bash script to use ocrmypdf to convert scanned pdfs without to pdfs with searchable text

convert-pdf-ocr.sh

#!/bin/bash

set -ex

function convertPdf(){
	local file=$1
	
	set -e
	
	echo $file
	local suffix=".backup_ocrmypdf.pdf"
	local backup="$file$suffix"
	
	echo 
	echo 
	echo 
	
	if [[ $file =~ $suffix$ ]]; then
		echo "skip backup file"
		return
	fi
	
	if [[ -f "$backup" ]]; then
		echo "INFO seems already ocred"
	else
		if [[ $(exiftool "$file" | grep -q ^"Creator Tool" | grep -q ": ocrmypdf") ]]; then
			echo "INFO: $file already ocred"
		else
			echo "NEED ocr: $file"
			local ocrTemp=$(dirname "$file")"/.temp_ocrmypdf_"$(basename "$file")
			
			local stdout=$(tempfile)
			local stderr=$(tempfile)
			
			set +e
			docker run --rm  -i \
				--user "$(id -u):$(id -g)" \
				--workdir /data \
				-v "$PWD:/data" \
				jbarlow83/ocrmypdf "$file" "$ocrTemp" 2> $stderr
			local exitCode=$?
			set -e
			
			# https://ocrmypdf.readthedocs.io/en/latest/advanced.html#return-code-policy
			echo $exitCode
			
			if [[ $exitCode == 0 ]]; then
				ln -v "$file" "$backup" || cp -v "$file" "$backup"
				mv -v "$ocrTemp" "$file"	
				return
			fi
			
			if [[ $exitCode == 6 ]]; then
				echo "already ocred"
				
			else
				cat $stderr
			fi
			
			if [[ -f "$ocrTemp" ]]; then
				rm -v "$ocrTemp"
			fi
		fi
	fi
}
export -f convertPdf

cd beruflich
find -type f -name "*.pdf" | xargs -P5 -n1 -I{} bash -c "convertPdf '{}'"