add CHs immediate submit

Author: chrishah

bin/immediate_submit_CH.py

@@ -0,0 +1,56 @@
+#!/usr/bin/env python3
+
+import os
+import sys
+
+from snakemake.utils import read_job_properties
+
+# last command-line argument is the job script
+jobscript = sys.argv[-1]
+
+# all other command-line arguments are the dependencies
+dependencies = set(sys.argv[1:-1])
+
+# parse the job script for the job properties that are encoded by snakemake within
+# this also includes the information contained in the cluster-config file as job_properties["cluster"]
+job_properties = read_job_properties(jobscript)
+
+sample = ""
+#find the sample name from the Snakemake rule properties (this assumes that in the Snakefile there exists a wildcard 'sample')
+if "sample" in job_properties["wildcards"]:
+        sample = job_properties["wildcards"]["sample"]
+
+#add a unit to the sample name (assumes that there is a wildcard 'unit', in our snakefile this is the number of the partition)
+if "unit" in job_properties["wildcards"]:
+	sample = sample+"-"+job_properties["wildcards"]["unit"]
+
+#add the sample name (this assumes that in the Snakefile there exists a wildcard 'sample') to the Jobname specified in the cluster_config file
+if len(sample) > 0:
+	job_properties["cluster"]["J"] = job_properties["cluster"]["J"]+"-"+sample
+
+#add rule name to log files (assumes that the log files, as specified in the cluster_config file ends in 'stdout.txt' or 'stderr.txt')
+job_properties["cluster"]["output"] = job_properties["cluster"]["output"].replace("stdout.txt", job_properties["rule"]+"."+sample+".stdout.txt")
+job_properties["cluster"]["error"] = job_properties["cluster"]["error"].replace("stderr.txt", job_properties["rule"]+"."+sample+".stderr.txt")
+
+#determine threads from the Snakemake profile, i.e. as determined in the Snakefile and the main config file respectively
+job_properties["cluster"]["ntasks"] = job_properties["threads"]
+job_properties["cluster"]["ntasks-per-node"] = job_properties["threads"]
+
+# create list with command line arguments
+cmdline = ["sbatch"]
+
+# create string for slurm submit options for rule
+slurm_args = "--partition={partition} --qos={qos} --mem={mem} --ntasks={ntasks} --ntasks-per-node={ntasks-per-node} --time={time} --hint={hint} --output={output} --error={error} -N {N} -J {J}".format(**job_properties["cluster"])
+cmdline.append(slurm_args)
+
+# now work on dependencies
+if dependencies:
+    cmdline.append("--dependency")
+    # only keep numbers (which are the jobids) in dependencies list. this is necessary because slurm returns more than the jobid. For other schedulers this could be different!
+    dependencies = [x for x in dependencies if x.isdigit()]
+    cmdline.append("afterok:" + ":".join(dependencies))
+
+cmdline.append(jobscript)
+
+#now write final commandback to the system
+os.system(" ".join(cmdline))