class OWDataSets(widget.OWWidget):
name = "Data Sets"
description = "Load a data set from an online repository"
icon = "icons/DataSets.svg"
priority = 20
replaces = ["orangecontrib.prototypes.widgets.owdatasets.OWDataSets"]
# The following constants can be overridden in a subclass
# to reuse this widget for a different repository
# Take care when refactoring! (used in e.g. single-cell)
INDEX_URL = "http://datasets.orange.biolab.si/"
DATASET_DIR = "datasets"
class Error(widget.OWWidget.Error):
no_remote_datasets = Msg("Could not fetch data set list")
class Warning(widget.OWWidget.Warning):
only_local_datasets = Msg("Could not fetch data sets list, only local "
"cached data sets are shown")
class Outputs:
data = Output("Data", Orange.data.Table)
#: Selected data set id
selected_id = settings.Setting(None) # type: Optional[str]
auto_commit = settings.Setting(False) # type: bool
#: main area splitter state
splitter_state = settings.Setting(b'') # type: bytes
header_state = settings.Setting(b'') # type: bytes
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)
@Slot(object)
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)
def __update_cached_state(self):
model = self.view.model().sourceModel()
localinfo = self.list_local()
assert isinstance(model, QStandardItemModel)
allinfo = []
for i in range(model.rowCount()):
item = model.item(i, 0)
info = item.data(Qt.UserRole)
info.islocal = (info.prefix, info.filename) in localinfo
item.setData(" " if info.islocal else "", Qt.DisplayRole)
allinfo.append(info)
self.infolabel.setText(format_info(
model.rowCount(), sum(info.islocal for info in allinfo)))
def selected_dataset(self):
"""
Return the current selected data set info or None if not selected
Returns
-------
info : Optional[namespace]
"""
rows = self.view.selectionModel().selectedRows(0)
assert 0 <= len(rows) <= 1
current = rows[0] if rows else None # type: Optional[QModelIndex]
if current is not None:
info = current.data(Qt.UserRole)
assert isinstance(info, namespace)
else:
info = None
return info
def filter(self):
filter_string = self.filterLineEdit.text().strip()
proxyModel = self.view.model()
if proxyModel:
proxyModel.setFilterFixedString(filter_string)
def __on_selection(self):
# Main data sets view selection has changed
rows = self.view.selectionModel().selectedRows(0)
assert 0 <= len(rows) <= 1
current = rows[0] if rows else None # type: Optional[QModelIndex]
if current is not None:
current = self.view.model().mapToSource(current)
di = current.data(Qt.UserRole)
text = description_html(di)
self.descriptionlabel.setText(text)
self.selected_id = os.path.join(di.prefix, di.filename)
else:
self.descriptionlabel.setText("")
self.selected_id = None
self.commit()
def commit(self):
"""
Commit a dataset to the output immediately (if available locally) or
schedule download background and an eventual send.
During the download the widget is in blocking state
(OWWidget.isBlocking)
"""
di = self.selected_dataset()
if di is not None:
self.Error.clear()
if self.__awaiting_state is not None:
# disconnect from the __commit_complete
self.__awaiting_state.watcher.done.disconnect(
self.__commit_complete)
# .. and connect to update_cached_state
# self.__awaiting_state.watcher.done.connect(
# self.__update_cached_state)
# TODO: There are possible pending __progress_advance queued
self.__awaiting_state.pb.advance.disconnect(
self.__progress_advance)
self.progressBarFinished(processEvents=None)
self.__awaiting_state = None
if not di.islocal:
pr = progress()
callback = lambda pr=pr: pr.advance.emit()
pr.advance.connect(self.__progress_advance, Qt.QueuedConnection)
self.progressBarInit(processEvents=None)
self.setStatusMessage("Fetching...")
self.setBlocking(True)
f = self._executor.submit(
ensure_local, self.INDEX_URL, di.prefix, di.filename,
self.local_cache_path, force=di.outdated,
progress_advance=callback)
w = FutureWatcher(f, parent=self)
w.done.connect(self.__commit_complete)
self.__awaiting_state = _FetchState(f, w, pr)
else:
self.setStatusMessage("")
self.setBlocking(False)
self.commit_cached(di.prefix, di.filename)
else:
self.Outputs.data.send(None)
@Slot(object)
def __commit_complete(self, f):
# complete the commit operation after the required file has been
# downloaded
assert QThread.currentThread() is self.thread()
assert self.__awaiting_state is not None
assert self.__awaiting_state.future is f
if self.isBlocking():
self.progressBarFinished(processEvents=None)
self.setBlocking(False)
self.setStatusMessage("")
self.__awaiting_state = None
try:
path = f.result()
except Exception as ex:
log.exception("Error:")
self.error(format_exception(ex))
path = None
self.__update_cached_state()
if path is not None:
data = Orange.data.Table(path)
else:
data = None
self.Outputs.data.send(data)
def commit_cached(self, prefix, filename):
path = LocalFiles(self.local_cache_path).localpath(prefix, filename)
self.Outputs.data.send(Orange.data.Table(path))
@Slot()
def __progress_advance(self):
assert QThread.currentThread() is self.thread()
self.progressBarAdvance(1, processEvents=None)
def onDeleteWidget(self):
super().onDeleteWidget()
if self.__awaiting_state is not None:
self.__awaiting_state.watcher.done.disconnect(self.__commit_complete)
self.__awaiting_state.pb.advance.disconnect(self.__progress_advance)
self.__awaiting_state = None
def sizeHint(self):
return QSize(900, 600)
def closeEvent(self, event):
self.splitter_state = bytes(self.splitter.saveState())
self.header_state = bytes(self.view.header().saveState())
super().closeEvent(event)
def list_remote(self):
# type: () -> Dict[Tuple[str, str], dict]
client = ServerFiles(server=self.INDEX_URL)
return client.allinfo()
def list_local(self):
# type: () -> Dict[Tuple[str, str], dict]
return LocalFiles(self.local_cache_path).allinfo()
class OWGeneInfo(widget.OWWidget):
name = "Gene Info"
description = "Displays gene information from NCBI and other sources."
icon = "../widgets/icons/GeneInfo.svg"
priority = 2010
inputs = [("Data", Orange.data.Table, "setData")]
outputs = [("Data Subset", Orange.data.Table)]
settingsHandler = settings.DomainContextHandler()
organism_index = settings.ContextSetting(0)
taxid = settings.ContextSetting("9606")
gene_attr = settings.ContextSetting(0)
auto_commit = settings.Setting(False)
search_string = settings.Setting("")
useAttr = settings.ContextSetting(False)
useAltSource = settings.ContextSetting(False)
def __init__(
self,
parent=None,
):
super().__init__(self, parent)
self.selectionChangedFlag = False
self.__initialized = False
self.initfuture = None
self.itemsfuture = None
self.infoLabel = gui.widgetLabel(
gui.widgetBox(self.controlArea, "Info", addSpace=True),
"Initializing\n")
self.organisms = None
self.organismBox = gui.widgetBox(self.controlArea,
"Organism",
addSpace=True)
self.organismComboBox = gui.comboBox(
self.organismBox,
self,
"organism_index",
callback=self._onSelectedOrganismChanged)
# For now only support one alt source, with a checkbox
# In the future this can be extended to multiple selections
self.altSourceCheck = gui.checkBox(self.organismBox,
self,
"useAltSource",
"Show information from dictyBase",
callback=self.onAltSourceChange)
self.altSourceCheck.hide()
box = gui.widgetBox(self.controlArea, "Gene names", addSpace=True)
self.geneAttrComboBox = gui.comboBox(box,
self,
"gene_attr",
"Gene attribute",
callback=self.updateInfoItems)
self.geneAttrComboBox.setEnabled(not self.useAttr)
cb = gui.checkBox(box,
self,
"useAttr",
"Use attribute names",
callback=self.updateInfoItems)
cb.toggled[bool].connect(self.geneAttrComboBox.setDisabled)
gui.auto_commit(self.controlArea, self, "auto_commit", "Commit")
# A label for dictyExpress link (Why oh god why???)
self.dictyExpressBox = gui.widgetBox(self.controlArea, "Dicty Express")
self.linkLabel = gui.widgetLabel(self.dictyExpressBox, "")
self.linkLabel.setOpenExternalLinks(False)
self.linkLabel.linkActivated.connect(self.onDictyExpressLink)
self.dictyExpressBox.hide()
gui.rubber(self.controlArea)
gui.lineEdit(self.mainArea,
self,
"search_string",
"Filter",
callbackOnType=True,
callback=self.searchUpdate)
self.treeWidget = QTreeView(self.mainArea,
selectionMode=QTreeView.ExtendedSelection,
rootIsDecorated=False,
uniformRowHeights=True,
sortingEnabled=True)
self.treeWidget.setItemDelegate(
gui.LinkStyledItemDelegate(self.treeWidget))
self.treeWidget.viewport().setMouseTracking(True)
self.mainArea.layout().addWidget(self.treeWidget)
box = gui.widgetBox(self.mainArea, "", orientation="horizontal")
gui.button(box, self, "Select Filtered", callback=self.selectFiltered)
gui.button(box,
self,
"Clear Selection",
callback=self.treeWidget.clearSelection)
self.geneinfo = []
self.cells = []
self.row2geneinfo = {}
self.data = None
# : (# input genes, # matches genes)
self.matchedInfo = 0, 0
self.setBlocking(True)
self.executor = ThreadExecutor(self)
self.progressBarInit()
task = Task(
function=partial(taxonomy.ensure_downloaded,
callback=methodinvoke(self, "advance", ())))
task.resultReady.connect(self.initialize)
task.exceptionReady.connect(self._onInitializeError)
self.initfuture = self.executor.submit(task)
def sizeHint(self):
return QSize(1024, 720)
@Slot()
def advance(self):
assert self.thread() is QThread.currentThread()
self.progressBarSet(self.progressBarValue + 1, processEvents=None)
def initialize(self):
if self.__initialized:
# Already initialized
return
self.__initialized = True
self.organisms = sorted(
set([
name.split(".")[-2]
for name in serverfiles.listfiles("NCBI_geneinfo")
] + gene.NCBIGeneInfo.common_taxids()))
self.organismComboBox.addItems(
[taxonomy.name(tax_id) for tax_id in self.organisms])
if self.taxid in self.organisms:
self.organism_index = self.organisms.index(self.taxid)
else:
self.organism_index = 0
self.taxid = self.organisms[self.organism_index]
self.altSourceCheck.setVisible(self.taxid == DICTY_TAXID)
self.dictyExpressBox.setVisible(self.taxid == DICTY_TAXID)
self.infoLabel.setText("No data on input\n")
self.initfuture = None
self.setBlocking(False)
self.progressBarFinished(processEvents=None)
def _onInitializeError(self, exc):
sys.excepthook(type(exc), exc, None)
self.error(0, "Could not download the necessary files.")
def _onSelectedOrganismChanged(self):
assert 0 <= self.organism_index <= len(self.organisms)
self.taxid = self.organisms[self.organism_index]
self.altSourceCheck.setVisible(self.taxid == DICTY_TAXID)
self.dictyExpressBox.setVisible(self.taxid == DICTY_TAXID)
if self.data is not None:
self.updateInfoItems()
def setData(self, data=None):
if not self.__initialized:
self.initfuture.result()
self.initialize()
if self.itemsfuture is not None:
raise Exception("Already processing")
self.closeContext()
self.data = data
if data is not None:
self.geneAttrComboBox.clear()
self.attributes = \
[attr for attr in data.domain.variables + data.domain.metas
if isinstance(attr, (Orange.data.StringVariable,
Orange.data.DiscreteVariable))]
for var in self.attributes:
self.geneAttrComboBox.addItem(*gui.attributeItem(var))
self.taxid = data_hints.get_hint(self.data, "taxid", self.taxid)
self.useAttr = data_hints.get_hint(self.data, "genesinrows",
self.useAttr)
self.openContext(data)
self.gene_attr = min(self.gene_attr, len(self.attributes) - 1)
if self.taxid in self.organisms:
self.organism_index = self.organisms.index(self.taxid)
else:
self.organism_index = 0
self.taxid = self.organisms[self.organism_index]
self.updateInfoItems()
else:
self.clear()
def infoSource(self):
""" Return the current selected info source getter function from
INFO_SOURCES
"""
org = self.organisms[min(self.organism_index, len(self.organisms) - 1)]
if org not in INFO_SOURCES:
org = "default"
sources = INFO_SOURCES[org]
name, func = sources[min(self.useAltSource, len(sources) - 1)]
return name, func
def inputGenes(self):
if self.useAttr:
genes = [attr.name for attr in self.data.domain.attributes]
elif self.attributes:
attr = self.attributes[self.gene_attr]
genes = [
str(ex[attr]) for ex in self.data if not math.isnan(ex[attr])
]
else:
genes = []
return genes
def updateInfoItems(self):
self.warning(0)
if self.data is None:
return
genes = self.inputGenes()
if self.useAttr:
genes = [attr.name for attr in self.data.domain.attributes]
elif self.attributes:
attr = self.attributes[self.gene_attr]
genes = [
str(ex[attr]) for ex in self.data if not math.isnan(ex[attr])
]
else:
genes = []
if not genes:
self.warning(0, "Could not extract genes from input dataset.")
