def getTrackView(self, region):
#print 'get tv for reg: ',region
#print str(type(self._origRegion)) + " and " + str(type(region))
if Config.USE_SLOW_DEFENSIVE_ASSERTS:
assert (not isIter(self._origRegion) and self._origRegion == region) or \
(isIter(self._origRegion) and region in self._origRegion)
#if self._cachedTV is None:
self._origTrack.addFormatReq(self._trackFormatReq)
origTV = self._origTrack.getTrackView(region)
self._checkTrackFormat(origTV)
assert(not origTV.allowOverlaps)
assert(origTV.borderHandling == 'crop')
assert region == origTV.genomeAnchor
starts, ends, vals, strands, ids, edges, weights, extras = \
self._createRandomizedNumpyArrays(len(origTV.genomeAnchor), origTV.startsAsNumpyArray(), \
origTV.endsAsNumpyArray(), origTV.valsAsNumpyArray(), \
origTV.strandsAsNumpyArray(), origTV.idsAsNumpyArray(), \
origTV.edgesAsNumpyArray(), origTV.weightsAsNumpyArray(), \
origTV.allExtrasAsDictOfNumpyArrays(), origTV.trackFormat, region)
from gtrackcore_memmap.util.CommonFunctions import getClassName
self._cachedTV = TrackView(origTV.genomeAnchor, \
(starts + origTV.genomeAnchor.start if starts is not None else None), \
(ends + origTV.genomeAnchor.start if ends is not None else None), \
vals, strands, ids, edges, weights, origTV.borderHandling, origTV.allowOverlaps, extraLists=extras)
assert self._trackFormatReq.isCompatibleWith(self._cachedTV.trackFormat), 'Incompatible track-format: '\
+ str(self._trackFormatReq) + ' VS ' + str(self._cachedTV.trackFormat)
return self._cachedTV
def isCompBin(region):
if isIter(region):
return False
offsetOK = (CompBinManager.getOffset( region.start, CompBinManager.getBinNumber(region.start) ) == 0)
lengthOK = (len(region) == min(CompBinManager.getCompBinSize(), GenomeInfo.getChrLen(region.genome, region.chr) - region.start))
return offsetOK and lengthOK
def _createExtraLists(target, dtype, size):
if type(target) == bool:
if target == True:
return OrderedDict([('extra1', getRandExtraList(size)), ('extra2', getRandExtraList(size))])
else:
return OrderedDict()
elif isinstance(target, dict):
return OrderedDict([(name, array(content, dtype=dtype)) for name, content in target.items()])
elif isIter(target):
return OrderedDict([(name, getRandExtraList(size)) for name in target])
else:
raise ShouldNotOccurError
def _getNewContentsOfPrev(self, x, y, prefix):
if x is not None and y is not None:
if isinstance(x, numpy.ndarray):
x = x.tolist()
if isinstance(y, numpy.ndarray):
y = y.tolist()
if isIter(x):
if x != y:
if len(x) != len(y):
return []
else:
return [self._getMissingValue(prefix) for i,item in enumerate(x)]
else:
valDataType = self._geSource.getValDataType()
if prefix == 'val' and valDataType[0] == 'S' and valDataType != 'S1':
return self._getConcatContent(x, y, prefix, sorted=True)
elif prefix in ['id'] + self._prevEl.orderedExtraKeys:
return self._getConcatContent(x, y, prefix, sorted=False)
else:
if x != y:
return self._getMissingValue(prefix)
def __init__(self, segments=None, starts=True, ends=True, vals=True, strands=False, ids=False, edges=False, weights=False, \
extras=False, anchor=None, numElements=None, valDType='float64', borderHandling='crop', allowOverlaps=False):
if type(starts) != bool and ends == True:
ends = False
if type(ends) != bool and starts == True:
starts = False
assert not (starts==False and ends==False)
assert segments!=False and segments!=True
assert starts!=None and ends!=None and vals!=None and strands!=None
assert segments==None or (starts==True and ends==True)
assert not (isIter(weights) and not isIter(edges))
assert (any( type(x) not in [bool,type(None)] for x in [segments,starts,ends,vals,strands,ids,edges,weights,extras]) and numElements==None) \
or numElements!=None
#assert(( (type(segments)!=bool or type(starts)!=bool or type(ends)!=bool or \
# type(vals)!=bool or type(strands)!=bool) and numElements==None )\
# or numElements!=None)
#
if anchor==None:
anchor = [10,1000]
if segments != None:
starts = []
ends = []
for seg in segments:
starts.append(seg[0])
ends.append(seg[1])
if isIter(edges):
maxNumEdges = self._findMaxNumEls(edges)
edges = self._appendEmptyToEnd(edges, '', maxNumEdges)
if isIter(weights):
weights = self._appendEmptyToEnd(weights, numpy.nan, maxNumEdges)
[starts, ends, vals, strands, ids, edges, weights] + ([x for x in extras.values()] if isinstance(extras, dict) else [])
for list in [starts, ends, vals, strands, ids, edges, weights] + ([x for x in extras.values()] if isinstance(extras, dict) else []):
if type(list) != bool and numElements == None:
numElements = len(list)
assert(type(list) == bool or len(list) == numElements)
for coordList in [starts, ends]:
if type(coordList) != bool:
for j in range(len(coordList)):
coordList[j] += anchor[0]
randSegmentLists = getRandSegments(numElements, anchor[0], anchor[1])
starts = self._createList(starts, randSegmentLists[0], 'int32')
ends = self._createList(ends, randSegmentLists[1], 'int32')
vals = self._createList(vals, getRandValList(numElements, valDType), valDType)
strands = self._createList(strands, getRandStrandList(numElements), 'bool8')
randIds, randEdges, randWeights = getRandGraphLists(numElements)
ids = self._createList(ids, randIds, randIds.dtype)
edges = self._createList(edges, randEdges, randEdges.dtype)
weights = self._createList(weights, randWeights, 'float64')
if weights is not None and len(weights.shape) == 1:
weights = weights.reshape(weights.shape + (0,))
extras = self._createExtraLists(extras, 'S', numElements)
if starts == None:
if ends[0] != 0:
ends = numpy.append([anchor[0]], ends)
if vals != None:
vals = numpy.append([nan], vals)
if strands != None:
strands = numpy.append([True], strands)
if ends[-1] != anchor[1]:
ends[-1] = anchor[1]
# print (starts, ends, vals, strands, anchor)
TrackView.__init__(self, GenomeRegion('TestGenome', 'chr21', anchor[0], anchor[1]), starts, ends, vals, \
strands, ids, edges, weights, borderHandling, allowOverlaps, extraLists=extras)
