Merge pull request #2 from CCBR/rand-seed

feat: add --seed arg to set --random-source for shuf
CCBR · Dec 11, 2024 · e37780e · e37780e
2 parents 3793492 + 02756f9
commit e37780e
Show file tree

Hide file tree

Showing 3 changed files with 15 additions and 6 deletions.
diff --git a/.github/workflows/docker.yml b/.github/workflows/docker.yml
@@ -13,7 +13,7 @@ env:
   IMAGE_NAME: recap
   CONTEXT: "./"
   NAMESPACE: nciccbr
-  VERSION_TAG: 0.2.2
+  VERSION_TAG: 0.3.0
 
 jobs:
   build-docker:

diff --git a/RECAP_MACS.sh b/RECAP_MACS.sh
@@ -92,7 +92,7 @@ then
 fi
 
 # 1) Re-mix ChIP and control bed files
-bash $REMIX_PATH -i $INPUT_DIR -t $CHIP_NAME -c $CONTROL_NAME -o $OUTPUT_DIR -m unequal -b $BOOTSTRAP
+bash $REMIX_PATH -i $INPUT_DIR -t $CHIP_NAME -c $CONTROL_NAME -o $OUTPUT_DIR -m unequal -b $BOOTSTRAP -s $SEED
 
 # 2) Call original peaks using MACS
 # Please specify your own MACS parameters!
@@ -186,9 +186,10 @@ while [[ $1 = -?* ]]; do
    	-t|--treatment) shift; CHIP_NAME=${1} ;;
    	-c|--control)   shift; CONTROL_NAME=${1} ;;
    	-o|--output)    shift; OUTPUT_DIR=${1} ;;
-	-b|--bootstrap) shift; BOOTSTRAP=${1} ;;
-	-e|--header)    shift; HEADER=${1} ;;
-	-h|--help)      usage >&2; exit 0 ;;
+	  -b|--bootstrap) shift; BOOTSTRAP=${1} ;;
+		-s|--seed)      shift; SEED=${1} ;;
+	  -e|--header)    shift; HEADER=${1} ;;
+	  -h|--help)      usage >&2; exit 0 ;;
 		*)       echo "ERROR: Bad argument ${1}" ; exit 1 ;;
 	esac
 	shift

diff --git a/RECAP_Re-Mix.sh b/RECAP_Re-Mix.sh
@@ -44,6 +44,12 @@ normal=$(tput sgr0)
 # Default values are below
 # ===============================================================
 
+# source: https://www.gnu.org/software/coreutils/manual/html_node/Random-sources.html
+function get_seeded_random() {
+  seed="$1"
+  openssl enc -aes-256-ctr -pass pass:"$seed" -nosalt \
+    </dev/zero 2>/dev/null
+}
 
 function mainRemix() {
 ####################### Begin Script Here #######################
@@ -61,6 +67,7 @@ echo "Control library:    $CONTROL_NAME"
 echo "Output directory (absolute path):   $OUTPUT_DIR"
 echo "Re-mix method:      $METHOD_NAME"
 echo "Number of re-mixes: $BOOTSTRAP"
+echo "Random seed: $SEED"
 
 if [[ ! -d $INPUT_DIR ]]
 then
@@ -125,7 +132,7 @@ do
 	fi
 
 	echo "Re-mixing..."
-	shuf -o $TREATMENT_NAME_BASE.bootstrap_$BOOTSTRAP_COUNT.tmp < $TREATMENT_NAME_BASE.bootstrap_$BOOTSTRAP_COUNT.tmp
+	shuf --random-source=<(get_seeded_random $SEED) -o $TREATMENT_NAME_BASE.bootstrap_$BOOTSTRAP_COUNT.tmp < $TREATMENT_NAME_BASE.bootstrap_$BOOTSTRAP_COUNT.tmp
 	echo "Check!"
 
 	echo "Creating re-mixed treatment library"
@@ -221,6 +228,7 @@ while [[ $1 = -?* ]]; do
 							exit 1
 						fi ;;
 		-b|--bootstrap) shift; BOOTSTRAP=${1} ;;
+		-s|--seed)      shift; SEED=${1} ;;
 		-h|--help)      usage >&2; exit 0 ;;
 		*)              echo "ERROR: Bad argument ${1}" ; exit 1 ;;
 	esac