def joinBnxFiles(varsP, bnxFiles):
"""After image processing, merge results into all.bnx.
"""
#the old way was to use this fn which simply copies lines
# while this is fine most of the time, RefAligner is more sophisticated,
# so it should be more robust to use RefAligner
#molecule.joinBnxFiles(bnxFiles, self.bnxFile)
#this used to be called writeIntermediate
varsP.writeListToFile(bnxFiles, varsP.bnxTarget)
# args, jobName, expectedResultFile, uniqueString
args = [varsP.RefAlignerBin, "-if", varsP.bnxTarget, "-merge", "-bnx", "-o", varsP.bnxFile.replace(".bnx",""), "-f"]
if varsP.stdoutlog :
args.extend( ['-stdout', '-stderr'] )
#print "joinBnxFiles: args:", args
jobwrapper = mthread.jobWrapper(varsP, "joinBnxFiles")
jobwrapper.addJob( mthread.singleJob(args, "joinBnxFiles", varsP.bnxFile, "joinBnxFiles") )
jobwrapper.multiThreadRunJobs(1)
jobwrapper.doAllPipeReport()
success = jobwrapper.allResultsFound()
if not success :
varsP.updatePipeReport("ERROR in performImageAnalysis: joinBnxFiles failed. Check: "+varsP.bnxTarget+"\n")
# this is just putting the path of bnxFile in bnxTarget
# on second thought, if I don't do this, then SampleCharModule will run on each bnx individually
#if success :
# varsP.writeListToFile([varsP.bnxFile], varsP.bnxTarget)
#sense of return of allResultsFound is opposite of return of performImageAnalysis:
# allResultsFound is True for all jobs success, False for any jobs fail
# performImage analysis return is 1 for failure
return not success
def generateJobList(self) :
"""splitBNX.generateJobList: submit varsP.nPairwiseJobs number of split bnx jobs. """
sorted_file = self.varsP.sorted_file
if not util.checkFile(sorted_file+".bnx") :
err = "ERROR: splitBNX input file (%s) not found; exiting" % self.varsP.sorted_file
self.varsP.updatePipeReport(err+"\n")
util.LogError("critical", err)
util.LogStatus("progress", "pipeline", "failure")
raise RuntimeError
N = calculateNPairwise(self.varsP, sorted_file) #move back here (not sortBNX) bc needs to use sorted bnx
#N = self.varsP.nPairwiseJobs
self.varsP.updatePipeReport('Splitting BNX\n')
#splitJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('splitting'))
super(splitBNX, self).__init__(self.varsP, self.stageName, clusterArgs=self.varsP.getClusterArgs('splitting'))
#should skip the rest and return 1, like in sortBNX, here:
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
self.varsP.updatePipeReport("Splitting"+(" scan-scaled" if self.varsP.doScanScale else "")+" bnx file: %s.bnx\n\n" % self.varsP.sorted_file)
#calculate threads per job: used to be fixed at 1, now file size / 1.5 GB rounded up. This was too low, add 1.
threads = max(1, int(math.ceil( os.path.getsize(sorted_file+".bnx")/1.5e9 ))) + 1
if threads > 1 :
self.varsP.updatePipeReport("Using %i threads per job\n" % threads)
#the change in job partitioning breaks backward compatibility and was causing too many problems; make it conditional on refaligner version
#this is now obsolete: assume binaries are up-to-date
if False : #self.varsP.refaligner_version < 3995 :
for partial in range(1,N + 1):
output_file=self.varsP.bnxFile.replace(".bnx", "_%s_of_%s" %(partial, self.varsP.nPairwiseJobs))
cargs=[self.varsP.RefAlignerBin, '-f', '-i', sorted_file+".bnx", "-maxthreads", str(threads), "-merge", "-subsetbin", str(partial), str(N), "-bnx", "-o", output_file]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
#print('%d/%d' % (partial, N), cargs)
expectedResultFile=output_file+".bnx"
self.addJob(mthread.singleJob(cargs, self.stageName + str(partial), expectedResultFile, self.stageName + str(partial), maxThreads=threads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=output_file+".stdout"))
else :
#change above to single command with -subsetbin 0 N
output_file=self.varsP.bnxFile.replace(".bnx", "")
cargs=[self.varsP.RefAlignerBin, '-f', '-i', sorted_file+".bnx", "-maxthreads", str(threads), "-merge", "-subsetbin", "0", str(N), "-bnx", "-o", output_file]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
self.addJob(mthread.singleJob(cargs, self.stageName, output_file+".bnx", self.stageName, maxThreads=threads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=output_file+".stdout"))
def getDetectJobs(self, contingentJob=None):
self.molFile = self.localTiff + '.mol'
#if the molFile is already there, the image processing is already done; do not repeat
if os.path.exists(self.molFile):
return []
#print "remoteTiff:", self.remoteTiff #debug
#print "localTiff:", self.localTiff, os.path.exists(self.localTiff) #debug
#print "expectedOverlap", self.curExp.ExpectedOverlap, "minOverlap", minOverlap #debug
#there was an issue with self.curExp.ExpectedOverlap being incorrectly computed due to a bad value in the xml (see manageTargetLocation.py)
#hopefully this will work
if self.curExp.ExpectedOverlap > 100 :
oldoverlap = self.curExp.ExpectedOverlap
self.curExp.ExpectedOverlap = 15
#print "Warning: calculated expectedOverlap", oldoverlap, "too large; defaulting to", self.curExp.ExpectedOverlap
self.varsP.updateInfoReport("Warning: "+self.nameStr()+": calculated expectedOverlap %i too large; defaulting to %i\n" % (oldoverlap, self.curExp.ExpectedOverlap), printalso=True)
expolap = (self.curExp.ExpectedOverlap - 10 if self.curExp.ExpectedOverlap >= 10 else 0) # ExpectedOverlap - 10
minOverlap = '%d' % (expolap)
maxOverlap = '%d' % (self.curExp.ExpectedOverlap + 10) # ExpectedOverlap + 10
dmOverlapArgs = ['-o', minOverlap, '-O', maxOverlap]
dmArgs = self.varsP.argsListed('imgDetection')
dmArgs = util.argumentReplaceList(dmArgs, ['-x', str(self.curExp.ScanColumnCount)])
dmArgs = util.argumentReplaceList(dmArgs, ['-y', str(self.curExp.ScanRowCount)])
dmArgs = util.argumentReplaceList(dmArgs, ['-p', str(self.curExp.Pitch)])
nchan = (self.curExp.nColors - 1 if self.curExp.nColors >= 2 else 1) #must be at least 1
colorArgs = ['-n', str(nchan)]
sJobCpName = 'cp ' + shorten(self.remoteTiff) + ' to ' + shorten(self.localTiff)
#print "cp\n"+self.remoteTiff, "\n"+self.localTiff #debug
sJobCp = mthread.singleJob(['cp', self.remoteTiff, self.localTiff], sJobCpName, self.localTiff, 'cpTiff', throttleClass = True)
if contingentJob:
sJobCp.addContingentJob(contingentJob)
sJobDMName = 'Detect ' + shorten(self.localTiff)
curArgs = [self.varsP.DMstaticBin] + dmOverlapArgs + dmArgs + colorArgs + [self.localTiff]
argumentString = " ".join(curArgs) + '\n'
print " ".join(curArgs) #debug
sJobDM = mthread.singleJob(curArgs, sJobDMName, self.molFile, 'Detect')
sJobDM.addContingentJob(sJobCp)
sJobDM.bpp = self.curExp.basesPerPixel
sJobDM.molTag = self.molTag
sJobDM.numLabelChannels = self.numLabelChannels
#inputMoleculesReport += ' ' + self.molTag + ' ' + self.remoteTiff + '\n'
dorm = True #default True (False for debug)
joblist = [sJobCp, sJobDM]
if dorm :
sJobRmImgName = 'Detect Complete, rm ' + shorten(self.localTiff)
sJobRmImg = mthread.singleJob(['rm', self.localTiff], sJobRmImgName, '', 'rmFile')
sJobRmImg.addContingentJob(sJobDM)
joblist.append( sJobRmImg )
return joblist
def generateJobListLinear(self):
"""Pairwise.generateJobListLinear: This method is the old way of doing pairwise
comparison of all molecules. It uses the -partial option to RefAligner. This
option is _incompatible_ with the various hashing options to RefAligner.
