
self.view.header().restoreState(self.header_state)
self.setBlocking(True)
self.setStatusMessage("Initializing")
self._executor = ThreadPoolExecutor(max_workers=1)
f = self._executor.submit(self.list_remote)
w = FutureWatcher(f, parent=self)
w.done.connect(self.__set_index)
def __on_enrichment_finished(self, results):
assert QThread.currentThread() is self.thread()
self.__state &= ~OWSetEnrichment.RunningEnrichment
query, reference, results = results
if self.annotationsChartView.model():
self.annotationsChartView.model().clear()
nquery = len(query)
nref = len(reference)
maxcount = max((len(e.query_mapped) for _, e in results),
default=1)
maxrefcount = max((len(e.reference_mapped) for _, e in results),
default=1)
nspaces = int(math.ceil(math.log10(maxcount or 1)))
refspaces = int(math.ceil(math.log(maxrefcount or 1)))
query_fmt = "%" + str(nspaces) + "s (%.2f%%)"
ref_fmt = "%" + str(refspaces) + "s (%.2f%%)"
def fmt_count(fmt, count, total):
return fmt % (count, 100.0 * count / (total or 1))
fmt_query_count = partial(fmt_count, query_fmt)
fmt_ref_count = partial(fmt_count, ref_fmt)
linkFont = QFont(self.annotationsChartView.viewOptions().font)
linkFont.setUnderline(True)
def item(value=None, tooltip=None, user=None):
si = QStandardItem()
if value is not None:
si.setData(value, Qt.DisplayRole)
if tooltip is not None:
si.setData(tooltip, Qt.ToolTipRole)
if user is not None:
si.setData(user, Qt.UserRole)
else:
si.setData(value, Qt.UserRole)
return si
model = QStandardItemModel()
model.setSortRole(Qt.UserRole)
model.setHorizontalHeaderLabels(
["Category", "Term", "Count", "Reference count", "p-value",
"FDR", "Enrichment"])
for i, (gset, enrich) in enumerate(results):
if len(enrich.query_mapped) == 0:
continue
nquery_mapped = len(enrich.query_mapped)
nref_mapped = len(enrich.reference_mapped)
row = [
item(", ".join(gset.hierarchy)),
item(gsname(gset), tooltip=gset.link),
item(fmt_query_count(nquery_mapped, nquery),
tooltip=nquery_mapped, user=nquery_mapped),
item(fmt_ref_count(nref_mapped, nref),
tooltip=nref_mapped, user=nref_mapped),
item(fmtp(enrich.p_value), user=enrich.p_value),
item(), # column 5, FDR, is computed in filterAnnotationsChartView
item(enrich.enrichment_score,
tooltip="%.3f" % enrich.enrichment_score,
user=enrich.enrichment_score)
]
row[0].geneset = gset
row[0].enrichment = enrich
row[1].setData(gset.link, gui.LinkRole)
row[1].setFont(linkFont)
row[1].setForeground(QColor(Qt.blue))
model.appendRow(row)
self.annotationsChartView.setModel(model)
self.annotationsChartView.selectionModel().selectionChanged.connect(
self.commit
)
if not model.rowCount():
self.warning(0, "No enriched sets found.")
else:
self.warning(0)
allnames = set(gsname(geneset)
for geneset, (count, _, _, _) in results if count)
allnames |= reduce(operator.ior,
(set(word_split(name)) for name in allnames),
set())
self.filterCompleter.setModel(None)
self.completerModel = QStringListModel(sorted(allnames))
self.filterCompleter.setModel(self.completerModel)
if results:
max_score = max((e.enrichment_score for _, e in results
if np.isfinite(e.enrichment_score)),
default=1)
self.annotationsChartView.setItemDelegateForColumn(
6, BarItemDelegate(self, scale=(0.0, max_score))
)
self.annotationsChartView.setItemDelegateForColumn(
1, gui.LinkStyledItemDelegate(self.annotationsChartView)
)
header = self.annotationsChartView.header()
for i in range(model.columnCount()):
sh = self.annotationsChartView.sizeHintForColumn(i)
sh = max(sh, header.sectionSizeHint(i))
self.annotationsChartView.setColumnWidth(i, max(min(sh, 300), 30))
# self.annotationsChartView.resizeColumnToContents(i)
self.filterAnnotationsChartView()
self.progressBarFinished()
self.setStatusMessage("")
class ScoreTable(OWComponent, QObject):
shown_scores = \
Setting(set(chain(*BUILTIN_SCORERS_ORDER.values())))
shownScoresChanged = Signal()
class ItemDelegate(QStyledItemDelegate):
def sizeHint(self, *args):
size = super().sizeHint(*args)
return QSize(size.width(), size.height() + 6)
def displayText(self, value, locale):
if isinstance(value, float):
return f"{value:.3f}"
else:
return super().displayText(value, locale)
def __init__(self, master):
QObject.__init__(self)
OWComponent.__init__(self, master)
self.view = gui.TableView(wordWrap=True,
editTriggers=gui.TableView.NoEditTriggers)
header = self.view.horizontalHeader()
header.setSectionResizeMode(QHeaderView.ResizeToContents)
header.setDefaultAlignment(Qt.AlignCenter)
header.setStretchLastSection(False)
header.setContextMenuPolicy(Qt.CustomContextMenu)
header.customContextMenuRequested.connect(self.show_column_chooser)
self.model = QStandardItemModel(master)
self.model.setHorizontalHeaderLabels(["Method"])
self.sorted_model = ScoreModel()
self.sorted_model.setSourceModel(self.model)
self.view.setModel(self.sorted_model)
self.view.setItemDelegate(self.ItemDelegate())
def _column_names(self):
return (self.model.horizontalHeaderItem(section).data(Qt.DisplayRole)
for section in range(1, self.model.columnCount()))
def show_column_chooser(self, pos):
# pylint doesn't know that self.shown_scores is a set, not a Setting
# pylint: disable=unsupported-membership-test
def update(col_name, checked):
if checked:
self.shown_scores.add(col_name)
else:
self.shown_scores.remove(col_name)
self._update_shown_columns()
menu = QMenu()
header = self.view.horizontalHeader()
for col_name in self._column_names():
action = menu.addAction(col_name)
action.setCheckable(True)
action.setChecked(col_name in self.shown_scores)
action.triggered.connect(partial(update, col_name))
menu.exec(header.mapToGlobal(pos))
def _update_shown_columns(self):
# pylint doesn't know that self.shown_scores is a set, not a Setting
# pylint: disable=unsupported-membership-test
header = self.view.horizontalHeader()
for section, col_name in enumerate(self._column_names(), start=1):
header.setSectionHidden(section, col_name not in self.shown_scores)
self.view.resizeColumnsToContents()
self.shownScoresChanged.emit()
def update_header(self, scorers):
# Set the correct horizontal header labels on the results_model.
self.model.setColumnCount(3 + len(scorers))
self.model.setHorizontalHeaderItem(0, QStandardItem("Model"))
self.model.setHorizontalHeaderItem(1, QStandardItem("Train time [s]"))
self.model.setHorizontalHeaderItem(2, QStandardItem("Test time [s]"))
for col, score in enumerate(scorers, start=3):
item = QStandardItem(score.name)
item.setToolTip(score.long_name)
self.model.setHorizontalHeaderItem(col, item)
self._update_shown_columns()
class OWTestLearners(OWWidget):
name = "Test & Score"
description = "Cross-validation accuracy estimation."
