def checkUndirectedEdges(genome, trackName, allowOverlaps):
collector = TrackInfoDataCollector(genome, trackName)
if not (collector.getTrackFormat().isLinked() and collector.hasUndirectedEdges()):
return
complementEdgeWeightDict = {}
for chr in collector.getPreProcessedChrs(allowOverlaps):
trackSource = TrackSource()
trackData = trackSource.getTrackData(trackName, genome, chr, allowOverlaps)
ids = trackData['id']
edges = trackData['edges']
weights = trackData.get('weights')
for i, id in enumerate(ids):
edgesAttr = edges[i][edges[i] != '']
weightsAttr = weights[i][edges[i] != ''] if weights is not None else None
PreProcessUtils._adjustComplementaryEdgeWeightDict(complementEdgeWeightDict, id, edgesAttr, weightsAttr)
if len(complementEdgeWeightDict) != 0:
unmatchedPairs = []
for toId in complementEdgeWeightDict:
for fromId in complementEdgeWeightDict[toId]:
unmatchedPairs.append((fromId, toId, complementEdgeWeightDict[toId][fromId]))
raise InvalidFormatError("Error: All edges are not undirected. The following edges specifications " +\
"are not matched by an opposite edge with equal weight:" + os.linesep +\
os.linesep.join(["from '%s' to '%s'" % (fromId, toId) + \
(" with weight '%s'" % weight if weight != '' else '') \
for fromId, toId, weight in unmatchedPairs]))
def _getNumTrackElements(self, trackName, chr, allowOverlaps):
trackSource = TrackSource()
trackData = trackSource.getTrackData(trackName, self.GENOME, chr,
allowOverlaps)
brShelve = trackData.boundingRegionShelve
if brShelve:
chrBoundingRegions = brShelve.getAllBoundingRegionsForChr(chr)
numTrackElements = sum(
self._getTrackView(trackName, br,
allowOverlaps).getNumElements()
for br in chrBoundingRegions)
assert numTrackElements == brShelve.getTotalElementCountForChr(chr)
else: # chr-specific folders have not been merged
numTrackElements = self._getNumTrackElementsFromMemmap(
trackName, chr, allowOverlaps)
return numTrackElements
def checkIfEdgeIdsExist(genome, trackName, allowOverlaps):
collector = PreProcMetaDataCollector(genome, trackName)
if not collector.getTrackFormat().isLinked():
return
uniqueIds = numpy.array([], dtype='S')
uniqueEdgeIds = numpy.array([], dtype='S')
for chr in collector.getPreProcessedChrs(allowOverlaps):
trackSource = TrackSource()
trackData = trackSource.getTrackData(trackName, genome, chr, allowOverlaps)
uniqueIds = numpy.unique(numpy.concatenate((uniqueIds, trackData['id'][:])))
uniqueEdgeIds = numpy.unique(numpy.concatenate((uniqueEdgeIds, trackData['edges'][:].flatten())))
uniqueIds = uniqueIds[uniqueIds != '']
uniqueEdgeIds = uniqueEdgeIds[uniqueEdgeIds != '']
unmatchedIds = set(uniqueEdgeIds) - set(uniqueIds)
if len(unmatchedIds) > 0:
raise InvalidFormatError("Error: the following ids specified in the 'edges' column do not exist in the dataset: " + ', '.join(sorted(unmatchedIds)))
def checkIfEdgeIdsExist(genome, trackName, allowOverlaps):
collector = TrackInfoDataCollector(genome, trackName)
if not collector.getTrackFormat().isLinked():
return
uniqueIds = numpy.array([], dtype='S')
uniqueEdgeIds = numpy.array([], dtype='S')
for chr in collector.getPreProcessedChrs(allowOverlaps):
trackSource = TrackSource()
trackData = trackSource.getTrackData(trackName, genome, chr, allowOverlaps)
uniqueIds = numpy.unique(numpy.concatenate((uniqueIds, trackData['id'][:])))
uniqueEdgeIds = numpy.unique(numpy.concatenate((uniqueEdgeIds, trackData['edges'][:].flatten())))
uniqueIds = uniqueIds[uniqueIds != '']
uniqueEdgeIds = uniqueEdgeIds[uniqueEdgeIds != '']
unmatchedIds = set(uniqueEdgeIds) - set(uniqueIds)
if len(unmatchedIds) > 0:
raise InvalidFormatError("Error: the following ids specified in the 'edges' column do not exist in the dataset: " + ', '.join(sorted(unmatchedIds)))
class Track(object):
IS_MEMOIZABLE = True
def __new__(cls, trackName):
if trackName == [] or trackName is None:
return None
else:
if ExternalTrackManager.isVirtualTrack(trackName):
return VirtualMinimalTrack.__new__(VirtualMinimalTrack)
else:
return object.__new__(cls)
def __init__(self, trackName):
self.trackName = trackName
self._trackSource = TrackSource()
self._trackViewLoader = TrackViewLoader()
self._trackFormatReq = NeutralTrackFormatReq()
self.formatConverters = None
self._trackId = None
def _getRawTrackView(self, region, borderHandling, allowOverlaps):
trackData = self._trackSource.getTrackData(self.trackName, region.genome, region.chr, allowOverlaps)
return self._trackViewLoader.loadTrackView(trackData, region, borderHandling, allowOverlaps, self.trackName)
def getTrackView(self, region):
