From 3c87d90aac30e9420806de37ec8752892d9b7b61 Mon Sep 17 00:00:00 2001
From: slsevilla <slsevilla@gmail.com>
Date: Sun, 31 Mar 2024 21:06:59 +0000
Subject: [PATCH] refactor: ica handling, filtering #147 #149

---
 modules/local/get_trimd_stats.nf | 26 ++++++++++++--------------
 1 file changed, 12 insertions(+), 14 deletions(-)

diff --git a/modules/local/get_trimd_stats.nf b/modules/local/get_trimd_stats.nf
index a14d4653..6ea2a575 100644
--- a/modules/local/get_trimd_stats.nf
+++ b/modules/local/get_trimd_stats.nf
@@ -13,41 +13,39 @@ process GET_TRIMD_STATS {
 
     output:
     tuple val(meta), path('*_trimmed_read_counts.txt'),          emit: fastp_total_qc
-    tuple val(meta), path('*_summary.txt'),                      emit: outcome
-    path('*_summaryline.tsv'),                    optional:true, emit: summary_line
-    tuple val(meta), path('*_summary_old_3.txt'),                emit: outcome_to_edit
-    tuple val(meta), path('*.synopsis'),          optional:true, emit: synopsis
+    path('*_summaryline.tsv'),                                   optional:true, emit: summary_line
+    tuple val(meta), path('*_summary_fastp.txt'),                emit: outcome
+    tuple val(meta), path('*.synopsis'),                         optional:true, emit: synopsis
     path("versions.yml"),                                        emit: versions
 
     script: // This script is bundled with the pipeline, in cdcgov/phoenix/bin/
-    // Adding if/else for if running on ICA it is a requirement to state where the script is, however, this causes CLI users to not run the pipeline from any directory.
-    if (params.ica==false) { ica = "" } 
-    else if (params.ica==true) { ica = "python ${workflow.launchDir}/bin/" }
-    else { error "Please set params.ica to either \"true\" if running on ICA or \"false\" for all other methods." }
     // define variables
     def prefix = task.ext.prefix ?: "${meta.id}"
     def busco_parameter = busco_val ? "--busco" : ""
     def container_version = "base_v2.1.0"
     def container = task.container.toString() - "quay.io/jvhagey/phoenix@"
+    def script_fastp = params.ica ? "python ${params.ica_path}/FastP_QC.py" : "FastP_QC.py"
+    def script_fairy = params.ica ? "python ${params.ica_path}/fairy.py" : "fairy.py"
     """
-    ${ica}FastP_QC.py \\
+    ${script_fastp} \\
       --trimmed_json ${fastp_trimd_json} \\
       --single_json ${fastp_singles_json} \\
       --name ${prefix}
 
-    # Check that there are still reads in R1 and R2 before fastqc. If there aren't reads then fastqc dies.
+    # making a copy of the summary file - this avoids writing to the previous file
+    cp ${fairy_outcome} ${prefix}_input.txt
 
     # Output check for messages indicating there are no trimmed reads after filtering.
-    ${ica}fairy.py -r ${raw_qc} -f ${fairy_outcome} -t ${prefix}_trimmed_read_counts.txt ${busco_parameter}
+    ${script_fairy} -r ${raw_qc} -f ${prefix}_input.txt -t ${prefix}_trimmed_read_counts.txt ${busco_parameter}
 
     #making a copy of the summary file to pass to BBMAP_REFORMAT to handle file names being the same
-    cp ${prefix}_summary.txt ${prefix}_summary_old_3.txt
+    mv ${prefix}_summary.txt ${prefix}_summary_fastp.txt
 
     cat <<-END_VERSIONS > versions.yml
     "${task.process}":
         python: \$(python --version | sed 's/Python //g')
-        fairy.py: \$( ${ica}fairy.py --version )
-        FastP_QC.py: \$(${ica}FastP_QC.py --version )
+        fairy.py: \$( ${script_fairy} --version )
+        FastP_QC.py: \$(${script_fastp} --version )
         phoenix_base_container_tag: ${container_version}
         phoenix_base_container: ${container}
     END_VERSIONS