def __init__(self,
region,
track,
track2,
rawStatistic=None,
normalizationType='zeroToOne',
minimal=False,
**kwArgs):
if minimal == True:
self._globalSource = MinimalBinSource(region.genome)
else:
from gold.application.StatRunner import StatJob
assert StatJob.USER_BIN_SOURCE is not None
self._globalSource = StatJob.USER_BIN_SOURCE
Statistic.__init__(self,
region,
track,
track2,
rawStatistic=rawStatistic,
normalizationType=normalizationType,
minimal=minimal,
**kwArgs)
if type(rawStatistic) is str:
from gold.statistic.AllStatistics import STAT_CLASS_DICT
rawStatistic = STAT_CLASS_DICT[rawStatistic]
self._rawStatistic = rawStatistic
self._normalizationType = normalizationType
def _determineStatClass(self, flushMemoized=True):
assert( hasattr(self, '_track') )
assert( hasattr(self, '_track2') )
dummyGESource = MinimalBinSource(self._genome)
if len(self._statClassList) == 0:
# if self._reversed:
logMessage('Stat class list is empty, for analysisDef: ' + self._analysisLine, level = logging.WARNING)
if DebugConfig.PASS_ON_VALIDSTAT_EXCEPTIONS:
raise ShouldNotOccurError('Stat class list is empty. Analysisdef: ' + self._analysisLine)
for statClass in self._statClassList:
if DebugConfig.VERBOSE:
logMessage('Checking validity of stat class "{}" for analysisDef "{}".'.format(statClass.__name__, self.getDefAfterChoices()))
trackA, trackB = self._track, self._track2
if trackA is None:
continue
try:
StatJob(dummyGESource, trackA, trackB, statClass, minimal=True,
**self.getAllChoices(filterByActivation=True)).run(False, flushMemoized=flushMemoized)
except IncompatibleTracksError, e:
if DebugConfig.VERBOSE:
logException(e, level=logging.DEBUG,
messagePrefix='Warning: error in _determineStatClass for stat: %s' % statClass.__name__)
if DebugConfig.PASS_ON_VALIDSTAT_EXCEPTIONS:
raise
except (AssertionError, IncompatibleAssumptionsError, IdenticalTrackNamesError), e:
if DebugConfig.VERBOSE:
logException(e, level=logging.DEBUG,
messagePrefix='Warning: error in _determineStatClass for stat: %s' % statClass.__name__)
if DebugConfig.PASS_ON_VALIDSTAT_EXCEPTIONS:
raise
def _init(self, globalSource='', minimal=False):
if isIter(self._region):
raise SplittableStatNotAvailableError()
if minimal:
self._globalSource = MinimalBinSource(self._region.genome)
elif globalSource == 'test':
self._globalSource = UserBinSource(
'TestGenome:chr21:10000000-15000000', '1000000')
else:
self._globalSource = GlobalBinSource(self._region.genome)
def __init__(self, region, track, track2, globalSource='', minimal=False, **kwArgs):
if isIter(region):
raise SplittableStatNotAvailableError()
if minimal == True:
self._globalSource = MinimalBinSource(region.genome)
elif globalSource == 'test':
self._globalSource = UserBinSource('TestGenome:chr21:10000000-15000000','1000000')
else:
self._globalSource = GlobalBinSource(region.genome)
super(self.__class__, self).__init__(region, track, track2, globalSource=globalSource, minimal=minimal, **kwArgs)
def __init__(self, genome, trackName, allowOverlaps):
assert allowOverlaps in [False, True]
self._genome = genome
self._trackName = trackName
self._fn = createDirPath(
trackName, genome,
allowOverlaps=allowOverlaps) + os.sep + BR_SHELVE_FILE_NAME
self._contents = {} #None
self._updatedChrs = set([])
from quick.application.UserBinSource import MinimalBinSource
minimalBinList = MinimalBinSource(genome)
self._minimalRegion = minimalBinList[
0] if minimalBinList is not None else None
def getGlobalSource(globalSourceStr, genome, minimal):
if minimal == True:
return MinimalBinSource(genome)
