Given a broken URL, one that doesn't work anymore, finds the lasting working URL (you can choose how far back to go) using Wayback Machine.

healthy-links.sh

#!/usr/bin/env sh
#
# Helps links on your blog/docs to stay healthy.
#
# vim: ft=sh sw=4 ts=4 noet:


APP_NAME=$(basename "$0")
APP_VERSION="1.0.0"
APP_AUTHOR="[email protected]"

DEFAULT_NOW=$(date -u +"%Y%m%d%H%M%S")
OPTION_TIMESTAMP="${WAYBACK_TIMESTAMP:-$DEFAULT_NOW}"

trap cleanup INT TERM EXIT

cleanup()
{
	echo "" > /dev/null
}

sanity_check()
{
	for dependency in jq curl
	do
		command_path="$(command -v $dependency)"
		if [ -z "$command_path" ]
		then
			echo "Error: $dependency is required." >&2
		fi
	done
}

display_help()
{
	cat <<-EOF
	Usage: $APP_NAME [options] <link> [link2 [link3] ...]
	       $APP_NAME --help
	       $APP_NAME --version
	
	Options:
	  -h | --help                  Displays help and exits.
	  -v | --version               Displays version and exits.
	  -x | --debug                 Show commands (-x) before executing them.
	  -t | --timestamp <YYYYmmdd>  Fetch site before given date/time, you
	                               can also specify time with HHiiss format.
	                               (Default: $OPTION_TIMESTAMP)
	
	Environment variables
	  WAYBACK_TIMESTAMP            Used on --timestamp option. If not set,
	                               defaults to "now".
	                               (Default: $DEFAULT_NOW)

	Suggestion and bugs to $APP_AUTHOR.
	EOF
}

parse_options()
{
	while :;
	do
		case ${1:-} in
			-h|--help)
				display_help
				exit 0
				;;
			-x|--debug)
				set -x
				;;
			-v|--version)
				echo "$APP_VERSION"
				exit 0
				;;
			-t|--timestamp)
				OPTION_TIMESTAMP="$2"
				shift
				;;
			--|*)
				break
				;;
		esac
		shift
	done
}

main()
{
	while [ $# -gt 0 ] 
	do
		case "${1:-}" in
			-?*)
				# "--option" parsed elsewhere
				shift
				;;
			*)
				url="$1"
				fetch_archived_url "$url"
				shift
				;;
		esac
	done
}

# Uses Archive.Org to find a working <URL> on Wayback Machine.
# Timestamp is in format <YYYYMMDD>[hhmmss].
#
# Usage: fetch_archived_url <url> [timestamp]
# Example: fetch_archived_url "https://example.org" "20112131245959"
fetch_archived_url()
{
	site="$1"
	timestamp="${2:-$OPTION_TIMESTAMP}"

	url="https://archive.org/wayback/available?url=${site}&timestamp=${timestamp}" 
	curl --silent --location "$url" \
		| jq '.archived_snapshots.closest.url' \
		| sed 's/"//g'
}


sanity_check
parse_options "$@"
main "$@"