diff baseline/wrapper.sh @ 81:b6f9a640e098 draft

Uploaded
author davidvanzessen
date Fri, 19 Feb 2021 15:10:54 +0000
parents
children d714f5ea83d7
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/baseline/wrapper.sh	Fri Feb 19 15:10:54 2021 +0000
@@ -0,0 +1,92 @@
+#!/bin/bash
+dir="$(cd "$(dirname "$0")" && pwd)"
+
+testID=$1
+species=$2
+substitutionModel=$3
+mutabilityModel=$4
+clonal=$5
+fixIndels=$6
+region=$7
+inputs=$8
+inputs=($inputs)
+IDs=$9
+IDs=($IDs)
+ref=${10}
+output=${11}
+selection=${12}
+output_table=${13}
+outID="result"
+
+echo "$PWD"
+
+echo "testID = $testID"
+echo "species = $species"
+echo "substitutionModel = $substitutionModel"
+echo "mutabilityModel = $mutabilityModel"
+echo "clonal = $clonal"
+echo "fixIndels = $fixIndels"
+echo "region = $region"
+echo "inputs = ${inputs[@]}"
+echo "IDs = ${IDs[@]}"
+echo "ref = $ref"
+echo "output = $output"
+echo "outID = $outID"
+
+fasta="$PWD/baseline.fasta"
+
+
+count=0
+for current in ${inputs[@]}
+do
+	f=$(file $current)
+	zipType="Zip archive"
+	if [[ "$f" == *"Zip archive"* ]] || [[ "$f" == *"XZ compressed data"* ]]
+	then
+		id=${IDs[$count]}
+		echo "id=$id"
+		if [[ "$f" == *"Zip archive"* ]] ; then
+			echo "Zip archive"
+			echo "unzip $input -d $PWD/files/"
+			unzip $current -d "$PWD/$id/"
+		elif [[ "$f" == *"XZ compressed data"* ]] ; then
+			echo "ZX archive"
+			echo "tar -xJf $input -C $PWD/files/"
+			mkdir -p "$PWD/$id/files"
+			tar -xJf $current -C "$PWD/$id/files/"
+		fi
+		filtered="$PWD/filtered_${id}.txt"
+		imgt_1_file="`find $PWD/$id -name '1_*.txt'`"
+		imgt_2_file="`find $PWD/$id -name '2_*.txt'`"
+		echo "1_Summary file: ${imgt_1_file}"
+		echo "2_IMGT-gapped file: ${imgt_2_file}"
+		echo "filter.r for $id"
+		Rscript $dir/filter.r ${imgt_1_file} ${imgt_2_file} "$selection" $filtered 2>&1
+		
+		final="$PWD/final_${id}.txt"
+		cat $filtered | cut -f2,4,7 > $final
+		python $dir/script_imgt.py --input $final --ref $ref --output $fasta --id $id
+	else
+		python $dir/script_xlsx.py --input $current --ref $ref --output $fasta
+	fi
+	count=$((count+1))
+done
+workdir="$PWD"
+cd $dir
+echo "file: ${inputs[0]}"
+#Rscript --verbose $dir/Baseline_Main.r $testID $species $substitutionModel $mutabilityModel $clonal $fixIndels $region ${inputs[0]} $workdir/ $outID 2>&1
+Rscript --verbose $dir/Baseline_Main.r $testID $species $substitutionModel $mutabilityModel $clonal $fixIndels $region $fasta $workdir/ $outID 2>&1
+
+echo "$workdir/${outID}.txt"
+
+rows=`tail -n +2 $workdir/${outID}.txt | grep -v "All sequences combined" | grep -n 'Group' | grep -Eoh '^[0-9]+' | tr '\n' ' '`
+rows=($rows)
+#unset rows[${#rows[@]}-1]
+
+cd $dir
+Rscript --verbose $dir/comparePDFs.r $workdir/${outID}.RData $output ${rows[@]} 2>&1
+cp $workdir/result.txt ${output_table}
+
+
+
+