allowOverlaps = self._trackFormatReq.allowOverlaps()
borderHandling = self._trackFormatReq.borderHandling()
assert(allowOverlaps is not None)
assert(borderHandling is not None)
origTrackView = self._getRawTrackView(region, borderHandling, allowOverlaps)
if self.formatConverters is None:
self.formatConverters = getFormatConverters(origTrackView.trackFormat, self._trackFormatReq)
if self.formatConverters == []:
raise IncompatibleTracksError(prettyPrintTrackName(self.trackName) + ' with format: '\
+ str(origTrackView.trackFormat) +
('(' + origTrackView.trackFormat._val + ')' if origTrackView.trackFormat._val else '') + \
' does not satisfy ' + str(self._trackFormatReq))
if not self.formatConverters[0].canHandle(origTrackView.trackFormat, self._trackFormatReq):
raise IncompatibleTracksError(getClassName(self.formatConverters[0]) +\
' does not support conversion from ' + str(origTrackView.trackFormat) + \
' to ' + str(self._trackFormatReq))
return self.formatConverters[0].convert(origTrackView)
def addFormatReq(self, requestedTrackFormat):
prevFormatReq = self._trackFormatReq
self._trackFormatReq = TrackFormatReq.merge(self._trackFormatReq, requestedTrackFormat)
if self._trackFormatReq is None:
raise IncompatibleTracksError(str(prevFormatReq ) + \
' is incompatible with additional ' + str(requestedTrackFormat))
def setFormatConverter(self, converterClassName):
assert( self.formatConverters is None )
if converterClassName is not None:
self.formatConverters = [getFormatConverterByName(converterClassName)]
def getUniqueKey(self, genome):
assert self.formatConverters is not None and len(self.formatConverters) == 1, 'FC: '+str(self.formatConverters)
assert( not None in [self._trackFormatReq.allowOverlaps(), \
self._trackFormatReq.borderHandling()] )
if not self._trackId:
self._trackId = TrackInfo(genome, self.trackName).id
return hash((tuple(self.trackName), self._trackId, getClassName(self.formatConverters[0]), \
self.formatConverters[0].VERSION, self._trackFormatReq.allowOverlaps(), \
self._trackFormatReq.borderHandling()))
class Track(object):
IS_MEMOIZABLE = True
def __new__(cls, trackName, trackTitle=None):
if trackName == [] or trackName is None:
return None
else:
if ExternalTrackManager.isVirtualTrack(trackName):
return VirtualMinimalTrack.__new__(VirtualMinimalTrack)
else:
return object.__new__(cls)
def __init__(self, trackName, trackTitle=None):
self.trackName = trackName
self.trackTitle = trackTitle
self._trackSource = TrackSource()
self._trackViewLoader = TrackViewLoader()
self._trackFormatReq = NeutralTrackFormatReq()
self.formatConverters = None
self._trackId = None
self._randIndex = None
def _getRawTrackView(self, region, borderHandling, allowOverlaps):
trackData = self._trackSource.getTrackData(self.trackName, region.genome, region.chr, allowOverlaps)
return self._trackViewLoader.loadTrackView(trackData, region, borderHandling, allowOverlaps, self.trackName)
def getTrackView(self, region):
allowOverlaps = self._trackFormatReq.allowOverlaps()
borderHandling = self._trackFormatReq.borderHandling()
assert(allowOverlaps is not None)
assert(borderHandling is not None)
origTrackView = self._getRawTrackView(region, borderHandling, allowOverlaps)
if self.formatConverters is None:
self.formatConverters = getFormatConverters(origTrackView.trackFormat, self._trackFormatReq)
if self.formatConverters == []:
raise IncompatibleTracksError(prettyPrintTrackName(self.trackName) + ' with format: '\
+ str(origTrackView.trackFormat) +
('(' + origTrackView.trackFormat._val + ')' if origTrackView.trackFormat._val else '') + \
' does not satisfy ' + str(self._trackFormatReq))
if not self.formatConverters[0].canHandle(origTrackView.trackFormat, self._trackFormatReq):
raise IncompatibleTracksError(getClassName(self.formatConverters[0]) +\
' does not support conversion from ' + str(origTrackView.trackFormat) + \
' to ' + str(self._trackFormatReq))
return self.formatConverters[0].convert(origTrackView)
def addFormatReq(self, requestedTrackFormat):
prevFormatReq = self._trackFormatReq
self._trackFormatReq = TrackFormatReq.merge(self._trackFormatReq, requestedTrackFormat)
if self._trackFormatReq is None:
raise IncompatibleTracksError(str(prevFormatReq ) + \
' is incompatible with additional ' + str(requestedTrackFormat))
def setFormatConverter(self, converterClassName):
assert( self.formatConverters is None )
if converterClassName is not None:
self.formatConverters = [getFormatConverterByName(converterClassName)]
def getUniqueKey(self, genome):
if not self._trackId:
self._trackId = TrackInfo(genome, self.trackName).id
