Shell Script to download youtube files from playlist

Question

I'm trying to write a bash script that will download all of the youtube videos from a playlist and save them to a specific file name based on the title of the youtube video itself. So far I have two separate pieces of code that do what I want but I don't know how to combine them together to function as a unit.

This piece of code finds the titles of all of the youtube videos on a given page:

curl -s "$1" | grep '<span class="title video-title "' | cut -d\> -f2 | cut -d\< -f1

And this piece of code downloads the files to a filename given by the youtube video id (e.g. the filename given by youtube.com/watch?v=CsBVaJelurE&feature=relmfu would be CsBVaJelurE.flv)

curl -s "$1" | grep "watch?" | cut -d\" -f4| while read video; 
do youtube-dl "http://www.youtube.com$video";
done

I want a script that will output the youtube .flv file to a filename given by the title of the video (in this case BASH lesson 2.flv) rather than simply the video id name. Thanks in advance for all the help.

Shooter

This is hands down the best Youtube download script out there: bitbucket.org/rg3/youtube-dl/wiki/Home Since all these scripts relay on scraping pages, it needs to stay up to date on the site structure. It can include video title in the filename. – Hydroxide

Good call Niklas on handing out the dl link for the script; here's a dedicated link with more info, documentation, and developer data: rg3.github.com/youtube-dl. People should definitely download, it's a great script – Shooter

Apparently I wasn't the only one to have this idea github.com/ArielAleksandrus/PlaylistDownloader – Ventricle

Answer 1

OK so after further research and updating my version of youtube-dl, it turns out that this functionality is now built directly into the program, negating the need for a shell script to solve the playlist download issue on youtube. The full documentation can be found here: (http://rg3.github.com/youtube-dl/documentation.html) but the simple solution to my original question is as follows:

1) youtube-dl will process a playlist link automatically, there is no need to individually feed it the URLs of the videos that are contained therein (this negates the need to use grep to search for "watch?" to find the unique video id

2) there is now an option included to format the filename with a variety of options including:

• id: The sequence will be replaced by the video identifier.
• url: The sequence will be replaced by the video URL.
• uploader: The sequence will be replaced by the nickname of the person who uploaded the video.
• upload_date: The sequence will be replaced by the upload date in YYYYMMDD format.
• title: The sequence will be replaced by the literal video title.
• ext: The sequence will be replaced by the appropriate extension (like flv or mp4).
• epoch: The sequence will be replaced by the Unix epoch when creating the file.
• autonumber: The sequence will be replaced by a five-digit number that will be increased with each download, starting at zero.

the syntax for this output option is as follows (where NAME is any of the options shown above):

youtube-dl -o '%(NAME)s' http://www.youtube.com/your_video_or_playlist_url

As an example, to answer my original question, the syntax is as follows:

youtube-dl -o '%(title)s.%(ext)s' http://www.youtube.com/playlist?list=PL2284887FAE36E6D8&feature=plcp

Thanks again to those who responded to my question, your help is greatly appreciated.

Shooter

This is the best answer for sure. Thanks for sharing. Looks like 'stitle' isn't supported anymore. Use 'title' instead; I updated your answer. – Miscreance

Answer 2

If you want to use the title from youtube page as a filename, you could use -t option of youtube-dl. If you want to use the title from your "video list" page and you sure that there is exactly one watch? URL for every <span class="title video-title" title, then you can use something like this:

#!/bin/bash

TMPFILE=/tmp/downloader-$$

onexit() {
  rm -f $TMPFILE
}

trap onexit EXIT

curl -s "$1" -o $TMPFILE

i=0
grep '<span class="title video-title "' $TMPFILE | cut -d\> -f2 | cut -d\< -f1 | while read title; do
  titles[$i]=$title
  ((i++))
done

i=0
grep "watch?" $TMPFILE | cut -d\" -f4 | while read url; do
  urls[$i]="http://www.youtube.com$url"
  ((i++))
done

i=0; while (( i < ${#urls[@]} )); do
  youtube-dl -o "${titles[$i]}.%(ext)" "${urls[$i]}"
  ((i++))
done

I did not tested it because I have no "video list" page example.

