linux/.config/bash/module/Archive.sh

export _archive_pattern="_[0-9]{12}-[[:alnum:]]{40}.t[xg]z$"
export _archive_pattern_fast="_[0-9]{12}-[[:alnum:]]{40}.tgz$"

# Archive directories.
# All directories by default.
# Supports .archiveignore exclude file.
# Usage: archive [DIRS]
function archive() {
	local IFS=$'\n'
	local targets=(${@})
	[[ "${targets}" = "" ]] && targets=($(_ls_dir))

	process() {
		local date=$(_archive_date)

		# Parse name.
		local name=$(parse_pascal ${target})

		# Exclude support.
		local exclude=""
		[[ -f ".archiveignore" ]] && exclude="--exclude-from=.archiveignore"
		[[ -f "${target}/.archiveignore" ]] && exclude="--exclude-from=${target}/.archiveignore"

		# create archive.
		local hash=$(tar ${exclude} -c ${target} | pv -s $(/usr/bin/env du -sb ${target} | awk '{print $1}') | xz -9e | tee ${name}.txz | sha1sum | cut -d\  -f1)

		# append hash to target name.
		local new_name="${name}_${date}-${hash}.txz"
		mv -- ${name}.txz ${new_name} && echo ${new_name}
	}

	_iterate_targets process ${targets[@]}
}

# Archive using multiple threads. Uses 75% of free RAM.
# All directories by default.
# Supports .archiveignore exclude file.
# Usage: archive_mt [DIRS]
function archive_mt() {
	local IFS=$'\n'
	local targets=(${@})
	[[ "${targets}" = "" ]] && targets=($(_ls_dir))

	process() {
		local date=$(_archive_date)

		# Parse name.
		local name=$(parse_pascal ${target})

		# Exclude support.
		local exclude=""
		[[ -f ".archiveignore" ]] && exclude="--exclude-from=.archiveignore"
		[[ -f "${target}/.archiveignore" ]] && exclude="--exclude-from=${target}/.archiveignore"

		# Determine memory limit.
		local mem_free=$(_mem_free)
		local mem_limit=$((mem_free*3/4))

		# create archive.
		local hash=$(tar ${exclude} -c ${target} | pv -s $(/usr/bin/env du -sb ${target} | awk '{print $1}') | xz -9e --threads=0 --memlimit=${mem_limit}MiB | tee ${name}.txz | sha1sum | cut -d\  -f1)

		# append hash to target name.
		local new_name="${name}_${date}-${hash}.txz"
		mv -- ${name}.txz ${new_name} && echo ${new_name}
	}

	_iterate_targets process ${targets[@]}
}

# Archive directories with fast compression.
# All directories by default.
# Supports .archiveignore exclude file.
# Usage: archive_fast [DIRS]
function archive_fast() {
	local IFS=$'\n'
	local targets=("${@}")
	[[ "${targets}" = "" ]] && targets=($(_ls_dir))

	process() {
		# Start timestamp.
		local date=$(_archive_date)

		# Parse name.
		local name=$(parse_pascal "${target}")

		# Exclude support.
		local exclude=""
		[[ -f ".archiveignore" ]] && exclude="--exclude-from=.archiveignore"
		[[ -f "${target}/.archiveignore" ]] && exclude="--exclude-from=${target}/.archiveignore"

		# create archive.
		local hash=$(tar ${exclude} -c "${target}" | pv -s $(/usr/bin/env du -sb "${target}" | awk '{print $1}') | gzip -1 | tee "${name}".tgz | sha1sum | cut -d\  -f1)

		# append hash to target name.
		local new_name="${name}_${date}-${hash}.tgz"
		mv -- "${name}".tgz ${new_name} && echo ${new_name}
	}

	_iterate_targets process ${targets[@]}
}

# Check archives integrity.
# Checks all archives by default.
# Usage: archive_check [FILES]
function archive_check() {
	local IFS=$'\n'
	local targets=(${@})
	[[ "${targets}" = "" ]] && targets=($(_ls_archive))

	process() {
		_archive_check "${target}"
	}

	_iterate_targets process ${targets[@]}
}

# Delete old versions of an archive.
# All archives with 1 version by default.
# Usage: archive_prune [NAME] [VERSIONS]
function archive_prune() {
	local IFS=$'\n'
	local targets=(${1})
	local versions=${2}

	[[ "${targets}" = "" ]] && targets=($(_archive_names))
	[[ "${versions}" = "" ]] && versions=1

	if [[ ${#} -gt 2 ]]; then
		help archive_prune
		return 2
	fi

	process() {
		local prune=($(ls | grep -E "^${target}${_archive_pattern}" | sort -r | sed -e "1,${versions}d"))

		for archive in ${prune[@]}; do
			rm -- "${archive}" && echo "${archive}"
		done
	}

	_iterate_targets process ${targets[@]}
}

# Delete specified or all archive files.
# Usage: archive_rm [FILES]
function archive_rm() {
	local IFS=$'\n'
	local targets=(${@})
	[[ "${targets}" = "" ]] && targets=($(_ls_archive))

	process() {
		rm -- "${target}" && echo "${target}"
	}

	_iterate_targets process ${targets[@]}
}

# Recompress previously created archive_fast with better compression.
# Usage: archive_xz [FILES]
function archive_xz() {
	local IFS=$'\n'
	local targets=(${@})
	[[ "${targets}" = "" ]] && targets=$(_ls_archive_fast)

	process() {
		local data=($(_archive_parse "${target}"))
		local tmp="${data[0]}.txz"

