rofi_youtube

#!/bin/sh

useragent=${useragent-'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/88.0.4324.152 Safari/537.36'}
tab_space=$(printf '\t')
mpv_args="-shuffle --ytdl-raw-options=sub-lang=en,write-auto-sub=,yes-playlist= --force-window=immediate"
menu_command="rofi -dmenu -i -fuzzy -width 1500 -p \"YouTube ❯\""
new_line="
"

function_exists () {
        type "$1" > /dev/null 2>&1 && return 0 || return 1
}

if ! function_exists 'video_info_text'; then
        video_info_text () {
                printf "%s" "$title"
                printf "%-${title_len}.${title_len}s\t" "$title"
                printf "%-${channel_len}.${channel_len}s\t" "$channel"
                printf "%-${dur_len}.${dur_len}s\t" "$duration"
                printf "%-${view_len}.${view_len}s\t" "$views"
                printf "%-${date_len}.${date_len}s\t" "$date"
                printf "%s" "$shorturl"
                printf "\n"
        }
fi

format_video_data () {
        while IFS=$tab_space read -r title channel views duration date shorturl; do
            video_info_text
        done << EOF
        $*
EOF
        unset title channel duration views date shorturl
}

video_menu () {
        #take input format it to the appropriate format, then pipe it into the menu
        format_video_data "$*" |  eval "$menu_command"
}

get_video_data () {
        # outputs tab and pipe separated fields: title, channel, view count, video length, video upload date, and the video id/url
        # from the videos_json
        printf "%s" "$*" |\
            jq -r '.[]| "\(.title)'"$tab_space"'|\(.channel)'"$tab_space"'|\(.views)'"$tab_space"'|\(.duration)'"$tab_space"'|\(.date)'"$tab_space"'|\(.videoID)"'
}

get_yt_json () {
        # scrapes the json embedded in the youtube html page
        printf "%s" "$*" | sed -n '/var *ytInitialData/,$p' | tr -d '\n' |\
        sed -E ' s_^.*var ytInitialData ?=__ ; s_;</script>.*__ ;'
}

get_yt_html () {
    link=$1
    query=$2
    printf "%s" "$(
        curl "$link" -s \
          -G --data-urlencode "search_query=$query" \
          -G --data-urlencode "sp=$sp" \
          -H 'Authority: www.youtube.com' \
          -H "User-Agent: $useragent" \
          -H 'Accept-Language: en-US,en;q=0.9' \
          -L \
          --compressed
    )"
    unset link query
}

scrape_youtube () {
        yt_html=$(get_yt_html "https://www.youtube.com/results" "$*")

        if [ -z "$yt_html" ]; then
                print_error "ERROR[#01]: Couldn't curl website. Please check your network and try again.\n"
                exit 1
        fi

        yt_json=$(get_yt_json "$yt_html")

        #if the data couldn't be found
        if [ -z "$yt_json" ]; then
                print_error "ERROR[#02]: Couldn't find data on site.\n"
                exit 1
        fi

        #gets a list of videos
        videos_json=$(printf "%s" "$yt_json" | jq '[ .contents|
        ..|.videoRenderer? |
        select(. !=null) |
                {
                        title: .title.runs[0].text,
                        channel: .longBylineText.runs[0].text,
                        duration:.lengthText.simpleText,
                        views: .shortViewCountText.simpleText,
                        date: .publishedTimeText.simpleText,
                        description: .detailedMetadataSnippets[0].snippetText.runs[0].text,
                        videoID: .videoId,
                        thumbs: .thumbnail.thumbnails[0].url
                }
        ]')

        playlist_json=$(printf "%s" "$yt_json" | jq '[ .contents|
        ..|.playlistRenderer? |
        select(. !=null) |
                {
                        title: "[Playlist] \(.title.simpleText)",
                        channel: .longBylineText.runs[0].text,
                        duration: "\(.videoCount) videos",
                        views: "playlist",
                        date: "playlist",
                        videoID: .playlistId,
                        thumbs: .thumbnails[0].thumbnails[0].url
                }
        ]')

        videos_json=$playlist_json$videos_json
        #printf "%s" "$videos_json" > "$tmp_video_json_file"
        printf "%s" "$videos_json" > videos_json.json

        #checks if it's empty in case it was defined in a config function eg: on_get_search
        [ -z "$videos_data" ] && videos_data=$(get_video_data "$videos_json")
        #if there aren't videos
        if [ -z "$videos_data" ]; then
            printf "No results found. Try different keywords.\n" >&2
            exit 1
        fi
        unset playlist_json yt_json yt_html
}

get_playlists() {
    while read -r LINE
    do 
        name_title="$(echo $LINE | gawk -F\| '{print($1)}')"
        echo $name_title
    done < "$videos_data"
}

choice_selection () {

    # Get choice from rofi
    selected_data=$( video_menu "$videos_data")

}

format_data () {
        #for each url append the full url to the $urls string
        #through this loop, the selected data which was truncated by formatting is retrived.
        selected_key="${selected_data%%${new_line}*}"
        selected_urls=
        while read data_line; do
                surl=${data_line##*|}
                [ -z "$surl" ] && continue # to ignore the separators
                #literally just check if there is :// to make sure it's a url (just in case the url doesn't use https or http)
                #peertube urls can differ a lot
                case ${#surl} in
                                        # youtube video
                        11) selected_urls=$selected_urls$new_line'https://www.youtube.com/watch?v='$surl ;;
                        34) selected_urls=$selected_urls$new_line'https://www.youtube.com/playlist?list='$surl ;;
                        36)
                            selected_urls=$selected_urls$new_line"$(printf "%s" "$videos_json" | jq '.[].url' | grep -F "$surl" | tr -d '"')" ;;
                        *) continue ;;
                esac
                refined_selected_data=$refined_selected_data$new_line$(printf '%s' "$videos_data" | grep "|$surl" )
        done << EOF
        $selected_data
EOF

        selected_urls=${selected_urls#$new_line}
        [ -z "$selected_urls" ] && exit # exit if no valid url

        #sometimes % shows up in selected data, could throw an error if it's an invalid directive
        selected_data=${refined_selected_data#$new_line}
        unset shorturls refined_selected_data
}

print_info () {
    printf "$*"
}

main() {

  # Grab the search query from user
  search_query=$(eval "$menu_command")

  # Get the required data using curl
  scrape_youtube "$search_query"

  # Display the data through rofi and capture the selected video option 
  choice_selection

  # Format the video data and extract the youtube link
  format_data

  # Send notification
  notify-send "YouTube - Loading:" "$choice"

  # Spawn mpv
  mpv $mpv_args "$selected_urls" 2>/dev/null 1>/dev/null
}

main &

exit 0