icon = "icons/TestLearners1.svg"
priority = 100
class Inputs:
train_data = Input("Data", Table, default=True)
test_data = Input("Test Data", Table)
learner = Input("Learner", Learner, multiple=True)
preprocessor = Input("Preprocessor", Preprocess)
class Outputs:
predictions = Output("Predictions", Table)
evaluations_results = Output("Evaluation Results", Results)
settings_version = 3
UserAdviceMessages = [
widget.Message(
"Click on the table header to select shown columns",
"click_header")]
settingsHandler = settings.PerfectDomainContextHandler(metas_in_res=True)
#: Resampling/testing types
KFold, FeatureFold, ShuffleSplit, LeaveOneOut, TestOnTrain, TestOnTest \
= 0, 1, 2, 3, 4, 5
#: Numbers of folds
NFolds = [2, 3, 5, 10, 20]
#: Number of repetitions
NRepeats = [2, 3, 5, 10, 20, 50, 100]
#: Sample sizes
SampleSizes = [5, 10, 20, 25, 30, 33, 40, 50, 60, 66, 70, 75, 80, 90, 95]
#: Selected resampling type
resampling = settings.Setting(0)
#: Number of folds for K-fold cross validation
n_folds = settings.Setting(3)
#: Stratified sampling for K-fold
cv_stratified = settings.Setting(True)
#: Number of repeats for ShuffleSplit sampling
n_repeats = settings.Setting(3)
#: ShuffleSplit sample size
sample_size = settings.Setting(9)
#: Stratified sampling for Random Sampling
shuffle_stratified = settings.Setting(True)
# CV where nr. of feature values determines nr. of folds
fold_feature = settings.ContextSetting(None)
fold_feature_selected = settings.ContextSetting(False)
TARGET_AVERAGE = "(Average over classes)"
class_selection = settings.ContextSetting(TARGET_AVERAGE)
BUILTIN_ORDER = {
DiscreteVariable: ("AUC", "CA", "F1", "Precision", "Recall"),
ContinuousVariable: ("MSE", "RMSE", "MAE", "R2")}
shown_scores = \
settings.Setting(set(chain(*BUILTIN_ORDER.values())))
class Error(OWWidget.Error):
train_data_empty = Msg("Train data set is empty.")
test_data_empty = Msg("Test data set is empty.")
class_required = Msg("Train data input requires a target variable.")
too_many_classes = Msg("Too many target variables.")
class_required_test = Msg("Test data input requires a target variable.")
too_many_folds = Msg("Number of folds exceeds the data size")
class_inconsistent = Msg("Test and train data sets "
"have different target variables.")
memory_error = Msg("Not enough memory.")
only_one_class_var_value = Msg("Target variable has only one value.")
class Warning(OWWidget.Warning):
missing_data = \
Msg("Instances with unknown target values were removed from{}data.")
test_data_missing = Msg("Missing separate test data input.")
scores_not_computed = Msg("Some scores could not be computed.")
test_data_unused = Msg("Test data is present but unused. "
"Select 'Test on test data' to use it.")
class Information(OWWidget.Information):
data_sampled = Msg("Train data has been sampled")
test_data_sampled = Msg("Test data has been sampled")
def __init__(self):
super().__init__()
self.data = None
self.test_data = None
self.preprocessor = None
self.train_data_missing_vals = False
self.test_data_missing_vals = False
self.scorers = []
#: An Ordered dictionary with current inputs and their testing results.
self.learners = OrderedDict() # type: Dict[Any, Input]
self.__state = State.Waiting
# Do we need to [re]test any learners, set by _invalidate and
# cleared by __update
self.__needupdate = False
self.__task = None # type: Optional[Task]
self.__executor = ThreadExecutor()
sbox = gui.vBox(self.controlArea, "Sampling")
rbox = gui.radioButtons(
sbox, self, "resampling", callback=self._param_changed)
gui.appendRadioButton(rbox, "Cross validation")
ibox = gui.indentedBox(rbox)
gui.comboBox(
ibox, self, "n_folds", label="Number of folds: ",
items=[str(x) for x in self.NFolds], maximumContentsLength=3,
orientation=Qt.Horizontal, callback=self.kfold_changed)
gui.checkBox(
ibox, self, "cv_stratified", "Stratified",
callback=self.kfold_changed)
gui.appendRadioButton(rbox, "Cross validation by feature")
ibox = gui.indentedBox(rbox)
self.feature_model = DomainModel(
order=DomainModel.METAS, valid_types=DiscreteVariable)
self.features_combo = gui.comboBox(
ibox, self, "fold_feature", model=self.feature_model,
orientation=Qt.Horizontal, callback=self.fold_feature_changed)
gui.appendRadioButton(rbox, "Random sampling")
ibox = gui.indentedBox(rbox)
gui.comboBox(
ibox, self, "n_repeats", label="Repeat train/test: ",
items=[str(x) for x in self.NRepeats], maximumContentsLength=3,
orientation=Qt.Horizontal, callback=self.shuffle_split_changed)
gui.comboBox(
ibox, self, "sample_size", label="Training set size: ",
items=["{} %".format(x) for x in self.SampleSizes],
maximumContentsLength=5, orientation=Qt.Horizontal,
callback=self.shuffle_split_changed)
gui.checkBox(
ibox, self, "shuffle_stratified", "Stratified",
callback=self.shuffle_split_changed)
gui.appendRadioButton(rbox, "Leave one out")
gui.appendRadioButton(rbox, "Test on train data")
gui.appendRadioButton(rbox, "Test on test data")
self.cbox = gui.vBox(self.controlArea, "Target Class")
self.class_selection_combo = gui.comboBox(
self.cbox, self, "class_selection", items=[],
sendSelectedValue=True, valueType=str,
callback=self._on_target_class_changed,
contentsLength=8)
gui.rubber(self.controlArea)
self.view = gui.TableView(
wordWrap=True,
)
header = self.view.horizontalHeader()
header.setSectionResizeMode(QHeaderView.ResizeToContents)
header.setDefaultAlignment(Qt.AlignCenter)
header.setStretchLastSection(False)
header.setContextMenuPolicy(Qt.CustomContextMenu)
header.customContextMenuRequested.connect(self.show_column_chooser)
self.result_model = QStandardItemModel(self)
self.result_model.setHorizontalHeaderLabels(["Method"])
self.view.setModel(self.result_model)
self.view.setItemDelegate(ItemDelegate())
box = gui.vBox(self.mainArea, "Evaluation Results")
box.layout().addWidget(self.view)
def sizeHint(self):
return QSize(780, 1)
def _update_controls(self):
self.fold_feature = None
self.feature_model.set_domain(None)
if self.data:
self.feature_model.set_domain(self.data.domain)
if self.fold_feature is None and self.feature_model:
self.fold_feature = self.feature_model[0]
enabled = bool(self.feature_model)
self.controls.resampling.buttons[
OWTestLearners.FeatureFold].setEnabled(enabled)
self.features_combo.setEnabled(enabled)
if self.resampling == OWTestLearners.FeatureFold and not enabled:
self.resampling = OWTestLearners.KFold
@Inputs.learner
def set_learner(self, learner, key):
"""
Set the input `learner` for `key`.
Parameters
----------
learner : Optional[Orange.base.Learner]
key : Any
"""
if key in self.learners and learner is None:
# Removed
self._invalidate([key])
del self.learners[key]
else:
self.learners[key] = InputLearner(learner, None, None)
self._invalidate([key])
@Inputs.train_data
def set_train_data(self, data):
"""
Set the input training dataset.
