Checking and save mode

[fireindark707]

Hi, overcast07

I borrowed a part of your code and combined it with lynx to make a script. The function of this script is to collect all the URLs in the website in a loop and query Wayback Machine's API. Only when Wayback Machine does not backup this link, the script will submit a request to Wayback Machine for backup. This makes my backup much faster, and maybe this particular mode can be added to your scripts ~

#!/usr/bin/bash
#spider the website using lynx to get sitemap and cache it using waybackmachine

auth=''
custom_dir=''
pattern=''
exclude_pattern=''

while getopts 'a:f:o:' flag; do
	case "${flag}" in
		a)	auth="$OPTARG" ;;
		f)	custom_dir="$OPTARG" ;;
		o)	pattern=$OPTARG ;;
		*)	echo "Usage: $0 [-a auth] [-f file] [-o pattern] [-x exclude_pattern] URL"
			exit 1 ;;
	esac
done
shift "$((OPTIND-1))"

if [[ -n "$custom_dir" ]]; then
	dir="$custom_dir"
	if [[ ! -d "$dir" ]]; then
		mkdir "$dir" || { echo "The folder $dir could not be created"; exit 1; }
		echo "Created data folder $dir"
	else
		echo "Using the existing data folder $dir"
	fi
	cd "$dir"
fi

# get links from input parameter
echo $1 > link.txt

:> cached_urls.txt
:> history.txt

if [[ -f daily_limit.txt ]]; then
    rm daily_limit.txt
elif [[ -f lock.txt ]]; then
    rm lock.txt
fi

export wait_time=1

# recursively crawl the website
while read line
do
    if [[ -f daily_limit.txt ]];
    then
        echo "daily limit reached"
        sleep 3600
    fi
    url_list=$(lynx -accept_all_cookies -dump -listonly -nonumbers $line | grep -E '^https?://.*' | grep -v -E '(#|error|\"|\.jpg|\.png|\.gif|comment|feed|\?share\=|[Ff]acebook|twitter|[Gg]oogle|public-api)|(^https://wordpress\.com)' | sort | uniq)
    echo "$line Crawled"
    for url in $url_list
    do
        if ! grep -q $url history.txt; then
            if ! [[ "$url" =~ $pattern ]]; then
                continue
            fi
            echo $url >> history.txt
            echo $url >> link.txt
            # use the wayback machine API to check if the urls has been cached
            line_encoded=$(python3 -c "import urllib.parse, sys; print(urllib.parse.quote(sys.argv[1]))" "$url")
            if ! curl -s "http://archive.org/wayback/available?url=$line_encoded" | grep -q "available"; then
                echo "$url not cached by Internet Archive yet, will be cached by us"
                while :; do
                    request=$(curl -s -m 60 -X POST --data-urlencode "url=$url" -H "Accept: application/json" -H "Authorization: LOW ${auth}" "https://web.archive.org/save/")
                    message=$(echo "$request" | grep -E -A 1 "<h2>" | tail -1 | sed -Ee 's|</?p>||g')
                    if [[ -z "$message" ]]; then
                        if [[ "$request" =~ "429 Too Many Requests" ]] || [[ "$request" == "" ]]; then
                            echo "$request"
                            if [[ ! -f lock$f.txt ]]; then
                                touch lock.txt
                                sleep 20
                                rm lock.txt
                            fi
                        elif [[ "$request" =~ "400 Bad Request" ]]; then
					        echo "$request"
                            echo "$(date -u '+%Y-%m-%d %H:%M:%S') [Job failed] $url"
                            echo "$(date -u '+%Y-%m-%d %H:%M:%S') $url" >> invalid.log
                            echo "$request" >> invalid.log
                        else
                            sleep 2
                            break
                        fi
                    else
                        echo "$message"
                        if ! [[ "$message" =~ "You have already reached the limit of active sessions" || "$message" =~ "Cannot start capture" || "$message" =~ "Crawling this host is paused" ]]; then
                            if [[ "$message" =~ "You cannot make more than "[1-9][0-9,]*" captures per day" ]]; then
                                touch daily_limit.txt
                                break
                            else
                                echo "$(date -u '+%Y-%m-%d %H:%M:%S') [Job failed] $url"
                                echo "$(date -u '+%Y-%m-%d %H:%M:%S') $url" >> invalid.log
                                echo "$message" >> invalid.log
                                break
                            fi
                        fi
                        if [[ ! -f lock.txt ]]; then
                            touch lock.txt
                            while [[ -f lock.txt ]]; do
                                # Retry the request until either the job is submitted or a different error is received
                                sleep 2
                                request=$(curl -s -m 60 -X POST --data-urlencode "url=$url" -H "Accept: application/json" -H "Authorization: LOW ${auth}" "https://web.archive.org/save/")
                                echo "$(date -u '+%Y-%m-%d %H:%M:%S') [Request failed] $url"
                                message=$(echo "$request" | grep -E -A 1 "<h2>" | tail -1 | sed -Ee 's|</?p>||g')
                                if [[ -z "$message" ]]; then
                                    if [[ "$request" =~ "429 Too Many Requests" ]] || [[ "$request" == "" ]]; then
                                        echo "$request"
                                        sleep 20
                                    else
                                        sleep 5
                                        rm lock.txt
                                        break
                                    fi
                                else
                                    echo "$message"
                                    if [[ "$message" =~ "You have already reached the limit of active sessions" || "$message" =~ "Cannot start capture" || "$message" =~ "Crawling this host is paused" ]]; then
                                        :
                                    elif [[ "$message" =~ "You cannot make more than "[1-9][0-9,]*" captures per day" ]]; then
                                        rm lock.txt
                                        touch daily_limit.txt
                                        break
                                    else
                                        rm lock.txt
                                        echo "$(date -u '+%Y-%m-%d %H:%M:%S') [Job failed] $url"
                                        echo "$(date -u '+%Y-%m-%d %H:%M:%S') $url" >> invalid.log
                                        echo "$message" >> invalid.log
                                        break
                                    fi
                                fi
                            done
                            break
                        fi
                    fi
                done
                echo "$url  $message" >> cached_urls.txt
            fi
            sleep 0.5
        fi
    done
    # sleep to avoid being banned
    if [ "$before" != "$line" ]; then
        sleep 3
    else
        sleep $wait_time
        wait_time=$(echo "$(($wait_time*2))")
        echo "wait for $wait_time"
    fi
    cnt=$(echo "$(($cnt+1))")
    before=$line
done < link.txt

sort history.txt | uniq > temp.txt
mv temp.txt history.txt

[overcast07]

I may consider adding an option not to capture URLs that have already been captured before, but I probably wouldn't include any behavior that's dependent on Lynx since it's not available by default in the macOS command line. In the meantime you could filter the list for URLs before running spn.sh, which might be faster anyway since you could usually get results for a whole domain with the CDX API.

[krackers]

option not to capture URLs that have already been captured before

Can't you get this behavior by using if_not_archived_within with a very long interval like 5 years?

[overcast07]

Can't you get this behavior by using if_not_archived_within with a very long interval like 5 years?

Yes, in theory you should be able to just do that. Sometimes when I use it, though, it causes the script to incorrectly mark the capture as failed because it gets the info in the initial JSON response when submitting the URL, which the script isn't expecting. I should probably fix this.