"""
baseArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('pairwise')
ct = 0
outputTarget = os.path.join(self.varsP.alignFolder, 'exp')
cArgs = [self.varsP.RefAlignerBin, '-i', self.varsP.bnxFile]
for jobNum in range(1,self.varsP.nPairwiseJobs + 1):
jobName = 'Pairwise %d of %d' % (jobNum, self.varsP.nPairwiseJobs)
outputString = 'pairwise%dof%d' % (jobNum,self.varsP.nPairwiseJobs)
expectedResultFile = outputTarget + outputString + '.align'
partialArgs = ['-partial', str(jobNum), str(self.varsP.nPairwiseJobs)]
currentArgs = cArgs + baseArgs + ['-o' , outputTarget + outputString]
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
if self.varsP.nPairwiseJobs > 1:
currentArgs += partialArgs
currentArgs += ['-maxthreads', str(self.varsP.maxthreads)]
if self.varsP.bnxStatsFile!=None:
currentArgs += ['-XmapStatRead', self.varsP.bnxStatsFile]
sJob = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
outputString,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=outputTarget + outputString+".stdout")
ct += 1
self.addJob(sJob)
self.logArguments()
def generateJobList(self):
"""Defines job parameters for merge. Updates variables for subsequent
completion test in mergeComplete()
"""
self.clearJobs()
self.prevPrefix = self.varsP.inputContigPrefix
#self.curPrefix = self.prefixIter.next()
self.curPrefix = self.stagePrefix + self.alphabet[self.iterCount]
self.groupName = self.stageName + self.alphabet[self.iterCount] #jobWrapper data member
utilities.LogStatus("progress", "stage_start", self.groupName)
self.varsP.updatePipeReport(' PREV PREFIX %s, CUR PREFIX %s' % (self.prevPrefix, self.curPrefix))
self.iterCount += 1
outputString = os.path.join(self.varsP.outputContigFolder, self.curPrefix)
currentArgs = [self.varsP.RefAlignerBin, '-o', outputString]
#if self.varsP.stdoutlog : #always use this here bc it's the only output which should always be there
currentArgs.extend( ['-f', '-stdout', '-stderr'] )
currentArgs += self.varsP.argsListed('merge')
currentArgs += ['-maxthreads', str(self.varsP.nThreads)]
contigsTextFile = os.path.join(self.varsP.inputContigFolder, 'mergeContigs.txt')
contigFiles, contigIDs = self.varsP.findContigs(self.varsP.inputContigFolder, self.prevPrefix, txtOutput=contigsTextFile) #this method creates the mergeContigs.txt file which is necessary for this job
self.varsP.prefixUsed.append(self.curPrefix)
fileArgs = ['-if', contigsTextFile]
#expoutput = outputString+".align" #don't know which contigs will disappear, but should always get an align file -- with new arg 'pairmergeRepeat', there's no .align; use stdout
expoutput = outputString+".stdout"
s1Job = mthread.singleJob(currentArgs + fileArgs, self.groupName, expoutput, self.groupName, maxThreads=self.varsP.nThreads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile = outputString + ".stdout")
self.addJob(s1Job)
self.logArguments()
def generateJobList(self):
baseArgs1 = self.varsP.argsListed(self.refineStage)
if self.refineStage != 'refineNGS' : #noise args are in refineNGS
baseArgs1 += self.varsP.argsListed('noise0')
contigFiles, contigIDs = self.varsP.findContigs(self.varsP.inputContigFolder, self.varsP.inputContigPrefix)
#nJobs = len(contigFiles)
bnx = self.varsP.sorted_file+".bnx" #was self.varsP.bnxFile, but need sorted bc ids are different after sorting
if self.refineStage == 'refineA' : #refineA uses assembler, all others use refaligner
r1args = [self.varsP.AssemblerBin, '-i', bnx] #need this before -contigs
r1args += ['-contigs', os.path.join(self.varsP.inputContigFolder, self.varsP.inputContigPrefix) + '.contigs']
else : #should be same for refineB/NGS/Final
r1args = [self.varsP.RefAlignerBin, '-i', bnx]
self.writeIDFile(len(contigFiles)) #nJobs)
output1String = os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix)
for contigID in contigIDs :
expectedOutputString = self.varsP.outputContigPrefix + '_contig' + contigID
expectedResultFile = os.path.join(self.varsP.outputContigFolder, expectedOutputString + '.cmap') #refineB
jobName = self.refineStage + ' %5s' % contigID
if self.refineStage == 'refineA' :
currentArgs = 2*[str(contigID)] #this must come after r1args because it's actually an argument to -contigs
else : #should be same for refineB/NGS/Final
r1_cmapFile = self.varsP.inputContigPrefix + '_contig' + str(contigID) + '.cmap'
r1_cmapFile = os.path.join(self.varsP.inputContigFolder, r1_cmapFile)
currentArgs = ['-maxthreads', str(self.varsP.maxthreads), '-ref', r1_cmapFile, '-id', contigID]
currentArgs = r1args + currentArgs + baseArgs1 + ['-o', output1String]
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
s1Job = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
expectedOutputString,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir)
self.addJob(s1Job)
self.logArguments()
def generateJobListLinear(self):
"""Pairwise.generateJobListLinear: This method is the old way of doing pairwise
comparison of all molecules. It uses the -partial option to RefAligner. This
option is _incompatible_ with the various hashing options to RefAligner.
"""
baseArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('pairwise')
ct = 0
outputTarget = os.path.join(self.varsP.alignFolder, 'exp')
cArgs = [self.varsP.RefAlignerBin, '-i', self.varsP.bnxFile]
for jobNum in range(1,self.varsP.nPairwiseJobs + 1):
jobName = 'Pairwise %d of %d' % (jobNum, self.varsP.nPairwiseJobs)
outputString = 'pairwise%dof%d' % (jobNum,self.varsP.nPairwiseJobs)
expectedResultFile = outputTarget + outputString + '.align'
partialArgs = ['-partial', str(jobNum), str(self.varsP.nPairwiseJobs)]
currentArgs = cArgs + baseArgs + ['-o' , outputTarget + outputString]
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
if self.varsP.nPairwiseJobs > 1:
currentArgs += partialArgs
currentArgs += ['-maxthreads', str(self.varsP.maxthreads)]
if self.varsP.bnxStatsFile!=None:
currentArgs += ['-XmapStatRead', self.varsP.bnxStatsFile]
sJob = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
outputString,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=outputTarget + outputString+".stdout")
ct += 1
self.addJob(sJob)
self.logArguments()
def generateJobList(self):
contigFiles, contigIDs = self.varsP.findContigs(self.varsP.inputContigFolder, self.varsP.inputContigPrefix)
curargs = [self.varsP.RefAlignerBin, '-i', self.varsP.sorted_file+".bnx"] #was bnxFile
baseArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('extension')
nJobs = contigFiles.__len__()
ct = 0
logArguments = "" #just in case the following loop isn't entered
for jobNum in range(1,nJobs + 1):
contigID = contigIDs[jobNum - 1]
#jobName = 'Extend ' + contigID + ', Job ' + str(jobNum) + ' of ' + str(nJobs)
expContigString = self.varsP.outputContigPrefix + '_contig' + contigID
outputString = os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix)
expectedResultFile = os.path.join(self.varsP.outputContigFolder, expContigString + '.cmap')# '_refined.cmap')
jobName = 'Ext %s' % expContigString# + ', Job ' + str(jobNum) + ' of ' + str(nJobs)
currentContig = contigFiles[jobNum - 1]
currentArgs = curargs + baseArgs
currentArgs += ['-maxthreads', str(self.varsP.maxthreads), '-o', outputString, '-id', contigID, '-ref', currentContig]
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
s1Job = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
expContigString,
maxThreads=self.varsP.maxthreads,
forceForward = currentContig,
clusterLogDir=self.varsP.clusterLogDir)
self.addJob(s1Job)
ct += 1
self.logArguments()
def generateJobListSubsample(self) :
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
sorted_file = self.varsP.sorted_file
nmols = 1000
seed = 1
self.subsampled=sorted_file+"_subsampled"
self.varsP.updatePipeReport('Subsampling %s\n' % (sorted_file))
jobName="SubsamplingBNX"
expectedResultFile=self.subsampled+".bnx"
# We use assembly section here because the memory usage is higher than pairwise, while the jobs are quite short.
#sortJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('assembly'))
super(autoNoise, self).__init__(self.varsP, jobName, clusterArgs=self.varsP.getClusterArgs("assembly"))
cargs=[self.varsP.RefAlignerBin, '-f', '-i', sorted_file+".bnx", "-maxthreads", str(self.varsP.maxthreads), "-merge", "-minlen", "200", "-randomize", str(seed), "-subset", "1", str(nmols), "-bnx", "-o", self.subsampled] + self.varsP.argsListed('bnx_sort')
if self.varsP.bnxStatsFile!=None:
cargs += ['-XmapStatRead', self.varsP.bnxStatsFile]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
self.addJob(mthread.singleJob(cargs, jobName, expectedResultFile, jobName, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=self.subsampled+".stdout"))
return 0 #success
def generateJobListChar(self, noise_in, input_file, optSection) :
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
self.varsP.updatePipeReport('%s\n' % (optSection))
self.output_folder=os.path.join(self.varsP.contigFolder, "auto_noise")
if not util.checkDir(self.output_folder) : #will make if not exist, only returns False if already exists or can't make
print "ERROR in autoNoise.varsPipeline.prepareContigIO: bad dir:", self.output_folder
self.output_file=os.path.join(self.output_folder, optSection)
expectedResultFile=self.output_file+".err"
# We use assembly section here because the memory usage is higher than pairwise, while the jobs are quite short.
#sortJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('assembly'))
super(autoNoise, self).__init__(self.varsP, self.stageName, clusterArgs=self.varsP.getClusterArgs("assembly"))
#cargs=[self.varsP.RefAlignerBin, '-f', '-i', input_file, "-ref", self.varsP.ref, "-maxthreads", str(self.varsP.maxthreads), "-o", self.output_file]
cargs=[self.varsP.RefAlignerBin, '-f', '-i', input_file, "-ref", self.varsP.ref, "-o", self.output_file] #remove maxthreads bc this is always running on its own
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
for v in noise_in.keys():
cargs.extend(["-"+v, str(noise_in[v])])
cargs.extend(self.varsP.argsListed(optSection))
if self.varsP.bnxStatsFile!=None:
cargs += ['-XmapStatWrite', self.varsP.bnxStatsFile]
self.addJob(mthread.singleJob(cargs, self.stageName, expectedResultFile, self.stageName, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=self.output_file+".stdout"))
return 0 #success
def generateJobListSubsample(self) :
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
sorted_file = self.varsP.sorted_file
nmols = 1000
seed = 1
self.subsampled=sorted_file+"_subsampled"
self.varsP.updatePipeReport('Subsampling %s\n' % (sorted_file))
jobName="SubsamplingBNX"
expectedResultFile=self.subsampled+".bnx"