Parameters
----------
data : Optional[Orange.data.Table]
"""
self.Information.data_sampled.clear()
self.Error.train_data_empty.clear()
self.Error.class_required.clear()
self.Error.too_many_classes.clear()
self.Error.only_one_class_var_value.clear()
if data is not None and not len(data):
self.Error.train_data_empty()
data = None
if data:
conds = [not data.domain.class_vars,
len(data.domain.class_vars) > 1,
data.domain.has_discrete_class and len(data.domain.class_var.values) == 1]
errors = [self.Error.class_required,
self.Error.too_many_classes,
self.Error.only_one_class_var_value]
for cond, error in zip(conds, errors):
if cond:
error()
data = None
break
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.train_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = HasClass()(data)
else:
self.Warning.missing_data.clear()
self.data = data
self.closeContext()
self._update_scorers()
self._update_controls()
if data is not None:
self._update_class_selection()
self.openContext(data.domain)
if self.fold_feature_selected and bool(self.feature_model):
self.resampling = OWTestLearners.FeatureFold
self._invalidate()
@Inputs.test_data
def set_test_data(self, data):
# type: (Orange.data.Table) -> None
"""
Set the input separate testing dataset.
Parameters
----------
data : Optional[Orange.data.Table]
"""
self.Information.test_data_sampled.clear()
self.Error.test_data_empty.clear()
if data is not None and not len(data):
self.Error.test_data_empty()
data = None
if data and not data.domain.class_var:
self.Error.class_required_test()
data = None
else:
self.Error.class_required_test.clear()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.test_data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.test_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = HasClass()(data)
else:
self.Warning.missing_data.clear()
self.test_data = data
if self.resampling == OWTestLearners.TestOnTest:
self._invalidate()
def _which_missing_data(self):
return {(True, True): " ", # both, don't specify
(True, False): " train ",
(False, True): " test "}[(self.train_data_missing_vals,
self.test_data_missing_vals)]
# List of scorers shouldn't be retrieved globally, when the module is
# loading since add-ons could have registered additional scorers.
# It could have been cached but
# - we don't gain much with it
# - it complicates the unit tests
def _update_scorers(self):
if self.data is None or self.data.domain.class_var is None:
self.scorers = []
return
class_var = self.data and self.data.domain.class_var
order = {name: i
for i, name in enumerate(self.BUILTIN_ORDER[type(class_var)])}
# 'abstract' is retrieved from __dict__ to avoid inheriting
usable = (cls for cls in scoring.Score.registry.values()
if cls.is_scalar and not cls.__dict__.get("abstract")
and isinstance(class_var, cls.class_types))
self.scorers = sorted(usable, key=lambda cls: order.get(cls.name, 99))
@Inputs.preprocessor
def set_preprocessor(self, preproc):
"""
Set the input preprocessor to apply on the training data.
"""
self.preprocessor = preproc
self._invalidate()
def handleNewSignals(self):
"""Reimplemented from OWWidget.handleNewSignals."""
self._update_class_selection()
self._update_header()
self._update_stats_model()
if self.__needupdate:
self.__update()
def kfold_changed(self):
self.resampling = OWTestLearners.KFold
self._param_changed()
def fold_feature_changed(self):
self.resampling = OWTestLearners.FeatureFold
self._param_changed()
def shuffle_split_changed(self):
self.resampling = OWTestLearners.ShuffleSplit
self._param_changed()
def _param_changed(self):
self._invalidate()
self.__update()
def _update_header(self):
# Set the correct horizontal header labels on the results_model.
model = self.result_model
model.setColumnCount(1 + len(self.scorers))
for col, score in enumerate(self.scorers):
item = QStandardItem(score.name)
item.setToolTip(score.long_name)
model.setHorizontalHeaderItem(col + 1, item)
self._update_shown_columns()
def _update_shown_columns(self):
# pylint doesn't know that self.shown_scores is a set, not a Setting
# pylint: disable=unsupported-membership-test
model = self.result_model
header = self.view.horizontalHeader()
for section in range(1, model.columnCount()):
col_name = model.horizontalHeaderItem(section).data(Qt.DisplayRole)
header.setSectionHidden(section, col_name not in self.shown_scores)
def _update_stats_model(self):
# Update the results_model with up to date scores.
# Note: The target class specific scores (if requested) are
# computed as needed in this method.
model = self.view.model()
# clear the table model, but preserving the header labels
for r in reversed(range(model.rowCount())):
model.takeRow(r)
target_index = None
if self.data is not None:
class_var = self.data.domain.class_var
if self.data.domain.has_discrete_class and \
self.class_selection != self.TARGET_AVERAGE:
target_index = class_var.values.index(self.class_selection)
else:
class_var = None
errors = []
has_missing_scores = False
for key, slot in self.learners.items():
name = learner_name(slot.learner)
head = QStandardItem(name)
head.setData(key, Qt.UserRole)
if isinstance(slot.results, Try.Fail):
head.setToolTip(str(slot.results.exception))
head.setText("{} (error)".format(name))
head.setForeground(QtGui.QBrush(Qt.red))
errors.append("{name} failed with error:\n"
"{exc.__class__.__name__}: {exc!s}"
.format(name=name, exc=slot.results.exception))
row = [head]
if class_var is not None and class_var.is_discrete and \
target_index is not None:
if slot.results is not None and slot.results.success:
ovr_results = results_one_vs_rest(
slot.results.value, target_index)
# Cell variable is used immediatelly, it's not stored
# pylint: disable=cell-var-from-loop
stats = [Try(scorer_caller(scorer, ovr_results))
for scorer in self.scorers]
else:
stats = None
else:
stats = slot.stats
if stats is not None:
for stat in stats:
item = QStandardItem()
if stat.success:
item.setText("{:.3f}".format(stat.value[0]))
else:
item.setToolTip(str(stat.exception))
has_missing_scores = True
row.append(item)
model.appendRow(row)
self.error("\n".join(errors), shown=bool(errors))
self.Warning.scores_not_computed(shown=has_missing_scores)
def _update_class_selection(self):
self.class_selection_combo.setCurrentIndex(-1)
self.class_selection_combo.clear()
if not self.data:
return
if self.data.domain.has_discrete_class:
self.cbox.setVisible(True)
class_var = self.data.domain.class_var
items = [self.TARGET_AVERAGE] + class_var.values
self.class_selection_combo.addItems(items)
class_index = 0
if self.class_selection in class_var.values:
class_index = class_var.values.index(self.class_selection) + 1
self.class_selection_combo.setCurrentIndex(class_index)
self.class_selection = items[class_index]
else:
self.cbox.setVisible(False)
def _on_target_class_changed(self):
self._update_stats_model()
def _invalidate(self, which=None):
self.fold_feature_selected = \
self.resampling == OWTestLearners.FeatureFold
# Invalidate learner results for `which` input keys
# (if None then all learner results are invalidated)
if which is None:
which = self.learners.keys()
model = self.view.model()
statmodelkeys = [model.item(row, 0).data(Qt.UserRole)
for row in range(model.rowCount())]
for key in which:
self.learners[key] = \
self.learners[key]._replace(results=None, stats=None)
if key in statmodelkeys:
row = statmodelkeys.index(key)
for c in range(1, model.columnCount()):
item = model.item(row, c)
if item is not None:
item.setData(None, Qt.DisplayRole)
item.setData(None, Qt.ToolTipRole)
self.__needupdate = True
def show_column_chooser(self, pos):
# pylint doesn't know that self.shown_scores is a set, not a Setting
# pylint: disable=unsupported-membership-test
def update(col_name, checked):
if checked:
self.shown_scores.add(col_name)
else:
self.shown_scores.remove(col_name)
self._update_shown_columns()
menu = QMenu()
model = self.result_model
header = self.view.horizontalHeader()
for section in range(1, model.columnCount()):
col_name = model.horizontalHeaderItem(section).data(Qt.DisplayRole)
action = menu.addAction(col_name)
action.setCheckable(True)
action.setChecked(col_name in self.shown_scores)
action.triggered.connect(partial(update, col_name))
menu.exec(header.mapToGlobal(pos))
def commit(self):
"""
Commit the results to output.