self.warning(1)
org = self.organisms[min(self.organism_index, len(self.organisms) - 1)]
source_name, info_getter = self.infoSource()
self.error(0)
self.updateDictyExpressLink(genes, show=org == DICTY_TAXID)
self.altSourceCheck.setVisible(org == DICTY_TAXID)
self.progressBarInit()
self.setBlocking(True)
self.setEnabled(False)
self.infoLabel.setText("Retrieving info records.\n")
self.genes = genes
task = Task(function=partial(
info_getter, org, genes, advance=methodinvoke(self, "advance", (
))))
self.itemsfuture = self.executor.submit(task)
task.finished.connect(self._onItemsCompleted)
def _onItemsCompleted(self):
self.setBlocking(False)
self.progressBarFinished()
self.setEnabled(True)
try:
schema, geneinfo = self.itemsfuture.result()
finally:
self.itemsfuture = None
self.geneinfo = geneinfo = list(zip(self.genes, geneinfo))
self.cells = cells = []
self.row2geneinfo = {}
links = []
for i, (_, gi) in enumerate(geneinfo):
if gi:
row = []
for _, item in zip(schema, gi):
if isinstance(item, Link):
# TODO: This should be handled by delegates
row.append(item.text)
links.append(item.link)
else:
row.append(item)
cells.append(row)
self.row2geneinfo[len(cells) - 1] = i
model = TreeModel(cells, [str(col) for col in schema], None)
model.setColumnLinks(0, links)
proxyModel = QSortFilterProxyModel(self)
proxyModel.setSourceModel(model)
self.treeWidget.setModel(proxyModel)
self.treeWidget.selectionModel().selectionChanged.connect(self.commit)
for i in range(7):
self.treeWidget.resizeColumnToContents(i)
self.treeWidget.setColumnWidth(
i, min(self.treeWidget.columnWidth(i), 200))
self.infoLabel.setText("%i genes\n%i matched NCBI's IDs" %
(len(self.genes), len(cells)))
self.matchedInfo = len(self.genes), len(cells)
def clear(self):
self.infoLabel.setText("No data on input\n")
self.treeWidget.setModel(
TreeModel([], [
"NCBI ID", "Symbol", "Locus Tag", "Chromosome", "Description",
"Synonyms", "Nomenclature"
], self.treeWidget))
self.geneAttrComboBox.clear()
self.send("Data Subset", None)
def commit(self):
if self.data is None:
self.send("Data Subset", None)
return
model = self.treeWidget.model()
selection = self.treeWidget.selectionModel().selection()
selection = model.mapSelectionToSource(selection)
selectedRows = list(
chain(*(range(r.top(),
r.bottom() + 1) for r in selection)))
model = model.sourceModel()
selectedGeneids = [self.row2geneinfo[row] for row in selectedRows]
selectedIds = [self.geneinfo[i][0] for i in selectedGeneids]
selectedIds = set(selectedIds)
gene2row = dict((self.geneinfo[self.row2geneinfo[row]][0], row)
for row in selectedRows)
isselected = selectedIds.__contains__
if self.useAttr:
def is_selected(attr):
return attr.name in selectedIds
attrs = [
attr for attr in self.data.domain.attributes
if isselected(attr.name)
]
domain = Orange.data.Domain(attrs, self.data.domain.class_vars,
self.data.domain.metas)
newdata = self.data.from_table(domain, self.data)
self.send("Data Subset", newdata)
elif self.attributes:
attr = self.attributes[self.gene_attr]
gene_col = [
attr.str_val(v) for v in self.data.get_column_view(attr)[0]
]
gene_col = [(i, name) for i, name in enumerate(gene_col)
if isselected(name)]
indices = [i for i, _ in gene_col]
# Add a gene info columns to the output
headers = [
str(model.headerData(i, Qt.Horizontal, Qt.DisplayRole))
for i in range(model.columnCount())
]
metas = [Orange.data.StringVariable(name) for name in headers]
domain = Orange.data.Domain(self.data.domain.attributes,
self.data.domain.class_vars,
self.data.domain.metas + tuple(metas))
newdata = self.data.from_table(domain, self.data)[indices]
model_rows = [gene2row[gene] for _, gene in gene_col]
for col, meta in zip(range(model.columnCount()), metas):
col_data = [
str(model.index(row, col).data(Qt.DisplayRole))
for row in model_rows
]
col_data = np.array(col_data, dtype=object, ndmin=2).T
newdata[:, meta] = col_data
if not len(newdata):
newdata = None
self.send("Data Subset", newdata)
else:
self.send("Data Subset", None)
def rowFiltered(self, row):
searchStrings = self.search_string.lower().split()
row = " ".join(self.cells[row]).lower()
return not all([s in row for s in searchStrings])
def searchUpdate(self):
if not self.data:
return
searchStrings = self.search_string.lower().split()
index = self.treeWidget.model().sourceModel().index
mapFromSource = self.treeWidget.model().mapFromSource
for i, row in enumerate(self.cells):
row = " ".join(row).lower()
self.treeWidget.setRowHidden(
mapFromSource(index(i, 0)).row(), QModelIndex(),
not all([s in row for s in searchStrings]))
def selectFiltered(self):
if not self.data:
return
itemSelection = QItemSelection()
index = self.treeWidget.model().sourceModel().index
mapFromSource = self.treeWidget.model().mapFromSource
for i, row in enumerate(self.cells):
if not self.rowFiltered(i):
itemSelection.select(mapFromSource(index(i, 0)),
mapFromSource(index(i, 0)))
self.treeWidget.selectionModel().select(
itemSelection,
QItemSelectionModel.Select | QItemSelectionModel.Rows)
def updateDictyExpressLink(self, genes, show=False):
def fix(ddb):
if ddb.startswith("DDB"):
if not ddb.startswith("DDB_G"):
ddb = ddb.replace("DDB", "DDB_G")
return ddb
return None
if show:
genes = [fix(gene) for gene in genes if fix(gene)]
link1 = '<a href="http://dictyexpress.biolab.si/run/index.php?gene=%s">Microarray profile</a>'
link2 = '<a href="http://dictyexpress.biolab.si/run/index.php?gene=%s&db=rnaseq">RNA-Seq profile</a>'
self.linkLabel.setText(link1 + "<br/>" + link2)
show = any(genes)
if show:
self.dictyExpressBox.show()
else:
self.dictyExpressBox.hide()
def onDictyExpressLink(self, link):
if not self.data:
return
selectedIndexes = self.treeWidget.selectedIndexes()
if not len(selectedIndexes):
QMessageBox.information(self, "No gene ids selected",
"Please select some genes and try again.")
return
model = self.treeWidget.model()
mapToSource = model.mapToSource
selectedRows = self.treeWidget.selectedIndexes()
selectedRows = [mapToSource(index).row() for index in selectedRows]
model = model.sourceModel()
selectedGeneids = [self.row2geneinfo[row] for row in selectedRows]
selectedIds = [self.geneinfo[i][0] for i in selectedGeneids]
selectedIds = set(selectedIds)
def fix(ddb):
if ddb.startswith("DDB"):
if not ddb.startswith("DDB_G"):
ddb = ddb.replace("DDB", "DDB_G")
return ddb
return None
genes = [fix(gene) for gene in selectedIds if fix(gene)]
url = str(link) % " ".join(genes)
QDesktopServices.openUrl(QUrl(url))
def onAltSourceChange(self):
self.updateInfoItems()
def onDeleteWidget(self):
# try to cancel pending tasks
if self.initfuture:
self.initfuture.cancel()
if self.itemsfuture:
self.itemsfuture.cancel()
self.executor.shutdown(wait=False)
super().onDeleteWidget()
class OWGEODatasets(OWWidget):
name = "GEO Data Sets"
description = DESCRIPTION
icon = "../widgets/icons/GEODataSets.svg"
priority = PRIORITY
inputs = []
outputs = [("Expression Data", Orange.data.Table)]
settingsList = [
"outputRows", "mergeSpots", "gdsSelectionStates", "splitterSettings",
"currentGds", "autoCommit", "datasetNames"
]
outputRows = Setting(True)
mergeSpots = Setting(True)
gdsSelectionStates = Setting({})
currentGds = Setting(None)
datasetNames = Setting({})
splitterSettings = Setting((
b'\x00\x00\x00\xff\x00\x00\x00\x00\x00\x00\x00\x02\x00\x00\x01\xea\x00\x00\x00\xd7\x01\x00\x00\x00\x07\x01\x00\x00\x00\x02',
b'\x00\x00\x00\xff\x00\x00\x00\x00\x00\x00\x00\x02\x00\x00\x01\xb5\x00\x00\x02\x10\x01\x00\x00\x00\x07\x01\x00\x00\x00\x01'
))
autoCommit = Setting(False)
def __init__(self, parent=None, signalManager=None, name=" GEO Data Sets"):
OWWidget.__init__(self, parent, signalManager, name)
self.selectionChanged = False
self.filterString = ""
self.datasetName = ""
## GUI
box = gui.widgetBox(self.controlArea, "Info", addSpace=True)
self.infoBox = gui.widgetLabel(box, "Initializing\n\n")
box = gui.widgetBox(self.controlArea, "Output", addSpace=True)
gui.radioButtonsInBox(box,
self,
"outputRows",
["Genes in rows", "Samples in rows"],
"Rows",
callback=self.commitIf)
gui.checkBox(box,
self,
"mergeSpots",
"Merge spots of same gene",
callback=self.commitIf)
gui.separator(box)
self.nameEdit = gui.lineEdit(
box,
self,
"datasetName",
"Data set name",
tooltip="Override the default output data set name",
callback=self.onNameEdited)
self.nameEdit.setPlaceholderText("")
if sys.version_info < (3, ):
box = gui.widgetBox(self.controlArea, "Commit", addSpace=True)
self.commitButton = gui.button(box,
self,
"Commit",
callback=self.commit)
cb = gui.checkBox(box, self, "autoCommit", "Commit on any change")
gui.setStopper(self, self.commitButton, cb, "selectionChanged",
self.commit)
else:
gui.auto_commit(self.controlArea,
self,
"autoCommit",
"Commit",
box="Commit")
self.commitIf = self.commit
gui.rubber(self.controlArea)
gui.widgetLabel(self.mainArea, "Filter")
self.filterLineEdit = QLineEdit(textChanged=self.filter)
self.completer = TokenListCompleter(self,
caseSensitivity=Qt.CaseInsensitive)
self.filterLineEdit.setCompleter(self.completer)
self.mainArea.layout().addWidget(self.filterLineEdit)
splitter = QSplitter(Qt.Vertical, self.mainArea)
self.mainArea.layout().addWidget(splitter)
self.treeWidget = QTreeView(splitter)
self.treeWidget.setSelectionMode(QTreeView.SingleSelection)
self.treeWidget.setRootIsDecorated(False)
self.treeWidget.setSortingEnabled(True)
