def getValues(cls, inputArgs):
Log.info("input arguments: {}".format(inputArgs))
options = "f:"
longOptions = ["file="]
try:
opts, args = getopt.getopt(inputArgs, options, longOptions)
except getopt.GetoptError as err:
Log.error(err)
Log.exit()
fileFlag = False
filePath = None
fileName = None
for opt, arg in opts:
Log.info("processing option: {} with arguments: {}".format(
opt, arg))
if opt in ("-f", "--file"):
if fileFlag:
Log.error(
"Input Error. Can't pass one argument twice. Exiting the application.."
)
Log.exit()
else:
fileFlag = True
filePath = arg
fileName = cls.fileInput(arg)
if fileFlag is False:
Log.error(
"Input Error. You must specify a valid file. Exiting the application.."
)
Log.exit()
return filePath, fileName
def command(cls, args_list):
Log.info('Running system command: {0}'.format(' '.join(args_list)))
proc = subprocess.Popen(args_list,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
s_output, s_err = proc.communicate()
s_return = proc.returncode
return s_return, s_output, s_err
def main():
Log.info('------------------')
Log.info('Split Task Starts')
Log.info('------------------')
inputArgs = sys.argv
args = inputArgs[1:]
filePath, fileName = Input.getValues(args)
Log.info('Path:')
Log.info(filePath)
Log.info('File name:')
Log.info(fileName)
try:
with open(filePath) as f:
lines = sum(1 for _ in f)
except EnvironmentError as e:
Log.error('Input file error:')
Log.error(e)
Log.exit()
Log.info('Number of sequences:')
Log.info(lines / 2)
Log.info('The dataset is being split in files of 50 sequences.')
Log.info('Total files:')
filesNumber = math.ceil(lines / 100)
Log.info(filesNumber)
filesList = []
count = 0
Log.info('File names:')
while count < filesNumber:
count += 1
filesList.append('in-' + str(count) + '-' + fileName)
for file in filesList:
Log.info(file)
try:
lineCount = 0
filesCount = 0
input = open(filePath, "r")
for line in input:
if lineCount % 100 == 0:
try:
output.close()
except:
Log.info('Start splitting the dataset..')
output = open('../data/input/' + filesList[filesCount], "a")
filesCount = filesCount + 1
output.write(line)
lineCount = lineCount + 1
input.close()
Log.info('Finish splitting the dataset..')
except EnvironmentError as e:
Log.error('Unkown error:')
Log.error(e)
Log.exit()
Log.info('---------------------')
Log.info('Split Task Completes')
Log.info('---------------------')
def main():
Log.info('------------------')
Log.info('Split task Starts')
Log.info('------------------')
inputArgs = sys.argv
args = inputArgs[1:]
filePath, fileName = Input.getValues(args)
Log.info('Path:')
Log.info(filePath)
Log.info('File name:')
Log.info(fileName)
try:
with open(filePath) as f:
lines = sum(1 for _ in f)
except EnvironmentError as e:
Log.error('Input file error:')
Log.error(e)
Log.exit()
Log.info('Number of sequences:')
Log.info(lines/2)
Log.info('The dataset is being splittted in files of 50 sequences.')
Log.info('Total files:')
filesNumber = math.ceil(lines/100)
Log.info(filesNumber)
filesList = []
count = 0
Log.info('File names:')
Log.info('---------------------')
Log.info('Split task Completes')
Log.info('---------------------')
def main():
Log.info('---------------------')
Log.info('Profiler Task Starts')
Log.info('---------------------')
files = Filesystem.getFiles()
Log.info("List of afas files:")
for file in files:
Log.info(file)
Log.info('Merging the files with MUSCLE profile option')
filesNumber = len(files)
count = 0
while count < filesNumber:
if count == 0:
Log.info('Profiling the first two files..')
print(files[count], files[count + 1], '../data/profile/temp-' + str(count+1) + '.afas')
(ret, out, err) = muscle(files[count], files[count + 1], '../data/profile/temp-' + str(count+1) + '.afas')
Log.info("return, {}".format(ret))
Log.info("output, {}".format(out))
Log.error("error, {}".format(err))
elif filesNumber == count + 1:
Log.info('Profiling the last file..')
print(files[count], '../data/profile/temp-' + str(count - 1) + '.afas','../data/final/result.afas')
(ret, out, err) = muscle(files[count], '../data/profile/temp-' + str(count - 1) + '.afas','../data/final/result.afas')
Log.info("return, {}".format(ret))
Log.info("output, {}".format(out))
Log.error("error, {}".format(err))
elif filesNumber >= count + 3:
Log.info('Profiling a single file..')
print(files[count+1], '../data/profile/temp-' + str(count) + '.afas', '../data/profile/temp-' + str(count+1) + '.afas')
(ret, out, err) = muscle(files[count+1], '../data/profile/temp-' + str(count) + '.afas', '../data/profile/temp-' + str(count+1) + '.afas')
Log.info("return, {}".format(ret))
Log.info("output, {}".format(out))
Log.error("error, {}".format(err))
count = count + 1
Log.info('Please see the results in folder data/final')
Log.info('------------------------')
Log.info('Profiler Task Completes')
Log.info('------------------------')
def main():
log = Log(''.join(choice(ascii_uppercase) for i in range(12)))
Log.info('------------------------------')
Log.info('MUSCLE Process Starts')
Log.info('------------------------------')
inputArgs = sys.argv
args = inputArgs[1:]
filePath, fileName = Input.getValues(args)
Log.info('Path:')
Log.info(filePath)
Log.info('Filename:')
Log.info(fileName)
try:
with open(filePath) as f:
lines = sum(1 for _ in f)
except EnvironmentError as e:
Log.error('Input file error:')
Log.error(e)
Log.exit()
Log.info('Number of sequences:')
Log.info(lines / 2)
outputFilePath = filePath.replace('in', 'out')
outputFilePath = outputFilePath.replace('fas', 'afas')
Log.info('Starting MUSCLE..')
print(filePath, outputFilePath)
(ret, out, err) = muscle(filePath, outputFilePath)
Log.info("return, {}".format(ret))
Log.info("output, {}".format(out))
Log.error("error, {}".format(err))
Log.info('------------------------')
Log.info('MUSCLE Process Completes')
Log.info('------------------------')