def writeCombineBAMJobsFromSAI(outDir, fastqDir, fastaPath, lockDirPath):
combinedBAMJobsName = "combinedBAMFrmSingleSAIJob.jobs"
combinedBAMJobsPath = os.path.join(outDir, combinedBAMJobsName)
tools = workerForBam.getToolsDict()
# contained all fastas against which to map the fastqs
fastaFilePaths = workerForBam.getFastaFilePaths(fastaPath)
# contained all fastqs
fastqFilePaths = workerForBam.getFastqFilePaths(fastqDir)
print "fastaFilePaths: ", fastaFilePaths
with open(combinedBAMJobsPath, "w") as combinedBAMJobsFile:
for fastqPath in fastqFilePaths:
for fastaFilePath in fastaFilePaths:
workerForBam.writeSAIIToBAMJob(combinedBAMJobsFile, fastaFilePath, fastqPath, lockDirPath, tools)
return combinedBAMJobsPath
def writeCombineSAIJobs(outDir, fastqDir, fastaPath, lockDirPath, numThreads):
combinedSAIJobsName = 'combinedSAIJob.jobs'
combinedSAIJobsPath = os.path.join(outDir, combinedSAIJobsName)
tools = workerForBam.getToolsDict()
#contained all fastas against which to map the fastqs
fastaFilePaths = workerForBam.getFastaFilePaths(fastaPath)
#contained all fastqs
fastqFilePaths = workerForBam.getFastqFilePaths(fastqDir)
print 'fastaFilePaths: ', fastaFilePaths
with open(combinedSAIJobsPath, 'w') as combinedSAIJobsFile:
for fastqPath in fastqFilePaths:
for fastaFilePath in fastaFilePaths:
workerForBam.writeSAIJob(combinedSAIJobsFile, fastaFilePath,\
fastqPath, lockDirPath, tools,\
numThreads)
return combinedSAIJobsPath
def writeCombineBAMJobsFromSAI(outDir, fastqDir, fastaPath, lockDirPath):
combinedBAMJobsName = 'combinedBAMFrmPairedSAIsJob.jobs'
combinedBAMJobsPath = os.path.join(outDir, combinedBAMJobsName)
tools = workerForBam.getToolsDict()
#contained all fastas against which to map the fastqs
fastaFilePaths = workerForBam.getFastaFilePaths(fastaPath)
print 'fastaFilePaths: ', fastaFilePaths
with open(combinedBAMJobsPath, 'w') as combinedBAMJobsFile:
pairedReads = getPairedReads(fastqDir)
print 'pairedReads: ', pairedReads
for pairedReadTuple in pairedReads:
for fastaFilePath in fastaFilePaths:
workerForBam.writePairedSAIToBAMJob(combinedBAMJobsFile, \
fastaFilePath,\
pairedReadTuple,\
lockDirPath, tools)
return combinedBAMJobsPath
def main():
logger = multiprocessing.log_to_stderr()
logger.setLevel(multiprocessing.SUBDEBUG)
if len(sys.argv) >= 4:
#directory containing fastq library
fastqsDir = workerForBam.getAbsPath(sys.argv[1])
#directory containing other directories with fasta names
fastaDir = workerForBam.getAbsPath(sys.argv[2])
#directory containing file locks
lockDirPath = workerForBam.getAbsPath(sys.argv[3])
#directory containing temp output -> fastQ's, jobsFile
outDir = workerForBam.getAbsPath(sys.argv[4])
#write all fastq's processing in job file
#combineJobPath = writeCombineFastqJobs(outDir, fastqsDir, fastaDir,\
# lockDirPath)
tools = workerForBam.getToolsDict()
"""retcode = workerForBam.callParallelDrone(combineJobPath,\
tools['PARALLEL_DRONE'])
if retcode != 0:
#error occured while calling parallel drone
print "parallel drone erred, in executing combined jobs"
return -1
"""
#now for all scaffolds combined bams and look for SNPs
parallelSNPsFinder.snpsFinder(fastaDir, outDir, lockDirPath)
else:
print 'err: files missing'
def snpsFinder(fastaPath, outDir, lockDirPath):
#get all tools
tools = workerForBam.getToolsDict()
#get all scaffold folder names inside fasta dir
fastaDirs = workerForBam.getAllFastas(fastaPath)
#file containing snpfinder job for each scaffold
snpsFinderJobFileName = fastaPath.split('/')[-1]+"SNPFinderJobs.jobs"
print snpsFinderJobFileName + ' job file to be created '
try:
#open the snpsFinderJob file
snpsFinderJobFile = open(outDir + snpsFinderJobFileName, 'w')
for fastaDir in fastaDirs:
#get fasta file path
fastaFilePath = fastaPath + fastaDir + "/" + fastaDir + ".fasta"
writeJob(snpsFinderJobFile, fastaFilePath, lockDirPath, tools)
snpsFinderJobFile.close()
except IOError as (errno, strerror):
print "I/O error({0}): {1}".format(errno, strerror)
return -1
def writeCombineFastqJobs(outDir, fastqDir, fastaPath, lockDirPath):
combinedBAMJobsName = 'combinedBAMJob.jobs'
combinedBAMJobsPath = os.path.join(outDir, combinedBAMJobsName)
tools = workerForBam.getToolsDict()
#contained all fastas against which to map the fastqs
fastaFilePaths = []
#fastaPath contains all .fasta inside a dir with same name as fasta
#get all fasta name without ext ".fasta" in a list
fastaDirs = workerForBam.getAllFastas(fastaPath)
for fastaDir in fastaDirs:
#get fasta File Path
fastaFilePath = fastaPath + fastaDir + "/" + fastaDir + ".fasta"
fastaFilePaths.append(fastaFilePath)
print 'fastaFilePaths: ', fastaFilePaths
with open(combinedBAMJobsPath, 'w') as combinedBAMJobsFile:
dirContents = os.listdir(fastqDir)
for fileName in dirContents:
fastqPath = os.path.join(fastqDir, fileName)
if os.path.isfile(fastqPath) and\
fileName.endswith('fastq'):
for fastaFilePath in fastaFilePaths:
workerForBam.writeJob(combinedBAMJobsFile, fastaFilePath,\
fastqPath, lockDirPath, tools)
return combinedBAMJobsPath