self.treeWidget.setAlternatingRowColors(True)
self.treeWidget.setUniformRowHeights(True)
self.treeWidget.setEditTriggers(QTreeView.NoEditTriggers)
linkdelegate = LinkStyledItemDelegate(self.treeWidget)
self.treeWidget.setItemDelegateForColumn(1, linkdelegate)
self.treeWidget.setItemDelegateForColumn(8, linkdelegate)
self.treeWidget.setItemDelegateForColumn(
0, gui.IndicatorItemDelegate(self.treeWidget, role=Qt.DisplayRole))
proxyModel = MySortFilterProxyModel(self.treeWidget)
self.treeWidget.setModel(proxyModel)
self.treeWidget.selectionModel().selectionChanged.connect(
self.updateSelection)
self.treeWidget.viewport().setMouseTracking(True)
splitterH = QSplitter(Qt.Horizontal, splitter)
box = gui.widgetBox(splitterH, "Description")
self.infoGDS = gui.widgetLabel(box, "")
self.infoGDS.setWordWrap(True)
gui.rubber(box)
box = gui.widgetBox(splitterH, "Sample Annotations")
self.annotationsTree = QTreeWidget(box)
self.annotationsTree.setHeaderLabels(
["Type (Sample annotations)", "Sample count"])
self.annotationsTree.setRootIsDecorated(True)
box.layout().addWidget(self.annotationsTree)
self.annotationsTree.itemChanged.connect(
self.annotationSelectionChanged)
self._annotationsUpdating = False
self.splitters = splitter, splitterH
for sp, setting in zip(self.splitters, self.splitterSettings):
sp.splitterMoved.connect(self.splitterMoved)
sp.restoreState(setting)
self.searchKeys = [
"dataset_id", "title", "platform_organism", "description"
]
self.gds = []
self.gds_info = None
self.resize(1000, 600)
self.setBlocking(True)
self.setEnabled(False)
self.progressBarInit()
self._executor = ThreadExecutor()
func = partial(get_gds_model,
methodinvoke(self, "_setProgress", (float, )))
self._inittask = Task(function=func)
self._inittask.finished.connect(self._initializemodel)
self._executor.submit(self._inittask)
self._datatask = None
@Slot(float)
def _setProgress(self, value):
self.progressBarValue = value
def _initializemodel(self):
assert self.thread() is QThread.currentThread()
model, self.gds_info, self.gds = self._inittask.result()
model.setParent(self)
proxy = self.treeWidget.model()
proxy.setFilterKeyColumn(0)
proxy.setFilterRole(TextFilterRole)
proxy.setFilterCaseSensitivity(False)
proxy.setFilterFixedString(self.filterString)
proxy.setSourceModel(model)
proxy.sort(0, Qt.DescendingOrder)
self.progressBarFinished()
self.setBlocking(False)
self.setEnabled(True)
filter_items = " ".join(gds[key] for gds in self.gds
for key in self.searchKeys)
tr_chars = ",.:;!?(){}[]_-+\\|/%#@$^&*<>~`"
tr_table = str.maketrans(tr_chars, " " * len(tr_chars))
filter_items = filter_items.translate(tr_table)
filter_items = sorted(set(filter_items.split(" ")))
filter_items = [item for item in filter_items if len(item) > 3]
self.completer.setTokenList(filter_items)
if self.currentGds:
current_id = self.currentGds["dataset_id"]
gdss = [(i, qunpack(proxy.data(proxy.index(i, 1), Qt.DisplayRole)))
for i in range(proxy.rowCount())]
current = [i for i, data in gdss if data and data == current_id]
if current:
current_index = proxy.index(current[0], 0)
self.treeWidget.selectionModel().select(
current_index,
QItemSelectionModel.Select | QItemSelectionModel.Rows)
self.treeWidget.scrollTo(current_index,
QTreeView.PositionAtCenter)
for i in range(8):
self.treeWidget.resizeColumnToContents(i)
self.treeWidget.setColumnWidth(
1, min(self.treeWidget.columnWidth(1), 300))
self.treeWidget.setColumnWidth(
2, min(self.treeWidget.columnWidth(2), 200))
self.updateInfo()
def updateInfo(self):
gds_info = self.gds_info
text = ("%i datasets\n%i datasets cached\n" %
(len(gds_info),
len(glob.glob(serverfiles.localpath("GEO") + "/GDS*"))))
filtered = self.treeWidget.model().rowCount()
if len(self.gds) != filtered:
text += ("%i after filtering") % filtered
self.infoBox.setText(text)
def updateSelection(self, *args):
current = self.treeWidget.selectedIndexes()
mapToSource = self.treeWidget.model().mapToSource
current = [mapToSource(index).row() for index in current]
if current:
self.currentGds = self.gds[current[0]]
self.setAnnotations(self.currentGds)
self.infoGDS.setText(self.currentGds.get("description", ""))
self.nameEdit.setPlaceholderText(self.currentGds["title"])
self.datasetName = \
self.datasetNames.get(self.currentGds["dataset_id"], "")
else:
self.currentGds = None
self.nameEdit.setPlaceholderText("")
self.datasetName = ""
self.commitIf()
def setAnnotations(self, gds):
self._annotationsUpdating = True
self.annotationsTree.clear()
annotations = defaultdict(set)
subsetscount = {}
for desc in gds["subsets"]:
annotations[desc["type"]].add(desc["description"])
subsetscount[desc["description"]] = str(len(desc["sample_id"]))
for type, subsets in annotations.items():
key = (gds["dataset_id"], type)
subsetItem = QTreeWidgetItem(self.annotationsTree, [type])
subsetItem.setFlags(subsetItem.flags() | Qt.ItemIsUserCheckable
| Qt.ItemIsTristate)
subsetItem.setCheckState(
0, self.gdsSelectionStates.get(key, Qt.Checked))
subsetItem.key = key
for subset in subsets:
key = (gds["dataset_id"], type, subset)
item = QTreeWidgetItem(
subsetItem, [subset, subsetscount.get(subset, "")])
item.setFlags(item.flags() | Qt.ItemIsUserCheckable)
item.setCheckState(
0, self.gdsSelectionStates.get(key, Qt.Checked))
item.key = key
self._annotationsUpdating = False
self.annotationsTree.expandAll()
for i in range(self.annotationsTree.columnCount()):
self.annotationsTree.resizeColumnToContents(i)
def annotationSelectionChanged(self, item, column):
if self._annotationsUpdating:
return
for i in range(self.annotationsTree.topLevelItemCount()):
item = self.annotationsTree.topLevelItem(i)
self.gdsSelectionStates[item.key] = item.checkState(0)
for j in range(item.childCount()):
child = item.child(j)
self.gdsSelectionStates[child.key] = child.checkState(0)
def filter(self):
filter_string = unicode(self.filterLineEdit.text())
proxyModel = self.treeWidget.model()
if proxyModel:
strings = filter_string.lower().strip().split()
proxyModel.setFilterFixedStrings(strings)
self.updateInfo()
def selectedSamples(self):
"""
Return the currently selected sample annotations.
The return value is a list of selected (sample type, sample value)
tuples.
.. note:: if some Sample annotation type has no selected values.
this method will return all values for it.
"""
samples = []
unused_types = []
used_types = []
for stype in childiter(self.annotationsTree.invisibleRootItem()):
selected_values = []
all_values = []
for sval in childiter(stype):
value = (str(stype.text(0)), str(sval.text(0)))
if self.gdsSelectionStates.get(sval.key, True):
selected_values.append(value)
all_values.append(value)
if selected_values:
samples.extend(selected_values)
used_types.append(str(stype.text(0)))
else:
# If no sample of sample type is selected we don't filter
# on it.
samples.extend(all_values)
unused_types.append(str(stype.text(0)))
return samples, used_types
def commitIf(self):
if self.autoCommit:
self.commit()
else:
self.selectionChanged = True
@Slot(int, int)
def progressCompleted(self, value, total):
if total > 0:
self.progressBarSet(100. * value / total, processEvents=False)
else:
pass
# TODO: report 'indeterminate progress'
def commit(self):
if self.currentGds:
self.error(0)
sample_type = None
self.progressBarInit(processEvents=None)
_, groups = self.selectedSamples()
if len(groups) == 1 and self.outputRows:
sample_type = groups[0]
self.setEnabled(False)
self.setBlocking(True)
progress = methodinvoke(self, "progressCompleted", (int, int))
def get_data(gds_id, report_genes, transpose, sample_type, title):
gds_ensure_downloaded(gds_id, progress)
gds = geo.GDS(gds_id)
data = gds.getdata(report_genes=report_genes,
transpose=transpose,
sample_type=sample_type)
data.name = title
return data
get_data = partial(get_data,
self.currentGds["dataset_id"],
report_genes=self.mergeSpots,
transpose=self.outputRows,
sample_type=sample_type,
title=self.datasetName
or self.currentGds["title"])
self._datatask = Task(function=get_data)
self._datatask.finished.connect(self._on_dataready)
self._executor.submit(self._datatask)
def _on_dataready(self):
self.setEnabled(True)
self.setBlocking(False)
self.progressBarFinished(processEvents=False)
try:
data = self._datatask.result()
except urlrequest.URLError as error:
self.error(0, ("Error while connecting to the NCBI ftp server! "
"'%s'" % error))
sys.excepthook(type(error), error, getattr(error, "__traceback__"))
return
finally:
self._datatask = None
data_name = data.name
samples, _ = self.selectedSamples()
self.warning(0)
message = None
if self.outputRows:
def samplesinst(ex):
out = []
for meta in data.domain.metas:
out.append((meta.name, ex[meta].value))
if data.domain.class_var.name != 'class':
out.append((data.domain.class_var.name,
ex[data.domain.class_var].value))
return out
samples = set(samples)
mask = [samples.issuperset(samplesinst(ex)) for ex in data]
data = data[numpy.array(mask, dtype=bool)]
if len(data) == 0:
message = "No samples with selected sample annotations."
else:
samples = set(samples)
domain = Orange.data.Domain([
attr for attr in data.domain.attributes
if samples.issuperset(attr.attributes.items())
], data.domain.class_var, data.domain.metas)
# domain.addmetas(data.domain.getmetas())
if len(domain.attributes) == 0:
message = "No samples with selected sample annotations."