Fons

Thanks for the answer praetorian, the -t option works for the time-being but the script you provided has some errors in it that I'll need to check out. Currently the files themselves do not download at all but I haven't yet had the time to test your script and see why. I'll keep this thread updated if I find anything out. – Shooter

Answer 3

this following method work and play you titanic from youtube

youtube-downloader.sh youtube-video-url.sh

#!/bin/bash
decode() {
    to_decode='s:%([0-9A-Fa-f][0-9A-Fa-f]):\\x\1:g'
    printf "%b" `echo $1 | sed 's:&:\n:g' | grep "^$2" | cut -f2 -d'=' | sed -r $to_decode`
}
data=`wget http://www.youtube.com/get_video_info?video_id=$1\&hl=pt_BR -q -O-`
url_encoded_fmt_stream_map=`decode $data 'url_encoded_fmt_stream_map' | cut -f1 -d','`
signature=`decode $url_encoded_fmt_stream_map 'sig'`
url=`decode $url_encoded_fmt_stream_map 'url'`
test $2 && name=$2 || name=`decode $data 'title' | sed 's:+: :g;s:/:-:g'`
test "$name" = "-" && name=/dev/stdout || name="$name.vid"
wget "${url}&signature=${signature}" -O "$name"






#!/usr/bin/env /bin/bash
function youtube-video-url {
    local field=
    local data=
    local split="s:&:\n:g"
    local decode_str='s:%([0-9A-Fa-f][0-9A-Fa-f]):\\x\1:g'
    local yt_url="http://www.youtube.com/get_video_info?video_id=$1"
    local grabber=`command -v curl`
    local args="-sL"
    if [ ! "$grabber" ]; then
        grabber=`command -v wget`
        args="-qO-"
    fi
    if [ ! "$grabber" ]; then
        echo 'No downloader available.' >&2
        test x"${BASH_SOURCE[0]}" = x"$0" && exit 1 || return 1
    fi
    function decode {
        data="`echo $1`"
        field="$2"
        if [ ! "$field" ]; then
            field="$1"
            data="`cat /dev/stdin`"
        fi
        data=`echo $data | sed $split | grep "^$field" | cut -f2 -d'=' | sed -r $decode_str`
        printf "%b" $data
    }
    local map=`$grabber $args $yt_url | decode 'url_encoded_fmt_stream_map' | cut -f1 -d','`
    echo `decode $map 'url'`\&signature=`decode $map 'sig'`
}
[ $SHLVL != 1 ] && export -f youtube-video-url

bash youtube-player.sh saalGKY7ifU

#!/bin/bash
decode() {
    to_decode='s:%([0-9A-Fa-f][0-9A-Fa-f]):\\x\1:g'
    printf "%b" `echo $1 | sed 's:&:\n:g' | grep "^$2" | cut -f2 -d'=' | sed -r $to_decode`
}


data=`wget http://www.youtube.com/get_video_info?video_id=$1\&hl=pt_BR -q -O-`


url_encoded_fmt_stream_map=` decode $data 'url_encoded_fmt_stream_map' | cut -f1 -d','`

signature=` decode $url_encoded_fmt_stream_map 'sig'`

url=`decode $url_encoded_fmt_stream_map 'url'`
test $2 && name=$2 || name=`decode $data 'title' | sed 's:+: :g;s:/:-:g'`


test "$name" = "-" && name=/dev/stdout || name="$name.mp4"

# // wget "${url}&signature=${signature}" -O "$name"

mplayer -zoom -fs "${url}&signature=${signature}"

It uses decode and bash, that you may have installed.