"""
self.Error.memory_error.clear()
valid = [slot for slot in self.learners.values()
if slot.results is not None and slot.results.success]
combined = None
predictions = None
if valid:
# Evaluation results
combined = results_merge([slot.results.value for slot in valid])
combined.learner_names = [learner_name(slot.learner)
for slot in valid]
# Predictions & Probabilities
try:
predictions = combined.get_augmented_data(combined.learner_names)
except MemoryError:
self.Error.memory_error()
self.Outputs.evaluations_results.send(combined)
self.Outputs.predictions.send(predictions)
def send_report(self):
"""Report on the testing schema and results"""
if not self.data or not self.learners:
return
if self.resampling == self.KFold:
stratified = 'Stratified ' if self.cv_stratified else ''
items = [("Sampling type", "{}{}-fold Cross validation".
format(stratified, self.NFolds[self.n_folds]))]
elif self.resampling == self.LeaveOneOut:
items = [("Sampling type", "Leave one out")]
elif self.resampling == self.ShuffleSplit:
stratified = 'Stratified ' if self.shuffle_stratified else ''
items = [("Sampling type",
"{}Shuffle split, {} random samples with {}% data "
.format(stratified, self.NRepeats[self.n_repeats],
self.SampleSizes[self.sample_size]))]
elif self.resampling == self.TestOnTrain:
items = [("Sampling type", "No sampling, test on training data")]
elif self.resampling == self.TestOnTest:
items = [("Sampling type", "No sampling, test on testing data")]
else:
items = []
if self.data.domain.has_discrete_class:
items += [("Target class", self.class_selection.strip("()"))]
if items:
self.report_items("Settings", items)
self.report_table("Scores", self.view)
@classmethod
def migrate_settings(cls, settings_, version):
if version < 2:
if settings_["resampling"] > 0:
settings_["resampling"] += 1
if version < 3:
# Older version used an incompatible context handler
settings_["context_settings"] = [
c for c in settings_.get("context_settings", ())
if not hasattr(c, 'classes')]
@Slot(float)
def setProgressValue(self, value):
self.progressBarSet(value, processEvents=False)
def __update(self):
self.__needupdate = False
assert self.__task is None or self.__state == State.Running
if self.__state == State.Running:
self.cancel()
self.Warning.test_data_unused.clear()
self.Warning.test_data_missing.clear()
self.warning()
self.Error.class_inconsistent.clear()
self.Error.too_many_folds.clear()
self.error()
# check preconditions and return early
if self.data is None:
self.__state = State.Waiting
self.commit()
return
if not self.learners:
self.__state = State.Waiting
self.commit()
return
if self.resampling == OWTestLearners.KFold and \
len(self.data) < self.NFolds[self.n_folds]:
self.Error.too_many_folds()
self.__state = State.Waiting
self.commit()
return
elif self.resampling == OWTestLearners.TestOnTest:
if self.test_data is None:
if not self.Error.test_data_empty.is_shown():
self.Warning.test_data_missing()
self.__state = State.Waiting
self.commit()
return
elif self.test_data.domain.class_var != self.data.domain.class_var:
self.Error.class_inconsistent()
self.__state = State.Waiting
self.commit()
return
elif self.test_data is not None:
self.Warning.test_data_unused()
rstate = 42
common_args = dict(
store_data=True,
preprocessor=self.preprocessor,
)
# items in need of an update
items = [(key, slot) for key, slot in self.learners.items()
if slot.results is None]
learners = [slot.learner for _, slot in items]
# deepcopy all learners as they are not thread safe (by virtue of
# the base API). These will be the effective learner objects tested
# but will be replaced with the originals on return (see restore
# learners bellow)
learners_c = [copy.deepcopy(learner) for learner in learners]
if self.resampling == OWTestLearners.KFold:
folds = self.NFolds[self.n_folds]
test_f = partial(
Orange.evaluation.CrossValidation,
self.data, learners_c, k=folds,
random_state=rstate, **common_args)
elif self.resampling == OWTestLearners.FeatureFold:
test_f = partial(
Orange.evaluation.CrossValidationFeature,
self.data, learners_c, self.fold_feature,
**common_args
)
elif self.resampling == OWTestLearners.LeaveOneOut:
test_f = partial(
Orange.evaluation.LeaveOneOut,
self.data, learners_c, **common_args
)
elif self.resampling == OWTestLearners.ShuffleSplit:
train_size = self.SampleSizes[self.sample_size] / 100
test_f = partial(
Orange.evaluation.ShuffleSplit,
self.data, learners_c,
n_resamples=self.NRepeats[self.n_repeats],
train_size=train_size, test_size=None,
stratified=self.shuffle_stratified,
random_state=rstate, **common_args
)
elif self.resampling == OWTestLearners.TestOnTrain:
test_f = partial(
Orange.evaluation.TestOnTrainingData,
self.data, learners_c, **common_args
)
elif self.resampling == OWTestLearners.TestOnTest:
test_f = partial(
Orange.evaluation.TestOnTestData,
self.data, self.test_data, learners_c, **common_args
)
else:
assert False, "self.resampling %s" % self.resampling
def replace_learners(evalfunc, *args, **kwargs):
res = evalfunc(*args, **kwargs)
assert all(lc is lo for lc, lo in zip(learners_c, res.learners))
res.learners[:] = learners
return res
test_f = partial(replace_learners, test_f)
self.__submit(test_f)
def __submit(self, testfunc):
# type: (Callable[[Callable[float]], Results]) -> None
"""
Submit a testing function for evaluation
MUST not be called if an evaluation is already pending/running.
Cancel the existing task first.
Parameters
----------
testfunc : Callable[[Callable[float]], Results])
Must be a callable taking a single `callback` argument and
returning a Results instance
"""
assert self.__state != State.Running
# Setup the task
task = Task()
def progress_callback(finished):
if task.cancelled:
raise UserInterrupt()
QMetaObject.invokeMethod(
self, "setProgressValue", Qt.QueuedConnection,
Q_ARG(float, 100 * finished)
)
def ondone(_):
QMetaObject.invokeMethod(
self, "__task_complete", Qt.QueuedConnection,
Q_ARG(object, task))
testfunc = partial(testfunc, callback=progress_callback)
task.future = self.__executor.submit(testfunc)
task.future.add_done_callback(ondone)
self.progressBarInit(processEvents=None)
self.setBlocking(True)
self.setStatusMessage("Running")
self.__state = State.Running
self.__task = task
@Slot(object)
def __task_complete(self, task):
# handle a completed task
assert self.thread() is QThread.currentThread()
if self.__task is not task:
assert task.cancelled
log.debug("Reaping cancelled task: %r", "<>")
return
self.setBlocking(False)
self.progressBarFinished(processEvents=None)
self.setStatusMessage("")
result = task.future
assert result.done()
self.__task = None
try:
results = result.result() # type: Results
learners = results.learners # type: List[Learner]
except Exception as er:
log.exception("testing error (in __task_complete):",
exc_info=True)
self.error("\n".join(traceback.format_exception_only(type(er), er)))
self.__state = State.Done
return
self.__state = State.Done
learner_key = {slot.learner: key for key, slot in
self.learners.items()}
assert all(learner in learner_key for learner in learners)
# Update the results for individual learners
class_var = results.domain.class_var
for learner, result in zip(learners, results.split_by_model()):
stats = None
if class_var.is_primitive():
ex = result.failed[0]
if ex:
stats = [Try.Fail(ex)] * len(self.scorers)
result = Try.Fail(ex)
else:
stats = [Try(scorer_caller(scorer, result))
for scorer in self.scorers]
result = Try.Success(result)
key = learner_key.get(learner)
self.learners[key] = \
self.learners[key]._replace(results=result, stats=stats)
self._update_header()
self._update_stats_model()
self.commit()
def cancel(self):
"""
Cancel the current/pending evaluation (if any).