stypes = set(s[0] for s in samples)
for attr in domain.attributes:
attr.attributes = dict(
(key, value) for key, value in attr.attributes.items()
if key in stypes)
data = Orange.data.Table(domain, data)
if message is not None:
self.warning(0, message)
data_hints.set_hint(data, "taxid", self.currentGds.get("taxid", ""),
10.0)
data_hints.set_hint(data, "genesinrows", self.outputRows, 10.0)
data.name = data_name
self.send("Expression Data", data)
model = self.treeWidget.model().sourceModel()
row = self.gds.index(self.currentGds)
model.setData(model.index(row, 0), " ", Qt.DisplayRole)
self.updateInfo()
self.selectionChanged = False
def splitterMoved(self, *args):
self.splitterSettings = [
bytes(sp.saveState()) for sp in self.splitters
]
def send_report(self):
self.report_items("GEO Dataset",
[("ID", self.currentGds['dataset_id']),
("Title", self.currentGds['title']),
("Organism", self.currentGds['sample_organism'])])
self.report_items("Data",
[("Samples", self.currentGds['sample_count']),
("Features", self.currentGds['feature_count']),
("Genes", self.currentGds['gene_count'])])
self.report_name("Sample annotations")
subsets = defaultdict(list)
for subset in self.currentGds['subsets']:
subsets[subset['type']].append(
(subset['description'], len(subset['sample_id'])))
self.report_html += "<ul>"
for type in subsets:
self.report_html += "<b>" + type + ":</b></br>"
for desc, count in subsets[type]:
self.report_html += 9 * " " + "<b>{}:</b> {}</br>".format(
desc, count)
self.report_html += "</ul>"
def onDeleteWidget(self):
if self._inittask:
self._inittask.future().cancel()
self._inittask.finished.disconnect(self._initializemodel)
if self._datatask:
self._datatask.future().cancel()
self._datatask.finished.disconnect(self._on_dataready)
self._executor.shutdown(wait=False)
super(OWGEODatasets, self).onDeleteWidget()
def onNameEdited(self):
if self.currentGds:
gds_id = self.currentGds["dataset_id"]
self.datasetNames[gds_id] = unicode(self.nameEdit.text())
self.commitIf()
class OWGEODatasets(OWWidget):
name = "GEO Data Sets"
description = DESCRIPTION
icon = "../widgets/icons/GEODataSets.svg"
priority = PRIORITY
inputs = []
outputs = [("Expression Data", Orange.data.Table)]
settingsList = ["outputRows", "mergeSpots", "gdsSelectionStates",
"splitterSettings", "currentGds", "autoCommit",
"datasetNames"]
outputRows = Setting(True)
mergeSpots = Setting(True)
gdsSelectionStates = Setting({})
currentGds = Setting(None)
datasetNames = Setting({})
splitterSettings = Setting(
(b'\x00\x00\x00\xff\x00\x00\x00\x00\x00\x00\x00\x02\x00\x00\x01\xea\x00\x00\x00\xd7\x01\x00\x00\x00\x07\x01\x00\x00\x00\x02',
b'\x00\x00\x00\xff\x00\x00\x00\x00\x00\x00\x00\x02\x00\x00\x01\xb5\x00\x00\x02\x10\x01\x00\x00\x00\x07\x01\x00\x00\x00\x01')
)
autoCommit = Setting(False)
def __init__(self, parent=None, signalManager=None, name=" GEO Data Sets"):
OWWidget.__init__(self, parent, signalManager, name)
self.selectionChanged = False
self.filterString = ""
self.datasetName = ""
## GUI
box = gui.widgetBox(self.controlArea, "Info", addSpace=True)
self.infoBox = gui.widgetLabel(box, "Initializing\n\n")
box = gui.widgetBox(self.controlArea, "Output", addSpace=True)
gui.radioButtonsInBox(box, self, "outputRows",
["Genes in rows", "Samples in rows"], "Rows",
callback=self.commitIf)
gui.checkBox(box, self, "mergeSpots", "Merge spots of same gene",
callback=self.commitIf)
gui.separator(box)
self.nameEdit = gui.lineEdit(
box, self, "datasetName", "Data set name",
tooltip="Override the default output data set name",
callback=self.onNameEdited
)
self.nameEdit.setPlaceholderText("")
if sys.version_info < (3, ):
box = gui.widgetBox(self.controlArea, "Commit", addSpace=True)
self.commitButton = gui.button(
box, self, "Commit", callback=self.commit)
cb = gui.checkBox(box, self, "autoCommit", "Commit on any change")
gui.setStopper(self, self.commitButton, cb, "selectionChanged",
self.commit)
else:
gui.auto_commit(self.controlArea, self, "autoCommit", "Commit",
box="Commit")
self.commitIf = self.commit
gui.rubber(self.controlArea)
gui.widgetLabel(self.mainArea, "Filter")
self.filterLineEdit = QLineEdit(
textChanged=self.filter
)
self.completer = TokenListCompleter(
self, caseSensitivity=Qt.CaseInsensitive
)
self.filterLineEdit.setCompleter(self.completer)
self.mainArea.layout().addWidget(self.filterLineEdit)
splitter = QSplitter(Qt.Vertical, self.mainArea)
self.mainArea.layout().addWidget(splitter)
self.treeWidget = QTreeView(splitter)
self.treeWidget.setSelectionMode(QTreeView.SingleSelection)
self.treeWidget.setRootIsDecorated(False)
self.treeWidget.setSortingEnabled(True)
self.treeWidget.setAlternatingRowColors(True)
self.treeWidget.setUniformRowHeights(True)
self.treeWidget.setEditTriggers(QTreeView.NoEditTriggers)
linkdelegate = LinkStyledItemDelegate(self.treeWidget)
self.treeWidget.setItemDelegateForColumn(1, linkdelegate)
self.treeWidget.setItemDelegateForColumn(8, linkdelegate)
self.treeWidget.setItemDelegateForColumn(
0, gui.IndicatorItemDelegate(self.treeWidget,
role=Qt.DisplayRole))
proxyModel = MySortFilterProxyModel(self.treeWidget)
self.treeWidget.setModel(proxyModel)
self.treeWidget.selectionModel().selectionChanged.connect(
self.updateSelection
)
self.treeWidget.viewport().setMouseTracking(True)
splitterH = QSplitter(Qt.Horizontal, splitter)
box = gui.widgetBox(splitterH, "Description")
self.infoGDS = gui.widgetLabel(box, "")
self.infoGDS.setWordWrap(True)
gui.rubber(box)
box = gui.widgetBox(splitterH, "Sample Annotations")
self.annotationsTree = QTreeWidget(box)
self.annotationsTree.setHeaderLabels(
["Type (Sample annotations)", "Sample count"]
)
self.annotationsTree.setRootIsDecorated(True)
box.layout().addWidget(self.annotationsTree)
self.annotationsTree.itemChanged.connect(
self.annotationSelectionChanged
)
self._annotationsUpdating = False
self.splitters = splitter, splitterH
for sp, setting in zip(self.splitters, self.splitterSettings):
sp.splitterMoved.connect(self.splitterMoved)
sp.restoreState(setting)
self.searchKeys = ["dataset_id", "title", "platform_organism",
"description"]
self.gds = []
self.gds_info = None
self.resize(1000, 600)
self.setBlocking(True)
self.setEnabled(False)
self.progressBarInit()
self._executor = ThreadExecutor()
func = partial(get_gds_model,
methodinvoke(self, "_setProgress", (float,)))
self._inittask = Task(function=func)
self._inittask.finished.connect(self._initializemodel)
self._executor.submit(self._inittask)
self._datatask = None
@Slot(float)
def _setProgress(self, value):
self.progressBarValue = value
def _initializemodel(self):
assert self.thread() is QThread.currentThread()
model, self.gds_info, self.gds = self._inittask.result()
model.setParent(self)
proxy = self.treeWidget.model()
proxy.setFilterKeyColumn(0)
proxy.setFilterRole(TextFilterRole)
proxy.setFilterCaseSensitivity(False)
proxy.setFilterFixedString(self.filterString)
proxy.setSourceModel(model)
proxy.sort(0, Qt.DescendingOrder)
self.progressBarFinished()
self.setBlocking(False)
self.setEnabled(True)
filter_items = " ".join(
gds[key] for gds in self.gds for key in self.searchKeys
)
tr_chars = ",.:;!?(){}[]_-+\\|/%#@$^&*<>~`"
tr_table = str.maketrans(tr_chars, " " * len(tr_chars))
filter_items = filter_items.translate(tr_table)
filter_items = sorted(set(filter_items.split(" ")))
filter_items = [item for item in filter_items if len(item) > 3]
self.completer.setTokenList(filter_items)
if self.currentGds:
current_id = self.currentGds["dataset_id"]
gdss = [(i, qunpack(proxy.data(proxy.index(i, 1), Qt.DisplayRole)))
for i in range(proxy.rowCount())]
current = [i for i, data in gdss if data and data == current_id]
if current:
current_index = proxy.index(current[0], 0)
self.treeWidget.selectionModel().select(
current_index,
QItemSelectionModel.Select | QItemSelectionModel.Rows
)
self.treeWidget.scrollTo(
current_index, QTreeView.PositionAtCenter)
for i in range(8):
self.treeWidget.resizeColumnToContents(i)
self.treeWidget.setColumnWidth(
1, min(self.treeWidget.columnWidth(1), 300))
self.treeWidget.setColumnWidth(
2, min(self.treeWidget.columnWidth(2), 200))
self.updateInfo()
def updateInfo(self):
gds_info = self.gds_info
text = ("%i datasets\n%i datasets cached\n" %
(len(gds_info),
len(glob.glob(serverfiles.localpath("GEO") + "/GDS*"))))
filtered = self.treeWidget.model().rowCount()
if len(self.gds) != filtered:
text += ("%i after filtering") % filtered
self.infoBox.setText(text)
def updateSelection(self, *args):
current = self.treeWidget.selectedIndexes()
mapToSource = self.treeWidget.model().mapToSource
current = [mapToSource(index).row() for index in current]
if current:
self.currentGds = self.gds[current[0]]
self.setAnnotations(self.currentGds)
self.infoGDS.setText(self.currentGds.get("description", ""))
self.nameEdit.setPlaceholderText(self.currentGds["title"])
self.datasetName = \
self.datasetNames.get(self.currentGds["dataset_id"], "")
else:
self.currentGds = None
self.nameEdit.setPlaceholderText("")
self.datasetName = ""
self.commitIf()
def setAnnotations(self, gds):
self._annotationsUpdating = True
self.annotationsTree.clear()
annotations = defaultdict(set)
subsetscount = {}
for desc in gds["subsets"]:
annotations[desc["type"]].add(desc["description"])
subsetscount[desc["description"]] = str(len(desc["sample_id"]))
for type, subsets in annotations.items():
key = (gds["dataset_id"], type)
subsetItem = QTreeWidgetItem(self.annotationsTree, [type])
subsetItem.setFlags(subsetItem.flags() | Qt.ItemIsUserCheckable |
Qt.ItemIsTristate)
subsetItem.setCheckState(
0, self.gdsSelectionStates.get(key, Qt.Checked)
)
subsetItem.key = key
for subset in subsets:
key = (gds["dataset_id"], type, subset)
item = QTreeWidgetItem(
subsetItem, [subset, subsetscount.get(subset, "")]
)
item.setFlags(item.flags() | Qt.ItemIsUserCheckable)
item.setCheckState(
0, self.gdsSelectionStates.get(key, Qt.Checked)
)
item.key = key
self._annotationsUpdating = False
self.annotationsTree.expandAll()
for i in range(self.annotationsTree.columnCount()):
self.annotationsTree.resizeColumnToContents(i)
def annotationSelectionChanged(self, item, column):
if self._annotationsUpdating:
return
for i in range(self.annotationsTree.topLevelItemCount()):
item = self.annotationsTree.topLevelItem(i)
self.gdsSelectionStates[item.key] = item.checkState(0)
for j in range(item.childCount()):
child = item.child(j)
self.gdsSelectionStates[child.key] = child.checkState(0)
def filter(self):
filter_string = unicode(self.filterLineEdit.text())
proxyModel = self.treeWidget.model()
if proxyModel:
strings = filter_string.lower().strip().split()
proxyModel.setFilterFixedStrings(strings)
self.updateInfo()
def selectedSamples(self):
"""
Return the currently selected sample annotations.
The return value is a list of selected (sample type, sample value)
tuples.
.. note:: if some Sample annotation type has no selected values.
this method will return all values for it.
"""
samples = []
unused_types = []
used_types = []
for stype in childiter(self.annotationsTree.invisibleRootItem()):
selected_values = []
all_values = []
for sval in childiter(stype):
value = (str(stype.text(0)), str(sval.text(0)))
if self.gdsSelectionStates.get(sval.key, True):
selected_values.append(value)
all_values.append(value)
if selected_values:
samples.extend(selected_values)
used_types.append(str(stype.text(0)))
else:
# If no sample of sample type is selected we don't filter
# on it.
samples.extend(all_values)
unused_types.append(str(stype.text(0)))
return samples, used_types
def commitIf(self):
if self.autoCommit:
self.commit()
else:
self.selectionChanged = True
@Slot(int, int)
def progressCompleted(self, value, total):
if total > 0:
self.progressBarSet(100. * value / total, processEvents=False)
else:
pass
# TODO: report 'indeterminate progress'
def commit(self):
if self.currentGds:
self.error(0)
sample_type = None
self.progressBarInit(processEvents=None)
_, groups = self.selectedSamples()
if len(groups) == 1 and self.outputRows:
sample_type = groups[0]
self.setEnabled(False)
self.setBlocking(True)
progress = methodinvoke(self, "progressCompleted", (int, int))
def get_data(gds_id, report_genes, transpose, sample_type, title):
gds_ensure_downloaded(gds_id, progress)
gds = geo.GDS(gds_id)
data = gds.getdata(
report_genes=report_genes, transpose=transpose,
sample_type=sample_type
)
data.name = title
return data
get_data = partial(
get_data, self.currentGds["dataset_id"],
report_genes=self.mergeSpots,
transpose=self.outputRows,
sample_type=sample_type,
title=self.datasetName or self.currentGds["title"]
)
self._datatask = Task(function=get_data)
self._datatask.finished.connect(self._on_dataready)
self._executor.submit(self._datatask)
def _on_dataready(self):
self.setEnabled(True)
self.setBlocking(False)
self.progressBarFinished(processEvents=False)
try:
data = self._datatask.result()
except urlrequest.URLError as error:
self.error(0, ("Error while connecting to the NCBI ftp server! "
"'%s'" % error))
sys.excepthook(type(error), error, getattr(error, "__traceback__"))
return
finally:
self._datatask = None
data_name = data.name
samples, _ = self.selectedSamples()
self.warning(0)
message = None
if self.outputRows:
def samplesinst(ex):
out = []
for meta in data.domain.metas:
out.append((meta.name, ex[meta].value))
if data.domain.class_var.name != 'class':
out.append((data.domain.class_var.name,
ex[data.domain.class_var].value))
return out
samples = set(samples)
mask = [samples.issuperset(samplesinst(ex)) for ex in data]
data = data[numpy.array(mask, dtype=bool)]
if len(data) == 0:
message = "No samples with selected sample annotations."