		# Check that old format.
		if [[ "${data[3]}" != "tgz" ]]; then
			_error "Not in .tgz format!"
			return 1
		fi

		# Check integrity.
		_archive_check "${target}" || return 1

		# Recompress.
		local hash=$(pv "${target}" | gzip -d | xz -9e | tee "${tmp}" | sha1sum | cut -d\  -f1)

		# Rename.
		local new_name="${data[0]}_${data[1]}-${hash}.txz"
		mv -- ${tmp} ${new_name} && rm ${target} && echo ${new_name}
	}

	_iterate_targets process ${targets[@]}
}

# Rename archives.
# If no name specified, it simplifies archive's name.
# If no archives specified, apply to all archives.
# Usage: archive_name [ARCHIVE] [NAME]
function archive_name() {
	local IFS=$'\n'
	local targets="${1}"
	local name="${2}"
	[[ "${targets}" = "" ]] && targets=($(_ls_archive))

	process() {
		# simplify name by default.
		if [[ "${name}" = "" || ${count} -gt 1 ]]; then
			name="${target%_*}"
			name="$(parse_pascal ${name})"
		fi

		# remove old name.
		local data="${target##*_}"
		local new_name="${name}_${data}"

		# check for the same name.
		[[ "${target}" = "${new_name}" ]] && return 0

		# check for existing target.
		if [[ -f "${new_name}" ]]; then
			_error "${new_name}: Already exists!"
			return 1
		fi

		# rename.
		mv -- ${target} ${new_name} && echo ${new_name}
	}

	_iterate_targets process ${targets[@]}
}

# Extract previously created archive with checksum validation.
# Usage: unarchive [FILES]
function unarchive() {
	local IFS=$'\n'
	local targets=(${@})
	[[ "${targets}" = "" ]] && targets=$(_ls_archive)

	process() {
		# Validate.
		# _archive_check "${target}" || return 1

		# Extract.
		case "${target##*.}" in
			"txz")
				pv ${target} | xz -d | tar -xf -
				;;
			"tgz")
				pv ${target} | gzip -d | tar -xf -
				;;
		esac
	}

	_iterate_targets process ${targets[@]}
}

# Change archive's filesystem time to match creation date.
# Usage: archive_touch [FILES]
function archive_touch() {
	local IFS=$'\n'
	local targets=(${@})
	[[ "${targets}" = "" ]] && targets=$(_ls_archive)

	process() {
		local data=($(_archive_parse "${target}"))
		local date=${data[1]}
		touch -t ${date} -- ${target}
	}

	_iterate_targets process ${targets[@]}
}

# Parse archive file name to get: name, date, hash and format.
# Usage: _archive_parse <FILENAME>
function _archive_parse() {
	local input="${1}"
	local name="${input%_*}"
	local format="${input##*.}"
	local data="${input##*_}"; data="${data%.*}"
	local date="${data%%-*}"
	local hash="${data##*-}"

	echo "${name}"
	echo "${date}"
	echo "${hash}"
	echo "${format}"
}

# Autocomplete for archive_name function.
# First arg is the archives list, second one is selected archive's current name.
function _comp_archive_name() {
	local IFS=$'\n'
	COMPREPLY=()

	local cur="${COMP_WORDS[COMP_CWORD]}"
	local prev="${COMP_WORDS[COMP_CWORD-1]}"
	local command="${COMP_WORDS[0]}"

	if [[ "${prev}" = "${command}" ]]; then
		COMPREPLY=( $(compgen -W "$(ls | grep -E ${_archive_pattern})" -- ${cur}) )
		return 0
	else
		local data=($(_archive_parse ${prev}))
		local name="${data[0]}"
		COMPREPLY=( $(compgen -W "${name}" -- ${cur}) )
		return 0
	fi
}

# Autocomplete with archives in current dir.
function _comp_archive_grep() {
	_autocomplete_grep ${_archive_pattern}
}

# Autocomplete with fast archives in current dir.
function _comp_archive_grep_fast() {
	_autocomplete_grep ${_archive_pattern_fast}
}

# Get date for a new archive.
function _archive_date() {
	date +%Y%m%d%H%M
}

# Get names of all archives.
function _archive_names() {
	local IFS=$'\n'
	local archives=($(_ls_archive))
	local names=()

	for archive in ${archives[@]}; do
		local data=($(_archive_parse ${archive}))
		names+=(${data[0]})
	done

	# Remove copies.
	names=($(printf '%s\n' "${names[@]}" | sort -u))

	printf '%s\n' "${names[@]}"
}

# Autocomplete with names of all archives.
function _comp_archive_names() {
	_autocomplete_first $(_archive_names)
}

# Check if file is an archive.
function _is_archive() {
	local out=$(echo "${*}" | grep -E ${_archive_pattern})

	[[ "${out}" != "" ]]
}

# List all archives.
function _ls_archive() {
	ls | grep -E ${_archive_pattern}
}

# List fast archives.
function _ls_archive_fast() {
	ls | grep -E ${_archive_pattern_fast}
}

# Filter input for archives only.
function _filter_archive() {
	grep -E ${_archive_pattern}
}

function _archive_check() {
	# extract hash from name.
	local data=($(_archive_parse ${target}))
	local saved=${data[2]}

	# calculate actual hash.
	local actual=$(pv ${target} | sha1sum | cut -d\  -f1)

	# compare hashes.
	[[ "${actual}" = "${saved}" ]]
}

complete -o filenames -F _comp_archive_grep        archive_check unarchive archive_rm archive_touch
complete -o filenames -F _comp_archive_grep_fast   archive_xz
complete -o filenames -F _comp_archive_name        archive_name
complete -o filenames -F _comp_archive_names       archive_prune