"""
if self.__task is not None:
assert self.__state == State.Running
self.__state = State.Cancelled
task, self.__task = self.__task, None
task.cancel()
assert task.future.done()
def onDeleteWidget(self):
self.cancel()
super().onDeleteWidget()
class OWTestLearners(OWWidget):
name = "Test & Score"
description = "Cross-validation accuracy estimation."
icon = "icons/TestLearners1.svg"
priority = 100
class Inputs:
train_data = Input("Data", Table, default=True)
test_data = Input("Test Data", Table)
learner = Input("Learner", Learner, multiple=True)
preprocessor = Input("Preprocessor", Preprocess)
class Outputs:
predictions = Output("Predictions", Table)
evaluations_results = Output("Evaluation Results", Results)
settings_version = 3
UserAdviceMessages = [
widget.Message("Click on the table header to select shown columns",
"click_header")
]
settingsHandler = settings.PerfectDomainContextHandler()
#: Resampling/testing types
KFold, FeatureFold, ShuffleSplit, LeaveOneOut, TestOnTrain, TestOnTest \
= 0, 1, 2, 3, 4, 5
#: Numbers of folds
NFolds = [2, 3, 5, 10, 20]
#: Number of repetitions
NRepeats = [2, 3, 5, 10, 20, 50, 100]
#: Sample sizes
SampleSizes = [5, 10, 20, 25, 30, 33, 40, 50, 60, 66, 70, 75, 80, 90, 95]
#: Selected resampling type
resampling = settings.Setting(0)
#: Number of folds for K-fold cross validation
n_folds = settings.Setting(3)
#: Stratified sampling for K-fold
cv_stratified = settings.Setting(True)
#: Number of repeats for ShuffleSplit sampling
n_repeats = settings.Setting(3)
#: ShuffleSplit sample size
sample_size = settings.Setting(9)
#: Stratified sampling for Random Sampling
shuffle_stratified = settings.Setting(True)
# CV where nr. of feature values determines nr. of folds
fold_feature = settings.ContextSetting(None)
fold_feature_selected = settings.ContextSetting(False)
TARGET_AVERAGE = "(Average over classes)"
class_selection = settings.ContextSetting(TARGET_AVERAGE)
BUILTIN_ORDER = {
DiscreteVariable: ("AUC", "CA", "F1", "Precision", "Recall"),
ContinuousVariable: ("MSE", "RMSE", "MAE", "R2")
}
shown_scores = \
settings.Setting(set(chain(*BUILTIN_ORDER.values())))
class Error(OWWidget.Error):
train_data_empty = Msg("Train data set is empty.")
test_data_empty = Msg("Test data set is empty.")
class_required = Msg("Train data input requires a target variable.")
too_many_classes = Msg("Too many target variables.")
class_required_test = Msg(
"Test data input requires a target variable.")
too_many_folds = Msg("Number of folds exceeds the data size")
class_inconsistent = Msg("Test and train data sets "
"have different target variables.")
memory_error = Msg("Not enough memory.")
only_one_class_var_value = Msg("Target variable has only one value.")
class Warning(OWWidget.Warning):
missing_data = \
Msg("Instances with unknown target values were removed from{}data.")
test_data_missing = Msg("Missing separate test data input.")
scores_not_computed = Msg("Some scores could not be computed.")
test_data_unused = Msg("Test data is present but unused. "
"Select 'Test on test data' to use it.")
class Information(OWWidget.Information):
data_sampled = Msg("Train data has been sampled")
test_data_sampled = Msg("Test data has been sampled")
def __init__(self):
super().__init__()
self.data = None
self.test_data = None
self.preprocessor = None
self.train_data_missing_vals = False
self.test_data_missing_vals = False
self.scorers = []
#: An Ordered dictionary with current inputs and their testing results.
self.learners = OrderedDict() # type: Dict[Any, Input]
self.__state = State.Waiting
# Do we need to [re]test any learners, set by _invalidate and
# cleared by __update
self.__needupdate = False
self.__task = None # type: Optional[Task]
self.__executor = ThreadExecutor()
sbox = gui.vBox(self.controlArea, "Sampling")
rbox = gui.radioButtons(sbox,
self,
"resampling",
callback=self._param_changed)
gui.appendRadioButton(rbox, "Cross validation")
ibox = gui.indentedBox(rbox)
gui.comboBox(ibox,
self,
"n_folds",
label="Number of folds: ",
items=[str(x) for x in self.NFolds],
maximumContentsLength=3,
orientation=Qt.Horizontal,
callback=self.kfold_changed)
gui.checkBox(ibox,
self,
"cv_stratified",
"Stratified",
callback=self.kfold_changed)
gui.appendRadioButton(rbox, "Cross validation by feature")
ibox = gui.indentedBox(rbox)
self.feature_model = DomainModel(order=DomainModel.METAS,
valid_types=DiscreteVariable)
self.features_combo = gui.comboBox(ibox,
self,
"fold_feature",
model=self.feature_model,
orientation=Qt.Horizontal,
callback=self.fold_feature_changed)
gui.appendRadioButton(rbox, "Random sampling")
ibox = gui.indentedBox(rbox)
gui.comboBox(ibox,
self,
"n_repeats",
label="Repeat train/test: ",
items=[str(x) for x in self.NRepeats],
maximumContentsLength=3,
orientation=Qt.Horizontal,
callback=self.shuffle_split_changed)
gui.comboBox(ibox,
self,
"sample_size",
label="Training set size: ",
items=["{} %".format(x) for x in self.SampleSizes],
maximumContentsLength=5,
orientation=Qt.Horizontal,
callback=self.shuffle_split_changed)
gui.checkBox(ibox,
self,
"shuffle_stratified",
"Stratified",
callback=self.shuffle_split_changed)
gui.appendRadioButton(rbox, "Leave one out")
gui.appendRadioButton(rbox, "Test on train data")
gui.appendRadioButton(rbox, "Test on test data")
self.cbox = gui.vBox(self.controlArea, "Target Class")
self.class_selection_combo = gui.comboBox(
self.cbox,
self,
"class_selection",
items=[],
sendSelectedValue=True,
valueType=str,
callback=self._on_target_class_changed,
contentsLength=8)
gui.rubber(self.controlArea)
self.view = gui.TableView(wordWrap=True, )
header = self.view.horizontalHeader()
header.setSectionResizeMode(QHeaderView.ResizeToContents)
header.setDefaultAlignment(Qt.AlignCenter)
header.setStretchLastSection(False)
header.setContextMenuPolicy(Qt.CustomContextMenu)
header.customContextMenuRequested.connect(self.show_column_chooser)
self.result_model = QStandardItemModel(self)
self.result_model.setHorizontalHeaderLabels(["Method"])
self.view.setModel(self.result_model)
self.view.setItemDelegate(ItemDelegate())
box = gui.vBox(self.mainArea, "Evaluation Results")
box.layout().addWidget(self.view)
def sizeHint(self):
return QSize(780, 1)
def _update_controls(self):
self.fold_feature = None
self.feature_model.set_domain(None)
if self.data:
self.feature_model.set_domain(self.data.domain)
if self.fold_feature is None and self.feature_model:
self.fold_feature = self.feature_model[0]
enabled = bool(self.feature_model)
self.controls.resampling.buttons[
OWTestLearners.FeatureFold].setEnabled(enabled)
self.features_combo.setEnabled(enabled)
if self.resampling == OWTestLearners.FeatureFold and not enabled:
self.resampling = OWTestLearners.KFold
@Inputs.learner
def set_learner(self, learner, key):
"""
Set the input `learner` for `key`.
Parameters
----------
learner : Optional[Orange.base.Learner]
key : Any
"""
if key in self.learners and learner is None:
# Removed
self._invalidate([key])
del self.learners[key]
else:
self.learners[key] = InputLearner(learner, None, None)
self._invalidate([key])
@Inputs.train_data
def set_train_data(self, data):
"""
Set the input training dataset.