# We use assembly section here because the memory usage is higher than pairwise, while the jobs are quite short.
#sortJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('assembly'))
super(autoNoise, self).__init__(self.varsP, jobName, clusterArgs=self.varsP.getClusterArgs("assembly"))
cargs=[self.varsP.RefAlignerBin, '-f', '-i', sorted_file+".bnx", "-maxthreads", str(self.varsP.maxthreads), "-merge", "-minlen", "200", "-randomize", str(seed), "-subset", "1", str(nmols), "-bnx", "-o", self.subsampled] + self.varsP.argsListed('bnx_sort')
if self.varsP.bnxStatsFile!=None:
cargs += ['-XmapStatRead', self.varsP.bnxStatsFile]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
self.addJob(mthread.singleJob(cargs, jobName, expectedResultFile, jobName, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=self.subsampled+".stdout"))
return 0 #success
def generateJobList(self):
""" Instantiate job wrapper class with queue of single jobs for assembly
"""
if self.varsP.pairwiseTriangleMode :
AssemblerInputFlag="-if"
AssemblerInputFile=self.varsP.bnxFileList
else:
AssemblerInputFlag="-i"
AssemblerInputFile=self.varsP.bnxFile
cargs = [self.varsP.AssemblerBin, AssemblerInputFlag, AssemblerInputFile, '-af', self.varsP.alignTarget]
if self.varsP.bnxStatsFile!=None:
cargs += ['-XmapStatRead', self.varsP.bnxStatsFile]
baseArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('assembly')
logFile = os.path.join(self.varsP.localRoot, 'AssemblyLog.txt')
errFile = os.path.join(self.varsP.localRoot, 'AssemblyLog_stderr.txt')
outFile = os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix) #no suffix for -o arg of Assembler
self.contigsFile = os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix+".contigs") #Assembler will append this suffix
currentArgs = cargs + baseArgs + ['-o', outFile]
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
logArguments = " ".join(currentArgs) + 2 * '\n'
jobName = 'Assembly'
#sJob = mthread.singleJob(currentArgs, jobName, self.contigsFile, jobName, stdOutFile=logFile, stdErrOutFile=errFile,clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=outFile+".stdout")
sJob = mthread.singleJob(currentArgs, jobName, self.contigsFile, jobName, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=outFile+".stdout")
self.addJob(sJob)
self.logArguments()
def generateJobListChar(self, noise_in, input_file, optSection) :
self.output_file=os.path.join(self.output_folder, optSection) #must assign before return bc used in constructor
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
self.varsP.updatePipeReport('%s\n' % (optSection))
#move to constructor
#self.output_folder=os.path.join(self.varsP.contigFolder, "auto_noise")
#if not util.checkDir(self.output_folder) : #will make if not exist, only returns False if already exists or can't make
# print "ERROR in autoNoise.varsPipeline.prepareContigIO: bad dir:", self.output_folder
expectedResultFile=self.output_file+".err"
# We use assembly section here because the memory usage is higher than pairwise, while the jobs are quite short.
#sortJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('assembly'))
#move to constructor
#super(autoNoise, self).__init__(self.varsP, self.stageName, clusterArgs=self.varsP.getClusterArgs("assembly"))
#cargs=[self.varsP.RefAlignerBin, '-f', '-i', input_file, "-ref", self.varsP.ref, "-maxthreads", str(self.varsP.maxthreads), "-o", self.output_file]
cargs=[self.varsP.RefAlignerBin, '-f', '-i', input_file, "-ref", self.varsP.ref, "-o", self.output_file] #remove maxthreads bc this is always running on its own
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
cargs.extend( ['-output-veto-filter', 'intervals.txt$'] )
for v in noise_in.keys():
cargs.extend(["-"+v, str(noise_in[v])])
cargs.extend(self.varsP.argsListed(optSection))
if self.varsP.bnxStatsFile!=None:
cargs += ['-XmapStatWrite', self.varsP.bnxStatsFile]
self.addJob(mthread.singleJob(cargs, self.stageName, expectedResultFile, self.stageName, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=self.output_file+".stdout"))
return 0 #success
def generateJobList(self,argset=-1):
if not self.varsP.ref : #no jobs if no ref
return
jobargs = [self.varsP.RefAlignerBin, '-ref', self.varsP.ref]
if argset == -1 and self.varsP.argData.has_key('characterizeDefault') : # don't use nominal default
opta = self.varsP.argsListed('characterizeDefault')
elif argset == 1 and self.varsP.argData.has_key('characterizeFinal') : #extend (on default) -- make this default
opta = self.varsP.argsListed('characterizeFinal')
else : #this is an error
self.varsP.updatePipeReport("ERROR in CharacterizeModule.generateJobList: invalid argset %s\n" % str(argset))
return
for i, curCharacterizeCmap in enumerate(self.varsP.curCharacterizeCmaps):
if self.varsP.numCharacterizeJobs == 1:
jobName = 'Char'+self.argStr+'_%s' % self.varsP.stageComplete
else:
jobName = 'Char'+self.argStr+'_%s_%d' % (self.varsP.stageComplete, i+1)
outFileName = os.path.split(curCharacterizeCmap)[-1].replace(".cmap", "")
outfile = os.path.join(self.varsP.contigAlignTarget,outFileName)
self.curCharacterizeFileRoots.append(outfile)
expectedResultFile = outfile+".xmap"
self.xmapTarget = expectedResultFile
currentArgs = jobargs + ["-i", curCharacterizeCmap, "-o", outfile]
stdoutf = None
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
stdoutf = outfile+".stdout"
currentArgs += ['-maxthreads', str(self.varsP.nThreads)]
currentArgs += ['-output-veto-filter', '_intervals.txt$']
currentArgs += opta
s1Job = mthread.singleJob(currentArgs, jobName, expectedResultFile, jobName.replace(' ',''),maxThreads=self.varsP.nThreads,clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=stdoutf)
self.addJob(s1Job)
if i==0:
self.logArguments()
def generateJobList(self) :
"""splitBNX.generateJobList: submit varsP.nPairwiseJobs number of split bnx jobs. """
sorted_file = self.varsP.sorted_file
if not util.checkFile(sorted_file+".bnx") :
err = "ERROR: splitBNX input file (%s) not found; exiting" % self.varsP.sorted_file
self.varsP.updatePipeReport(err+"\n")
util.LogError("critical", err)
util.LogStatus("progress", "pipeline", "failure")
raise RuntimeError
N = calculateNPairwise(self.varsP, sorted_file) #move back here (not sortBNX) bc needs to use sorted bnx
#N = self.varsP.nPairwiseJobs
self.varsP.updatePipeReport('Splitting BNX\n')
#splitJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('splitting'))
super(splitBNX, self).__init__(self.varsP, self.stageName, clusterArgs=self.varsP.getClusterArgs('splitting'))
#should skip the rest and return 1, like in sortBNX, here:
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
self.varsP.updatePipeReport("Splitting"+(" scan-scaled" if self.varsP.doScanScale else "")+" bnx file: %s.bnx\n\n" % self.varsP.sorted_file)
#calculate threads per job: used to be fixed at 1, now file size / 1.5 GB rounded up. This was too low, add 1.
threads = max(1, int(math.ceil( os.path.getsize(sorted_file+".bnx")/1.5e9 ))) + 1
if threads > 1 :
self.varsP.updatePipeReport("Using %i threads per job\n" % threads)
#the change in job partitioning breaks backward compatibility and was causing too many problems; make it conditional on refaligner version
if self.varsP.refaligner_version < 3995 :
for partial in range(1,N + 1):
output_file=self.varsP.bnxFile.replace(".bnx", "_%s_of_%s" %(partial, self.varsP.nPairwiseJobs))
cargs=[self.varsP.RefAlignerBin, '-f', '-i', sorted_file+".bnx", "-maxthreads", str(threads), "-merge", "-subsetbin", str(partial), str(N), "-bnx", "-o", output_file]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
#print('%d/%d' % (partial, N), cargs)
expectedResultFile=output_file+".bnx"
self.addJob(mthread.singleJob(cargs, self.stageName + str(partial), expectedResultFile, self.stageName + str(partial), maxThreads=threads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=output_file+".stdout"))
else :
#change above to single command with -subsetbin 0 N
output_file=self.varsP.bnxFile.replace(".bnx", "")
cargs=[self.varsP.RefAlignerBin, '-f', '-i', sorted_file+".bnx", "-maxthreads", str(threads), "-merge", "-subsetbin", "0", str(N), "-bnx", "-o", output_file]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
self.addJob(mthread.singleJob(cargs, self.stageName, output_file+".bnx", self.stageName, maxThreads=threads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=output_file+".stdout"))
def getLambdaMapJob(self, snrCutoff=0, verbose=False):
#Note, verbose will print once per job, so use for debugging only
# add lambda alignment band to this
lambdaFilter = self.varsP.argData['lambdaFilter']
lamMinLen = float(lambdaFilter[lambdaFilter.index('-minlen') +1]) if '-minlen' in lambdaFilter else 40.
lamMaxLen = float(lambdaFilter[lambdaFilter.index('-maxlen') +1]) if '-maxlen' in lambdaFilter else 60.
lamMinLab = float(lambdaFilter[lambdaFilter.index('-minsites')+1]) if '-minsites' in lambdaFilter else 6.
lamMaxLab = float(lambdaFilter[lambdaFilter.index('-maxsites')+1]) if '-maxsites' in lambdaFilter else 10.
