comparator

#!/bin/bash

#To be able to efficiently use all the files, we have to take 1 pattern line from
#metadata file A and compare it with all the lines in the metadata file B. If there
#is a match, then the file is conserved and therefore, not written into the
#difference.txt file containing the modifications. If there are mo matches,
#The missing file is written into the .txt
#Consider : A=databaseA B=databaseB S=databaseS R=databaseRAM

#Usage : compare databaseA databaseS databaseB

compare()
{
	if [ $# -lt 3 ] ; then
		echo "Not enough arguments, 4 files are required"
   		echo "Usage : compare databaseA DB databaseB"
   		echo "DB is the database of the last sync"
	fi
	if [ ! -f "$1" ] || [ ! -f "$2" ] || [ ! -f "$3" ]; then
		echo "Not a file"
	fi

	#Path to temp databaseRAM (equivalent of databaseS)
	pathToDataBaseRAM="/dev/shm/syncTmpDataBaseS"
	pathToA=`sed '1q;d' $2`
	pathToB=`sed '2q;d' $2`

	echo "DB A :"
	cat $1
	echo "DB B :"
	cat $3

	#Remove the 2 first lines (which contains the path to the files in A and B)
	tail -n +3 $2 > "$pathToDataBaseRAM"
	#Iterate in R
	while IFS= read -r line2; do
		echo " "
		echo "DB RAM : ""$line2"
		#if [[ "$(echo "$line2" | awk '{print $2}')" -eq "directory" ]]; then
		#	line2=$(echo "$line2" | awk '{print $2}')
		#fi
		#get line number of match btw A and R (actual from while)
		nline1=$(awk -v var="$line2" '$0 == var {print NR}' $1)
		nline1=$((0 + nline1)) #add a 0 in case of no return

		nline3=$(awk -v var="$line2" '$0 == var {print NR}' $3)
		nline3=$((0 + nline3)) #add a 0 in case of no return

		echo nline1 = $nline1 nline3 = $nline3

		if [[ $nline1 -ne 0 ]] && [[ $nline3 -ne 0 ]]; then #no sync to do
			echo Already sync
			sed -i -e "${nline1}d" $1 #deleting the corresponding line in databaseA
			sed -i -e "${nline3}d" $3 #deleting the corresponding line in databaseB
		elif [[ $nline1 -eq 0 ]] && [[ $nline3 -eq 0 ]]; then #both are equal to 0 : conflict
			echo "File has changed in A and B"

			#searching if the file has been modified or deleted :
			#get the file name + path
			fileline2=`echo $line2 | head -n1 | awk '{print $1}'`
			#search if the file name + path is in B
			nline3=`awk -v var="$fileline2" '$1 == var{print NR}' $3`
			nline3=$((0 + nline3)) #add a 0 in case of no return
			#searching if the file has been modified or deleted :
			#search if the file name + path is in A
			nline1=`awk -v var="$fileline2" '$1 == var{print NR}' $1`
			nline1=$((0 + nline3)) #add a 0 in case of no return
			echo nline1 $nline1 nline3 $nline3

			if [[ $nline1 -eq 0 ]] && [[ $nline3 -eq 0 ]]; then #both deleted
				#Remove file in #!/bin/sh
				sed -i "/$line2/d" $2 #delete the line in S
			elif [[ $nline1 -eq 0 ]]; then #file deleted in A and modified in B
				echo "file deleted in A and modified in B"
				echo "Copy file B to A"
				#Copy file B to A + upate S
				cp -rp "$pathToB""$fileline2" "$pathToA""$fileline2" #> /dev/null #don't display error from cp
				#(usualy file not found) ???
				sed -i "/$line2/d" $2 #delete the line in S
				sed "${nline3}q;d" $3 >> "$2" #save in S the new file with new metadata
			elif [[ $nline3 -eq 0 ]]; then #file deleted in B and modified in A
				echo "file deleted in B and modified in A"
				echo "Copy file A to B"
				#Copy file B to A + upate S
				cp -rp "$pathToA""$fileline2" "$pathToB""$fileline2" #> /dev/null #don't display error from cp
				#(usualy file not found) ???
				sed -i "/$line2/d" $2 #delete the line in S
				sed "${nline1}q;d" $1 >> "$2" #save in S the new file with new metadata
			else
				echo "Conflict"
			  #name, nameline, line, databaseA, databaseS, databaseB, pathToA, pathToB
			  modDateA=$(cat $1 | awk -v var=$nline1 '{if (NR==var) print}' | awk '{print $7}' | sed "s/-//g")
			  modDateB=$(cat $3 | awk -v var=$nline3 '{if (NR==var) print}' | awk '{print $7}' | sed "s/-//g")
			  echo modDateA: $modDateA
			  echo modDateB: $modDateB

