From bf27de1528550549e36022296d1fdaaf4e85a188 Mon Sep 17 00:00:00 2001
From: Eric Coissac <eric.coissac@metabarcoding.org>
Date: Tue, 16 May 2023 13:39:01 +0200
Subject: [PATCH] Correction of go_rps12 for not passing anymore the sequence
 as a variable

Former-commit-id: 0f9bb9472a53aa16a91a9cab5106ee66ee781c34
Former-commit-id: 016607c59e62105850d1d25f29bfe214943abc5c
---
 detectors/cds/bin/do_rps12.sh        |  3 +--
 detectors/cds/lib/rps12_filter_3.awk | 33 +++++++++++++++++++++++++++-
 2 files changed, 33 insertions(+), 3 deletions(-)

diff --git a/detectors/cds/bin/do_rps12.sh b/detectors/cds/bin/do_rps12.sh
index 130912b..ac34f62 100755
--- a/detectors/cds/bin/do_rps12.sh
+++ b/detectors/cds/bin/do_rps12.sh
@@ -118,10 +118,9 @@ blastx \
         | $AwkCmd -f $LIB_DIR/rps12_filter_2.awk \
         | $AwkCmd -v delta="$DELTA" \
                   -v seqlen="$SEQLEN" \
-                  -v chloro="$SEQUENCE" \
+                  -v chloro="${QUERY}" \
                   -f $LIB_DIR/rps12_filter_3.awk
 
-
     nrps12=$(ls -1 rps12_fragments_*.fasta | wc -l)
 
     if (( nrps12 > 1 )) ; then
diff --git a/detectors/cds/lib/rps12_filter_3.awk b/detectors/cds/lib/rps12_filter_3.awk
index d25fbab..c549ad9 100644
--- a/detectors/cds/lib/rps12_filter_3.awk
+++ b/detectors/cds/lib/rps12_filter_3.awk
@@ -6,7 +6,7 @@ function min(a,b) {return (a <= b) ? a:b }
                     x=x substr(s,i,1)
                 return x
               }    
-                     
+
 function swapchar(s,a,b) {
                 gsub(a,"@",s)
                 gsub(b,a,s)
@@ -31,6 +31,37 @@ function revcomp(s) {
                 s = swapchar(s,"d","h")
                 return rev(s)
               }
+
+function extractFirstSequence(file) {
+    # Variables
+    sequence = ""
+    inSequence = 0
+
+    # Lecture du fichier
+    while ((getline line < file) > 0) {
+        # Ignorer les lignes commençant par ">"
+        if (substr(line, 1, 1) == ">") {
+            if (inSequence == 1)
+                break;  # Sortir si nous avons déjà extrait la première séquence
+            else
+                inSequence = 1;  # Commencer à extraire la séquence
+        } else if (inSequence == 1) {
+            # Supprimer les espaces et retours chariot de la séquence
+            gsub(/[[:space:]]/, "", line);
+            sequence = sequence line;
+        }
+    }
+
+    # Fermer le fichier
+    close(file);
+
+    # Retourner la première séquence
+    return sequence;
+}
+
+BEGIN {
+    chloro = extractFirstSequence(chloro)
+}
     {   from = max(1,$1 - delta)
         to   = min($2 + delta,seqlen)
         sequence = substr(chloro,from,to-from+1)