jip bash [–help]
Wraps a bash command and either executes it directly or submits it to a compute cluster
Please not that this command is indented to work on single file input/output. You can specify more that one input file and the command will run independently on all inputs. The ‘output’ options is used for pipes explicitly. If you do not want to pipe your output, but handle output yourself, use the ‘outfile’ (-f/–outfile) option. Here is a quick example:
jip bash -n 'LC ${input}' --input A.txt B.txt \
-f '${input|ext}.count' -c 'wc -l ${input} > ${outfile}'
This will run the following two jobs:
wc -l A.txt > A.count
and
wc -l B.txt > B.count
Note that you can use the job options also in the jobs name, which might be usefull if you run the job on a compute cluster.
| -P <profile>, --profile <profile> | |
| Select a job profile for resubmission | |
| -t <time>, --time <time> | |
| Max wallclock time for the job | |
| -q <queue>, --queue <queue> | |
| Job queue | |
| -p <priority>, --priority <priority> | |
| Job priority | |
| -A <account>, --account <account> | |
| The account to use for submission | |
| -C <threads>, --threads <cpus> | |
| Number of CPU’s assigned to the job | |
| -m <mem>, --mem <mem> | |
| Max memory assigned to the job | |
| -n <name>, --name <name> | |
| Job name | |
| -R <reload>, --reload | |
| Reload and rerender the job command | |
| -E <err>, --log <err> | |
| Jobs stderr log file | |
| -O <out>, --out <out> | |
| Jobs stdout log file | |
| -s, --submit | Submit as job to the cluster |
| --hold | Put job on hold after submission |
| --keep | Keep output also in case of failure |
| --dry | Show a dry run |
| --show | Show the command that will be executed |
| --force | Force execution/submission |
| -i <input>, --input <input> | |
| The scripts input [default: stdin] | |
| -o <output>, --output <output> | |
| The scripts output [default: stdout] | |
| -s, --submit | Submit as job to the cluster |
| -h, --help | Show this help message |