			  if (( $(echo "$modDateA > $modDateB" | bc -l) )); then
			    mv "$pathToA""$fileline2" "$pathToA""$fileline2""_conflict"
			    #Rename the file with the most recent mod date to file_conflict in the metadata
			    conflictLine="${1/"$fileline2"/"$fileline2"_conflict}"
			    #Update the DB with both metadata, the file_conflict and the file.
			    echo "$conflictLine" >> $pathToDataBaseRAM
			    awk '{if (NR=='$nline3') print $0}' $3 >> $pathToDataBaseRAM
			  elif (( $(echo "$modDateA < $modDateB" | bc -l) )); then
			    #Rename the file with the most recent mod date in its arb
			    mv "$pathToB""$fileline2" "$pathToB""$fileline2"_conflict
			    #Rename the file with the most recent mod date to file_conflict in the metadata
			    conflictLine="${3/"$fileline2"/"$fileline2"_conflict}"
			    #Update the DB with both metadata, the file_conflict and the file.
			    echo "$conflictLine" >> $pathToDataBaseRAM
			    awk '{if (NR=='$nline1') print $0}' $1 >> $pathToDataBaseRAM
			  else
			    modTimeA=$(cat $1 | awk -v var=$nline1 '{if (NR==var) print}' | awk '{print $8}' | sed "s/://g")
			    echo modTimeA: $modTimeA
			    #Find the string at the line $nameline inside B
			    modTimeB=$(cat $3 | awk -v var=$nline3 '{if (NR==var) print}' | awk '{print $8}' | sed "s/://g")
			    echo modTimeB: $modTimeB
			    if (( $(echo "$modTimeA > $modTimeB" | bc -l) )); then
			      #Rename the file with the most recent mod date in its arb
			      mv "$pathToA""$fileline2" "$pathToA""$fileline2""_conflict"
			      #Rename the file with the most recent mod date to file_conflict in the metadata
						#Change the filename in metadata
						sed -i "s/"$fileline2"/"$fileline2"_conflict/g" $1
						sed "${nline1}q;d" $1 >> "$2"
			      awk '{if (NR=='$nline3') print $0}' $3 >> $pathToDataBaseRAM
			    else
			      #Rename the file with the most recent mod date in its arb
			      mv "$pathToB""$fileline2" "$pathToB""$fileline2""_conflict"
			      #Rename the file with the most recent mod date to file_conflict in the metadata
						sed -i "s/"$fileline2"/"$fileline2"_conflict/g" $3
						sed "${nline1}q;d" $3 >> "$2"
			      awk '{if (NR=='$nline1') print $0}' $1 >> $pathToDataBaseRAM
			    fi
			  fi
		  fi
		elif [[ $nline3 -eq 0 ]]; then #file has changed in B (A == R)
			echo "File has changed in B"
			#searching if the file has been modified or deleted :
			#get the file name + path
			fileline2=`echo $line2 | head -n1 | awk '{print $1}'`
			#search if the file name + path is in B
			nline3=`awk -v var="$fileline2" '$1 == var{print NR}' $3`
			nline3=$((0 + nline3)) #add a 0 in case of no return
			if [[ $nline3 -ne 0 ]]; then #file + path found ==> B has been modified
				echo "The file still exist but has changed"
				#Copy file B to A + upate S
				cp -rp "$pathToB""$fileline2" "$pathToA""$fileline2" #> /dev/null #don't display error from cp
				#(usualy file not found) ???
				sed -i "/$line2/d" $2 #delete the line in S
				sed "${nline3}q;d" $3 >> "$2" #save in S the new file with new metadata
			else
				echo "The file has been deleted"
				#Delete the file in A + update S
				rm -r "$pathToA""$fileline2" #> /dev/null #don't display error from rm
				#(usualy file not found)
				sed -i "/$line2/d" $2 #delete the line in S
			fi
			fileline2="$fileline2"" "
			sed -i "/$fileline2/d" $1 #delete the line in DB A
			sed -i "/$fileline2/d" $3 #delete the line in DB B
		elif [[ $nline1 -eq 0 ]]; then #file has changed in A (B == R)
			echo "File has changed in A"
			#searching if the file has been modified or deleted :
			#get the file name + path
			fileline2=`echo $line2 | head -n1 | awk '{print $1}'`
			#search if the file name + path is in A
			nline1=`awk -v var="$fileline2" '$1 == var{print NR}' $1`
			nline1=$((0 + nline1)) #add a 0 in case of no return
			if [[ $nline1 -ne 0 ]]; then #file + path found ==> A has been modified
				echo "The file still exist but has changed"
				#TODO : copy file A to B + upate S
				cp -rp "$pathToA""$fileline2" "$pathToB""$fileline2" #> /dev/null #don't display error from cp
				#(usualy file not found) ???
				sed -i "/$line2/d" $2 #delete the line in S
				sed "${nline1}q;d" $1 >> "$2" #save in S the new file with new metadata
			else
				echo "The file has been deleted"
				#Delete the file in B + update S
				rm -r "$pathToB""$fileline2" #> /dev/null #don't display error from rm
				#(usualy file not found)
				sed -i "/$line2/d" $2 #delete the line in S
			fi
			fileline2="$fileline2"" "
			sed -i "/$fileline2/d" $1 #delete the line in DB A
			sed -i "/$fileline2/d" $3 #delete the line in DB B
		fi

