Merge pull request #41 from ENCODE-DCC/PIP-428_v1.1.3

PIP-428 v1.1.3

.circleci/config.yml

@@ -51,7 +51,7 @@ jobs:
           name: build image
           command: |
             source ${BASH_ENV}
-            export DOCKER_CACHE_TAG=v1.1.2
+            export DOCKER_CACHE_TAG=v1.1.3
             echo "pulling ${DOCKER_CACHE_TAG}!"
             docker pull quay.io/encode-dcc/chip-seq-pipeline:${DOCKER_CACHE_TAG}
             docker login -u=${QUAY_ROBOT_USER} -p=${QUAY_ROBOT_USER_TOKEN} quay.io

backends/backend.conf

@@ -4,6 +4,63 @@ backend {
   default = "Local"
   providers {
 
+    pbs {
+      actor-factory = "cromwell.backend.impl.sfs.config.ConfigBackendLifecycleActorFactory"
+      config {
+        script-epilogue = "sleep 30 && sync"
+        concurrent-job-limit = 50
+        runtime-attributes = """
+          Int cpu = 1
+          Int? gpu
+          Int? time
+          Int? memory_mb
+        """
+        submit = """
+          qsub \
+          -N ${job_name} \
+          -o ${out} \
+          -e ${err} \
+          ${"-lselect=1:ncpus=" + cpu + ":mem=" + memory_mb/1024 + "gb"} \
+          ${"-lwalltime=" + time + ":0:0"} \
+          ${if gpu>1 then "-lngpus=" + gpu else ""} \
+          -V \
+          ${script}
+        """
+        kill = "qdel ${job_id}"
+        check-alive = "qstat ${job_id}"
+        job-id-regex = "(\\d+).+"
+      }      
+    }
+
+    pbs_singularity {
+      actor-factory = "cromwell.backend.impl.sfs.config.ConfigBackendLifecycleActorFactory"
+      config {
+        script-epilogue = "sleep 30 && sync"
+        concurrent-job-limit = 50
+        runtime-attributes = """
+          Int cpu = 1
+          Int? gpu
+          Int? time
+          Int? memory_mb
+          String singularity_container
+          String? singularity_bindpath
+        """
+        submit = """
+          ls ${singularity_container} $(echo ${singularity_bindpath} | tr , ' ') 1>/dev/null && (echo "chmod u+x ${script} && SINGULARITY_BINDPATH=$(echo ${cwd} | sed 's/cromwell-executions/\n/g' | head -n1),${singularity_bindpath} singularity exec --home ${cwd} ${if defined(gpu) then '--nv' else ''} ${singularity_container} ${script}" | qsub \
+          -N ${job_name} \
+          -o ${out} \
+          -e ${err} \
+          ${"-lselect=1:ncpus=" + cpu + ":mem=" + memory_mb/1024 + "gb"} \
+          ${"-lwalltime=" + time + ":0:0"} \
+          ${if gpu>1 then "-lngpus=" + gpu else ""} \
+          -V)
+        """
+        kill = "qdel ${job_id}"
+        check-alive = "qstat -j ${job_id}"
+        job-id-regex = "(\\d+)"
+      }
+    }
+
     slurm_singularity {
       actor-factory = "cromwell.backend.impl.sfs.config.ConfigBackendLifecycleActorFactory"
       config {

chip.wdl

@@ -1518,10 +1518,10 @@ task qc_report {
 task read_genome_tsv {
 	File genome_tsv
 	command {
-		cat ${genome_tsv}
+		cat ${genome_tsv} > 'tmp.tsv'
 	}
 	output {
-		Map[String,String] genome = read_map(stdout())
+		Map[String,String] genome = read_map('tmp.tsv')
 	}
 	runtime {
 		cpu : 1
@@ -1536,16 +1536,17 @@ task rounded_mean {
 	command <<<
 		python <<CODE
 		arr = [${sep=',' ints}]
-		if len(arr):
-		    sum_ = sum(arr)
-		    mean_ = sum(arr)/float(len(arr))
-		    print(int(round(mean_)))
-		else:
-		    print(0)
+		with open('tmp.txt','w') as fp:
+			if len(arr):
+			    sum_ = sum(arr)
+			    mean_ = sum(arr)/float(len(arr))
+			    fp.write('{}'.format(int(round(mean_))))
+			else:
+			    fp.write('0')
 		CODE
 	>>>
 	output {
-		Int rounded_mean = read_int(stdout())
+		Int rounded_mean = read_int('tmp.txt')
 	}
 	runtime {
 		cpu : 1

docs/dev.md

@@ -1,6 +1,16 @@
 Dev
 ===
 
+## Command line for version change
+```
+PREV_VER=v1.1.2
+NEW_VER=v1.1.3
+for f in $(grep -rl ${PREV_VER} --include=*.{json,md,sh,yml})
+do
+  sed -i "s/${PREV_VER}/${NEW_VER}/g" ${f}
+done
+```
+
 ## Building templates on DX for each genome
 
 Make sure that you have [`dxWDL-0.77.jar`](https://github.com/dnanexus/dxWDL/releases/download/0.77/dxWDL-0.77.jar) on your `$HOME`. Install [DNANexus Platform SDK](https://wiki.dnanexus.com/downloads) with `pip install dxpy`. Log-in on DNANexus with `dx login` and choose "ENCODE Uniform Processing Pipelines" (name of our official DNANexus project for pipelines).
@@ -9,7 +19,7 @@ Run the following command line locally to build out DX workflows for this pipeli
 
 ```
 # version
-VER=v1.1.2
+VER=v1.1.3
 
 # general
 java -jar ~/dxWDL-0.77.jar compile chip.wdl -project "ENCODE Uniform Processing Pipelines" -extras workflow_opts/docker.json -f -folder /ChIP-seq2/workflows/$VER/general -defaults examples/dx/template_general.json
@@ -42,7 +52,7 @@ java -jar ~/dxWDL-0.77.jar compile chip.wdl -project "ENCODE Uniform Processing
 ## DX Azure
 ```
 # version
-VER=v1.1.2
+VER=v1.1.3
 
 # general
 java -jar ~/dxWDL-0.77.jar compile chip.wdl -project "ENCODE Uniform Processing Pipelines Azure" -extras workflow_opts/docker.json -f -folder /ChIP-seq2/workflows/$VER/general -defaults examples/dx_azure/template_general.json

docs/tutorial_dx_cli.md

@@ -58,6 +58,6 @@ This document describes instruction for the item 1).
 
 12. Click on "Run as Analysis..." and you will be automatically redirected to the "Monitor" tab.
 
-13. It will take about an hour. You will be able to find all outputs on your output folder. Final QC report (`qc.html`)/JSON (`qc.json`) will be found on it.
+13. It will take about 6 hours. You will be able to find all outputs on your output folder. Final QC report (`qc.html`)/JSON (`qc.json`) will be found on it.
 
 14. See full specification for [input JSON file](input.md).