#old format below (dict, not list)
#lamMinLen = int(lambdaFilter['-minlen' ]) # 40
#lamMaxLen = int(lambdaFilter['-maxlen' ]) # 60
#lamMinLab = int(lambdaFilter['-minsites']) # 6
#lamMaxLab = int(lambdaFilter['-maxsites']) # 10
if verbose :
self.varsP.updateInfoReport("lamMinLen = %.0f\n" % lamMinLen, printalso=True)
self.varsP.updateInfoReport("lamMaxLen = %.0f\n" % lamMaxLen, printalso=True)
self.varsP.updateInfoReport("lamMinLab = %.0f\n" % lamMinLab, printalso=True)
self.varsP.updateInfoReport("lamMaxLab = %.0f\n" % lamMaxLab, printalso=True)
#need mol file to do this; if doesn't exist, return with warning
if not(os.path.exists(self.molFile)):
print "Skipping map lambda job", self.molTag, "because mol file missing:", self.molFile
self.lambdaErrFile = None
return
bnxFileLambda = '%s_lambda.bnx' % self.molTag
bnxFileLambda = os.path.join(os.path.split(self.molFile)[0], bnxFileLambda)
#if lambda bnx exists, skip the isolation step
if os.path.exists(bnxFileLambda) :
print "Using lambda bnx", bnxFileLambda
else :
print ' Isolating Lambda %s' % self.molTag
lab2File = self.molFile.replace('.mol', '.0.lab')
scanDset = molecule.moleculeDataset(self.curExp.basesPerPixel, molTag=int(self.molTag))
scanDset.readMolFile(self.molFile)
scanDset.annotateLabels(lab2File)
# Introduce optArguments for Lambda Band
scanDsetLambda = molecule.filteredSubset(scanDset,snrCutoff,lamMinLab,lamMaxLab,lamMinLen,lamMaxLen,True)
scanDsetLambda.writeBnxFile(bnxFileLambda, quality=self.quality)
self.lambdaBnx = bnxFileLambda
baseArgs = self.varsP.argsListed('mapLambda')
outputTarget = bnxFileLambda.replace('.bnx', '')
curArgs = [self.varsP.RefAlignerBin, '-i', bnxFileLambda, '-o', outputTarget, '-ref', self.varsP.lambdaRef] + baseArgs
if self.varsP.stdoutlog :
curArgs.extend( ['-stdout', '-stderr'] )
jobTag = self.molTag + '_lambda'
self.lambdaErrFile = outputTarget + '.err'
#if the err file exists, no need to process
if os.path.exists(self.lambdaErrFile) :
print "Skipping map lambda job ", jobTag, "because err file exists", self.lambdaErrFile
return
return mthread.singleJob(curArgs, jobTag, self.lambdaErrFile, jobTag)
def generateJobList(self):
baseArgs1 = self.varsP.argsListed(self.refineStage)
if self.refineStage != 'refineNGS': #noise args are in refineNGS
baseArgs1 += self.varsP.argsListed('noise0')
contigFiles, contigIDs = self.varsP.findContigs(
self.varsP.inputContigFolder, self.varsP.inputContigPrefix)
#nJobs = len(contigFiles)
bnx = self.varsP.sorted_file + ".bnx" #was self.varsP.bnxFile, but need sorted bc ids are different after sorting
if self.refineStage == 'refineA': #refineA uses assembler, all others use refaligner
r1args = [self.varsP.AssemblerBin, '-i',
bnx] #need this before -contigs
r1args += [
'-contigs',
os.path.join(self.varsP.inputContigFolder,
self.varsP.inputContigPrefix) + '.contigs'
]
else: #should be same for refineB/NGS/Final
r1args = [self.varsP.RefAlignerBin, '-i', bnx]
self.writeIDFile(len(contigFiles)) #nJobs)
output1String = os.path.join(self.varsP.outputContigFolder,
self.varsP.outputContigPrefix)
for contigID in contigIDs:
expectedOutputString = self.varsP.outputContigPrefix + '_contig' + contigID
expectedResultFile = os.path.join(self.varsP.outputContigFolder,
expectedOutputString +
'.cmap') #refineB
jobName = self.refineStage + ' %5s' % contigID
if self.refineStage == 'refineA':
currentArgs = 2 * [
str(contigID)
] #this must come after r1args because it's actually an argument to -contigs
else: #should be same for refineB/NGS/Final
r1_cmapFile = self.varsP.inputContigPrefix + '_contig' + str(
contigID) + '.cmap'
r1_cmapFile = os.path.join(self.varsP.inputContigFolder,
r1_cmapFile)
currentArgs = [
'-maxthreads',
str(self.varsP.maxthreads), '-ref', r1_cmapFile, '-id',
contigID
]
currentArgs = r1args + currentArgs + baseArgs1 + [
'-o', output1String
]
if self.varsP.stdoutlog:
currentArgs.extend(['-stdout', '-stderr'])
s1Job = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
expectedOutputString,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir)
self.addJob(s1Job)
self.logArguments()
def getTargetJobs(self, dormdir=False):
localDataLocation = os.path.join(self.varsP.localRoot, self.expTag + '/')
#print "localDataLocation:", localDataLocation #debug
if dormdir :
sJobRmName = 'Pre-Remove Folder: ' + shorten(localDataLocation)
sJobRm = mthread.singleJob(['rm', '-f', '-r', localDataLocation], sJobRmName, '', 'rmDir')
sJobMkdirName = 'Make Folder: ' + shorten(localDataLocation)
sJobMkdir = mthread.singleJob(['mkdir', localDataLocation], sJobMkdirName, localDataLocation, 'mkDir')
sJobMkdir.addContingentJob(sJobRm)
allJobs = [sJobRm, sJobMkdir]
contingentjob = sJobMkdir
else :
util.checkDir(localDataLocation) #will make dir localDataLocation
allJobs = []
contingentjob = None
for scan in self.scans:
scanjobs = scan.getDetectJobs(contingentjob)
if not scanjobs : #no scan jobs means the scan has already been processed--clear all jobs
self.varsP.updatePipeReport("Device.getTargetJobs: skipping path "+scan.nameStr()+"\n") #localDataLocation
else :
allJobs += scanjobs
return allJobs
def __init__(self, varsP) :
jobName = "reference_process"
opta_section = "referenceSvdetect"
default_mres = "2.9"
mres = "-mres"
self.varsP = varsP
usedefault = False
if self.varsP.argData.has_key(opta_section) : #check if in optargs
opta = self.varsP.argsListed(opta_section)
if not mres in opta : #must have mres
self.varsP.updatePipeReport("Warning in referenceProcess: "+mres+" missing in optArguments section "+opta_section+"\n")
usedefault = True
else :
self.varsP.updatePipeReport("Warning in referenceProcess: optArguments section "+opta_section+" missing\n")
usedefault = True
if usedefault :
opta = [mres, default_mres]
mresstr = opta[opta.index(mres)+1] #get string for mres value for output name
mresstr = mresstr.replace(".","")
if not util.checkDir(self.varsP.refFolder) :
self.varsP.updatePipeReport( "ERROR in referenceProcess: could not make output dir %s\n" % self.varsP.refFolder )
return None
refpref = os.path.basename(self.varsP.ref[:self.varsP.ref.rfind(".")]) + "_res" + mresstr
outarg = os.path.join(self.varsP.refFolder, refpref) #refFolder is new output folder for this job
expectedResultFile = outarg+".cmap" #if ref is spots, is this spots?
