def checkUndirectedEdges(genome, trackName, allowOverlaps):
collector = PreProcMetaDataCollector(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 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 merge(genome, trackName, allowOverlaps):
path = createDirPath(trackName, genome, allowOverlaps=allowOverlaps)
collector = PreProcMetaDataCollector(genome, trackName)
chrList = collector.getPreProcessedChrs(allowOverlaps)
if not collector.getTrackFormat().reprIsDense():
chrList = sorted(chrList)
existingChrList = [
chr
for chr in ChrMemmapFolderMerger._existingChrIter(path, chrList)
]
if len(existingChrList) == 0:
raise EmptyGESourceError(
'No data lines has been read from source file (probably because it is empty).'
)
firstChrTrackData = TrackSource().getTrackData(trackName,
genome,
existingChrList[0],
allowOverlaps,
forceChrFolders=True)
arrayList = firstChrTrackData.keys()
for arrayName in arrayList:
mergedArray = firstChrTrackData[arrayName][:]
elementDim, dtypeDim = parseMemmapFileFn(
firstChrTrackData[arrayName].filename)[1:3]
del firstChrTrackData[arrayName]
for chr in existingChrList[1:]:
chrTrackData = TrackSource().getTrackData(trackName,
genome,
chr,
allowOverlaps,
forceChrFolders=True)
mergedArray = ChrMemmapFolderMerger.mergeArrays(
mergedArray, np.array(chrTrackData[arrayName][:]))
elementDimNew, dtypeDimNew = parseMemmapFileFn(
chrTrackData[arrayName].filename)[1:3]
elementDim = max(elementDim, elementDimNew)
dtypeDim = max(dtypeDim, dtypeDimNew)
del chrTrackData[arrayName]
mergedFn = createMemmapFileFn(path, arrayName, elementDim,
dtypeDim, str(mergedArray.dtype))
f = np.memmap(mergedFn,
dtype=mergedArray.dtype,
mode='w+',
shape=mergedArray.shape)
f[:] = mergedArray
f.flush()
del f
del mergedArray
def createBoundingRegionShelve(genome, trackName, allowOverlaps):
collector = PreProcMetaDataCollector(genome, trackName)
boundingRegionTuples = collector.getBoundingRegionTuples(allowOverlaps)
if not collector.getTrackFormat().reprIsDense():
boundingRegionTuples = sorted(boundingRegionTuples)
geChrList = collector.getPreProcessedChrs(allowOverlaps)
brShelve = BoundingRegionShelve(genome, trackName, allowOverlaps)
brShelve.storeBoundingRegions(boundingRegionTuples, geChrList, not collector.getTrackFormat().reprIsDense())
#Sanity check
if brShelve.getTotalElementCount() != collector.getNumElements(allowOverlaps):
raise ShouldNotOccurError("Error: The total element count for all bounding regions is not equal to the total number of genome elements. %s != %s" % \
(brShelve.getTotalElementCount(), collector.getNumElements(allowOverlaps)) )