subs/subs

#!/bin/sh
#                     █████            
#                    ▒▒███             
#   █████  █████ ████ ▒███████   █████ 
#  ███▒▒  ▒▒███ ▒███  ▒███▒▒███ ███▒▒  
# ▒▒█████  ▒███ ▒███  ▒███ ▒███▒▒█████ 
#  ▒▒▒▒███ ▒███ ▒███  ▒███ ▒███ ▒▒▒▒███
#  ██████  ▒▒████████ ████████  ██████ 
# ▒▒▒▒▒▒    ▒▒▒▒▒▒▒▒ ▒▒▒▒▒▒▒▒  ▒▒▒▒▒▒  
#
# Watch your youtube subscriptions without a youtube account
# via curl, dmenu, mpv and basic unix commands.
#
# The $SUBS_FILE is a text file containing usernames or channel IDs
# comments and blank lines are ignored.
#
# For more information and examples, see:
# http://github.com/mitchweaver/subs
#
# >> note: this is highly experimental / janky, it can and will break <<
#
# -/-/-/-/- Settings -/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
: "${SUBS_FILE:=~/files/subs.txt}"
: "${SUBS_MENU_PROG:=dmenu -l 18 -p Subs:}"
: "${SUBS:=${XDG_CACHE_HOME:-~/.cache}/subs}"
: "${SUBS_LINKS:=$SUBS/links}"
: "${SUBS_CACHE:=$SUBS/cache}"
: "${SUBS_SLEEP_VALUE:=1}" # raise this if you experience problems
: "${SUBS_DAEMON_INTERVAL:=600}"
# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
SEP=^^^^^ # shouldn't need to change this
# -/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/
export LC_ALL=C    # this speeds things up a bit but can cause
                   # issues for titles in foreign languages
# -/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/-/

die() {
    >&2 printf '%s\n' "$*"
    exit 1
}

usage() {
    die 'Usage: subs [-m no-video] [-g gen-links] [-u update-cache] [-d daemonize] [-n dont-play]'
}

# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
# Synopsis: $SUBS_FILE [txt] -> $SUBS_LINKS [xml links]
#
# Updates local cache of xml subscription links from the
# subscription file containing either usernames or channel ids.
# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
gen_links() {
    :>"$SUBS_LINKS"

    count=0
    total=$(sed -e '/^$/d' -e '/^#/d' <"$SUBS_FILE" | wc -l)
    total=${total##* }

    while read -r line ; do

        # ignore comments and blank lines
        case $line in ''|' '|'#'*) continue ;; esac

        # strip off in-line comments and any trailing whitespace
        line=${line%%#*}
        line=${line%% *}

        count=$(( count + 1 ))

        case $line in
            UC*)
                # YT channel IDs always begin with 'UC' and are 24 chars long
                if [ ${#line} -eq 24 ] ; then
                    printf "[%s/%s] using channel ID '%s' for xml link\n" "$count" "$total" "$line"

                    printf 'https://youtube.com/feeds/videos.xml?%s\n' \
                        "channel_id=$line" >>"$SUBS_LINKS"
                else
                    >&2 printf 'Error: cannot determine channel for %s\n' "$line"
                fi
                ;;
            *)
                # otherwise we are given a username, we must find out its channel ID
                printf "Fetching channel ID for %s..." "$line"

                data=$(curl -sL --retry 10 "https://youtube.com/user/$line/about")
                if printf '%s\n' "$data" | grep '404 Not Found' >/dev/null ; then
                    >&2 printf '\n[ERROR]: Could not determine channel for %s... 404\n' "$line"
                    >&2 printf '[%s] %s\n' "$(date)" "$line" >> "$SUBS"/ERRORS.log
                    return 1
                fi

                printf '%s\n' "$data" | \
                while read -r line ; do
                    case $line in
                        *channel/UC??????????????????????*)
                            printf ' Found!\n'
                            line=${line##*channel/}
                            line=${line%%\"*}
                            printf "[%s/%s] using channel ID '%s' for xml link\n" "$count" "$total" "$line"
                            printf 'https://youtube.com/feeds/videos.xml?channel_id=%s\n' \
                                "$line" >>"$SUBS_LINKS"
                            break
                    esac
                done
        esac

    done <"$SUBS_FILE"
}

# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
# Synopsis: $1 [LINK] -> $SUBS_CACHE/$chan_name/concat [CHANNEL INFO]
#
# Takes a channel rss feed link and creates a file
# with a line of its videos dates, titles, and urls.
# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
get_vids() {
    data=$(curl -sL --retry 15 "$1")

    if printf '%s\n' "$data" | grep '404 Not Found' >/dev/null ; then
        >&2 printf '[ERROR]: Could not get vids for %s... 404\n' "$1"
        >&2 printf '[%s] %s\n' "$(date)" "$1" >> "$SUBS"/ERRORS.log
        return 1
    fi