args = [self.varsP.RefAlignerBin, '-o', outarg, '-i', self.varsP.ref, '-f', '-merge'] + opta
stdoutf = None
if self.varsP.stdoutlog :
args.extend( ['-stdout', '-stderr'] )
stdoutf = outarg+".stdout"
args += ['-maxthreads', str(self.varsP.nThreads)]
super(referenceProcess, self).__init__(self.varsP, jobName, clusterArgs=self.varsP.getClusterArgs("assembly"))
job = mthread.singleJob(args, jobName, expectedResultFile, jobName, maxThreads=self.varsP.nThreads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=stdoutf)
self.addJob(job)
util.LogStatus("progress", "stage_start", jobName)
self.varsP.runJobs(self, "referenceProcess")
self.doAllPipeReport()
if not self.allResultsFound() : #this is an error, but we'll continue processing without SV detect
err = "ERROR in referenceProcess: job failed, disabling SV detect"
self.varsP.updatePipeReport( err+"\n" )
util.LogError("error", err)
#self.varsP.runSV = False #no need since this class is used in SVModule
else :
self.varsP.refDeresed = expectedResultFile #store good result for SV detect
self.varsP.updatePipeReport( "referenceProcess: using reference %s for svdetect\n" % self.varsP.refDeresed )
util.LogStatus("progress", "stage_complete", jobName)
def generateJobList(self, argset=-1):
if not self.varsP.ref: #no jobs if no ref
return
jobargs = [self.varsP.RefAlignerBin, '-ref', self.varsP.ref]
if argset == -1 and self.varsP.argData.has_key(
'characterizeDefault'): # don't use nominal default
opta = self.varsP.argsListed('characterizeDefault')
elif argset == 1 and self.varsP.argData.has_key(
'characterizeFinal'
): #extend (on default) -- make this default
opta = self.varsP.argsListed('characterizeFinal')
else: #this is an error
self.varsP.updatePipeReport(
"ERROR in CharacterizeModule.generateJobList: invalid argset %s\n"
% str(argset))
return
for i, curCharacterizeCmap in enumerate(
self.varsP.curCharacterizeCmaps):
if self.varsP.numCharacterizeJobs == 1:
jobName = 'Char' + self.argStr + '_%s' % self.varsP.stageComplete
else:
jobName = 'Char' + self.argStr + '_%s_%d' % (
self.varsP.stageComplete, i + 1)
outFileName = os.path.split(curCharacterizeCmap)[-1].replace(
".cmap", "")
outfile = os.path.join(self.varsP.contigAlignTarget, outFileName)
self.curCharacterizeFileRoots.append(outfile)
expectedResultFile = outfile + ".xmap"
self.xmapTarget = expectedResultFile
currentArgs = jobargs + ["-i", curCharacterizeCmap, "-o", outfile]
stdoutf = None
if self.varsP.stdoutlog:
currentArgs.extend(['-stdout', '-stderr'])
stdoutf = outfile + ".stdout"
currentArgs += ['-maxthreads', str(self.varsP.nThreads)]
currentArgs += ['-output-veto-filter', '_intervals.txt$']
currentArgs += opta
s1Job = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
jobName.replace(' ', ''),
maxThreads=self.varsP.nThreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=stdoutf)
self.addJob(s1Job)
if i == 0:
self.logArguments()
def generateJobListTriangle(self):
baseArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('pairwise')
cArgs = [self.varsP.RefAlignerBin, '-i', self.varsP.bnxFile]
ct = 0
outputTarget = os.path.join(self.varsP.alignFolder, 'exp')
njobs=self.varsP.nPairwiseJobs*(self.varsP.nPairwiseJobs+1)/2
BNX_list=[]
for i in range(1,self.varsP.nPairwiseJobs + 1):
file1=self.varsP.bnxFile.replace(".bnx", "_%s_of_%s.bnx" %(i, self.varsP.nPairwiseJobs))
BNX_list.append(file1+"\n")
for j in range(i,self.varsP.nPairwiseJobs + 1):
file2=self.varsP.bnxFile.replace(".bnx", "_%s_of_%s.bnx" %(j, self.varsP.nPairwiseJobs))
jobName = 'Pairwise %d of %d' % (ct+1, njobs)
outputString = 'pairwise%dof%d' % (ct+1, njobs)
expectedResultFile = outputTarget + outputString + '.align'
if i==j :
currentArgs = [self.varsP.RefAlignerBin, '-i', file1] + ['-o' , outputTarget + outputString] + baseArgs
else :
currentArgs = [self.varsP.RefAlignerBin, "-first", "-1", "-i", file1, "-i", file2] + ['-o' , outputTarget + outputString] + baseArgs
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
#if self.varsP.nPairwiseJobs > 1:
#currentArgs += partialArgs
currentArgs += ['-maxthreads', str(self.varsP.maxthreads)]
if self.varsP.bnxStatsFile!=None:
currentArgs += ['-XmapStatRead', self.varsP.bnxStatsFile]
#if ct == 0: #redundant with logArguments below
# self.pipeReport += " ".join(currentArgs) + 2 * '\n'
sJob = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
outputString,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=outputTarget + outputString+".stdout",
)#, shell=True)
ct += 1
self.addJob(sJob)
self.varsP.bnxFileList=self.varsP.bnxFile.replace(".bnx", ".list")
f=open(self.varsP.bnxFileList, "w")
f.writelines(BNX_list)
f.close()
self.logArguments()
def generateJobListTriangle(self):
baseArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('pairwise')
cArgs = [self.varsP.RefAlignerBin, '-i', self.varsP.bnxFile]
ct = 0
outputTarget = os.path.join(self.varsP.alignFolder, 'exp')
njobs=self.varsP.nPairwiseJobs*(self.varsP.nPairwiseJobs+1)/2
BNX_list=[]
for i in range(1,self.varsP.nPairwiseJobs + 1):
file1=self.varsP.bnxFile.replace(".bnx", "_%s_of_%s.bnx" %(i, self.varsP.nPairwiseJobs))
BNX_list.append(file1+"\n")
for j in range(i,self.varsP.nPairwiseJobs + 1):
file2=self.varsP.bnxFile.replace(".bnx", "_%s_of_%s.bnx" %(j, self.varsP.nPairwiseJobs))
jobName = 'Pairwise %d of %d' % (ct+1, njobs)
outputString = 'pairwise%dof%d' % (ct+1, njobs)
expectedResultFile = outputTarget + outputString + '.align'
if i==j :
currentArgs = [self.varsP.RefAlignerBin, '-i', file1] + ['-o' , outputTarget + outputString] + baseArgs
else :
currentArgs = [self.varsP.RefAlignerBin, "-first", "-1", "-i", file1, "-i", file2] + ['-o' , outputTarget + outputString] + baseArgs
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
#if self.varsP.nPairwiseJobs > 1:
#currentArgs += partialArgs
currentArgs += ['-maxthreads', str(self.varsP.maxthreads)]
if self.varsP.bnxStatsFile!=None:
currentArgs += ['-XmapStatRead', self.varsP.bnxStatsFile]
#if ct == 0: #redundant with logArguments below
# self.pipeReport += " ".join(currentArgs) + 2 * '\n'
sJob = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
outputString,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=outputTarget + outputString+".stdout",
)#, shell=True)
ct += 1
self.addJob(sJob)
self.varsP.bnxFileList=self.varsP.bnxFile.replace(".bnx", ".list")
f=open(self.varsP.bnxFileList, "w")
f.writelines(BNX_list)
f.close()
self.logArguments()
def generateJobList(self):
"""Defines job parameters for merge. Updates variables for subsequent
completion test in mergeComplete()
"""
self.clearJobs()
self.prevPrefix = self.varsP.inputContigPrefix
#self.curPrefix = self.prefixIter.next()
self.curPrefix = self.stagePrefix + self.alphabet[self.iterCount]
self.groupName = self.stageName + self.alphabet[
self.iterCount] #jobWrapper data member
utilities.LogStatus("progress", "stage_start", self.groupName)
self.varsP.updatePipeReport(' PREV PREFIX %s, CUR PREFIX %s' %
(self.prevPrefix, self.curPrefix))
self.iterCount += 1
outputString = os.path.join(self.varsP.outputContigFolder,
self.curPrefix)
currentArgs = [self.varsP.RefAlignerBin, '-o', outputString]
#if self.varsP.stdoutlog : #always use this here bc it's the only output which should always be there
currentArgs.extend(['-f', '-stdout', '-stderr'])
currentArgs += self.varsP.argsListed('merge')
currentArgs += ['-maxthreads', str(self.varsP.nThreads)]
contigsTextFile = os.path.join(self.varsP.inputContigFolder,
'mergeContigs.txt')
contigFiles, contigIDs = self.varsP.findContigs(
self.varsP.inputContigFolder,
self.prevPrefix,
txtOutput=contigsTextFile
) #this method creates the mergeContigs.txt file which is necessary for this job
self.varsP.prefixUsed.append(self.curPrefix)
fileArgs = ['-if', contigsTextFile]
#expoutput = outputString+".align" #don't know which contigs will disappear, but should always get an align file -- with new arg 'pairmergeRepeat', there's no .align; use stdout
expoutput = outputString + ".stdout"
s1Job = mthread.singleJob(currentArgs + fileArgs,
self.groupName,
expoutput,
self.groupName,
maxThreads=self.varsP.nThreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=outputString + ".stdout")
self.addJob(s1Job)
self.logArguments()
def generateJobList(self):
contigFiles, contigIDs = self.varsP.findContigs(
self.varsP.inputContigFolder, self.varsP.inputContigPrefix)
curargs = [
self.varsP.RefAlignerBin, '-i', self.varsP.sorted_file + ".bnx"
] #was bnxFile
baseArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed(
'extension')
nJobs = contigFiles.__len__()
ct = 0
logArguments = "" #just in case the following loop isn't entered
for jobNum in range(1, nJobs + 1):
contigID = contigIDs[jobNum - 1]
#jobName = 'Extend ' + contigID + ', Job ' + str(jobNum) + ' of ' + str(nJobs)
expContigString = self.varsP.outputContigPrefix + '_contig' + contigID
outputString = os.path.join(self.varsP.outputContigFolder,
self.varsP.outputContigPrefix)
expectedResultFile = os.path.join(self.varsP.outputContigFolder,
expContigString +
'.cmap') # '_refined.cmap')
jobName = 'Ext %s' % expContigString # + ', Job ' + str(jobNum) + ' of ' + str(nJobs)
currentContig = contigFiles[jobNum - 1]
currentArgs = curargs + baseArgs
currentArgs += [
'-maxthreads',
str(self.varsP.maxthreads), '-o', outputString, '-id',
contigID, '-ref', currentContig
]
if self.varsP.stdoutlog:
currentArgs.extend(['-stdout', '-stderr'])
s1Job = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
expContigString,
maxThreads=self.varsP.maxthreads,
forceForward=currentContig,
clusterLogDir=self.varsP.clusterLogDir)
self.addJob(s1Job)
ct += 1
self.logArguments()
def generateJobList(self):
curArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('sampleChar')
if util.checkFile(self.varsP.bnxTarget) : #file exists only if image processing was run
bnxFiles = parseExperimentFile(self.varsP.bnxTarget)
if not bnxFiles : #check that you got at least one
errstr = "ERROR in SampleChar.generateJobList: no bnx files found in: "+self.varsP.bnxTarget
print errstr
self.varsP.updatePipeReport(errstr+"\n\n")
return
basepath = "" #os.path.split(bnxFiles[0])[0] #don't use basepath for this case
else : #otherwise, assume this is the only bnx file
bnxFiles = [self.varsP.bnxFile]
#here, make a dir for the results--should really check results of checkEmptyDir for errors
basepath = os.path.join(self.varsP.localRoot, "sampleChar")
if self.varsP.wipe and os.path.isdir(basepath) :
shutil.rmtree(basepath)
#util.checkEmptyDir(basepath) #will make if not exist, but if it does, will remove and re-make -- this fn doesn't exist...
