adding sge/uge profile

workflow/profiles/sge/cluster.yaml

@@ -0,0 +1,16 @@
+__resources__:
+  v100: 
+    - "gpu"
+    - "v100"  
+    - "nvidia_gpu"
+
+__options__:
+  jc:
+    - "jc"
+    - "jclass"
+    - "job_class"
+
+__default__:
+  pe: "def_slot " 
+  output: logs/{rulename}/{rulename}.{jobid}.out
+  error: logs/{rulename}/{rulename}.{jobid}.err

workflow/profiles/sge/config.yaml

@@ -0,0 +1,80 @@
+restart-times: 0
+jobscript: sge-jobscript.sh
+cluster: "sge-submit.py"
+cluster-status: "sge-status.py"
+cluster-cancel: "sge-cancel.py"
+cluster-cancel-nargs: 20
+max-jobs-per-second: 0.1
+max-status-checks-per-second: 0.017
+latency-wait: 180
+local-cores: 1
+
+configfile: TriosCompass_v2/config/bam_config.yaml
+snakefile: TriosCompass_v2/workflow/Snakefile
+verbose: True
+skip-script-cleanup: True
+reason: True
+show-failed-logs: True
+rerun-incomplete: True
+keep-incomplete: True
+keep-going: True
+jobs: 500
+use-conda: True
+use-singularity: True
+
+# Job resources
+set-resources:
+  - bwa_index:mem_mb=16000
+  - fastp:mem_mb=40000
+  - fastqc:mem_mb=10000
+  - fastq_screen:mem_mb=50000
+  - fq2bam:mem_mb=40000
+  - fq2bam:runtime=8:00:00
+  - fq2bam:v100=1
+  - flagstat:mem_mb=80000
+  - collectwgsmetrics:runtime=24:00:00
+  - collectwgsmetrics:mem_mb=64000
+  - collectmultiplemetrics:mem_mb=8000
+  - collectmultiplemetrics:v100=1
+  - gatkhc_pb:mem_mb=8000
+  - gatkhc_pb:runtime=24:00:00
+  - gatkhc_pb:v100=1
+  - gatk_combine_gvcf:mem_mb=80000
+  - gatk_combine_gvcf:runtime=144:00:00
+  - gatk_cgp:mem_mb=40000
+  - gatk_cgp:runtime=24:00:00
+  - gatk_genotype_gvcf_pb:runtime=24:00:00
+  - gatk_genotype_gvcf_pb:mem_mb=80000
+  - gatk_genotype_gvcf_pb:v100=1
+  - deepvariant_pb:mem_mb=80000
+  - deepvariant_pb:runtime=24:00:00
+  - deepvariant_pb:v100=1
+  - glnexus_dv:mem_mb=80000
+  - glnexus_dv:runtime=24:00:00
+  - call_dnm_dv:mem_mb=20000
+  - call_dnm_dv:runtime=24:00:00
+  - call_dnm_hc:mem_mb=20000
+  - call_dnm_hc:runtime=24:00:00
+  - call_JIGV:mem_mb=60000
+  - replace_rg:mem_mb=60000
+  - replace_rg:s_vmem=60000  
+  - replace_rg:runtime=24:00:00
+  - phase_child:mem_mb=40000
+  - phase_trios:mem_mb=40000
+  - split_bed_hipstr:mem_mb=10000
+  - hipstr:mem_mb=40000
+  - hipstr:runtime=200:00:00
+  - dumpstr_call_hipstr:mem_mb=40000
+  - dumpstr_locus:mem_mb=40000
+  - dumpstr_locus:runtime=20:00:00
+  - monstr:mem_mb=40000
+  - monstr:runtime=20:00:00
+  - merge_monstr:mem_mb=40000
+  - merge_monstr:runtime=20:00:00
+  - monstr_filter:mem_mb=20000
+  - hipstr_recall:mem_mb=10000
+
+# For some reasons time needs quotes to be read by snakemake
+default-resources:
+  - mem_mb=2000
+  - runtime="10:00:00"    

workflow/profiles/sge/sge-cancel.py

@@ -0,0 +1,9 @@
+#!/usr/bin/env python3
+import subprocess as sp
+import shlex
+import sys
+
+jobid_list = ', '.join(sys.argv[1:])
+
+sp.check_call(shlex.split(f"qdel {jobid_list}"))
+

workflow/profiles/sge/sge-jobscript.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+# properties = {properties}
+
+# exit on first error
+set -o errexit
+
+{exec_job}

workflow/profiles/sge/sge-status.py

@@ -0,0 +1,69 @@
+#!/usr/bin/env python3
+import re
+import subprocess as sp
+import shlex
+import sys
+import time
+import logging
+
+logger = logging.getLogger("__name__")
+logger.setLevel(40)
+
+STATUS_ATTEMPTS = 20
+
+jobid = int(sys.argv[1])
+job_status = "running"
+
+# WARNING this currently has no support for task array jobs
+
+for i in range(STATUS_ATTEMPTS):
+    # first try qstat to see if job is running
+    # we can use `qstat -s pr -u "*"` to check for all running and pending jobs
+    try:
+        qstat_res = sp.check_output(shlex.split(f"qstat -s pr")).decode().strip()
+
+        # skip the header using [2:]
+        res = {
+            int(x.split()[0]) : x.split()[4] for x in qstat_res.splitlines()[2:]
+        }
+
+        # job is in an unspecified error state
+        if "E" in res[jobid]:
+            job_status = "failed"
+            break
+
+        job_status = "running"
+        break
+
+    except sp.CalledProcessError as e:
+        logger.error("qstat process error")
+        logger.error(e)
+    except KeyError as e:
+        # if the job has finished it won't appear in qstat and we should check qacct
+        # this will also provide the exit status (0 on success, 128 + exit_status on fail)
+        # Try getting job with scontrol instead in case sacct is misconfigured
+        try:
+            qacct_res = sp.check_output(shlex.split(f"qacct -j {jobid}"))
+
+            exit_code = int(re.search("exit_status\s*([0-9]+)", qacct_res.decode()).group(1))
+
+            if exit_code == 0:
+                job_status = "success"
+                break
+
+            if exit_code != 0:
+                job_status = "failed"
+                break
+
+        except sp.CalledProcessError as e:
+            logger.warning("qacct process error")
+            logger.warning(e)
+            if i >= STATUS_ATTEMPTS - 1:
+                job_status = "failed"
+                break
+            else:
+                # qacct can be quite slow to update on large servers
+                time.sleep(30)
+        pass
+
+print(job_status)