return hash((tuple(self.trackName),
self._trackId if self._trackId else '',
getClassName(self.formatConverters[0]) if self.formatConverters else '',
self.formatConverters[0].VERSION if self.formatConverters else '',
self._trackFormatReq.allowOverlaps() if self._trackFormatReq.allowOverlaps() else '',
self._trackFormatReq.borderHandling() if self._trackFormatReq.borderHandling() else ''))
def resetTrackSource(self):
self._trackSource = TrackSource()
def setRandIndex(self, randIndex):
pass #used only by TsBasedRandomTrack
class Track(object):
IS_MEMOIZABLE = True
def __new__(cls, trackName, trackTitle=None):
if trackName == [] or trackName is None:
return None
else:
if ExternalTrackManager.isVirtualTrack(trackName):
return VirtualMinimalTrack.__new__(VirtualMinimalTrack)
else:
return object.__new__(cls)
def __init__(self, trackName, trackTitle=None):
self.trackName = trackName
self.trackTitle = trackTitle
self._trackSource = TrackSource()
self._trackViewLoader = TrackViewLoader()
self._trackFormatReq = NeutralTrackFormatReq()
self.formatConverters = None
self._trackId = None
self._hasBeenRead = False
def _getRawTrackView(self, region, borderHandling, allowOverlaps):
trackData = self._trackSource.getTrackData(self.trackName, region.genome, region.chr, allowOverlaps)
return self._trackViewLoader.loadTrackView(trackData, region, borderHandling, allowOverlaps, self.trackName)
def getTrackView(self, region):
allowOverlaps = self._trackFormatReq.allowOverlaps()
borderHandling = self._trackFormatReq.borderHandling()
assert(allowOverlaps is not None)
assert(borderHandling is not None)
origTrackView = self._getRawTrackView(region, borderHandling, allowOverlaps)
if self.formatConverters is None:
self.formatConverters = getFormatConverters(origTrackView.trackFormat, self._trackFormatReq)
self._hasBeenRead = True
if self.formatConverters == []:
raise IncompatibleTracksError(prettyPrintTrackName(self.trackName) + ' with format: '\
+ str(origTrackView.trackFormat) +
('(' + origTrackView.trackFormat._val + ')' if origTrackView.trackFormat._val else '') + \
' does not satisfy ' + str(self._trackFormatReq))
if not self.formatConverters[0].canHandle(origTrackView.trackFormat, self._trackFormatReq):
raise IncompatibleTracksError(getClassName(self.formatConverters[0]) +\
' does not support conversion from ' + str(origTrackView.trackFormat) + \
' to ' + str(self._trackFormatReq))
return self.formatConverters[0].convert(origTrackView)
def addFormatReq(self, requestedTrackFormat):
prevFormatReq = self._trackFormatReq
self._trackFormatReq = TrackFormatReq.merge(self._trackFormatReq, requestedTrackFormat)
if self._trackFormatReq is None:
raise IncompatibleTracksError(str(prevFormatReq ) + \
' is incompatible with additional ' + str(requestedTrackFormat))
# TODO: track.formatConverters needs a complete overhaul. It is currently used:
# 1) to link tracks with possible choices for track conversion
# 2) to store the choice of format converter made by the user
# 3) to manage the default choice of format converters [=always the first item in the list]
# 4) to hold the currently selected format converter class until needed by getTrackView
def setFormatConverter(self, converterClassName):
assert( self.formatConverters is None )
if converterClassName is not None:
self.formatConverters = [getFormatConverterByName(converterClassName)]
def hasBeenFlaggedAsRead(self):
return self._hasBeenRead
def getUniqueKey(self, genome):
itemsToBeHashed = [tuple(self.trackName)]
itemsToBeHashed.append(self._getTrackId(genome))
if self._trackFormatReq is not None:
if self._trackFormatReq.allowOverlaps() is not None:
itemsToBeHashed.append(self._trackFormatReq.allowOverlaps())
itemsToBeHashed.append(self._trackFormatReq.borderHandling())
itemsToBeHashed.append(getClassName(self.formatConverters[0]) if self.formatConverters else '')
# TODO: Move away from fixed VERSION, as these in practice are never updated. Also for statistics.
itemsToBeHashed.append(self.formatConverters[0].VERSION if self.formatConverters else '')
from config.DebugConfig import DebugConfig
if DebugConfig.VERBOSE:
from gold.application.LogSetup import logMessage
logMessage('Unique key items for track "{}": '.format(self.trackName) +
', '.join(str(_) for _ in itemsToBeHashed))
return hash(tuple(itemsToBeHashed))
def _getTrackId(self, genome):
if not self._trackId:
trackInfo = TrackInfo(genome, self.trackName)
self._trackId = trackInfo.id
return self._trackId if self._trackId else ''
def resetTrackSource(self):
self._trackSource = TrackSource()