Parameters
----------
data : Optional[Orange.data.Table]
"""
self.Information.data_sampled.clear()
self.Error.train_data_empty.clear()
self.Error.class_required.clear()
self.Error.too_many_classes.clear()
self.Error.only_one_class_var_value.clear()
if data is not None and not len(data):
self.Error.train_data_empty()
data = None
if data:
conds = [
not data.domain.class_vars,
len(data.domain.class_vars) > 1, data.domain.has_discrete_class
and len(data.domain.class_var.values) == 1
]
errors = [
self.Error.class_required, self.Error.too_many_classes,
self.Error.only_one_class_var_value
]
for cond, error in zip(conds, errors):
if cond:
error()
data = None
break
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.train_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = HasClass()(data)
else:
self.Warning.missing_data.clear()
self.data = data
self.closeContext()
self._update_scorers()
self._update_controls()
if data is not None:
self._update_class_selection()
self.openContext(data.domain)
if self.fold_feature_selected and bool(self.feature_model):
self.resampling = OWTestLearners.FeatureFold
self._invalidate()
@Inputs.test_data
def set_test_data(self, data):
# type: (Orange.data.Table) -> None
"""
Set the input separate testing dataset.
Parameters
----------
data : Optional[Orange.data.Table]
"""
self.Information.test_data_sampled.clear()
self.Error.test_data_empty.clear()
if data is not None and not len(data):
self.Error.test_data_empty()
data = None
if data and not data.domain.class_var:
self.Error.class_required_test()
data = None
else:
self.Error.class_required_test.clear()
if isinstance(data, SqlTable):
if data.approx_len() < AUTO_DL_LIMIT:
data = Table(data)
else:
self.Information.test_data_sampled()
data_sample = data.sample_time(1, no_cache=True)
data_sample.download_data(AUTO_DL_LIMIT, partial=True)
data = Table(data_sample)
self.test_data_missing_vals = \
data is not None and np.isnan(data.Y).any()
if self.train_data_missing_vals or self.test_data_missing_vals:
self.Warning.missing_data(self._which_missing_data())
if data:
data = HasClass()(data)
else:
self.Warning.missing_data.clear()
self.test_data = data
if self.resampling == OWTestLearners.TestOnTest:
self._invalidate()
def _which_missing_data(self):
return {
(True, True): " ", # both, don't specify
(True, False): " train ",
(False, True): " test "
}[(self.train_data_missing_vals, self.test_data_missing_vals)]
# List of scorers shouldn't be retrieved globally, when the module is
# loading since add-ons could have registered additional scorers.
# It could have been cached but
# - we don't gain much with it
# - it complicates the unit tests
def _update_scorers(self):
if self.data is None or self.data.domain.class_var is None:
self.scorers = []
return
class_var = self.data and self.data.domain.class_var
order = {
name: i
for i, name in enumerate(self.BUILTIN_ORDER[type(class_var)])
}
# 'abstract' is retrieved from __dict__ to avoid inheriting
usable = (cls for cls in scoring.Score.registry.values()
if cls.is_scalar and not cls.__dict__.get("abstract")
and isinstance(class_var, cls.class_types))
self.scorers = sorted(usable, key=lambda cls: order.get(cls.name, 99))
@Inputs.preprocessor
def set_preprocessor(self, preproc):
"""
Set the input preprocessor to apply on the training data.
"""
self.preprocessor = preproc
self._invalidate()
def handleNewSignals(self):
"""Reimplemented from OWWidget.handleNewSignals."""
self._update_class_selection()
self._update_header()
self._update_stats_model()
if self.__needupdate:
self.__update()
def kfold_changed(self):
self.resampling = OWTestLearners.KFold
self._param_changed()
def fold_feature_changed(self):
self.resampling = OWTestLearners.FeatureFold
self._param_changed()
def shuffle_split_changed(self):
self.resampling = OWTestLearners.ShuffleSplit
self._param_changed()
def _param_changed(self):
self._invalidate()
self.__update()
def _update_header(self):
# Set the correct horizontal header labels on the results_model.
model = self.result_model
model.setColumnCount(1 + len(self.scorers))
for col, score in enumerate(self.scorers):
item = QStandardItem(score.name)
item.setToolTip(score.long_name)
model.setHorizontalHeaderItem(col + 1, item)
self._update_shown_columns()
def _update_shown_columns(self):
# pylint doesn't know that self.shown_scores is a set, not a Setting
# pylint: disable=unsupported-membership-test
model = self.result_model
header = self.view.horizontalHeader()
for section in range(1, model.columnCount()):
col_name = model.horizontalHeaderItem(section).data(Qt.DisplayRole)
header.setSectionHidden(section, col_name not in self.shown_scores)
def _update_stats_model(self):
# Update the results_model with up to date scores.
# Note: The target class specific scores (if requested) are
# computed as needed in this method.
model = self.view.model()
# clear the table model, but preserving the header labels
for r in reversed(range(model.rowCount())):
model.takeRow(r)
target_index = None
if self.data is not None:
class_var = self.data.domain.class_var
if self.data.domain.has_discrete_class and \
self.class_selection != self.TARGET_AVERAGE:
target_index = class_var.values.index(self.class_selection)
else:
class_var = None
errors = []
has_missing_scores = False
for key, slot in self.learners.items():
name = learner_name(slot.learner)
head = QStandardItem(name)
head.setData(key, Qt.UserRole)
if isinstance(slot.results, Try.Fail):
head.setToolTip(str(slot.results.exception))
head.setText("{} (error)".format(name))
head.setForeground(QtGui.QBrush(Qt.red))
errors.append("{name} failed with error:\n"
"{exc.__class__.__name__}: {exc!s}".format(
name=name, exc=slot.results.exception))
row = [head]
if class_var is not None and class_var.is_discrete and \
target_index is not None:
if slot.results is not None and slot.results.success:
ovr_results = results_one_vs_rest(slot.results.value,
target_index)
# Cell variable is used immediatelly, it's not stored
# pylint: disable=cell-var-from-loop
stats = [
Try(scorer_caller(scorer, ovr_results))
for scorer in self.scorers
]
else:
stats = None
else:
stats = slot.stats
if stats is not None:
for stat in stats:
item = QStandardItem()
if stat.success:
item.setText("{:.3f}".format(stat.value[0]))
else:
item.setToolTip(str(stat.exception))
has_missing_scores = True
row.append(item)
model.appendRow(row)
self.error("\n".join(errors), shown=bool(errors))
self.Warning.scores_not_computed(shown=has_missing_scores)
def _update_class_selection(self):
self.class_selection_combo.setCurrentIndex(-1)
self.class_selection_combo.clear()
if not self.data:
return
if self.data.domain.has_discrete_class:
self.cbox.setVisible(True)
class_var = self.data.domain.class_var
items = [self.TARGET_AVERAGE] + class_var.values
self.class_selection_combo.addItems(items)
class_index = 0
if self.class_selection in class_var.values:
class_index = class_var.values.index(self.class_selection) + 1
self.class_selection_combo.setCurrentIndex(class_index)
self.class_selection = items[class_index]
else:
self.cbox.setVisible(False)
def _on_target_class_changed(self):
self._update_stats_model()
def _invalidate(self, which=None):
self.fold_feature_selected = \
self.resampling == OWTestLearners.FeatureFold
# Invalidate learner results for `which` input keys
# (if None then all learner results are invalidated)
if which is None:
which = self.learners.keys()
model = self.view.model()
statmodelkeys = [
model.item(row, 0).data(Qt.UserRole)
for row in range(model.rowCount())
]
for key in which:
self.learners[key] = \
self.learners[key]._replace(results=None, stats=None)
if key in statmodelkeys:
row = statmodelkeys.index(key)
for c in range(1, model.columnCount()):
item = model.item(row, c)
if item is not None:
item.setData(None, Qt.DisplayRole)
item.setData(None, Qt.ToolTipRole)
self.__needupdate = True
def show_column_chooser(self, pos):
# pylint doesn't know that self.shown_scores is a set, not a Setting
# pylint: disable=unsupported-membership-test
def update(col_name, checked):
if checked:
self.shown_scores.add(col_name)
else:
self.shown_scores.remove(col_name)
self._update_shown_columns()
menu = QMenu()
model = self.result_model
header = self.view.horizontalHeader()
for section in range(1, model.columnCount()):
col_name = model.horizontalHeaderItem(section).data(Qt.DisplayRole)
action = menu.addAction(col_name)
action.setCheckable(True)
action.setChecked(col_name in self.shown_scores)
action.triggered.connect(partial(update, col_name))
menu.exec(header.mapToGlobal(pos))
def commit(self):
"""
Commit the results to output.