Twofold

Answer 4

I use this bash script to download a given set of songs from a given youtube's playlist

#!/bin/bash
downloadDirectory = <directory where you want your videos to be saved>
playlistURL = <URL of the playlist>
for i in {<keyword 1>,<keyword 2>,...,<keyword n>}; do
    youtube-dl -o ${downloadDirectory}"/youtube-dl/%(title)s.%(ext)s" ${playlistURL} --match-title $i
done

Note: "keyword i" is the title (in whole or part; if part, it should be unique to that playlist) of a given video in that playlist.

Edit: You can install youtube-dl by pip install youtube-dl

Disintegrate

Answer 5

#!/bin/bash
# Coded by Biki Teron 
# String replace command in linux

echo "Enter youtube url:"
read url1

wget -c -O index.html $url1
################################### Linux string replace    ##################################################
sed -e 's/%3A%2F%2F/:\/\//g'  index.html > youtube.txt
sed -i 's/%2F/\//g' youtube.txt
sed -i 's/%3F/?/g' youtube.txt
sed -i 's/%3D/=/g' youtube.txt
sed -i 's/%26/\&/g'  youtube.txt
sed -i 's/%252/%2/g'  youtube.txt
sed -i 's/sig/&signature/g'  youtube.txt
## command to get filename
nawk '/<title>/,/<\/title>/' youtube.txt > filename.txt ## Print the line between   containing <title> and <\/title> .
sed -i 's/.*content="//g' filename.txt 
sed -i 's/">.*//g' filename.txt
sed -i 's/.*<title>//g' filename.txt
sed -i 's/<.*//g' filename.txt
######################################## Coding to get all itag list ########################################

nawk '/"fmt_list":/,//' youtube.txt > fmt.html ## Print the line containing "fmt_list": .
sed -i 's/.*"fmt_list"://g' fmt.html
sed -i 's/, "platform":.*//g' fmt.html
sed -i 's/, "title":.*//g' fmt.html

# String replace command in linux to get correct itag format
sed -i 's/\\\/1920x1080\\\/99\\\/0\\\/0//g'  fmt.html  ## Replace \/1920x1080\/99\/0\/0 by blank .
sed -i 's/\\\/1920x1080\\\/9\\\/0\\\/115//g' fmt.html  ## Replace \/1920x1080\/9\/0\/115 by blank.
sed -i 's/\\\/1280x720\\\/99\\\/0\\\/0//g'   fmt.html  ## Replace \/1280x720\/99\/0\/0 by blank.
sed -i 's/\\\/1280x720\\\/9\\\/0\\\/115//g'  fmt.html  ## Replace \/1280x720\/9\/0\/115 by blank.
sed -i 's/\\\/854x480\\\/99\\\/0\\\/0//g'    fmt.html  ## Replace \/854x480\/99\/0\/0 by blank.
sed -i 's/\\\/854x480\\\/9\\\/0\\\/115//g'   fmt.html  ## Replace \/854x480\/9\/0\/115 by blank.
sed -i 's/\\\/640x360\\\/99\\\/0\\\/0//g'    fmt.html  ## Replace \/640x360\/99\/0\/0 by blank.
sed -i 's/\\\/640x360\\\/9\\\/0\\\/115//g'   fmt.html  ## Replace \/640x360\/9\/0\/115 by blank.
sed -i 's/\\\/640x360\\\/9\\\/0\\\/115//g'   fmt.html  ## Replace \/640x360\/9\/0\/115 by blank.
sed -i 's/\\\/320x240\\\/7\\\/0\\\/0//g'     fmt.html  ## Replace \/320x240\/7\/0\/0 by blank.
sed -i 's/\\\/320x240\\\/99\\\/0\\\/0//g'    fmt.html  ## Replace \/320x240\/99\/0\/0 by blank.
sed -i 's/\\\/176x144\\\/99\\\/0\\\/0//g'    fmt.html  ## Replace \/176x144\/99\/0\/0 by blank.