		#cat $2

	done < "$pathToDataBaseRAM"
	echo ""
		#rm "$pathToDataBaseRAM"

	#read each line of A, get the filename, check if it exists in B.
	#If the file isn't in B, copy it to the DB and to B
	#If the file is in B, shows a conflict.

	while IFS= read -r line1; do #Read A
	  echo "databaseA" : "$line1"
	  name=$(echo "$line1" | head -n1 | awk '{print $1}')

	  nameline=$(awk -v var="$name" '$1 == var{print NR}' $3)

	  nameline=$((0+ nameline))
		if [ $nameline -eq 0 ]; then	#New file in abr1 not present in abr2
			echo New file "$name" detected
			echo "copying from A to B"
			cp -rp "$pathToA""$name" "$pathToB""$name"
			echo "$line1" >> "$2" #Save the line from A into S
			#sed -i "/$line1/d" "$1" #delete the line in A (OPTIONAL)
		else #New file is present in abr2
			echo "File also in B"
				nlineS=$(awk -v var="$line1" 'var == $1{print NR}' $3)
				nlineS=$((0 + nlineS))
				if [ $nlineS -ne 0 ]; then #new file exists in both abr1 and 2, same metadata.
					echo "$line1" >> $2
					#sed -i "/$line1/d" $1
					sed -i "/$nameline/d" $3
				else
					echo "Conflict"
					#name, nameline, line, databaseA, databaseS, databaseB, pathToA, pathToB
					modDateA=$(cat $1 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $7}' | sed "s/-//g")
					modDateB=$(cat $3 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $7}' | sed "s/-//g")
					echo modDateA: $modDateA
					echo modDateB: $modDateB

					if (( $(echo "$modDateA > $modDateB" | bc -l) )); then
						mv "$pathToA""$name" "$pathToA""$name""_conflict"
						#Rename the file with the most recent mod date to file_conflict in the metadata
						sed -i "s/"$name"/"$name"_conflict/g" $1
						sed "${nameline}q;d" $1 >> "$2"
						awk '{if (NR=='$nameline') print $0}' $3 >> $pathToDataBaseRAM
					elif (( $(echo "$modDateA > $modDateB" | bc -l) )); then
						#Rename the file with the most recent mod date in its arb
						mv "$pathToB""$name" "$pathToB""$name""_conflict"
						sed -i "s/"$name"/"$name"_conflict/g" $3
						sed "${nameline}q;d" $3 >> "$2"
						awk '{if (NR=='$nameline') print $0}' $1 >> $pathToDataBaseRAM
					else
						modTimeA=$(cat $1 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $8}' | sed "s/://g")
						echo modTimeA: $modTimeA
						#Find the string at the line $nameline inside B
						modTimeB=$(cat $3 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $8}' | sed "s/://g")
						echo modTimeB: $modTimeB
						if (( $(echo "$modTimeA > $modTimeB" | bc -l) )); then
							#Rename the file with the most recent mod date in its arb
							mv "$pathToA""$name" "$pathToA""$name""_conflict"
							#Rename the file with the most recent mod date to file_conflict in the metadata
							sed -i "s/"$name"/"$name"_conflict/g" $1
							sed "${nameline}q;d" $1 >> "$2"
							awk '{if (NR=='$nameline') print $0}' $3 >> $pathToDataBaseRAM
						else
							#Rename the file with the most recent mod date in its arb
							mv "$pathToB""$name" "$pathToB""$name"_conflict
							#Rename the file with the most recent mod date to file_conflict in the metadata
							conflictLine="${3/"$name"/"$name"_conflict}"
							sed -i "s/"$name"/"$name"_conflict/g" $3
							sed "${nameline}q;d" $3 >> "$2"
							awk '{if (NR=='$nameline') print $0}' $1 >> $pathToDataBaseRAM
						fi
					fi
				fi
		fi
		echo ""
	done < "$1"