elif globalSourceStr == 'test':
return UserBinSource('TestGenome:chr21:10000000-15000000',
'1000000')
elif globalSourceStr == 'chrs':
return GenomeInfo.getChrRegs(genome)
elif globalSourceStr == 'chrarms':
return GenomeInfo.getChrArmRegs(genome)
elif globalSourceStr == 'ensembl':
return GenomeInfo.getStdGeneRegs(genome)
elif globalSourceStr == 'userbins':
from gold.application.StatRunner import StatJob
assert StatJob.USER_BIN_SOURCE is not None
return StatJob.USER_BIN_SOURCE
#return kwArgs['userBins']
else:
raise ShouldNotOccurError('globalSource not recognized')
def _determineStatClass(self):
assert( hasattr(self, '_track') )
assert( hasattr(self, '_track2') )
dummyGESource = MinimalBinSource(self._genome)
if len(self._statClassList)==0:
#logging.getLogger(HB_LOGGER).warning('Stat class list is empty, for analysisDef: ' + self._analysisLine)
if self._reversed:
logMessage('Stat class list is empty, for analysisDef: ' + self._analysisLine, level = logging.WARNING)
if DebugConfig.PASS_ON_VALIDSTAT_EXCEPTIONS:
raise ShouldNotOccurError('Stat class list is empty. Analysisdef: '+self._analysisLine)
for statClass in self._statClassList:
if DebugConfig.VERBOSE:
logMessage(statClass.__name__ + ': Trying (' + self.getDefAfterChoices() + ')')
# print statClass.__name__ + ': Trying (' + self.getDefAfterChoices() + ')'
#for reversed, trackA, trackB in [(False, self._track, self._track2), (True, self._track2, self._track) ]:
trackA, trackB = self._track, self._track2
if trackA is None:
continue
try:
# The hackiest of all hacks!
# TODO: reimplement together with TrackStructure
job = StatJob(dummyGESource, trackA, trackB, statClass, minimal=True,
**self.getChoices(filterByActivation=True))
stat = job._getSingleResult(dummyGESource[0])[-1]
tracks = stat._tracks if hasattr(stat, '_tracks') else [trackA, trackB]
trackUniqueKeys = [Track(tr.trackName).getUniqueKey(self._genome) for tr in tracks
if tr is not None]
StatJob(dummyGESource, trackA, trackB, statClass, minimal=True,
**self.getChoices(filterByActivation=True)).run(False)
#In order not to mess up integration tests
initSeed()
for trackIndex, restTrackIndexes in allElementsVersusRest(xrange(len(tracks))):
track = tracks[trackIndex]
if track is not None and track.formatConverters is None:
uniqueKeyForRestTracks = \
set(trackUniqueKeys[i] for i in restTrackIndexes)
# If several tracks are the same, memory memoization will only result
# in one RawDataStat being created, for one Track object. This is a
# wanted optimization. In other cases, something is probably wrong if
# a track has not been touched. However, this rule may be revisited
# when track structure functionality is implemented.
if trackUniqueKeys[trackIndex] not in uniqueKeyForRestTracks:
raise IncompatibleTracksError(
'Track ' + prettyPrintTrackName(track.trackName) +
' was created, but not touched by statistic')
except IncompatibleTracksError, e:
if DebugConfig.VERBOSE:
logException(e, level=logging.DEBUG,
message='(Warning: error in _determineStatClass for stat: %s)' % statClass.__name__)
if DebugConfig.PASS_ON_VALIDSTAT_EXCEPTIONS:
raise
except (AssertionError, IncompatibleAssumptionsError, IdenticalTrackNamesError), e:
if DebugConfig.VERBOSE:
logException(e, level=logging.DEBUG,
message='(Warning: error in _determineStatClass for stat: %s)' % statClass.__name__)
if DebugConfig.PASS_ON_VALIDSTAT_EXCEPTIONS:
raise
def isMinimalBin(region):
from quick.application.UserBinSource import MinimalBinSource
if not isIter(region) and region == MinimalBinSource(region.genome)[0]:
return True