"""
self.Error.memory_error.clear()
valid = [
slot for slot in self.learners.values()
if slot.results is not None and slot.results.success
]
combined = None
predictions = None
if valid:
# Evaluation results
combined = results_merge([slot.results.value for slot in valid])
combined.learner_names = [
learner_name(slot.learner) for slot in valid
]
# Predictions & Probabilities
try:
predictions = combined.get_augmented_data(
combined.learner_names)
except MemoryError:
self.Error.memory_error()
self.Outputs.evaluations_results.send(combined)
self.Outputs.predictions.send(predictions)
def send_report(self):
"""Report on the testing schema and results"""
if not self.data or not self.learners:
return
if self.resampling == self.KFold:
stratified = 'Stratified ' if self.cv_stratified else ''
items = [("Sampling type", "{}{}-fold Cross validation".format(
stratified, self.NFolds[self.n_folds]))]
elif self.resampling == self.LeaveOneOut:
items = [("Sampling type", "Leave one out")]
elif self.resampling == self.ShuffleSplit:
stratified = 'Stratified ' if self.shuffle_stratified else ''
items = [
("Sampling type",
"{}Shuffle split, {} random samples with {}% data ".format(
stratified, self.NRepeats[self.n_repeats],
self.SampleSizes[self.sample_size]))
]
elif self.resampling == self.TestOnTrain:
items = [("Sampling type", "No sampling, test on training data")]
elif self.resampling == self.TestOnTest:
items = [("Sampling type", "No sampling, test on testing data")]
else:
items = []
if self.data.domain.has_discrete_class:
items += [("Target class", self.class_selection.strip("()"))]
if items:
self.report_items("Settings", items)
self.report_table("Scores", self.view)
@classmethod
def migrate_settings(cls, settings_, version):
if version < 2:
if settings_["resampling"] > 0:
settings_["resampling"] += 1
if version < 3:
# Older version used an incompatible context handler
settings_["context_settings"] = [
c for c in settings_.get("context_settings", ())
if not hasattr(c, 'classes')
]
@Slot(float)
def setProgressValue(self, value):
self.progressBarSet(value, processEvents=False)
def __update(self):
self.__needupdate = False
assert self.__task is None or self.__state == State.Running
if self.__state == State.Running:
self.cancel()
self.Warning.test_data_unused.clear()
self.Warning.test_data_missing.clear()
self.warning()
self.Error.class_inconsistent.clear()
self.Error.too_many_folds.clear()
self.error()
# check preconditions and return early
if self.data is None:
self.__state = State.Waiting
self.commit()
return
if not self.learners:
self.__state = State.Waiting
self.commit()
return
if self.resampling == OWTestLearners.KFold and \
len(self.data) < self.NFolds[self.n_folds]:
self.Error.too_many_folds()
self.__state = State.Waiting
self.commit()
return
elif self.resampling == OWTestLearners.TestOnTest:
if self.test_data is None:
if not self.Error.test_data_empty.is_shown():
self.Warning.test_data_missing()
self.__state = State.Waiting
self.commit()
return
elif self.test_data.domain.class_var != self.data.domain.class_var:
self.Error.class_inconsistent()
self.__state = State.Waiting
self.commit()
return
elif self.test_data is not None:
self.Warning.test_data_unused()
rstate = 42
common_args = dict(
store_data=True,
preprocessor=self.preprocessor,
)
# items in need of an update
items = [(key, slot) for key, slot in self.learners.items()
if slot.results is None]
learners = [slot.learner for _, slot in items]
# deepcopy all learners as they are not thread safe (by virtue of
# the base API). These will be the effective learner objects tested
# but will be replaced with the originals on return (see restore
# learners bellow)
learners_c = [copy.deepcopy(learner) for learner in learners]
if self.resampling == OWTestLearners.KFold:
folds = self.NFolds[self.n_folds]
test_f = partial(Orange.evaluation.CrossValidation,
self.data,
learners_c,
k=folds,
random_state=rstate,
**common_args)
elif self.resampling == OWTestLearners.FeatureFold:
test_f = partial(Orange.evaluation.CrossValidationFeature,
self.data, learners_c, self.fold_feature,
**common_args)
elif self.resampling == OWTestLearners.LeaveOneOut:
test_f = partial(Orange.evaluation.LeaveOneOut, self.data,
learners_c, **common_args)
elif self.resampling == OWTestLearners.ShuffleSplit:
train_size = self.SampleSizes[self.sample_size] / 100
test_f = partial(Orange.evaluation.ShuffleSplit,
self.data,
learners_c,
n_resamples=self.NRepeats[self.n_repeats],
train_size=train_size,
test_size=None,
stratified=self.shuffle_stratified,
random_state=rstate,
**common_args)
elif self.resampling == OWTestLearners.TestOnTrain:
test_f = partial(Orange.evaluation.TestOnTrainingData, self.data,
learners_c, **common_args)
elif self.resampling == OWTestLearners.TestOnTest:
test_f = partial(Orange.evaluation.TestOnTestData, self.data,
self.test_data, learners_c, **common_args)
else:
assert False, "self.resampling %s" % self.resampling
def replace_learners(evalfunc, *args, **kwargs):
res = evalfunc(*args, **kwargs)
assert all(lc is lo for lc, lo in zip(learners_c, res.learners))
res.learners[:] = learners
return res
test_f = partial(replace_learners, test_f)
self.__submit(test_f)
def __submit(self, testfunc):
# type: (Callable[[Callable[float]], Results]) -> None
"""
Submit a testing function for evaluation
MUST not be called if an evaluation is already pending/running.
Cancel the existing task first.