# Command to cut a part of a file between any two strings
nawk '/"url_encoded_fmt_stream_map":/,//' youtube.txt > url.txt
sed -i 's/.*url_encoded_fmt_stream_map"://g' url.txt

#Display video resolution information
echo ""
echo "Video resolution:"
echo "[46=1080(.webm)]--[37=1080(.mp4)]--[35=480(.flv)]--[36=180(.3gpp)]"
echo "[45=720 (.webm)]--[22=720 (.mp4)]--[34=360(.flv)]--[17=144(.3gpp)]"
echo "[44=480 (.webm)]--[18=360 (.mp4)]--[5=240 (.flv)]"
echo "[43=360 (.webm)]"
echo ""
echo "itag list= "`cat fmt.html`
echo "Enter itag number: "
read fmt

####################################### Coding to get required resolution #################################################
## cut itag=?

sed -e "s/.*,itag=$fmt//g" url.txt > "$fmt"_1.txt
sed -e 's/\u0026quality.*//g' "$fmt"_1.txt > "$fmt".txt
sed -i 's/.*u0026url=//g' "$fmt".txt ## Ignore all lines before \u0026url= but print all lines after \u0026url=.
sed -e 's/\u0026type.*//g' "$fmt".txt > "$fmt"url.txt ## Ignore all lines after \u0026type but print all lines before \u0026type.
sed -i 's/\\/\&/g' "$fmt"url.txt ## replace \ by &
sed -e 's/.*\u0026sig//g' "$fmt".txt > "$fmt"sig.txt ## Ignore all lines before \u0026sig but print all lines after \u0026sig.
sed -i 's/\\/\&ptk=machinima/g' "$fmt"sig.txt ## replace \ by &
echo `cat "$fmt"url.txt``cat "$fmt"sig.txt` > "$fmt"url.txt ## Add string at the end of a line
echo `cat "$fmt"url.txt` > link.txt ## url and signature content to 44url.txt
rm "$fmt"sig.txt
rm "$fmt"_1.txt
rm "$fmt".txt
rm "$fmt"url.txt
rm youtube.txt
########################################### Coding for filename with correct extension #####################################
if [ $fmt -eq 46 ]
 then 
    echo `cat filename.txt`.webm > filename.txt

elif [ $fmt -eq 45 ]
  then 
     echo `cat filename.txt`.webm > filename.txt

elif [ $fmt -eq 44 ]
  then 
     echo `cat filename.txt`.webm > filename.txt

elif [ $fmt -eq 43 ]
   then 
   echo `cat filename.txt`.webm > filename.txt

elif [ $fmt -eq 37 ]
   then 
   echo `cat filename.txt`.mp4 > filename.txt

elif [ $fmt -eq 22 ]
   then 
   echo `cat filename.txt`.mp4 > filename.txt

elif [ $fmt -eq 18 ]
   then 
  echo `cat filename.txt`.mp4 > filename.txt

elif [ $fmt -eq 35 ]
   then 
   echo `cat filename.txt`.flv > filename.txt

elif [ $fmt -eq 34 ]
   then 
   echo `cat filename.txt`.flv > filename.txt

elif [ $fmt -eq 5 ]
   then 
   echo `cat filename.txt`.flv > filename.txt

elif [ $fmt -eq 36 ]
   then 
   echo `cat filename.txt`.3gpp > filename.txt

 else
   echo `cat filename.txt`.3gpp > filename.txt

fi
rm fmt.html
rm url.txt
filename=`cat filename.txt`
linkdownload=`cat link.txt`
wget -c -O "$filename" $linkdownload
echo "Download Finished!"
read

Donyadoodad

I am learning bash script in downloading youtube videos this is based on latest youtube algorithm. It will display all the available video resolution.you can download .webm, .mp4, .flv, .3gpp file formats. – Donyadoodad

Sorry but this has nothing to do with the original question, -1. – Holder