    # hide the first <published> tag which is the channel
    # creation date
    data=${data#*\<\/published\>}

    # trim off outer <name> tags
    chan=${data%%</name*}
    chan=${chan##*name>}

    printf "%s\n" "$data" | \
    while read -r line ; do
        case $line in
            *'link rel='*)
                line=${line#*href=\"}
                line=${line%\"/\>}
                line=https://${line#*www.}
                url=$line
                ;;
            *'<published>'*)
                line=${line%+00:*}
                line=${line#*<published>}
                date=$line
                ;;
            *'<media:title>'*)
                line=${line%</*}
                line=${line#*:title>}
                title=$line
                printf '%s\n' \
                    "${date}${SEP}${chan}${SEP}${title}${SEP}${url}" \
                    >>"$SUBS_CACHE/$chan"
        esac
    done
}

# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
# Updates the local cache of subscriptions. ([-u] flag)
# shellcheck disable=2086
# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
update_subs() {
    [ -f "$SUBS_LINKS" ] || die 'Subs links have not been generated.'

    rm -r "${SUBS_CACHE:-?}" 2>/dev/null ||:
    mkdir -p "$SUBS_CACHE"

    total=$(wc -l <"$SUBS_LINKS")
    total=${total##* }

    count=0
    while read -r link ; do
        count=$(( count + 1 ))
        printf 'starting job [%s/%s] for %s\n' "$count" "$total" "$link"
        get_vids "$link" &
        sleep "${SUBS_SLEEP_VALUE:-0}"
    done <"$SUBS_LINKS"

    count=0
    max_retries=$total
    while [ "$count" -ne "$total" ] && [ "$max_retries" -ne "$total" ] ; do
        count=$(printf '%s\n' "$SUBS_CACHE"/* | wc -l)
        count=${count##* }
        printf "[%s/%s] waiting for fetch jobs to complete...\n"  "$count" "$total"
        sleep 1
        max_retries=$(( max_retries + 1 ))
    done

    printf '%s\n\n' 'done!'
}

# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
# Grab current cache of subscriptions, sort by date uploaded
# -*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
cat_subs() {
    sort -r "$SUBS_CACHE"/* | \
    while read -r line ; do
        chan=${line#*$SEP}
        chan=${chan%%$SEP*}
        title=${line#*$chan$SEP}
        title=${title%%$SEP*}
        date=${line%%$SEP*}
        date=${date#*-}
        date=${date%T*}
        printf '[%s %s] %s\n' "$date" "$chan" "$title"
    done
}

# Split the concatenated lines into entities, send to menu program.
# Finally, play the result with mpv.
get_sel() {
    if [ -d "$SUBS_CACHE" ] ; then
        # Pipe your subs feed into your desired menu and replace HTML escape codes (&quot; -> ")
        sel=$(cat_subs | sed -e 's/&quot;/"/g' -e 's/&amp;/\&/g' -e 's/&lt;/</g' -e 's/&gt;/>/g' | $SUBS_MENU_PROG)
    else
        die 'Subs cache has not been retrieved.'
    fi

    [ "$sel" ] || die Interrupted

    oldchan="${sel#* }"
    chan=$(printf '%s' "${oldchan%%] *}" | sed -e 's/\&/&amp;/g' -e 's/"/\&quot;/g'  -e 's/</\&lt;/g' -e 's/>/\&gt;/g' )
    title=$(printf '%s' "${sel#*"${oldchan%%] *}"\] }" | sed -e 's/\&/&amp;/g' -e 's/"/\&quot;/g'  -e 's/</\&lt;/g' -e 's/>/\&gt;/g' )
    while read -r line ; do
        case $line in
            *"$SEP$title$SEP"*)
                url=${line##*$SEP}
                if [ "$url" ] ; then
                    if [ "$DONT_PLAY" ] ; then
                        # just print the url
                        # (useful for piping to other programs)
                        printf '%s\n' "$url"
                    else
                        printf 'playing: %s\n' "$url"
                        # shellcheck disable=2086
                        exec mpv $MPV_OPTS "$url"
                    fi
                fi
                break
        esac
    done <"$SUBS_CACHE/$chan"
}

daemonize() {
    # create a cached copy of the subs file to check for changes
    # if changes occur, re-generate links automatically
    daemon_file=${XDG_CACHE_HOME:-~/.cache}/subs_daemon.cache
    if [ ! -f "$daemon_file" ] ; then
        cp -f "${SUBS_FILE:=~/files/subs.txt}" "$daemon_file"
        gen_links
    fi

    while true ; do
        if ! cmp "${SUBS_FILE:=~/files/subs.txt}" "$daemon_file" ; then
            gen_links
            cp -f "${SUBS_FILE:=~/files/subs.txt}" "$daemon_file"
        fi
        update_subs
        printf 'Sleeping for %s seconds...\n' "$SUBS_DAEMON_INTERVAL"
        sleep "$SUBS_DAEMON_INTERVAL"
    done
}

main() {
    mkdir -p "$SUBS"

    if [ "$1" ] ; then
        case ${1#-} in
            g)
                gen_links
                ;;
            u)
                update_subs
                ;;
            c)
                cat_subs
                ;;
            d)
                daemonize
                ;;
            m)
                MPV_OPTS="$MPV_OPTS --no-video" \
                get_sel
                ;;
            n)
                DONT_PLAY=true \
                get_sel
                ;;
            *)
                usage
        esac
    else
        get_sel
    fi
}

main "$@"