From 7f205d0a044b031655bbcbe2c3d8c21588a2f854 Mon Sep 17 00:00:00 2001
From: LouisLeNezet <louislenezet@gmail.com>
Date: Tue, 26 Mar 2024 10:14:01 +0100
Subject: [PATCH] Update table stats selection

---
 docs/development.md                              |  1 +
 modules/local/addcolumns/main.nf                 |  7 ++++++-
 .../tests/main.nf.test.snap                      | 16 +++-------------
 3 files changed, 10 insertions(+), 14 deletions(-)

diff --git a/docs/development.md b/docs/development.md
index 33051d56..5b358bec 100644
--- a/docs/development.md
+++ b/docs/development.md
@@ -41,6 +41,7 @@ All channel need to be identified by a meta map as follow:
 - T : tool used
 - G : reference genome used (is it needed ?)
 - S : simulation (depth or genotype array)
+
 ## Open questions
 
 How to use different schema ?
diff --git a/modules/local/addcolumns/main.nf b/modules/local/addcolumns/main.nf
index 54da94a6..2fbe882c 100644
--- a/modules/local/addcolumns/main.nf
+++ b/modules/local/addcolumns/main.nf
@@ -14,7 +14,12 @@ process ADD_COLUMNS {
     script:
     def prefix = task.ext.prefix ?: "${meta.id}"
     """
-    awk '(NR>=2) && (NR<=10)' $input | \\
+    # Find the header line
+    HEADER_STR="#Genotype concordance by allele frequency bin (Variants: SNPs + indels)"
+    HEADER_LINE=\$(grep -n -m 1 "^\${HEADER_STR}" $input | cut -d: -f1 )
+    HEADER_START=\$((HEADER_LINE + 1))
+
+    tail -n +\$HEADER_START $input | \\
     awk 'NR==1{\$(NF+1)="ID"} NR>1{\$(NF+1)="${meta.id}"}1' | \\
     awk 'NR==1{\$(NF+1)="Region"} NR>1{\$(NF+1)="${meta.region}"}1' | \\
     awk 'NR==1{\$(NF+1)="Depth"} NR>1{\$(NF+1)="${meta.depth}"}1' | \\
diff --git a/subworkflows/local/vcf_concordance_glimpse/tests/main.nf.test.snap b/subworkflows/local/vcf_concordance_glimpse/tests/main.nf.test.snap
index 608ceca5..94785ba1 100644
--- a/subworkflows/local/vcf_concordance_glimpse/tests/main.nf.test.snap
+++ b/subworkflows/local/vcf_concordance_glimpse/tests/main.nf.test.snap
@@ -3,23 +3,13 @@
         "content": [
             {
                 "0": [
-                    [
-                        {
-                            "id": "TestQuality"
-                        },
-                        "TestQuality.txt:md5,910b294df62dbe64e8f16379428d93ad"
-                    ]
+                    
                 ],
                 "1": [
                     
                 ],
                 "stats": [
-                    [
-                        {
-                            "id": "TestQuality"
-                        },
-                        "TestQuality.txt:md5,910b294df62dbe64e8f16379428d93ad"
-                    ]
+                    
                 ],
                 "versions": [
                     
@@ -30,6 +20,6 @@
             "nf-test": "0.8.4",
             "nextflow": "23.10.1"
         },
-        "timestamp": "2024-03-25T12:44:54.967846019"
+        "timestamp": "2024-03-26T10:07:33.413253799"
     }
 }
\ No newline at end of file