else:
samples = set(samples)
domain = Orange.data.Domain(
[attr for attr in data.domain.attributes
if samples.issuperset(attr.attributes.items())],
data.domain.class_var,
data.domain.metas
)
# domain.addmetas(data.domain.getmetas())
if len(domain.attributes) == 0:
message = "No samples with selected sample annotations."
stypes = set(s[0] for s in samples)
for attr in domain.attributes:
attr.attributes = dict(
(key, value) for key, value in attr.attributes.items()
if key in stypes
)
data = Orange.data.Table(domain, data)
if message is not None:
self.warning(0, message)
data_hints.set_hint(data, "taxid", self.currentGds.get("taxid", ""),
10.0)
data_hints.set_hint(data, "genesinrows", self.outputRows, 10.0)
data.name = data_name
self.send("Expression Data", data)
model = self.treeWidget.model().sourceModel()
row = self.gds.index(self.currentGds)
model.setData(model.index(row, 0), " ", Qt.DisplayRole)
self.updateInfo()
self.selectionChanged = False
def splitterMoved(self, *args):
self.splitterSettings = [bytes(sp.saveState()) for sp in self.splitters]
def send_report(self):
self.report_items("GEO Dataset",
[("ID", self.currentGds['dataset_id']), ("Title", self.currentGds['title']),
("Organism", self.currentGds['sample_organism'])])
self.report_items("Data", [("Samples", self.currentGds['sample_count']),
("Features", self.currentGds['feature_count']),
("Genes", self.currentGds['gene_count'])])
self.report_name("Sample annotations")
subsets = defaultdict(list)
for subset in self.currentGds['subsets']:
subsets[subset['type']].append((subset['description'], len(subset['sample_id'])))
self.report_html += "<ul>"
for type in subsets:
self.report_html += "<b>" + type + ":</b></br>"
for desc, count in subsets[type]:
self.report_html += 9 * " " + "<b>{}:</b> {}</br>".format(desc, count)
self.report_html += "</ul>"
def onDeleteWidget(self):
if self._inittask:
self._inittask.future().cancel()
self._inittask.finished.disconnect(self._initializemodel)
if self._datatask:
self._datatask.future().cancel()
self._datatask.finished.disconnect(self._on_dataready)
self._executor.shutdown(wait=False)
super(OWGEODatasets, self).onDeleteWidget()
def onNameEdited(self):
if self.currentGds:
gds_id = self.currentGds["dataset_id"]
self.datasetNames[gds_id] = unicode(self.nameEdit.text())
self.commitIf()
class OWDataSets(OWWidget):
name = "Datasets"
description = "Load a dataset from an online repository"
icon = "icons/DataSets.svg"
priority = 20
replaces = ["orangecontrib.prototypes.widgets.owdatasets.OWDataSets"]
keywords = ["online"]
# The following constants can be overridden in a subclass
# to reuse this widget for a different repository
# Take care when refactoring! (used in e.g. single-cell)
INDEX_URL = "https://datasets.biolab.si/"
DATASET_DIR = "datasets"
# override HEADER_SCHEMA to define new columns
# if schema is changed override methods: self.assign_delegates and
# self.create_model
HEADER_SCHEMA = [['islocal', {
'label': ''
}], ['title', {
'label': 'Title'
}], ['size', {
'label': 'Size'
}], ['instances', {
'label': 'Instances'
}], ['variables', {
'label': 'Variables'
}], ['target', {
'label': 'Target'
}], ['tags', {
'label': 'Tags'
}]] # type: List[str, dict]
class Error(OWWidget.Error):
no_remote_datasets = Msg("Could not fetch dataset list")
class Warning(OWWidget.Warning):
only_local_datasets = Msg("Could not fetch datasets list, only local "
"cached datasets are shown")
class Outputs:
data = Output("Data", Orange.data.Table)
#: Selected dataset id
selected_id = settings.Setting(None) # type: Optional[str]
auto_commit = settings.Setting(False) # type: bool
#: main area splitter state
splitter_state = settings.Setting(b'') # type: bytes
header_state = settings.Setting(b'') # type: bytes
def __init__(self):
super().__init__()
self.allinfo_local = {}
self.allinfo_remote = {}
self.local_cache_path = os.path.join(data_dir(), self.DATASET_DIR)
self._header_labels = [
header['label'] for _, header in self.HEADER_SCHEMA
]
self._header_index = namedtuple(
'_header_index', [info_tag for info_tag, _ in self.HEADER_SCHEMA])
self.Header = self._header_index(
*[index for index, _ in enumerate(self._header_labels)])
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,
uniformRowHeights=True,
)
# the method doesn't exists yet, pylint: disable=unnecessary-lambda
self.view.doubleClicked.connect(lambda: self.unconditional_commit())
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")
proxy = QSortFilterProxyModel()
proxy.setFilterKeyColumn(-1)
proxy.setFilterCaseSensitivity(False)
self.view.setModel(proxy)
if self.splitter_state:
self.splitter.restoreState(self.splitter_state)
self.assign_delegates()
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 assign_delegates(self):
# NOTE: All columns must have size hinting delegates.
# QTreeView queries only the columns displayed in the viewport so
# the layout would be different depending in the horizontal scroll
# position
self.view.setItemDelegate(UniformHeightDelegate(self))
self.view.setItemDelegateForColumn(
self.Header.islocal,
UniformHeightIndicatorDelegate(self, role=Qt.DisplayRole))
self.view.setItemDelegateForColumn(self.Header.size,
SizeDelegate(self))
self.view.setItemDelegateForColumn(self.Header.instances,
NumericalDelegate(self))
self.view.setItemDelegateForColumn(self.Header.variables,
NumericalDelegate(self))
self.view.resizeColumnToContents(self.Header.islocal)
def _parse_info(self, file_path):
if file_path in self.allinfo_remote:
info = self.allinfo_remote[file_path]
else:
info = self.allinfo_local[file_path]
islocal = file_path in self.allinfo_local
isremote = file_path in self.allinfo_remote
outdated = islocal and isremote and (
self.allinfo_remote[file_path].get('version', '') !=
self.allinfo_local[file_path].get('version', ''))
islocal &= not outdated
prefix = os.path.join('', *file_path[:-1])
filename = file_path[-1]
return Namespace(file_path=file_path,
prefix=prefix,
filename=filename,
islocal=islocal,
outdated=outdated,
**info)
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
@Slot(object)
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("")
self.allinfo_local = self.list_local()
try:
self.allinfo_remote = f.result()
except Exception: # anytying can happen, pylint: disable=broad-except
log.exception("Error while fetching updated index")
if not self.allinfo_local:
self.Error.no_remote_datasets()
else:
self.Warning.only_local_datasets()
self.allinfo_remote = {}
model, current_index = self.create_model()
self.view.model().setSourceModel(model)
self.view.selectionModel().selectionChanged.connect(
self.__on_selection)
self.view.resizeColumnToContents(0)
self.view.setColumnWidth(
1,
min(self.view.sizeHintForColumn(1),
self.view.fontMetrics().width("X" * 37)))
header = self.view.header()
header.restoreState(self.header_state)
# Update the info text
self.infolabel.setText(
format_info(model.rowCount(), len(self.allinfo_local)))
if current_index != -1:
selmodel = self.view.selectionModel()
selmodel.select(
self.view.model().mapFromSource(model.index(current_index, 0)),
QItemSelectionModel.ClearAndSelect | QItemSelectionModel.Rows)
def __update_cached_state(self):
model = self.view.model().sourceModel()
localinfo = self.list_local()
assert isinstance(model, QStandardItemModel)
allinfo = []
for i in range(model.rowCount()):
item = model.item(i, 0)
info = item.data(Qt.UserRole)
info.islocal = info.file_path in localinfo
item.setData(" " if info.islocal else "", Qt.DisplayRole)
allinfo.append(info)
self.infolabel.setText(
format_info(model.rowCount(),
sum(info.islocal for info in allinfo)))
def selected_dataset(self):
"""
Return the current selected dataset info or None if not selected
Returns
-------
info : Optional[Namespace]
"""
rows = self.view.selectionModel().selectedRows(0)
assert 0 <= len(rows) <= 1
current = rows[0] if rows else None # type: Optional[QModelIndex]
if current is not None:
info = current.data(Qt.UserRole)
assert isinstance(info, Namespace)
else:
info = None
return info
def filter(self):
filter_string = self.filterLineEdit.text().strip()
proxyModel = self.view.model()
if proxyModel:
proxyModel.setFilterFixedString(filter_string)
def __on_selection(self):
# Main datasets view selection has changed
rows = self.view.selectionModel().selectedRows(0)
assert 0 <= len(rows) <= 1
current = rows[0] if rows else None # type: Optional[QModelIndex]
if current is not None:
current = self.view.model().mapToSource(current)
di = current.data(Qt.UserRole)
text = description_html(di)
self.descriptionlabel.setText(text)
self.selected_id = os.path.join(di.prefix, di.filename)
else:
self.descriptionlabel.setText("")
self.selected_id = None
self.commit()
def commit(self):
"""
Commit a dataset to the output immediately (if available locally) or
schedule download background and an eventual send.
During the download the widget is in blocking state
(OWWidget.isBlocking)
"""
di = self.selected_dataset()
if di is not None:
self.Error.clear()
if self.__awaiting_state is not None:
# disconnect from the __commit_complete
self.__awaiting_state.watcher.done.disconnect(
self.__commit_complete)
# .. and connect to update_cached_state
# self.__awaiting_state.watcher.done.connect(
# self.__update_cached_state)
# TODO: There are possible pending __progress_advance queued
self.__awaiting_state.pb.advance.disconnect(
self.__progress_advance)
self.progressBarFinished(processEvents=None)
self.__awaiting_state = None
if not di.islocal:
pr = progress()
callback = lambda pr=pr: pr.advance.emit()
pr.advance.connect(self.__progress_advance,
Qt.QueuedConnection)
self.progressBarInit(processEvents=None)
self.setStatusMessage("Fetching...")
self.setBlocking(True)
f = self._executor.submit(ensure_local,
self.INDEX_URL,
di.file_path,
self.local_cache_path,
force=di.outdated,
progress_advance=callback)
w = FutureWatcher(f, parent=self)
w.done.connect(self.__commit_complete)
self.__awaiting_state = _FetchState(f, w, pr)
else:
self.setStatusMessage("")
self.setBlocking(False)
self.commit_cached(di.file_path)
else:
self.Outputs.data.send(None)
@Slot(object)
def __commit_complete(self, f):
# complete the commit operation after the required file has been
# downloaded
assert QThread.currentThread() is self.thread()
assert self.__awaiting_state is not None
assert self.__awaiting_state.future is f
if self.isBlocking():
self.progressBarFinished(processEvents=None)
self.setBlocking(False)
self.setStatusMessage("")
self.__awaiting_state = None
try:
path = f.result()
# anything can happen here, pylint: disable=broad-except
except Exception as ex:
log.exception("Error:")
self.error(format_exception(ex))
path = None
self.__update_cached_state()
if path is not None:
data = self.load_data(path)
else:
data = None
self.Outputs.data.send(data)
def commit_cached(self, file_path):
path = LocalFiles(self.local_cache_path).localpath(*file_path)
self.Outputs.data.send(self.load_data(path))
@Slot()
def __progress_advance(self):
assert QThread.currentThread() is self.thread()
self.progressBarAdvance(1, processEvents=None)
def onDeleteWidget(self):
super().onDeleteWidget()
if self.__awaiting_state is not None:
self.__awaiting_state.watcher.done.disconnect(
self.__commit_complete)
self.__awaiting_state.pb.advance.disconnect(
self.__progress_advance)
self.__awaiting_state = None
@staticmethod
def sizeHint():
return QSize(1100, 500)
def closeEvent(self, event):
self.splitter_state = bytes(self.splitter.saveState())
self.header_state = bytes(self.view.header().saveState())
super().closeEvent(event)
def load_data(self, path): # pylint: disable=no-self-use
return Orange.data.Table(path)
def list_remote(self):
# type: () -> Dict[Tuple[str, ...], dict]
client = ServerFiles(server=self.INDEX_URL)
return client.allinfo()
def list_local(self):
# type: () -> Dict[Tuple[str, ...], dict]
return LocalFiles(self.local_cache_path).allinfo()
class OWDataSets(OWWidget):
name = "Datasets"
description = "Load a dataset from an online repository"
icon = "icons/DataSets.svg"
priority = 20
replaces = ["orangecontrib.prototypes.widgets.owdatasets.OWDataSets"]
keywords = ["online"]
# The following constants can be overridden in a subclass
# to reuse this widget for a different repository
# Take care when refactoring! (used in e.g. single-cell)
INDEX_URL = "https://datasets.biolab.si/"
DATASET_DIR = "datasets"
# override HEADER_SCHEMA to define new columns
# if schema is changed override methods: self.assign_delegates and
# self.create_model
HEADER_SCHEMA = [
['islocal', {'label': ''}],
['title', {'label': 'Title'}],
['size', {'label': 'Size'}],
['instances', {'label': 'Instances'}],
['variables', {'label': 'Variables'}],
['target', {'label': 'Target'}],
['tags', {'label': 'Tags'}]
] # type: List[str, dict]
class Error(OWWidget.Error):
no_remote_datasets = Msg("Could not fetch dataset list")
class Warning(OWWidget.Warning):
only_local_datasets = Msg("Could not fetch datasets list, only local "
"cached datasets are shown")
class Outputs:
data = Output("Data", Orange.data.Table)
#: Selected dataset id
selected_id = settings.Setting(None) # type: Optional[str]
auto_commit = settings.Setting(False) # type: bool
#: main area splitter state
splitter_state = settings.Setting(b'') # type: bytes
header_state = settings.Setting(b'') # type: bytes
def __init__(self):
super().__init__()
self.allinfo_local = {}
self.allinfo_remote = {}
self.local_cache_path = os.path.join(data_dir(), self.DATASET_DIR)
self._header_labels = [
header['label'] for _, header in self.HEADER_SCHEMA]
self._header_index = namedtuple(
'_header_index', [info_tag for info_tag, _ in self.HEADER_SCHEMA])
self.Header = self._header_index(
*[index for index, _ in enumerate(self._header_labels)])
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,
uniformRowHeights=True,
)
# the method doesn't exists yet, pylint: disable=unnecessary-lambda
self.view.doubleClicked.connect(lambda: self.unconditional_commit())
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")
proxy = QSortFilterProxyModel()
proxy.setFilterKeyColumn(-1)
proxy.setFilterCaseSensitivity(False)
self.view.setModel(proxy)
if self.splitter_state:
self.splitter.restoreState(self.splitter_state)
self.assign_delegates()
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 assign_delegates(self):
