Backup cleanup algorithm test

backup_test.sh

#!/bin/bash
# This script will simulate an infinite backup cycle per day.
# Empty backup files are created in the current working directory
# and deleted acording to the cleanup algorithm parameters.
# One file per second is created, which is one simulated day.
# => It is best to run this script form inside an empty test directory.
# This script will not halt and must be aborted!
# The same algorithm is used in https://github.com/bit-team/backintime
set -o pipefail

KEEP_LAST=2
KEEP_DAY=7
KEEP_WEEK=4
KEEP_MONTH=24

function err() { echo "$@" 1>&2; }

function sscanf() {
    local str="$1"
    if [[ "$str" =~ ^([0-9]+)-([0-9]+)-([0-9]+)-([0-9]+)-([0-9]+).inc.backup$ ]]; then
        HOUR=${BASH_REMATCH[1]}
        DAY=${BASH_REMATCH[2]}
        MONTH=${BASH_REMATCH[3]}
        YEAR=${BASH_REMATCH[4]}
        WEEK=${BASH_REMATCH[5]}
        return
    fi
    false
}

function date_ts_from_ymd() {
    local YEAR="$1"
    local MONTH="$2"
    local DAY="$3"
    date -d "$YEAR-$MONTH-$DAY 00:00:00" "+%s"
}

function date_from_ts() {
    local TS="$1"
    local FORMAT="$2"
    local OTHER="$3"
    date -d "$(date -d @$TS "+%Y-%m-%d %H:%M:%S") $OTHER" "$FORMAT"
}

function date_in_range() {
    local TS="$1"
    local MIN="$2"
    local MAX="$3"
    (( MIN <= TS && TS < MAX ))
}

D=0
CURRENT_DATE=$(date "+%s")

while true; do

    D="$(( D + 1 ))"

    CURRENT_HOUR=$(date_from_ts "$CURRENT_DATE" "+%s" "$D days")
    CURRENT_DAY=$(date_from_ts "$CURRENT_DATE" "+%d" "$D days")
    CURRENT_WEEK=$(date_from_ts "$CURRENT_DATE" "+%V" "$D days")
    CURRENT_MONTH=$(date_from_ts "$CURRENT_DATE" "+%m" "$D days")
    CURRENT_YEAR=$(date_from_ts "$CURRENT_DATE" "+%Y" "$D days")

    FILE="$CURRENT_HOUR-$CURRENT_DAY-$CURRENT_MONTH-$CURRENT_YEAR-$CURRENT_WEEK.inc.backup"

    function date_current_ts() {
        local OTHER="$1"
        local FROM="$2"
        if [[ -z "$FROM" ]]; then
            FROM="$CURRENT_HOUR"
        fi
        date_from_ts "$FROM" "+%s" "$OTHER"
    }

    function find_snapshots() {
        local MIN_YEAR=9999
        while IFS= read -r -d $'\0' file; do
            THE_FILE=$(basename $file)
            if sscanf "$THE_FILE"; then
                if (( MIN_YEAR > YEAR )); then
                    MIN_YEAR=$YEAR
                fi
                echo "$THE_FILE"
            fi
        done < <(find $TARGET_DIR -maxdepth 1 -type f  -print0 | sort -Vz)
        echo $MIN_YEAR
    }

    function find_remove() {
        local TARGET_DIR="$1"
        local MIN=0
        local MAX=0
        local CURRENT=0
        local CURRENT2=0
        local SNAPSHOTS=$(find_snapshots)
        local MIN_YEAR=0
        local KEEP=""
        local MAP=""
        declare -A MAP
        
        function find_keep() {
            if (( KEEP_LAST > 0 )); then
                MIN=$(date_current_ts "$(( KEEP_LAST )) days ago")
                MAX=$(date_current_ts "1 days")
                for file in $SNAPSHOTS; do
                    sscanf "$file"
                    if date_in_range $HOUR $MIN $MAX; then
                        echo "$file"
                    fi
                    MIN_YEAR="$file"
                done
            fi
                
            if (( KEEP_DAY > 0 )); then
                for (( j=0; j < $KEEP_DAY; j++ )); do
                    MIN=$(date_current_ts "$j days ago")
                    MAX=$(date_current_ts "1 days" $MIN)
                    for file in $SNAPSHOTS; do
                        sscanf "$file"
                        if date_in_range $HOUR $MIN $MAX; then
                            echo "$file"
                            break
                        fi
                    done
                done
            fi
            
            if (( KEEP_WEEK > 0 )); then
                local WEEK_DAY=$(date_from_ts $(date_current_ts "") "+%u")
                CURRENT=$(date_current_ts "$WEEK_DAY days ago")
                for (( j=0; j < KEEP_WEEK; j++ )); do
                    MIN=$CURRENT
                    MAX=$(date_current_ts "8 days" $MIN)
                    for file in $SNAPSHOTS; do
                        sscanf "$file"
                        if date_in_range $HOUR $MIN $MAX; then
                            echo "$file"
                            break
                        fi
                    done
                    CURRENT=$(date_current_ts "7 days ago" $CURRENT)
                done
            fi
            
            if (( KEEP_MONTH > 0 )); then
                CURRENT=$(date_ts_from_ymd "$CURRENT_YEAR" "$CURRENT_MONTH" "1")
                CURRENT2=$(date_current_ts "1 months" $CURRENT)
                for (( j=0; j < KEEP_MONTH; j++ )); do
                    MIN=$CURRENT
                    MAX=$CURRENT2
                    for file in $SNAPSHOTS; do
                        sscanf "$file"
                        if date_in_range $HOUR $MIN $MAX; then
                            echo "$file"
                            break
                        fi
                    done
                    CURRENT2=$CURRENT
                    CURRENT=$(date_current_ts "1 months ago" $CURRENT)
                done
            fi
            
            for (( j=MIN_YEAR; j <= CURRENT_YEAR; j++ )); do
                MIN=$(date_ts_from_ymd "$j" "1" "1")
                MAX=$(date_ts_from_ymd "$(( j + 1))" "1" "1")
                for file in $SNAPSHOTS; do
                    sscanf "$file"
                    if date_in_range $HOUR $MIN $MAX; then
                        echo "$file"
                        break
                    fi
                done
            done
        }
        
        KEEP=$(find_keep | sort -u)
        
        for file in $KEEP; do
            MAP["$file"]=1
        done
        
        for file in $SNAPSHOTS; do
            if [[  ${MAP["$file"]} != 1  ]]; then
                if (( ${#file} > 10 )); then
                    echo "$file"
                fi
            fi
        done
    }

    OUT_FILE="$FILE"

    if [[ -f "$OUT_FILE" ]]; then
        err "$FILE exists, try later."
        exit 1
    fi

    #date_from_ts $(date_current_ts "") "+%Y-%m-%d %H:%M:%S"
    #date_from_ts $(date_current_ts "1 months") "+%Y-%m-%d %H:%M:%S"

    #echo "add $OUT_FILE"
    touch "$OUT_FILE"

    TO_REMOVE=$(find_remove .)
    for line in $TO_REMOVE; do
        #echo "remove: $line"
        rm "$line"
    done
    sleep 1
done