path: root/invidious-dl.sh

#!/bin/sh
# Copyright (c) 2021 Aleksei Kovura
#
# invidious-dl.sh is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# invidious-dl.sh is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with invidious-dl.sh.  If not, see <https://www.gnu.org/licenses/>.
#DBG=1
CACHE_INST="$HOME/.cache/invidious-dl/instances"
UAGENT="Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:53.0) Gecko/20100101 Firefox/53.0"
INSTCS_URL="https://api.invidious.io/instances.json"
OUTDIR=/tmp
[ -n "$INVDL_OUTD" ] && {
	mkdir -pv $INVDL_OUTD 2>/dev/null
	OUTDIR="${INVDL_OUTD%/}"
}
get_rnd_inst() {
	instance_url=""
	for u in $(shuf $CACHE_INST); do
		httpc=0
		printf "Trying instance %s\n" "$u" >&2
		tst_url="${u}/api/v1/trending"
		tst_url2="${u}/api/v1/videos/2l6JUNFAJ9o?fields=title,author"
		httpc=$(curl -k -m 3 -sS -o /dev/null -Iw '%{http_code}' "$tst_url")
		[ $? -gt 0 ] && continue
		httpc2=$(curl -k -m 3 -sS -o /dev/null -s -Iw '%{http_code}' "$tst_url2")
		tst=$(curl -s "$tst_url2")
		dbg "Test json = $tst"
		[ $httpc2 -eq 200 ] &&[ $httpc -eq 200 ] && [ "$tst" != "{}" ] && {
			instance_url="$u"
			break
		}
	done
	[ -z "$instance_url" ] && printf "No working instance\n" && exit 1
	dbg "Using instance = $instance_url"
	printf $instance_url
}
dbg() {
	[ -n "$DBG" ] && printf "%s\n" "$1" >&2
}
dbg $OUTDIR
deps=""
jq --version 1>/dev/null 2>&1 || deps=$deps",jq"
ffmpeg -version 1>/dev/null 2>&1 || deps=$deps",ffmpeg"
curl --version 1>/dev/null 2>&1 || deps=$deps",curl"
[ -n "$deps" ] && printf "Missing dependencies: %s\n" ${deps#,} && exit 1
mkdir -p /tmp/.invidious-dl $HOME/.cache/invidious-dl 2>/dev/null
dl_inst_cache=""
[ -s $CACHE_INST ] || {
	dbg "Instances cache is absent/empty, set dl_inst_cache=1"
	dl_inst_cache=1
}
[ -z "$dl_inst_cache" ] && {
	age=$(( $(date +'%s') - $(stat -c "%Y" --cached=never $CACHE_INST) ))
	age_days=$(( $age / 3600 / 24 ))
	[ $age_days -gt 7 ] && {
		dbg "Instances cache is over 7 days old, set dl_inst_cache=1"
		dl_inst_cache=1
	}
}
[ -n "$dl_inst_cache" ] && {
	printf "Getting a fresh list of instances\n"
	httpc=0
	httpc=$(curl -m 5 -ss -o /dev/null -iw '%{http_code}' "$INSTCS_URL")
	[ ! $httpc -eq 200 ] && {
		printf "Instances API is not accessible: %s\n" "$INSTCS_URL"
		exit 1
	}
	curl -sS "$INSTCS_URL" |
		jq -rM -c '.[] | select(.[1].type=="https" and .[1].stats.software.version) |
			.[1].uri | sub("/$"; "")' > $CACHE_INST
}
URL=""
# <Options handling
while getopts thu:f:Fl name; do
case $name in
	u)
		URL="$OPTARG"
		case "$URL" in
			https://*watch*=*)
				URL_TYPE="video"
				id=${URL#*=}
				;;
			https://youtu.be/*)
				URL_TYPE="video"
				id=${URL#https://youtu.be/}
				;;
			https://www.youtube.com/embed/*)
				URL_TYPE="video"
				id=${URL#https://www.youtube.com/embed/}
				;;
			https://*/playlist\?list=*)