# NOTE: All columns must have size hinting delegates.
# QTreeView queries only the columns displayed in the viewport so
# the layout would be different depending in the horizontal scroll
# position
self.view.setItemDelegate(UniformHeightDelegate(self))
self.view.setItemDelegateForColumn(
self.Header.islocal,
UniformHeightIndicatorDelegate(self, role=Qt.DisplayRole)
)
self.view.setItemDelegateForColumn(
self.Header.size,
SizeDelegate(self)
)
self.view.setItemDelegateForColumn(
self.Header.instances,
NumericalDelegate(self)
)
self.view.setItemDelegateForColumn(
self.Header.variables,
NumericalDelegate(self)
)
self.view.resizeColumnToContents(self.Header.islocal)
def _parse_info(self, file_path):
if file_path in self.allinfo_remote:
info = self.allinfo_remote[file_path]
else:
info = self.allinfo_local[file_path]
islocal = file_path in self.allinfo_local
isremote = file_path in self.allinfo_remote
outdated = islocal and isremote and (
self.allinfo_remote[file_path].get('version', '')
!= self.allinfo_local[file_path].get('version', '')
)
islocal &= not outdated
prefix = os.path.join('', *file_path[:-1])
filename = file_path[-1]
return Namespace(file_path=file_path, prefix=prefix, filename=filename,
islocal=islocal, outdated=outdated, **info)
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
@Slot(object)
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("")
self.allinfo_local = self.list_local()
try:
self.allinfo_remote = f.result()
except Exception: # anytying can happen, pylint: disable=broad-except
log.exception("Error while fetching updated index")
if not self.allinfo_local:
self.Error.no_remote_datasets()
else:
self.Warning.only_local_datasets()
self.allinfo_remote = {}
model, current_index = self.create_model()
self.view.model().setSourceModel(model)
self.view.selectionModel().selectionChanged.connect(
self.__on_selection
)
self.view.resizeColumnToContents(0)
self.view.setColumnWidth(
1, min(self.view.sizeHintForColumn(1),
self.view.fontMetrics().width("X" * 24)))
header = self.view.header()
header.restoreState(self.header_state)
# Update the info text
self.infolabel.setText(
format_info(model.rowCount(), len(self.allinfo_local)))
if current_index != -1:
selmodel = self.view.selectionModel()
selmodel.select(
self.view.model().mapFromSource(model.index(current_index, 0)),
QItemSelectionModel.ClearAndSelect | QItemSelectionModel.Rows)
def __update_cached_state(self):
model = self.view.model().sourceModel()
localinfo = self.list_local()
assert isinstance(model, QStandardItemModel)
allinfo = []
for i in range(model.rowCount()):
item = model.item(i, 0)
info = item.data(Qt.UserRole)
info.islocal = info.file_path in localinfo
item.setData(" " if info.islocal else "", Qt.DisplayRole)
allinfo.append(info)
self.infolabel.setText(format_info(
model.rowCount(), sum(info.islocal for info in allinfo)))
def selected_dataset(self):
"""
Return the current selected dataset info or None if not selected
Returns
-------
info : Optional[Namespace]
"""
rows = self.view.selectionModel().selectedRows(0)
assert 0 <= len(rows) <= 1
current = rows[0] if rows else None # type: Optional[QModelIndex]
if current is not None:
info = current.data(Qt.UserRole)
assert isinstance(info, Namespace)
else:
info = None
return info
def filter(self):
filter_string = self.filterLineEdit.text().strip()
proxyModel = self.view.model()
if proxyModel:
proxyModel.setFilterFixedString(filter_string)
def __on_selection(self):
# Main datasets view selection has changed
rows = self.view.selectionModel().selectedRows(0)
assert 0 <= len(rows) <= 1
current = rows[0] if rows else None # type: Optional[QModelIndex]
if current is not None:
current = self.view.model().mapToSource(current)
di = current.data(Qt.UserRole)
text = description_html(di)
self.descriptionlabel.setText(text)
self.selected_id = os.path.join(di.prefix, di.filename)
else:
self.descriptionlabel.setText("")
self.selected_id = None
self.commit()
def commit(self):
"""
Commit a dataset to the output immediately (if available locally) or
schedule download background and an eventual send.
During the download the widget is in blocking state
(OWWidget.isBlocking)
"""
di = self.selected_dataset()
if di is not None:
self.Error.clear()
if self.__awaiting_state is not None:
# disconnect from the __commit_complete
self.__awaiting_state.watcher.done.disconnect(
self.__commit_complete)
# .. and connect to update_cached_state
# self.__awaiting_state.watcher.done.connect(
# self.__update_cached_state)
# TODO: There are possible pending __progress_advance queued
self.__awaiting_state.pb.advance.disconnect(
self.__progress_advance)
self.progressBarFinished(processEvents=None)
self.__awaiting_state = None
if not di.islocal:
pr = progress()
callback = lambda pr=pr: pr.advance.emit()
pr.advance.connect(self.__progress_advance, Qt.QueuedConnection)
self.progressBarInit(processEvents=None)
self.setStatusMessage("Fetching...")
self.setBlocking(True)
f = self._executor.submit(
ensure_local, self.INDEX_URL, di.file_path,
self.local_cache_path, force=di.outdated,
progress_advance=callback)
w = FutureWatcher(f, parent=self)
w.done.connect(self.__commit_complete)
self.__awaiting_state = _FetchState(f, w, pr)
else:
self.setStatusMessage("")
self.setBlocking(False)
self.commit_cached(di.file_path)
else:
self.Outputs.data.send(None)
@Slot(object)
def __commit_complete(self, f):
# complete the commit operation after the required file has been
# downloaded
assert QThread.currentThread() is self.thread()
assert self.__awaiting_state is not None
assert self.__awaiting_state.future is f
if self.isBlocking():
self.progressBarFinished(processEvents=None)
self.setBlocking(False)
self.setStatusMessage("")
self.__awaiting_state = None
try:
path = f.result()
# anything can happen here, pylint: disable=broad-except
except Exception as ex:
log.exception("Error:")
self.error(format_exception(ex))
path = None
self.__update_cached_state()
if path is not None:
data = self.load_data(path)
else:
data = None
self.Outputs.data.send(data)
def commit_cached(self, file_path):
path = LocalFiles(self.local_cache_path).localpath(*file_path)
self.Outputs.data.send(self.load_data(path))
@Slot()
def __progress_advance(self):
assert QThread.currentThread() is self.thread()
self.progressBarAdvance(1, processEvents=None)
def onDeleteWidget(self):
super().onDeleteWidget()
if self.__awaiting_state is not None:
self.__awaiting_state.watcher.done.disconnect(self.__commit_complete)
self.__awaiting_state.pb.advance.disconnect(self.__progress_advance)
self.__awaiting_state = None
@staticmethod
def sizeHint():
return QSize(900, 600)
def closeEvent(self, event):
self.splitter_state = bytes(self.splitter.saveState())
self.header_state = bytes(self.view.header().saveState())
super().closeEvent(event)
def load_data(self, path): # pylint: disable=no-self-use
return Orange.data.Table(path)
def list_remote(self):
# type: () -> Dict[Tuple[str, ...], dict]
client = ServerFiles(server=self.INDEX_URL)
return client.allinfo()
def list_local(self):
# type: () -> Dict[Tuple[str, ...], dict]
return LocalFiles(self.local_cache_path).allinfo()
class OWGeneInfo(widget.OWWidget):
name = "Gene Info"
description = "Displays gene information from NCBI and other sources."
icon = "../widgets/icons/OWGeneInfo.svg"
priority = 5
class Inputs:
data = Input("Data", Orange.data.Table)
class Outputs:
selected_genes = Output("Selected Genes", Orange.data.Table)
data = Output("Data", Orange.data.Table)
settingsHandler = settings.DomainContextHandler()
organism_index = settings.ContextSetting(0)
taxid = settings.ContextSetting("9606")
gene_attr = settings.ContextSetting(0)
auto_commit = settings.Setting(False)
search_string = settings.Setting("")
useAttr = settings.ContextSetting(False)
useAltSource = settings.ContextSetting(False)
def __init__(
self,
parent=None,
):
super().__init__(self, parent)
self.selectionChangedFlag = False
self.__initialized = False
self.initfuture = None
self.itemsfuture = None
self.map_input_to_ensembl = None
self.infoLabel = gui.widgetLabel(
gui.widgetBox(self.controlArea, "Info", addSpace=True),
"Initializing\n")
self.organisms = None
self.organismBox = gui.widgetBox(self.controlArea,
"Organism",
addSpace=True)
self.organismComboBox = gui.comboBox(
self.organismBox,
self,
"organism_index",
callback=self._onSelectedOrganismChanged)
box = gui.widgetBox(self.controlArea, "Gene names", addSpace=True)
self.geneAttrComboBox = gui.comboBox(box,
self,
"gene_attr",
"Gene attribute",
callback=self.updateInfoItems)
self.geneAttrComboBox.setEnabled(not self.useAttr)
self.geneAttrCheckbox = gui.checkBox(box,
self,
"useAttr",
"Use column names",
callback=self.updateInfoItems)
self.geneAttrCheckbox.toggled[bool].connect(
self.geneAttrComboBox.setDisabled)
gui.auto_commit(self.controlArea, self, "auto_commit", "Commit")
gui.rubber(self.controlArea)
gui.lineEdit(self.mainArea,
self,
"search_string",
"Filter",
callbackOnType=True,
callback=self.searchUpdate)
self.treeWidget = QTreeView(self.mainArea)
self.treeWidget.setAlternatingRowColors(True)
self.treeWidget.setSortingEnabled(True)
self.treeWidget.setSelectionMode(QTreeView.ExtendedSelection)
self.treeWidget.setUniformRowHeights(True)
self.treeWidget.setRootIsDecorated(False)
self.treeWidget.setItemDelegateForColumn(
HEADER_SCHEMA['NCBI ID'],
gui.LinkStyledItemDelegate(self.treeWidget))
self.treeWidget.setItemDelegateForColumn(
HEADER_SCHEMA['Ensembl ID'],
gui.LinkStyledItemDelegate(self.treeWidget))
self.treeWidget.viewport().setMouseTracking(True)
self.mainArea.layout().addWidget(self.treeWidget)
box = gui.widgetBox(self.mainArea, "", orientation="horizontal")
gui.button(box, self, "Select Filtered", callback=self.selectFiltered)
gui.button(box,
self,
"Clear Selection",
callback=self.treeWidget.clearSelection)
self.geneinfo = []
self.cells = []
self.row2geneinfo = {}
self.data = None
# : (# input genes, # matches genes)
self.matchedInfo = 0, 0
self.setBlocking(True)
self.executor = ThreadExecutor(self)
self.progressBarInit()
task = Task(
function=partial(taxonomy.ensure_downloaded,
callback=methodinvoke(self, "advance", ())))
task.resultReady.connect(self.initialize)
task.exceptionReady.connect(self._onInitializeError)
self.initfuture = self.executor.submit(task)
def sizeHint(self):
return QSize(1024, 720)
@Slot()
def advance(self):
assert self.thread() is QThread.currentThread()
self.progressBarSet(self.progressBarValue + 1, processEvents=None)
def _get_available_organisms(self):
available_organism = sorted([(tax_id, taxonomy.name(tax_id))
for tax_id in taxonomy.common_taxids()],
key=lambda x: x[1])
self.organisms = [tax_id[0] for tax_id in available_organism]
self.organismComboBox.addItems(
[tax_id[1] for tax_id in available_organism])
def initialize(self):
if self.__initialized:
# Already initialized
return
self.__initialized = True
self._get_available_organisms()
self.organism_index = self.organisms.index(taxonomy.DEFAULT_ORGANISM)
self.taxid = self.organisms[self.organism_index]
self.infoLabel.setText("No data on input\n")
self.initfuture = None
self.setBlocking(False)
self.progressBarFinished(processEvents=None)
def _onInitializeError(self, exc):
sys.excepthook(type(exc), exc, None)
self.error(0, "Could not download the necessary files.")