#else :
util.checkDir(basepath) #will make if not exist, but won't remove anything
nJobs = len(bnxFiles)
#for i, bnxFile in enumerate(bnxFiles):
for bnxFile in bnxFiles :
#bnxGroupName = '%02d' % (i+1) #get this from the path, ie, bnxFiles
cargs = [self.varsP.RefAlignerBin, '-i', bnxFile]
bnxname = os.path.split(bnxFile)[1].replace(".bnx","")
jobname = 'Sample_Char_' + bnxname
#outputTarget = os.path.join(basepath, bnxGroupName)
if basepath : #bnx input
outputTarget = os.path.join(basepath, bnxname)
else : #image processing
outputTarget = bnxFile.replace(".bnx","") + "_sampleChar"
expectedResultFile = outputTarget + '.err' #this is used in checkResults
currentArgs = cargs + ['-ref', self.varsP.ref, '-o' , outputTarget, '-f']
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
currentArgs += ['-maxthreads', str(self.varsP.maxthreads)] + curArgs
sJob = mthread.singleJob(currentArgs, jobname, expectedResultFile, jobname, clusterLogDir=self.varsP.clusterLogDir) # peStr is deprecated in favor of clusterargs
#sJob.expTag = bnxGroupName #removed from checkResults
self.addJob(sJob)
self.logArguments()
def generateJobList(self):
curArgs = self.varsP.argsListed('noise0') + self.varsP.argsListed('sampleChar')
if util.checkFile(self.varsP.bnxTarget) : #file exists only if image processing was run
bnxFiles = parseExperimentFile(self.varsP.bnxTarget)
if not bnxFiles : #check that you got at least one
errstr = "ERROR in SampleChar.generateJobList: no bnx files found in: "+self.varsP.bnxTarget
print errstr
self.varsP.updatePipeReport(errstr+"\n\n")
return
basepath = "" #os.path.split(bnxFiles[0])[0] #don't use basepath for this case
else : #otherwise, assume this is the only bnx file
bnxFiles = [self.varsP.bnxFile]
#here, make a dir for the results--should really check results of checkEmptyDir for errors
basepath = os.path.join(self.varsP.localRoot, "sampleChar")
if self.varsP.wipe and os.path.isdir(basepath) :
shutil.rmtree(basepath)
#util.checkEmptyDir(basepath) #will make if not exist, but if it does, will remove and re-make -- this fn doesn't exist...
#else :
util.checkDir(basepath) #will make if not exist, but won't remove anything
nJobs = len(bnxFiles)
#for i, bnxFile in enumerate(bnxFiles):
for bnxFile in bnxFiles :
#bnxGroupName = '%02d' % (i+1) #get this from the path, ie, bnxFiles
cargs = [self.varsP.RefAlignerBin, '-i', bnxFile]
bnxname = os.path.split(bnxFile)[1].replace(".bnx","")
jobname = 'Sample_Char_' + bnxname
#outputTarget = os.path.join(basepath, bnxGroupName)
if basepath : #bnx input
outputTarget = os.path.join(basepath, bnxname)
else : #image processing
outputTarget = bnxFile.replace(".bnx","") + "_sampleChar"
expectedResultFile = outputTarget + '.err' #this is used in checkResults
currentArgs = cargs + ['-ref', self.varsP.ref, '-o' , outputTarget, '-f']
if self.varsP.stdoutlog :
currentArgs.extend( ['-stdout', '-stderr'] )
currentArgs += ['-maxthreads', str(self.varsP.maxthreads)] + curArgs
sJob = mthread.singleJob(currentArgs, jobname, expectedResultFile, jobname, clusterLogDir=self.varsP.clusterLogDir) # peStr is deprecated in favor of clusterargs
#sJob.expTag = bnxGroupName #removed from checkResults
self.addJob(sJob)
self.logArguments()
def generateJobList(self) :
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
sorted_file = self.varsP.sorted_file
self.varsP.updatePipeReport('Sorting %s into %s\n' % (self.varsP.bnxFile, sorted_file))
expectedResultFile=sorted_file+".bnx"
# We use assembly section here because the memory usage is higher than pairwise, while the jobs are quite short.
#sortJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('assembly'))
super(sortBNX, self).__init__(self.varsP, self.stageName, clusterArgs=self.varsP.getClusterArgs("assembly"))
cargs=[self.varsP.RefAlignerBin, '-f', '-i', self.varsP.bnxFile, "-maxthreads", str(self.varsP.maxthreads), "-merge", "-sort-idinc", "-bnx", "-o", sorted_file] + self.varsP.argsListed('bnx_sort')
if self.varsP.bnxStatsFile!=None:
cargs += ['-XmapStatWrite', self.varsP.bnxStatsFile]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
self.addJob(mthread.singleJob(cargs, self.stageName, expectedResultFile, self.stageName, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=sorted_file+".stdout"))
return 0 #success
def generateJobList(self) :
if not self.varsP.executeCurrentStage:
return 1 #tell self.__init__ not to continue processing
sorted_file = self.varsP.sorted_file
self.varsP.updatePipeReport('Sorting %s into %s\n' % (self.varsP.bnxFile, sorted_file))
expectedResultFile=sorted_file+".bnx"
# We use assembly section here because the memory usage is higher than pairwise, while the jobs are quite short.
#sortJobSet=mthread.jobWrapper(self.varsP,jobName,clusterArgs=self.varsP.getClusterArgs('assembly'))
super(sortBNX, self).__init__(self.varsP, self.stageName, clusterArgs=self.varsP.getClusterArgs("assembly"))
cargs=[self.varsP.RefAlignerBin, '-f', '-i', self.varsP.bnxFile, "-maxthreads", str(self.varsP.maxthreads), "-merge", "-sort-idinc", "-bnx", "-o", sorted_file] + self.varsP.argsListed('bnx_sort')
if self.varsP.bnxStatsFile!=None:
cargs += ['-XmapStatWrite', self.varsP.bnxStatsFile]
if self.varsP.stdoutlog :
cargs.extend( ['-stdout', '-stderr'] )
self.addJob(mthread.singleJob(cargs, self.stageName, expectedResultFile, self.stageName, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=sorted_file+".stdout"))
return 0 #success
def __init__(self, varsP):
jobName = "reference_process"
opta_section = "referenceSvdetect"
default_mres = "2.9"
mres = "-mres"
self.varsP = varsP
usedefault = False
if self.varsP.argData.has_key(opta_section): #check if in optargs
opta = self.varsP.argsListed(opta_section)
if not mres in opta: #must have mres
self.varsP.updatePipeReport(
"Warning in referenceProcess: " + mres +
" missing in optArguments section " + opta_section + "\n")
usedefault = True
else:
self.varsP.updatePipeReport(
"Warning in referenceProcess: optArguments section " +
opta_section + " missing\n")
usedefault = True
if usedefault:
opta = [mres, default_mres]
mresstr = opta[opta.index(mres) +
1] #get string for mres value for output name
mresstr = mresstr.replace(".", "")
if not util.checkDir(self.varsP.refFolder):
self.varsP.updatePipeReport(
"ERROR in referenceProcess: could not make output dir %s\n" %
self.varsP.refFolder)
return None
refpref = os.path.basename(
self.varsP.ref[:self.varsP.ref.rfind(".")]) + "_res" + mresstr
outarg = os.path.join(
self.varsP.refFolder,
refpref) #refFolder is new output folder for this job
expectedResultFile = outarg + ".cmap" #if ref is spots, is this spots?
args = [
self.varsP.RefAlignerBin, '-f', '-o', outarg, '-i', self.varsP.ref,
'-merge'
] + opta
stdoutf = None
if self.varsP.stdoutlog:
args.extend(['-stdout', '-stderr'])
stdoutf = outarg + ".stdout"
args += ['-maxthreads', str(self.varsP.nThreads)]
super(referenceProcess,
self).__init__(self.varsP,
jobName,
clusterArgs=self.varsP.getClusterArgs("assembly"))
job = mthread.singleJob(args,
jobName,
expectedResultFile,
jobName,
maxThreads=self.varsP.nThreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=stdoutf)
self.addJob(job)
util.LogStatus("progress", "stage_start", jobName)
self.varsP.runJobs(self, "referenceProcess")
self.doAllPipeReport()
if not self.allResultsFound(
): #this is an error, but we'll continue processing without SV detect
err = "ERROR in referenceProcess: job failed, disabling SV detect"
self.varsP.updatePipeReport(err + "\n")
util.LogError("error", err)
#self.varsP.runSV = False #no need since this class is used in SVModule
else:
self.varsP.refDeresed = expectedResultFile #store good result for SV detect
self.varsP.updatePipeReport(
"referenceProcess: using reference %s for svdetect\n" %
self.varsP.refDeresed)
util.LogStatus("progress", "stage_complete", jobName)
break
if case():
self.varsP.updatePipeReport(
"Internal error: cannot handle stage %s" %
(self.refineStage))
raise ValueError
if self.varsP.bnxStatsFile != None:
currentArgs.extend(['-XmapStatRead', self.varsP.bnxStatsFile])
s1Job = mthread.singleJob(
currentArgs,
jobName,
expectedResultFile,
expectedOutputString,
maxThreads=nthreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=expectedStdoutFile,
)
self.addJob(s1Job)
self.logArguments()
def checkResults(self, stageSuffix=""):
'''Call jobWrapper (self) .doAllPipeReport, and varsP.mergeIntoSingleCmap.
stageSuffix, if supplied, is appended to varsP.stageComplete in order to
fix the stage name reported by the CharacterizeModule in the informaticsReport.