				URL_TYPE="plist"
				id=${URL#*list=}
				;;
			*)
				printf "%s doesn't seem like invidious or youtube video url\n" "$url"
				exit 1
				;;
		esac
		id=${id%%\&*}
		id=${id%%\?*}
		#$url can be:
		#https://yewtu.be/watch?v=IwPsZOBSQTc
		#https://invidious.tube/watch?v=IwPsZOBSQTc
		#https://youtu.be/IwPsZOBSQTc
		#https://www.youtube.com/watch?v=IwPsZOBSQTc
		#https://www.youtube.com/watch?v=IwPsZOBSQTc&feature=youtu.be&t=28m4s
		#https://www.youtube.com/playlist?list=PLsGUcSbWFgXiESsUnTwZXizWixbJvACql
		#https://yewtu.be/playlist?list=PLBS34APvqtnGth6WHL3L-xGZ6ScJPl6Bg
		#https://www.youtube.com/embed/CkXbyklunp4
		;;
	l|F) MODE_LIST=1 ;;
	t) MODE_TITLE=1 ;;
	f)
		MODE_DL=1
		FMT_SPEC="$OPTARG"
		case "$FMT_SPEC" in
			[0-9]*[0-9]|[0-9]*[0-9]+[0-9]*[0-9]) ;;
			*)
				printf "%s format spec doesn't look legit\n" "$FMT_SPEC"
				exit 1
				;;
		esac
		;;
	h|?)
		printf "Usage: %s: [-l/-F]|[-f fmt] [-u URL]\n" ${0##*/}
		printf "  [-l/-F] : List formats\n"
		printf "  [-f] : Download specified formats\n"
		exit 2
		;;
esac
done
[ -z "$URL" ] && {
	printf "Need at least -u URL and either -l/-F or -f. -h to see usage.\n"
	exit 1
}
[ -n "$MODE_LIST" ] && [ -n "$MODE_DL" ] && {
	printf "Can't use both -l/-F and -f. -h to see usage.\n"
	exit 1
}
[ "$URL_TYPE" = "plist" ] && [ -z "$MODE_LIST" ] && {
	printf "Only list mode (-l/-F) is supported for playlists.\n"
	exit 1
}
[ -z "$MODE_DL" ] && [ -z "$MODE_TITLE" ] && [ -z "$MODE_LIST" ] && {
	printf "Pass at least -l/-F or -f or -t. -h to see usage.\n"
	exit 1
}
# Options handling >
dbg "vid = $id"
dbg "MODE_DL = $MODE_DL"
dbg "URL_TYPE = $URL_TYPE"
dbg "MODE_LIST = $MODE_LIST"
DL_JSON=""
jsonf=/tmp/.invidious-dl/invidious-dl_${id}.json
[ -s "$jsonf" ] || {
	dbg "JSON not cached, setting DL_JSON=1"
	DL_JSON=1
}
# vids only: check google URLs expiration date in already downloaded JSON
[ -z "$DL_JSON" ] && [ -s "$jsonf" ] && [ "$URL_TYPE" = "video" ] && {
	dbg "found cached ${jsonf}, checking expiration"
	exp=$(awk '{match($0, /.*expire=([0-9]+).*/, arr); print arr[1]; exit}' $jsonf)
	[ -z "$exp" ] && {
		printf "No expiration date in %s, unreleased video? Re-trying\n" $jsonf >&2
		DL_JSON=1
	}
	cur=$(date +'%s')
	dbg "Cur dt = ${cur}, exp dt = ${exp}"
	[ -z $DL_JSON ] && [ $exp -le $cur ] && {
		printf "Cached %s expired, re-downloading\n" $jsonf >&2
		DL_JSON=1
	}
}
[ -n "$DL_JSON" ] && {
	# Pick random working instance on every run to spread load
	instance_url=$(get_rnd_inst)
	dbg "instance_url = $instance_url"
	api_url=""
	dbg "URL_TYPE = $URL_TYPE"
	case "$URL_TYPE" in
		video)
			api_url="${instance_url}/api/v1/videos/"$id"?fields="
			api_url="$api_url""title,author,adaptiveFormats,formatStreams"
			;;
		plist)
			api_url="${instance_url}/api/v1/playlists/"$id"?fields=title,author,videos"
			;;