def _onSelectedOrganismChanged(self):
assert 0 <= self.organism_index <= len(self.organisms)
self.taxid = self.organisms[self.organism_index]
if self.data is not None:
self.updateInfoItems()
@Inputs.data
def setData(self, data=None):
if not self.__initialized:
self.initfuture.result()
self.initialize()
if self.itemsfuture is not None:
raise Exception("Already processing")
self.data = data
if data is not None:
self.geneAttrComboBox.clear()
self.attributes = [
attr for attr in data.domain.variables + data.domain.metas
if isinstance(attr, (Orange.data.StringVariable,
Orange.data.DiscreteVariable))
]
for var in self.attributes:
self.geneAttrComboBox.addItem(*gui.attributeItem(var))
self.taxid = str(self.data.attributes.get(TAX_ID, ''))
self.useAttr = self.data.attributes.get(GENE_AS_ATTRIBUTE_NAME,
self.useAttr)
self.gene_attr = min(self.gene_attr, len(self.attributes) - 1)
if self.taxid in self.organisms:
self.organism_index = self.organisms.index(self.taxid)
self.updateInfoItems()
else:
self.clear()
def updateInfoItems(self):
self.warning(0)
if self.data is None:
return
if self.useAttr:
genes = [attr.name for attr in self.data.domain.attributes]
elif self.attributes:
attr = self.attributes[self.gene_attr]
genes = [
str(ex[attr]) for ex in self.data if not math.isnan(ex[attr])
]
else:
genes = []
if not genes:
self.warning(0, "Could not extract genes from input dataset.")
self.warning(1)
org = self.organisms[min(self.organism_index, len(self.organisms) - 1)]
source_name, info_getter = ("NCBI Info", ncbi_info)
self.error(0)
self.progressBarInit()
self.setBlocking(True)
self.setEnabled(False)
self.infoLabel.setText("Retrieving info records.\n")
self.genes = genes
task = Task(function=partial(
info_getter, org, genes, advance=methodinvoke(self, "advance", (
))))
self.itemsfuture = self.executor.submit(task)
task.finished.connect(self._onItemsCompleted)
def _onItemsCompleted(self):
self.setBlocking(False)
self.progressBarFinished()
self.setEnabled(True)
try:
self.map_input_to_ensembl, geneinfo = self.itemsfuture.result()
finally:
self.itemsfuture = None
self.geneinfo = geneinfo
self.cells = cells = []
self.row2geneinfo = {}
for i, (input_name, gi) in enumerate(geneinfo):
if gi:
row = []
for item in gi:
row.append(item)
# parse synonyms
row[HEADER_SCHEMA['Synonyms']] = ','.join(
row[HEADER_SCHEMA['Synonyms']])
cells.append(row)
self.row2geneinfo[len(cells) - 1] = i
model = TreeModel(cells, list(HEADER_SCHEMA.keys()), None)
proxyModel = QSortFilterProxyModel(self)
proxyModel.setSourceModel(model)
self.treeWidget.setModel(proxyModel)
self.treeWidget.selectionModel().selectionChanged.connect(self.commit)
for i in range(len(HEADER_SCHEMA)):
self.treeWidget.resizeColumnToContents(i)
self.treeWidget.setColumnWidth(
i, min(self.treeWidget.columnWidth(i), 200))
self.infoLabel.setText("%i genes\n%i matched NCBI's IDs" %
(len(self.genes), len(cells)))
self.matchedInfo = len(self.genes), len(cells)
if self.useAttr:
new_data = self.data.from_table(self.data.domain, self.data)
for gene_var in new_data.domain.attributes:
gene_var.attributes['Ensembl ID'] = str(
self.map_input_to_ensembl[gene_var.name])
self.Outputs.data.send(new_data)
elif self.attributes:
ensembl_ids = []
for gene_name in self.data.get_column_view(
self.attributes[self.gene_attr])[0]:
if gene_name and gene_name in self.map_input_to_ensembl:
ensembl_ids.append(self.map_input_to_ensembl[gene_name])
else:
ensembl_ids.append('')
data_with_ensembl = append_columns(
self.data,
metas=[(Orange.data.StringVariable('Ensembl ID'), ensembl_ids)
])
self.Outputs.data.send(data_with_ensembl)
def clear(self):
self.infoLabel.setText("No data on input\n")
self.treeWidget.setModel(
TreeModel([], [
"NCBI ID", "Symbol", "Locus Tag", "Chromosome", "Description",
"Synonyms", "Nomenclature"
], self.treeWidget))
self.geneAttrComboBox.clear()
self.Outputs.selected_genes.send(None)
def commit(self):
if self.data is None:
self.Outputs.selected_genes.send(None)
self.Outputs.data.send(None)
return
model = self.treeWidget.model()
selection = self.treeWidget.selectionModel().selection()
selection = model.mapSelectionToSource(selection)
selectedRows = list(
chain(*(range(r.top(),
r.bottom() + 1) for r in selection)))
model = model.sourceModel()
selectedGeneids = [self.row2geneinfo[row] for row in selectedRows]
selectedIds = [self.geneinfo[i][0] for i in selectedGeneids]
selectedIds = set(selectedIds)
gene2row = dict((self.geneinfo[self.row2geneinfo[row]][0], row)
for row in selectedRows)
isselected = selectedIds.__contains__
if selectedIds:
if self.useAttr:
attrs = [
attr for attr in self.data.domain.attributes
if isselected(attr.name)
]
domain = Orange.data.Domain(attrs, self.data.domain.class_vars,
self.data.domain.metas)
newdata = self.data.from_table(domain, self.data)
self.Outputs.selected_genes.send(newdata)
elif self.attributes:
attr = self.attributes[self.gene_attr]
gene_col = [
attr.str_val(v) for v in self.data.get_column_view(attr)[0]
]
gene_col = [(i, name) for i, name in enumerate(gene_col)
if isselected(name)]
indices = [i for i, _ in gene_col]
# SELECTED GENES OUTPUT
selected_genes_metas = [
Orange.data.StringVariable(name)
for name in gene.GENE_INFO_HEADER_LABELS
]
selected_genes_domain = Orange.data.Domain(
self.data.domain.attributes, self.data.domain.class_vars,
self.data.domain.metas + tuple(selected_genes_metas))
selected_genes_data = self.data.from_table(
selected_genes_domain, self.data)[indices]
model_rows = [gene2row[gene_name] for _, gene_name in gene_col]
for col, meta in zip(range(model.columnCount()),
selected_genes_metas):
col_data = [
str(model.index(row, col).data(Qt.DisplayRole))
for row in model_rows
]
col_data = np.array(col_data, dtype=object, ndmin=2).T
selected_genes_data[:, meta] = col_data
if not len(selected_genes_data):
selected_genes_data = None
self.Outputs.selected_genes.send(selected_genes_data)
else:
self.Outputs.selected_genes.send(None)
def rowFiltered(self, row):
searchStrings = self.search_string.lower().split()
row = " ".join(self.cells[row]).lower()
return not all([s in row for s in searchStrings])
def searchUpdate(self):
if not self.data:
return
searchStrings = self.search_string.lower().split()
index = self.treeWidget.model().sourceModel().index
mapFromSource = self.treeWidget.model().mapFromSource
for i, row in enumerate(self.cells):
row = " ".join(row).lower()
self.treeWidget.setRowHidden(
mapFromSource(index(i, 0)).row(), QModelIndex(),
not all([s in row for s in searchStrings]))
def selectFiltered(self):
if not self.data:
return
itemSelection = QItemSelection()
index = self.treeWidget.model().sourceModel().index
mapFromSource = self.treeWidget.model().mapFromSource
for i, row in enumerate(self.cells):
if not self.rowFiltered(i):
itemSelection.select(mapFromSource(index(i, 0)),
mapFromSource(index(i, 0)))
self.treeWidget.selectionModel().select(
itemSelection,
QItemSelectionModel.Select | QItemSelectionModel.Rows)
def onAltSourceChange(self):
self.updateInfoItems()
def onDeleteWidget(self):
# try to cancel pending tasks
if self.initfuture:
self.initfuture.cancel()
if self.itemsfuture:
self.itemsfuture.cancel()
self.executor.shutdown(wait=False)
super().onDeleteWidget()
class OWGeneInfo(widget.OWWidget):
name = "Gene Info"
description = "Displays gene information from NCBI and other sources."
icon = "../widgets/icons/GeneInfo.svg"
priority = 2010
inputs = [("Data", Orange.data.Table, "setData")]
outputs = [("Data Subset", Orange.data.Table)]
settingsHandler = settings.DomainContextHandler()
organism_index = settings.ContextSetting(0)
taxid = settings.ContextSetting("9606")
gene_attr = settings.ContextSetting(0)
auto_commit = settings.Setting(False)
search_string = settings.Setting("")
useAttr = settings.ContextSetting(False)
useAltSource = settings.ContextSetting(False)
def __init__(self, parent=None, ):
super().__init__(self, parent)
self.selectionChangedFlag = False
self.__initialized = False
self.initfuture = None
self.itemsfuture = None
self.infoLabel = gui.widgetLabel(
gui.widgetBox(self.controlArea, "Info", addSpace=True),
"Initializing\n"
)
self.organisms = None
self.organismBox = gui.widgetBox(
self.controlArea, "Organism", addSpace=True)
self.organismComboBox = gui.comboBox(
self.organismBox, self, "organism_index",
callback=self._onSelectedOrganismChanged)
# For now only support one alt source, with a checkbox
# In the future this can be extended to multiple selections
self.altSourceCheck = gui.checkBox(
self.organismBox, self, "useAltSource",
"Show information from dictyBase",
callback=self.onAltSourceChange)
self.altSourceCheck.hide()
box = gui.widgetBox(self.controlArea, "Gene names", addSpace=True)
self.geneAttrComboBox = gui.comboBox(
box, self, "gene_attr",
"Gene attribute", callback=self.updateInfoItems
)
self.geneAttrComboBox.setEnabled(not self.useAttr)
cb = gui.checkBox(box, self, "useAttr", "Use attribute names",
callback=self.updateInfoItems)
cb.toggled[bool].connect(self.geneAttrComboBox.setDisabled)
gui.auto_commit(self.controlArea, self, "auto_commit", "Commit")