	while IFS= read -r line3; do	#Read B
		echo "databaseB" : "$line3"
		name=$(echo "$line3" | head -n1 | awk '{print $1}')
		nameline=$(awk -v var="$name" '$1 == var{print NR}' $1)
		nameline=$((0+ nameline))
		if [ $nameline -eq 0 ]; then	#New file in abr1 not present in abr2
			echo New file "$name" detected
			echo "copying from B to A"
			cp -rp "$pathToB""$name" "$pathToA""$name"

			echo $line3 >> "$2"#Save the line from B into S

			#sed -i "/{$line3}/d" "$3" #delete the line in B (OPTIONAL)

		else #New file is present in abr2
				echo "File also in A"
				nlineS=$(awk -v var="$line1" 'var == $1{print NR}' $3)
				nlineS=$((0 + nlineS))
				if [ $nlineS -ne 0 ]; then #new file exists in both abr1 and 2, same metadata.
					echo "$line3" >> $2
					sed -i "/$line3/d" "$1"
					#sed -i "/$line/d" "$3"
				else
					echo "Conflict"
					#name, nameline, line, databaseA, databaseS, databaseB, pathToA, pathToB
					modDateA=$(cat $1 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $7}' | sed "s/-//g")
					modDateB=$(cat $3 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $7}' | sed "s/-//g")
					echo modDateA: $modDateA
					echo modDateB: $modDateB
					if (( $(echo "$modDateA > $modDateB" | bc -l) )); then
						mv "$pathToA""$name" "$pathToA""$name""_conflict"
						sed -i "s/"$name"/"$name"_conflict/g" $1
						sed "${nameline}q;d" $1 >> "$2"
						awk '{if (NR=='$nameline') print $0}' $3 >> $pathToDataBaseRAM
					elif (( $(echo "$modDateA > $modDateB" | bc -l) )); then
						#Rename the file with the most recent mod date in its arb
						mv "$pathToB""$name" "$pathToB""$name"_conflict
						sed -i "s/"$name"/"$name"_conflict/g" $3
						sed "${nameline}q;d" $3 >> "$2"
						awk '{if (NR=='$nameline') print $0}' $1 >> $pathToDataBaseRAM
					else
						modTimeA=$(cat $1 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $8}' | sed "s/://g")
						echo modTimeA: $modTimeA
						#Find the string at the line $nameline inside B
						modTimeB=$(cat $3 | awk -v var=$nameline '{if (NR==var) print}' | awk '{print $8}' | sed "s/://g")
						echo modTimeB: $modTimeB
						if (( $(echo "$modTimeA > $modTimeB" | bc -l) )); then
							#Rename the file with the most recent mod date in its arb
							mv "$pathToA""$name" "$pathToA""$name""_conflict"
							#Rename the file with the most recent mod date to file_conflict in the metadata
							sed -i "s/"$name"/"$name"_conflict/g" $1
							sed "${nameline}q;d" $1 >> "$2"
							awk -v var="$nameline" '{if (NR==var) print $0}' $3 >> $pathToDataBaseRAM
						else
							#Rename the file with the most recent mod date in its arb
							mv "$pathToB""$name" "$pathToB""$name""_conflict"
							#Rename the file with the most recent mod date to file_conflict in the metadata
							sed -i "s/"$name"/"$name"_conflict/g" $3
							sed "${nameline}q;d" $3 >> "$2"
							awk -v var="$nameline" '{if (NR==var) print $0}' $1 >> $pathToDataBaseRAM
						fi
					fi
				fi
		fi
		echo ""
done < "$3"
}