From 52ca8f152836904803c499cdcd6d0726cfc0d28b Mon Sep 17 00:00:00 2001
From: "Thomas A. Christensen II" <25492070+MillironX@users.noreply.github.com>
Date: Sat, 23 Sep 2023 16:24:57 -0500
Subject: [PATCH] refactor: Replace in-workflow HAPLINK processes with imported
 modules

---
 main.nf         | 55 ++-----------------------------------------------
 nextflow.config |  8 +++++++
 2 files changed, 10 insertions(+), 53 deletions(-)

diff --git a/main.nf b/main.nf
index d6449ed..f6fbf59 100755
--- a/main.nf
+++ b/main.nf
@@ -1,6 +1,8 @@
 #!/usr/bin/env nextflow
 
 include { EFETCH } from './modules/efetch'
+include { HAPLINK_HAPLOTYPES as HAPLINK_ML_HAPLOTYPES } from './modules/haplink/haplotypes'
+include { HAPLINK_HAPLOTYPES as HAPLINK_RAW_HAPLOTYPES } from './modules/haplink/haplotypes'
 include { HAPLINK_VARIANTS } from './modules/haplink/variants'
 include { MINIMAP2 } from './modules/minimap2'
 include { NANOFILT } form './modules/nanofilt'
@@ -73,59 +75,6 @@ workflow {
     )
 }
 
-process HAPLINK_RAW_HAPLOTYPES {
-    cpus 2
-    memory '12.GB'
-
-    input:
-    tuple val(prefix), path(bam), path(bai), path(vcf)
-    path reference
-
-    output:
-    tuple val(prefix), path("*.yaml")
-
-    publishDir "results/raw-haplotypes", mode: 'copy'
-
-    script:
-    """
-    export JULIA_NUM_THREADS=${task.cpus}
-    haplink haplotypes \\
-            "${reference}" \\
-            "${vcf}" \\
-            "${bam}" \\
-        --frequency 0.01 \\
-        > "${prefix}.yaml"
-    """
-}
-
-process HAPLINK_ML_HAPLOTYPES {
-    cpus 8
-    memory '12.GB'
-
-    input:
-    tuple val(prefix), path(bam), path(bai), path(vcf)
-    path reference
-
-    output:
-    tuple val(prefix), path("*.yaml")
-
-    publishDir "results/ml-haplotypes", mode: 'copy'
-
-    script:
-    """
-    export JULIA_NUM_THREADS=${task.cpus}
-    haplink haplotypes \\
-            "${reference}" \\
-            "${vcf}" \\
-            "${bam}" \\
-        --simulated-reads \\
-        --overlap-min 20 \\
-        --overlap-max 8000 \\
-        --frequency 0.01 \\
-        > "${prefix}.yaml"
-    """
-}
-
 process HAPLINK_SEQUENCES {
     cpus 1
     memory '6.GB'
diff --git a/nextflow.config b/nextflow.config
index 84597cb..fce47a2 100644
--- a/nextflow.config
+++ b/nextflow.config
@@ -1,6 +1,14 @@
 process {
     errorStrategy = 'finish'
     time = '7d'
+
+    withName: 'HAPLINK_ML_HAPLOTYPES' {
+        ext.ml_args = """
+        --simulated-reads \\
+        --overlap-min 20 \\
+        --overlap-max 8000 \\
+        """
+    }
 }
 
 singularity.enabled = true