Parameters
----------
testfunc : Callable[[Callable[float]], Results])
Must be a callable taking a single `callback` argument and
returning a Results instance
"""
assert self.__state != State.Running
# Setup the task
task = Task()
def progress_callback(finished):
if task.cancelled:
raise UserInterrupt()
QMetaObject.invokeMethod(self, "setProgressValue",
Qt.QueuedConnection,
Q_ARG(float, 100 * finished))
def ondone(_):
QMetaObject.invokeMethod(self, "__task_complete",
Qt.QueuedConnection, Q_ARG(object, task))
testfunc = partial(testfunc, callback=progress_callback)
task.future = self.__executor.submit(testfunc)
task.future.add_done_callback(ondone)
self.progressBarInit(processEvents=None)
self.setBlocking(True)
self.setStatusMessage("Running")
self.__state = State.Running
self.__task = task
@Slot(object)
def __task_complete(self, task):
# handle a completed task
assert self.thread() is QThread.currentThread()
if self.__task is not task:
assert task.cancelled
log.debug("Reaping cancelled task: %r", "<>")
return
self.setBlocking(False)
self.progressBarFinished(processEvents=None)
self.setStatusMessage("")
result = task.future
assert result.done()
self.__task = None
try:
results = result.result() # type: Results
learners = results.learners # type: List[Learner]
except Exception as er:
log.exception("testing error (in __task_complete):", exc_info=True)
self.error("\n".join(traceback.format_exception_only(type(er),
er)))
self.__state = State.Done
return
self.__state = State.Done
learner_key = {
slot.learner: key
for key, slot in self.learners.items()
}
assert all(learner in learner_key for learner in learners)
# Update the results for individual learners
class_var = results.domain.class_var
for learner, result in zip(learners, results.split_by_model()):
stats = None
if class_var.is_primitive():
ex = result.failed[0]
if ex:
stats = [Try.Fail(ex)] * len(self.scorers)
result = Try.Fail(ex)
else:
stats = [
Try(scorer_caller(scorer, result))
for scorer in self.scorers
]
result = Try.Success(result)
key = learner_key.get(learner)
self.learners[key] = \
self.learners[key]._replace(results=result, stats=stats)
self._update_header()
self._update_stats_model()
self.commit()
def cancel(self):
"""
Cancel the current/pending evaluation (if any).
"""
if self.__task is not None:
assert self.__state == State.Running
self.__state = State.Cancelled
task, self.__task = self.__task, None
task.cancel()
assert task.future.done()
def onDeleteWidget(self):
self.cancel()
super().onDeleteWidget()
class OWBatchNorm(OWWidget):
name = "Batch Effect Removal"
description = "Batch effect normalization on Single Cell data set."
icon = "icons/BatchEffectRemoval.svg"
priority = 230
class Inputs:
data = Input("Data", Table)
class Outputs:
data = Output("Data", Table)
class Error(OWWidget.Error):
general_error = Msg({})
discrete_attributes = Msg("Data with discrete attributes "
"can not be processed.")
class Warning(OWWidget.Warning):
missing_values = Msg("Missing values have been replaced with 0.")
negative_values = Msg("Unable to use current settings due "
"to negative values in data.")
resizing_enabled = False
want_main_area = False
settingsHandler = PerfectDomainContextHandler()
batch_vars = ContextSetting([])
link_method = Setting(LinkMethod.IDENTITY_LINK)
skip_zeros = Setting(False)
auto_commit = Setting(True)
def __init__(self, parent=None):
super().__init__(parent)
self.data = None
# Info
infobox = gui.widgetBox(self.controlArea, "Info")
self.info_label = gui.widgetLabel(infobox, "No data on input.")
# Link method
method_box = gui.widgetBox(self.controlArea, "Method")
gui.comboBox(method_box,
self,
"link_method",
items=LinkMethod.items(),
callback=self.__link_method_changed)
gui.separator(method_box)
self.skip_zeros_check = gui.checkBox(
method_box,
self,
"skip_zeros",
"Skip zero expressions",
enabled=self.link_method != LinkMethod.LOG_LINK,
callback=lambda: self.commit())
# Batch Variable Selection
header_shema = (("selected", ""), ("variable", "Variable"),
("count", "#"), ("score", "Score"))
header_labels = labels = [label for _, label in header_shema]
header = namedtuple("header", [tag for tag, _ in header_shema])
self.Header = header(*[index for index, _ in enumerate(labels)])
batch_box = gui.widgetBox(self.controlArea, "Batch Variable Selection")
self.view = QTreeView()
self.model = QStandardItemModel()
self.model.itemChanged.connect(self.__selected_batch_vars_changed)
self.model.setHorizontalHeaderLabels(header_labels)
batch_box.layout().addWidget(self.view)
self._setup_view()
gui.auto_commit(self.controlArea, self, "auto_commit", "Apply",
"Apply Automatically")
def __link_method_changed(self):
enable = self.link_method != LinkMethod.LOG_LINK
self.skip_zeros_check.setEnabled(enable)
if not enable:
self.skip_zeros_check.setChecked(True)
self.commit()
def __selected_batch_vars_changed(self, item):
if item.checkState():
self.batch_vars.append(item.data(VariableRole))
else:
self.batch_vars.remove(item.data(VariableRole))
self.commit()
def _setup_view(self):
self.view.setModel(self.model)
self.view.setSelectionMode(QTreeView.NoSelection)
self.view.setSortingEnabled(True)
self.view.setRootIsDecorated(False)
self.view.setItemDelegateForColumn(self.Header.count,
IntegralDelegate(self))
self.view.setItemDelegateForColumn(self.Header.score,
RealDelegate(self))
self.view.header().setSectionResizeMode(QHeaderView.ResizeToContents)
self.view.header().setStretchLastSection(False)
self.view.header().setSectionResizeMode(self.Header.variable,
QHeaderView.Stretch)
self.view.setFocus()
@Inputs.data
def set_data(self, data):
self.closeContext()
self.clear()
self.data = data
self._setup_info_label()
self._check_data()
self.openContext(data)
if self.data is not None:
self.batch_vars = [data.domain[v.name] for v in self.batch_vars]
self._setup_model()
self.commit()
def clear(self):
self.batch_vars = []
if self.model:
n_rows = self.model.rowCount()
self.model.removeRows(0, n_rows)
def _setup_info_label(self):
text = "No data on input."
if self.data is not None:
domain, attrs = self.data.domain, self.data.domain.attributes
text = "{} cells, {} genes\n".format(len(self.data), len(attrs))
text += "{} meta features".format(len(domain.metas)) \
if len(domain.metas) else "(no meta features)"
self.info_label.setText(text)
def _check_data(self):
self.clear_messages()
if self.data and self.data.domain.has_discrete_attributes():
self.data = None
self.Error.discrete_attributes()
if self.data and np.isnan(self.data.X).any():
self.data.X = np.nan_to_num(self.data.X)
self.Warning.missing_values()
def _setup_model(self):
estimator = ScBatchScorer()
for var in self.data.domain.class_vars + self.data.domain.metas:
if not var.is_primitive():
continue
try:
score = float(estimator.score_data(self.data, var))
except Exception:
score = np.nan
self.model.appendRow([
self.__selected_item(var),
self.__variable_item(var),
self.__count_item(var),
self.__score_item(score)
])
def __selected_item(self, var):
item = QStandardItem()
item.setData(var, VariableRole)
item.setCheckable(True)
select = var in self.batch_vars
item.setCheckState(Qt.Checked if select else Qt.Unchecked)
item.setEditable(False)
return item
def __variable_item(self, var):
item = QStandardItem()
item.setData(var.name, Qt.DisplayRole)
item.setData(gui.attributeIconDict[var], Qt.DecorationRole)
item.setEditable(False)
return item
def __count_item(self, var):
item = QStandardItem()
if var.is_discrete:
item.setData(len(var.values), Qt.DisplayRole)
item.setEditable(False)
return item
def __score_item(self, score):
item = QStandardItem()
item.setData(score, Qt.DisplayRole)
item.setEditable(False)
return item
def commit(self):
data = None
self.Error.general_error.clear()
self.Warning.negative_values.clear()
if self.data is not None:
if (self.data.X < 0).any() and self.skip_zeros:
self.Warning.negative_values()
data = self.data
else:
try:
data = SCBatchNormalizer(
LinkMethod.items()[self.link_method], self.skip_zeros,
self.batch_vars)(self.data)
except Exception as e:
self.Error.general_error(str(e))
data = None
self.Outputs.data.send(data)
def send_report(self):
method = LinkMethod.items()[self.link_method]
if self.skip_zeros:
method += " (Skip zero expressions)"
variables = ", ".join([v.name for v in self.batch_vars]) \
if self.batch_vars else "None"
self.report_items("", [("Method", method),
("Batch variable selection", variables)])