# A label for dictyExpress link (Why oh god why???)
self.dictyExpressBox = gui.widgetBox(
self.controlArea, "Dicty Express")
self.linkLabel = gui.widgetLabel(self.dictyExpressBox, "")
self.linkLabel.setOpenExternalLinks(False)
self.linkLabel.linkActivated.connect(self.onDictyExpressLink)
self.dictyExpressBox.hide()
gui.rubber(self.controlArea)
gui.lineEdit(self.mainArea, self, "search_string", "Filter",
callbackOnType=True, callback=self.searchUpdate)
self.treeWidget = QTreeView(
self.mainArea,
selectionMode=QTreeView.ExtendedSelection,
rootIsDecorated=False,
uniformRowHeights=True,
sortingEnabled=True)
self.treeWidget.setItemDelegate(
gui.LinkStyledItemDelegate(self.treeWidget))
self.treeWidget.viewport().setMouseTracking(True)
self.mainArea.layout().addWidget(self.treeWidget)
box = gui.widgetBox(self.mainArea, "", orientation="horizontal")
gui.button(box, self, "Select Filtered", callback=self.selectFiltered)
gui.button(box, self, "Clear Selection",
callback=self.treeWidget.clearSelection)
self.geneinfo = []
self.cells = []
self.row2geneinfo = {}
self.data = None
# : (# input genes, # matches genes)
self.matchedInfo = 0, 0
self.setBlocking(True)
self.executor = ThreadExecutor(self)
self.progressBarInit()
task = Task(
function=partial(
taxonomy.ensure_downloaded,
callback=methodinvoke(self, "advance", ())
)
)
task.resultReady.connect(self.initialize)
task.exceptionReady.connect(self._onInitializeError)
self.initfuture = self.executor.submit(task)
def sizeHint(self):
return QSize(1024, 720)
@Slot()
def advance(self):
assert self.thread() is QThread.currentThread()
self.progressBarSet(self.progressBarValue + 1,
processEvents=None)
def initialize(self):
if self.__initialized:
# Already initialized
return
self.__initialized = True
self.organisms = sorted(
set([name.split(".")[-2] for name in
serverfiles.listfiles("NCBI_geneinfo")] +
gene.NCBIGeneInfo.common_taxids())
)
self.organismComboBox.addItems(
[taxonomy.name(tax_id) for tax_id in self.organisms]
)
if self.taxid in self.organisms:
self.organism_index = self.organisms.index(self.taxid)
else:
self.organism_index = 0
self.taxid = self.organisms[self.organism_index]
self.altSourceCheck.setVisible(self.taxid == DICTY_TAXID)
self.dictyExpressBox.setVisible(self.taxid == DICTY_TAXID)
self.infoLabel.setText("No data on input\n")
self.initfuture = None
self.setBlocking(False)
self.progressBarFinished(processEvents=None)
def _onInitializeError(self, exc):
sys.excepthook(type(exc), exc, None)
self.error(0, "Could not download the necessary files.")
def _onSelectedOrganismChanged(self):
assert 0 <= self.organism_index <= len(self.organisms)
self.taxid = self.organisms[self.organism_index]
self.altSourceCheck.setVisible(self.taxid == DICTY_TAXID)
self.dictyExpressBox.setVisible(self.taxid == DICTY_TAXID)
if self.data is not None:
self.updateInfoItems()
def setData(self, data=None):
if not self.__initialized:
self.initfuture.result()
self.initialize()
if self.itemsfuture is not None:
raise Exception("Already processing")
self.closeContext()
self.data = data
if data is not None:
self.geneAttrComboBox.clear()
self.attributes = \
[attr for attr in data.domain.variables + data.domain.metas
if isinstance(attr, (Orange.data.StringVariable,
Orange.data.DiscreteVariable))]
for var in self.attributes:
self.geneAttrComboBox.addItem(*gui.attributeItem(var))
self.taxid = data_hints.get_hint(self.data, "taxid", self.taxid)
self.useAttr = data_hints.get_hint(
self.data, "genesinrows", self.useAttr)
self.openContext(data)
self.gene_attr = min(self.gene_attr, len(self.attributes) - 1)
if self.taxid in self.organisms:
self.organism_index = self.organisms.index(self.taxid)
else:
self.organism_index = 0
self.taxid = self.organisms[self.organism_index]
self.updateInfoItems()
else:
self.clear()
def infoSource(self):
""" Return the current selected info source getter function from
INFO_SOURCES
"""
org = self.organisms[min(self.organism_index, len(self.organisms) - 1)]
if org not in INFO_SOURCES:
org = "default"
sources = INFO_SOURCES[org]
name, func = sources[min(self.useAltSource, len(sources) - 1)]
return name, func
def inputGenes(self):
if self.useAttr:
genes = [attr.name for attr in self.data.domain.attributes]
elif self.attributes:
attr = self.attributes[self.gene_attr]
genes = [str(ex[attr]) for ex in self.data
if not math.isnan(ex[attr])]
else:
genes = []
return genes
def updateInfoItems(self):
self.warning(0)
if self.data is None:
return
genes = self.inputGenes()
if self.useAttr:
genes = [attr.name for attr in self.data.domain.attributes]
elif self.attributes:
attr = self.attributes[self.gene_attr]
genes = [str(ex[attr]) for ex in self.data
if not math.isnan(ex[attr])]
else:
genes = []
if not genes:
self.warning(0, "Could not extract genes from input dataset.")
self.warning(1)
org = self.organisms[min(self.organism_index, len(self.organisms) - 1)]
source_name, info_getter = self.infoSource()
self.error(0)
self.updateDictyExpressLink(genes, show=org == DICTY_TAXID)
self.altSourceCheck.setVisible(org == DICTY_TAXID)
self.progressBarInit()
self.setBlocking(True)
self.setEnabled(False)
self.infoLabel.setText("Retrieving info records.\n")
self.genes = genes
task = Task(
function=partial(
info_getter, org, genes,
advance=methodinvoke(self, "advance", ()))
)
self.itemsfuture = self.executor.submit(task)
task.finished.connect(self._onItemsCompleted)
def _onItemsCompleted(self):
self.setBlocking(False)
self.progressBarFinished()
self.setEnabled(True)
try:
schema, geneinfo = self.itemsfuture.result()
finally:
self.itemsfuture = None
self.geneinfo = geneinfo = list(zip(self.genes, geneinfo))
self.cells = cells = []
self.row2geneinfo = {}
links = []
for i, (_, gi) in enumerate(geneinfo):
if gi:
row = []
for _, item in zip(schema, gi):
if isinstance(item, Link):
# TODO: This should be handled by delegates
row.append(item.text)
links.append(item.link)
else:
row.append(item)
cells.append(row)
self.row2geneinfo[len(cells) - 1] = i
model = TreeModel(cells, [str(col) for col in schema], None)
model.setColumnLinks(0, links)
proxyModel = QSortFilterProxyModel(self)
proxyModel.setSourceModel(model)
self.treeWidget.setModel(proxyModel)
self.treeWidget.selectionModel().selectionChanged.connect(self.commit)
for i in range(7):
self.treeWidget.resizeColumnToContents(i)
self.treeWidget.setColumnWidth(
i, min(self.treeWidget.columnWidth(i), 200)
)
self.infoLabel.setText("%i genes\n%i matched NCBI's IDs" %
(len(self.genes), len(cells)))
self.matchedInfo = len(self.genes), len(cells)
def clear(self):
self.infoLabel.setText("No data on input\n")
self.treeWidget.setModel(
TreeModel([], ["NCBI ID", "Symbol", "Locus Tag",
"Chromosome", "Description", "Synonyms",
"Nomenclature"], self.treeWidget))
self.geneAttrComboBox.clear()
self.send("Data Subset", None)
def commit(self):
if self.data is None:
self.send("Data Subset", None)
return
model = self.treeWidget.model()
selection = self.treeWidget.selectionModel().selection()
selection = model.mapSelectionToSource(selection)
selectedRows = list(
chain(*(range(r.top(), r.bottom() + 1) for r in selection))
)
model = model.sourceModel()
selectedGeneids = [self.row2geneinfo[row] for row in selectedRows]
selectedIds = [self.geneinfo[i][0] for i in selectedGeneids]
selectedIds = set(selectedIds)
gene2row = dict((self.geneinfo[self.row2geneinfo[row]][0], row)
for row in selectedRows)
isselected = selectedIds.__contains__
if self.useAttr:
def is_selected(attr):
return attr.name in selectedIds
attrs = [attr for attr in self.data.domain.attributes
if isselected(attr.name)]
domain = Orange.data.Domain(
attrs, self.data.domain.class_vars, self.data.domain.metas)
newdata = self.data.from_table(domain, self.data)
self.send("Data Subset", newdata)
elif self.attributes:
attr = self.attributes[self.gene_attr]
gene_col = [attr.str_val(v)
for v in self.data.get_column_view(attr)[0]]
gene_col = [(i, name) for i, name in enumerate(gene_col)
if isselected(name)]
indices = [i for i, _ in gene_col]
# Add a gene info columns to the output
headers = [str(model.headerData(i, Qt.Horizontal, Qt.DisplayRole))
for i in range(model.columnCount())]
metas = [Orange.data.StringVariable(name) for name in headers]
domain = Orange.data.Domain(
self.data.domain.attributes, self.data.domain.class_vars,
self.data.domain.metas + tuple(metas))
newdata = self.data.from_table(domain, self.data)[indices]
model_rows = [gene2row[gene] for _, gene in gene_col]
for col, meta in zip(range(model.columnCount()), metas):
col_data = [str(model.index(row, col).data(Qt.DisplayRole))
for row in model_rows]
col_data = np.array(col_data, dtype=object, ndmin=2).T
newdata[:, meta] = col_data
if not len(newdata):
newdata = None
self.send("Data Subset", newdata)
else:
self.send("Data Subset", None)
def rowFiltered(self, row):
searchStrings = self.search_string.lower().split()
row = " ".join(self.cells[row]).lower()
return not all([s in row for s in searchStrings])
def searchUpdate(self):
if not self.data:
return
searchStrings = self.search_string.lower().split()
index = self.treeWidget.model().sourceModel().index
mapFromSource = self.treeWidget.model().mapFromSource
for i, row in enumerate(self.cells):
row = " ".join(row).lower()
self.treeWidget.setRowHidden(
mapFromSource(index(i, 0)).row(),
QModelIndex(),
not all([s in row for s in searchStrings]))
def selectFiltered(self):
if not self.data:
return
itemSelection = QItemSelection()
index = self.treeWidget.model().sourceModel().index
mapFromSource = self.treeWidget.model().mapFromSource
for i, row in enumerate(self.cells):
if not self.rowFiltered(i):
itemSelection.select(mapFromSource(index(i, 0)),
mapFromSource(index(i, 0)))
self.treeWidget.selectionModel().select(
itemSelection,
QItemSelectionModel.Select | QItemSelectionModel.Rows)
def updateDictyExpressLink(self, genes, show=False):
def fix(ddb):
if ddb.startswith("DDB"):
if not ddb.startswith("DDB_G"):
ddb = ddb.replace("DDB", "DDB_G")
return ddb
return None
if show:
genes = [fix(gene) for gene in genes if fix(gene)]
link1 = '<a href="http://dictyexpress.biolab.si/run/index.php?gene=%s">Microarray profile</a>'
link2 = '<a href="http://dictyexpress.biolab.si/run/index.php?gene=%s&db=rnaseq">RNA-Seq profile</a>'
self.linkLabel.setText(link1 + "<br/>" + link2)
show = any(genes)
if show:
self.dictyExpressBox.show()
else:
self.dictyExpressBox.hide()
def onDictyExpressLink(self, link):
if not self.data:
return
selectedIndexes = self.treeWidget.selectedIndexes()
if not len(selectedIndexes):
QMessageBox.information(
self, "No gene ids selected",
"Please select some genes and try again."
)
return
model = self.treeWidget.model()
mapToSource = model.mapToSource
selectedRows = self.treeWidget.selectedIndexes()
selectedRows = [mapToSource(index).row() for index in selectedRows]
model = model.sourceModel()
selectedGeneids = [self.row2geneinfo[row] for row in selectedRows]
selectedIds = [self.geneinfo[i][0] for i in selectedGeneids]
selectedIds = set(selectedIds)
def fix(ddb):
if ddb.startswith("DDB"):
if not ddb.startswith("DDB_G"):
ddb = ddb.replace("DDB", "DDB_G")
return ddb
return None
genes = [fix(gene) for gene in selectedIds if fix(gene)]
url = str(link) % " ".join(genes)
QDesktopServices.openUrl(QUrl(url))
def onAltSourceChange(self):
self.updateInfoItems()
def onDeleteWidget(self):
# try to cancel pending tasks
if self.initfuture:
self.initfuture.cancel()
if self.itemsfuture:
self.itemsfuture.cancel()
self.executor.shutdown(wait=False)
super().onDeleteWidget()