	esac
	dbg "api_url = $api_url"
	curl --silent --show-error "$api_url" > $jsonf
	[ $? -eq 0 ] || exit 1
}
[ -n "$MODE_TITLE" ] && [ "$URL_TYPE" = "video" ] && {
	jq -crM '.title' $jsonf
	exit 0
}
[ -n "$MODE_LIST" ] && [ "$URL_TYPE" = "video" ] && {
	jq -crM '(.formatStreams[]| ""
		+ .itag + " "
		+ "combo" + " "
		+ (.qualityLabel//"unkn") + " "
		+ ((.bitrate//0|tonumber)/1024|round|tostring) + "_kbps" + " "
		+ ((.clen//0|tonumber)/1024/1024|round|tostring) + "_Mb" + " "
		+ (.type | gsub("\""; ""))),
		(.adaptiveFormats[]| ""
		+ .itag + " "
		+ "single-" + (if (.type|test("video.*")) then "video" else "audio" end) + " "
		#+ "single" + .type + " "
		+ (.qualityLabel//"n/a") + " "
		+ ((.bitrate//0|tonumber)/1024|round|tostring) + "_kbps" + " "
		+ ((.clen//0|tonumber)/1024/1024|round|tostring) + "_Mb" + " "
		+ (.type | gsub("\""; "")))' $jsonf
	exit 0
}
[ -n "$MODE_DL" ] && {
	itag1=${FMT_SPEC%%+*}
	dbg "itag1 = $itag1"
	[ "$itag1" != "$FMT_SPEC" ] && {
		itag2=${FMT_SPEC##*+}
		dbg "itag2 = $itag2"
	}
	title=$(jq -rM '.title' $jsonf)
	author=$(jq -rM '.author' $jsonf)
	dbg "Output dir = $OUTDIR"
	ofname_base="$OUTDIR"/$(printf "%s" "$title" |
		tr -s "[:space:]" "_" |
		tr -cd "[:alpha:][:digit:]_")"_id_${id}"
	dbg "ofname_base = $ofname_base"
	printf "author = %s\n" "$author"
	printf "title = %s\n" "$title"
	fmt1_url=$(jq -r --arg itag1 $itag1 '.formatStreams[],.adaptiveFormats[]|
		select(.itag == $itag1)|(.type | gsub("\""; "")) +"_"+ .url' $jsonf)
	printf "%s\n" "${fmt1_url%%_https://*}"
	case "${fmt1_url%%_https://*}" in # e.g. "video/webm; codecs=vp9"
		video/*) ofname=$ofname_base".mkv" ;;
		audio*codecs=mp4a*) ofname=$ofname_base".m4a" ;;
		audio*codecs=opus*) ofname=$ofname_base".ogg" ;;
	esac
	dbg "ofname = $ofname"
	if [ -z $itag2 ]; then {
		#I think sh doesn't have lookahead - can't (.*)https://
		printf "Downloading combo or audio format %s to %s...\n" $itag1 $ofname
		ffmpeg -loglevel info -user_agent "$UAGENT" -multiple_requests 1 \
			-i "https://${fmt1_url##*_https://}" \
			-c:v copy -c:a copy \
			-metadata title="$title" -metadata artist="$author" \
			"$ofname"
		exit $?
	}
	else
		fmt2_url=$(jq -r --arg itag2 $itag2 '.formatStreams[],.adaptiveFormats[]|
			select(.itag == $itag2)|(.type | gsub("\""; "")) +"_"+ .url' $jsonf)
		printf "Downloading 2 formats %s + %s to %s\n" $itag1 $itag2 $ofname
		ffmpeg -loglevel info -user_agent "$UAGENT" -multiple_requests 1 \
			-i "https://${fmt1_url##*_https://}" \
			-i "https://${fmt2_url##*_https://}" \
			-c:v copy -c:a copy \
			-metadata title="$title" -metadata artist="$author" -multiple_requests 1 \
			"$ofname"
		exit $?
	fi
}
[ "$URL_TYPE" = "plist" ] && [ -n "$MODE_LIST" ] && {
	dbg "Printing vids in plist"
	jq -r '.videos[]|"https://yewtu.be/watch?v="+ .videoId +" # "+ .title' $jsonf
}