diff --git a/src/run-submit-pol-job.py b/src/run-submit-pol-job.py
index be824cc8d760f1a31015446e1b7b2cfae44a7019..21be17f3ea8f0eb406f2d5f1883e570850efb231 100755
--- a/src/run-submit-pol-job.py
+++ b/src/run-submit-pol-job.py
@@ -15,6 +15,7 @@ def parse_args():
parser = argparse.ArgumentParser(description=description,
formatter_class=argparse.RawTextHelpFormatter
)
+
parser.add_argument('-o','--outdir', type=str,
help="Directory to store the policy output in",
default='/data/rc/gpfs-policy/data')
@@ -24,6 +25,7 @@ def parse_args():
"file name will have the policy type, the job ID, and the run date tagged on the end")
parser.add_argument('--with-dirs', action='store_true',
help="Include directories as entries in the policy output (Default: false)")
+
sbatch = parser.add_argument_group('sbatch parameters')
sbatch.add_argument('-N','--nodes',type=int,default=1,
help='Number of nodes to run job across')
@@ -35,10 +37,11 @@ def parse_args():
help='Time limit for job formatted as [D-]HH:MM:SS')
sbatch.add_argument('-m','--mem-per-cpu',type=str,default='8G',
help='Amount of RAM to allocate per core')
+
parser.add_argument('device',type=str,
help="GPFS fileset/directory apply the policy to. Can be specified as either the name of the"
"fileset or the full path to the directory. (Examples: scratch, /data/user/[username])')")
- args = parser.parse_args()
+ args = vars(parser.parse_args())
return args
# Validate that the string supplied to 'device' is either the name of a valid, predefined fileset (only 'scratch'
@@ -62,14 +65,14 @@ def validate_device(device):
if p.exists() and any([parent in p.parents for parent in valid_parents]):
return p
else:
- raise InvalidDeviceError('The path or fileset must be within /data or /scratch')
+ raise InvalidDeviceError(f'The path or fileset {device} does not exist within /data or /scratch')
def validate_time(time):
if not re.match(r'^(?:[0-6]-\d{2}|\d{1,3}):\d{2}:\d{2}$',time):
raise ValueError("Time must have format [[H]H]H:MM:SS or D-HH:MM:SS")
def validate_mem(mem):
- if not re.match(r'[\d]+[GM]?', mem):
+ if not re.fullmatch(r'^[\d]+[GM]?$', mem):
raise ValueError("Mem per CPU must be an integer. May be followed by M or G to denote units")
def validate_partition(partition):
@@ -77,11 +80,11 @@ def validate_partition(partition):
raise ValueError("Multiple partitions should be given as a comma-separated list")
partition = re.sub(r'\s',r'',partition)
- partitions = [partition.split(r',')]
+ partitions = partition.split(r',')
cmd=r'sinfo -h -o "%R"'
- avail_partitions=subprocess.run(cmd, shell=True, stdout=subprocess.PIPE, text=True).stdout.splitlines()
+ avail_partitions=subprocess.run(cmd, shell=True, stdout=subprocess.PIPE).stdout.decode('utf-8').splitlines()
- if ([p for p in partitions if p not in avail_partitions]):
+ if any([p for p in partitions if p not in avail_partitions]):
incorrect = [p for p in partitions if p not in avail_partitions]
raise ValueError(f"The following partition(s) do not exist: {', '.join(incorrect)}. To see a list of valid partitions, visit https://rc.uab.edu/")
@@ -95,14 +98,6 @@ def validate_ntasks(n):
if not isinstance(n,int) and n >= 1 and n <= 48:
raise ValueError('Ntasks must be an integer between 1 and 48')
-def validate_sbatch_options(args):
- validate_nodes(args['nodes'])
- validate_ntasks(args['ntasks'])
- validate_mem(args['mem_per_cpu'])
- validate_time(args['time'])
- args['partition'] = validate_partition(args['partition'])
- return args
-
# Need to validate that the output directory exists. This will not create a directory that does not already exist.
def validate_output_directory(outdir):
p = Path(outdir).resolve()
@@ -120,20 +115,24 @@ def create_default_outfile(device):
def main():
args = parse_args()
- args_dict = vars(args)
-
args['device'] = validate_device(args['device'])
- sbatch_args = {(k,v) for k,v in args_dict.items() if k in ['nodes','ntasks','partition','time','mem_per_cpu'] and v is not None}
- sbatch_args = validate_sbatch_options(sbatch_args)
+ validate_nodes(args['nodes'])
+ validate_ntasks(args['ntasks'])
+ validate_mem(args['mem_per_cpu'])
+ validate_time(args['time'])
+ args['partition'] = validate_partition(args['partition'])
args['outdir'] = validate_output_directory(args['outdir'])
- if 'outfile' not in args.keys():
+ if args['outfile'] is None:
args['outfile'] = create_default_outfile(args['device'])
- cmd = f"./submit-pol-job "
+ cmd = "./submit-pol-job -o {outdir} -f {outfile} -N {nodes} -n {ntasks} -p {partition} -t {time} -m {mem_per_cpu} {device}".format(**args)
+ if args['with_dirs']:
+ cmd = f'{cmd} --with-dirs'
+ print(cmd)
exit()
if __name__ == '__main__':