'''
self.doAllPipeReport()
self.varsP.stageComplete = self.refineStage + stageSuffix
if self.refineStage not in ['refineB0', 'refineFinal0', 'extension0']:
def generateJobList(self):
"""AlignModule.generateJobList: create RefAligner jobs for aligning molecules to contigs.
"""
#for runAlignMol, this method is called but not used: exit if RefAlignerBin is empty
if not self.varsP.RefAlignerBin:
return
#the contigs are obtained from varsP.latestMergedCmap--check its validity, a return will mean no jobs, and no jobs is now handled in multiThreadRunJobs.
if not self.doref and (
not self.varsP.latestMergedCmap
or not util.checkCmap(self.varsP.latestMergedCmap)):
err = "Error in AlignModule.generateJobList: varsP.latestMergedCmap is not set or not valid cmap; skipping %s" % self.stageName
self.varsP.updatePipeReport(err + "\n")
util.LogError("error", err)
return
#Note: noise parameters should be fixed becuase when bnx is split, -M
# would find different parameters for different contigs. Use noise0.
baseargs = [self.varsP.RefAlignerBin]
if not self.doref:
baseargs += ['-ref', self.varsP.latestMergedCmap
] #reference is latest merged cmap
mappref = os.path.split(self.varsP.latestMergedCmap)[1]
mappref = mappref[:mappref.find(".")]
else:
baseargs += ['-ref', self.varsP.ref]
mappref = self.stageName #use stageName also for output filename
noiseargs = self.varsP.argsListed('noise0')
haverefargs = False
try: #argsListed does not check key
refargs = self.varsP.argsListed(self.stageName) #'alignmolvref'
haverefargs = True
except KeyError: #this is same as old behavior
#refargs = self.varsP.argsListed('noise0') + self.varsP.argsListed(self.argStageName) #old
refargs = self.varsP.argsListed(self.argStageName) #new
#refargs = noiseargs + refargs
if haverefargs:
self.jobargs = refargs
#single job with bnxin (constructor)
if self.bnxin:
outarg = os.path.join(self.alignTarget, mappref)
self.outFileList.append(outarg) #file prefixes
jobargs = baseargs + ['-o', outarg]
jobargs += ['-i', self.bnxin]
stdoutf = None
if self.varsP.stdoutlog: #remember, these must be after -o
jobargs.extend(['-f', '-stdout', '-stderr'])
stdoutf = outarg + ".stdout"
jobargs += ['-maxthreads', str(self.varsP.maxthreads)]
#add noise0 before alignmol (stageName) so that the latter can override the former
jobargs += noiseargs
jobargs.extend(['-output-veto-filter', 'intervals.txt$'
]) #this feature not in old RefAligner
jobargs += refargs
s1Job = mthread.singleJob(jobargs,
self.stageName,
outarg + ".xmap",
self.stageName,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=stdoutf)
self.addJob(s1Job)
return #and this is the only job
#loop over the split bnxs, make one job per bnx
for idx in range(1, self.varsP.nPairwiseJobs + 1):
outarg = os.path.join(self.alignTarget, mappref + "_" + str(idx))
self.outFileList.append(outarg) #file prefixes
jobargs = baseargs + ['-o', outarg]
idxstr = "_%s_of_%s" % (idx, self.varsP.nPairwiseJobs)
jobargs += [
'-i',
self.varsP.bnxFile.replace(".bnx", idxstr + ".bnx")
]
stdoutf = None
if self.varsP.stdoutlog: #remember, these must be after -o
jobargs.extend(['-f', '-stdout', '-stderr'])
stdoutf = outarg + ".stdout"
jobargs += ['-maxthreads', str(self.varsP.maxthreads)]
#add noise0 before alignmol (stageName) so that the latter can override the former
jobargs += noiseargs
#if idx != 1 : #keep _r for first job only -- copied from SVModule
# jobargs.extend( ['-output-veto-filter', '_r.cmap$'] ) #need this for copy number; do NOT veto
jobargs.extend(['-output-veto-filter', 'intervals.txt$'
]) #this feature not in old RefAligner
jobargs += refargs
s1Job = mthread.singleJob(jobargs,
self.stageName + idxstr,
outarg + ".xmap",
self.stageName + idxstr,
maxThreads=self.varsP.maxthreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=stdoutf)
self.addJob(s1Job)
def generateJobList(self):
baseArgs1 = self.varsP.argsListed(self.refineStage)
for case in util.switch(self.refineStage):
if case("refine(B1|Final1)", regexp=True):
baseArgs1 += self.varsP.argsListed('noise0')
ContigGroupList=self.findGroupedContigs()
r1args = [self.varsP.RefAlignerBin]
break
if case("refine(B0|Final0)", regexp=True):
baseArgs1 += self.varsP.argsListed('noise0')
ContigGroupListFull=self.groupContigs()
setattr(self.varsP, "count_"+self.varsP.outputContigPrefix, (ContigGroupListFull))
#print self.varsP.outputContigPrefix, getattr(self.varsP, "count_"+self.varsP.outputContigPrefix)
#r1args = [self.varsP.RefAlignerBin, '-i', self.varsP.bnxFile]
#InputFileList=[self.varsP.bnxFile]
r1args = [self.varsP.RefAlignerBin]
ContigGroupList = zip(range(1,self.varsP.nPairwiseJobs + 1), range(1,self.varsP.nPairwiseJobs + 1), [self.varsP.bnxFile.replace(".bnx", "_%s_of_%s.bnx" %(x, self.varsP.nPairwiseJobs)) for x in range(1,self.varsP.nPairwiseJobs + 1)], [1]*self.varsP.nPairwiseJobs)
break
if case("refineA"):
baseArgs1 += self.varsP.argsListed('noise0')
ContigGroupList=self.groupContigs()
#print("Found %d groups for refineA" % (len(ContigGroupList)))
#r1args = [self.varsP.AssemblerBin, '-i', self.varsP.bnxFile.replace(".bnx", "_sorted.bnx")] #need this before -contigs -- can no longer use all_sorted.bnx due to scan scaling: must refer to varsP.sorted_file
#r1args = [self.varsP.AssemblerBin, '-i', self.varsP.sorted_file+".bnx"] #need this before -contigs
r1args = [self.varsP.AssemblerBin, '-if', self.varsP.bnxFileList] #need this before -contigs; use split files in case splitting changed (eg due to scan scaling producing labels at < 20 bp)
r1args += ['-contigs', os.path.join(self.varsP.inputContigFolder, self.varsP.inputContigPrefix) + '.contigs']
break
if case("refineNGS"):
r1args = [self.varsP.RefAlignerBin, '-i', self.varsP.bnxFile]
ContigGroupList=self.groupContigs()
break
if case("extension0"):
baseArgs1 += self.varsP.argsListed('noise0')
ContigGroupList=self.groupContigs()
setattr(self.varsP, "count_"+self.varsP.outputContigPrefix, (ContigGroupList))
#print self.varsP.outputContigPrefix, getattr(self.varsP, "count_"+self.varsP.outputContigPrefix), self.varsP.inputContigFolder, self.varsP.inputContigPrefix
#r1args = [self.varsP.RefAlignerBin, '-i', self.varsP.bnxFile]
#InputFileList=[self.varsP.bnxFile]
r1args = [self.varsP.RefAlignerBin]
ContigGroupList = zip(range(1,self.varsP.nPairwiseJobs + 1), range(1,self.varsP.nPairwiseJobs + 1), [self.varsP.bnxFile.replace(".bnx", "_%s_of_%s.bnx" %(x, self.varsP.nPairwiseJobs)) for x in range(1,self.varsP.nPairwiseJobs + 1)], [1]*self.varsP.nPairwiseJobs)
break;
if case("extension1"):
baseArgs1 += self.varsP.argsListed('noise0')
ContigGroupList=self.findGroupedContigs()
r1args = [self.varsP.RefAlignerBin]
break;
if case():
varsP.error += 1
varsP.message += ' Error: Refine stage name invalid: '+str(StageName)+'\n'
return
stdarg = []
if self.varsP.stdoutlog : #this is the same for all cases below
stdarg = ['-stdout', '-stderr']
#contigFiles, contigIDs = self.varsP.findContigs(self.varsP.inputContigFolder, self.varsP.inputContigPrefix)
#nJobs = len(contigFiles)
output1String = os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix)
#for jobNum in range(1,nJobs + 1):
#contigID = contigIDs[jobNum - 1]
for m in range(0, len(ContigGroupList)):
contigID=ContigGroupList[m][0]
rawContigID=ContigGroupList[m][1]
contig=ContigGroupList[m][2]
# Figure out desired number of threads to use
threadBoost=ceil(ContigGroupList[m][3])
if threadBoost<1:
threadBoost=1
minthreads=self.varsP.getClusterArgs(self.refineStage, category="MinThreads")
if minthreads:
minthreads=Template(minthreads).substitute(maxthreads=self.varsP.maxthreads)
else:
minthreads=self.varsP.maxthreads
nthreads=float(minthreads)
nthreads=int(round(nthreads*threadBoost))
if nthreads>self.varsP.maxthreads:
nthreads=self.varsP.maxthreads
# for contigID, contig in ContigGroupList :
jobName = self.refineStage + ' %5s' % contigID
for case in util.switch(self.refineStage):
if case("refineA"):
endId=int(rawContigID)+self.bunching-1
if m+1<len(ContigGroupList) :
endId=int(ContigGroupList[m+1][1])-1
currentArgs = [str(rawContigID), str(endId), '-maxthreads', str(nthreads)] #this must come after r1args because it's actually an argument to -contigs
#currentArgs = r1args + currentArgs + baseArgs1 + ['-id', str(contigID), '-i', contig+"_mapped.bnx", '-o', output1String]
currentArgs = r1args + currentArgs + ['-o', output1String] + stdarg + baseArgs1
expectedOutputString = self.varsP.outputContigPrefix + '_contig' + str(rawContigID)
expectedResultFile = os.path.join(self.varsP.outputContigFolder, expectedOutputString + '.cmap') #refineB
expectedStdoutFile = output1String + "_id"+str(rawContigID)+".stdout"
break
#if case("refineB1|refineFinal1|extension1", regexp=True):
## TODO: make thread number configurable from clusterArgs
#currentArgs = ['-maxthreads', str(16), self.ref_arg, contig]
#currentArgs = r1args + currentArgs + baseArgs1 + ['-id', str(contigID), '-i', contig+"_mapped.bnx", '-o', output1String]
#expectedOutputString = self.varsP.outputContigPrefix + '_contig' + str(contigID)
#expectedResultFile = os.path.join(self.varsP.outputContigFolder, expectedOutputString + '.cmap') #refineB
#break
if case("refineB1|refineFinal1|extension1", regexp=True):
Inputs=zip(["-i"]*self.varsP.nPairwiseJobs, [contig.replace("_group", "_group"+str(i)+"_mapped_group")+".bnx" for i in range(1,self.varsP.nPairwiseJobs + 1)])
Inputs=[x for t in Inputs for x in t]
#-id must come before -o, otherwise expectedStdoutFile is wrong
currentArgs = ['-maxthreads', str(nthreads), '-id', str(contigID), '-o', output1String, self.ref_arg, contig]
currentArgs = r1args + currentArgs + stdarg + baseArgs1 + Inputs
expectedOutputString = self.varsP.outputContigPrefix + '_contig' + str(rawContigID)
expectedResultFile = os.path.join(self.varsP.outputContigFolder, expectedOutputString + '.cmap') #refineB
expectedStdoutFile = output1String + "_id"+str(contigID)+".stdout"
break
#if case("refineB0|refineFinal0|extension0", regexp=True):
#currentArgs = ['-maxthreads', str(self.varsP.maxthreads), self.ref_arg, contig]
#currentArgs = r1args + currentArgs + baseArgs1 + ['-mapped-unsplit', '1', '-refine', '0', '-mapped', contig+"_mapped", "-o", "/dev/null"]
#expectedOutputString = self.refineStage + "contig"+str(contigID) + "_mapped.bnx"
#expectedResultFile = contig + "_mapped.bnx" #refineB
#break
if case("refineB0|refineFinal0|extension0", regexp=True):
currentArgs = [ '-maxthreads', str(nthreads), "-ref", os.path.join(self.varsP.inputContigFolder, util.uniquifyContigName(self.varsP.inputContigPrefix)+".cmap")]
outputfile = os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix+'_group'+str(contigID))
#-id must come before -o, otherwise expectedStdoutFile is wrong
currentArgs = r1args + ['-i', contig, '-id', str(contigID), '-o', outputfile] + stdarg + currentArgs + baseArgs1
currentArgs += ['-refine', '0', '-grouped', os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix+'_group_manifest'), '-mapped', os.path.join(self.varsP.outputContigFolder, self.varsP.outputContigPrefix+'_group'+str(contigID)+"_mapped"), '-output-filter', ".*.bnx"]
expectedOutputString = self.varsP.outputContigPrefix+'_group'+str(contigID)+"_mapped.bnx"
expectedResultFile = outputfile + "_mapped_group1.bnx"
expectedStdoutFile = outputfile + "_id"+str(contigID)+ ".stdout"
break
if case():
self.varsP.updatePipeReport("Internal error: cannot handle stage %s" % (self.refineStage))
raise ValueError
if self.varsP.bnxStatsFile!=None:
currentArgs.extend(['-XmapStatRead', self.varsP.bnxStatsFile])
s1Job = mthread.singleJob(currentArgs,
jobName,
expectedResultFile,
expectedOutputString,
maxThreads=nthreads,
clusterLogDir=self.varsP.clusterLogDir,
expectedStdoutFile=expectedStdoutFile,
)
self.addJob(s1Job)
self.logArguments()
def generateJobList(self):
"""AlignModule.generateJobList: create RefAligner jobs for aligning molecules to contigs.
"""
#for runAlignMol, this method is called but not used: exit if RefAlignerBin is empty
if not self.varsP.RefAlignerBin :
return
#the contigs are obtained from varsP.latestMergedCmap--check its validity, a return will mean no jobs, and no jobs is now handled in multiThreadRunJobs.
if not self.doref and ( not self.varsP.latestMergedCmap or
not util.checkCmap(self.varsP.latestMergedCmap) ) :
err = "Error in AlignModule.generateJobList: varsP.latestMergedCmap is not set or not valid cmap; skipping %s" % self.stageName
self.varsP.updatePipeReport(err+"\n")
util.LogError("error", err)
return
#Note: noise parameters should be fixed becuase when bnx is split, -M
# would find different parameters for different contigs. Use noise0.
baseargs = [self.varsP.RefAlignerBin]
if not self.doref :
baseargs += ['-ref', self.varsP.latestMergedCmap] #reference is latest merged cmap
mappref = os.path.split(self.varsP.latestMergedCmap)[1]
mappref = mappref[:mappref.find(".")]
else :
baseargs += ['-ref', self.varsP.ref]
mappref = self.stageName #use stageName also for output filename
noiseargs = self.varsP.argsListed('noise0')
haverefargs = False
try : #argsListed does not check key
refargs = self.varsP.argsListed(self.stageName) #'alignmolvref'
haverefargs = True
except KeyError : #this is same as old behavior
#refargs = self.varsP.argsListed('noise0') + self.varsP.argsListed(self.argStageName) #old
refargs = self.varsP.argsListed(self.argStageName) #new
#refargs = noiseargs + refargs
if haverefargs :
self.jobargs = refargs
#single job with bnxin (constructor)
if self.bnxin :
outarg = os.path.join(self.alignTarget, mappref)
self.outFileList.append( outarg ) #file prefixes
jobargs = baseargs + ['-o', outarg]
jobargs += ['-i', self.bnxin]
stdoutf = None
if self.varsP.stdoutlog : #remember, these must be after -o
jobargs.extend( ['-f', '-stdout', '-stderr'] )
stdoutf = outarg+".stdout"
jobargs += ['-maxthreads', str(self.varsP.maxthreads)]
#add noise0 before alignmol (stageName) so that the latter can override the former
jobargs += noiseargs
jobargs.extend( ['-output-veto-filter', 'intervals.txt$'] ) #this feature not in old RefAligner
jobargs += refargs
s1Job = mthread.singleJob(jobargs, self.stageName, outarg+".xmap", self.stageName, maxThreads=self.varsP.maxthreads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=stdoutf)
self.addJob(s1Job)
return #and this is the only job
#loop over the split bnxs, make one job per bnx
for idx in range(1,self.varsP.nPairwiseJobs+1) :
outarg = os.path.join(self.alignTarget, mappref+"_"+str(idx))
self.outFileList.append( outarg ) #file prefixes
jobargs = baseargs + ['-o', outarg]
idxstr = "_%s_of_%s" % (idx, self.varsP.nPairwiseJobs)
jobargs += ['-i', self.varsP.bnxFile.replace(".bnx", idxstr+".bnx")]
stdoutf = None
if self.varsP.stdoutlog : #remember, these must be after -o
jobargs.extend( ['-f', '-stdout', '-stderr'] )
stdoutf = outarg+".stdout"
jobargs += ['-maxthreads', str(self.varsP.maxthreads)]
#add noise0 before alignmol (stageName) so that the latter can override the former
jobargs += noiseargs
#if idx != 1 : #keep _r for first job only -- copied from SVModule
# jobargs.extend( ['-output-veto-filter', '_r.cmap$'] ) #need this for copy number; do NOT veto
jobargs.extend( ['-output-veto-filter', 'intervals.txt$'] ) #this feature not in old RefAligner
jobargs += refargs
s1Job = mthread.singleJob(jobargs, self.stageName+idxstr, outarg+".xmap", self.stageName+idxstr, maxThreads=self.varsP.maxthreads, clusterLogDir=self.varsP.clusterLogDir, expectedStdoutFile=stdoutf)
self.addJob(s1Job)