Script to tramsfer random songs to portable media device

[junior aspirin]

hi, i think this is the most approriate forum to ask this.

i have very little knowlage about scripts so i was wondering if some one could help me.

i want to fill my iriver t20 (flashed to ums http://www.mtp-ums.net/) with random songs from my collection, which has 496 MB of free space

* music stored in /home/paul/music/*album name*/*track*.ogg
* device mounts to /media/usbdisk
* music to be copied to /media/usbdisk/music/
* i have mp3's and ogg's
* can only store 496MB aof music

hope some one can help me, im sure its not hard, i think i could do it myself, its just i dnt know if its posible to make it stop copying files when the player is full.

thanks in advance. :p

[purpleturtle]

The easy way would be to check the return code from the copy (cp) command and as soon as an error occurs stop copying.

[purpleturtle]

Here's a script I've knocked up that should do the trick..

When you run it, it creates an index file with all your music filenames in it (goes to file ~/MusicIndex). It then randomly picks files from this index of files to copy over. When it has successfully copied the file, it removes the entry from the index. It should keep doing this until the copy fails. Then next time you run it, it will reuse the index it originally created to repeat the process, so this time you will not get the same songs included (as they were removed from this list).

Hope this helps :) I've not tested it much...

If anyone would like to critique my scripting please do.. I'm learing bash scripting at the moment, and thought this seemed like a good challange!

Apologies for the screwed up indentation.. I don't know how you stop the web site removing leading spaces...


#!/bin/bash

# randtracks
# Copy a random collection of songs from source to dest.

# Do not use ~ for these (TODO fix this)
sourced="/home/paul/music"
destd="/media/usbdisk/music"

index=~/MusicIndex
index_done=~/MusicIndexDone

function ErrorExit()
{
echo "Error: $*" >&2
exit 1
}

function Dbg
{
echo "[D] $*" >&2
}

# Make sure the source and directories exist

if [ ! -d "$sourced" ]; then
ErrorExit "Could not find source directory $sourced"
fi

if [ ! -d "$destd" ]; then
ErrorExit "Could not find destination directory $destd"
fi

if [ ! -e $index ]; then
echo "Creating index $index"
find "$sourced" -type f -name '*.mp3' -o -name '*.ogg' > "$index"
fi

track_count=$(wc -l < "$index")

echo "Have $track_count tracks in the collection"

# Now randomly pick tracks to be copied

max_loop=$(($track_count * 2))
for((i=0; i < $max_loop && $track_count > 0; i++))
do
line_num=$(($RANDOM%$track_count+1))

# Pick a line at random out of the index
src_this=$(sed -n "$line_num p" < "$index")

Dbg "Have chosen line $line_num/$track_count, Src=$src_this"

# Format up the destination path
destf="${destd}${src_this#$sourced}"

#Dbg "From : $src_this"
#Dbg "To : $destf"

# Try and create the dest directory
mkdir -p "$(dirname "$destf")"

if [ -e "$destf" ]; then
echo "$destf already exists, skipping"
# Maybe stop processing at this point
elif cp -v "$src_this" "$destf" ; then

Dbg "Removing line $line_num from $index"

# Remove the line from the index
if ! (sed "$line_num d" < "$index" > "$index.tmp" && mv "$index.tmp" "$index") ; then
ErrorExit "Removing line $line_num from $index"
fi

# Write the line we have randomly picked to another file for reference
# This is not needed
echo "$src_this" >> "$index_done"

# adjust counts
track_count=$(($track_count - 1))

else
Dbg "Failed to copy, exiting"
break
fi

done

[jerome bettis]

i've wanted to do this myself for a while.

here's a python script i came up with, it's working great over here. i hate python but it's good for stuff like this. i tried using bash but was struggling to do a few things ... i did this in 35 minutes, not bad \\:D/

#!/usr/bin/env python

import sys, os, random

src_path = "/home/paul/music"
dev_path = "/media/usbdisk"
dest_path = dev_path + "/music"

# when the player only has 3 megs left, we'll stop trying to copy
# you can lower this to 0 (or anything) so it will try to fit every
# song in your collection, but consider the player might have 2 bytes
# left and you have 800000 songs
player_is_almost_full = 3000000

# returns whether or not dev_path is mounted
def player_is_mounted() : #{
return (os.popen("mount | awk '{print $3}' | grep " + \
dev_path).read().replace("\n", "") == dev_path)
#}

# returns the # of bytes avail on the player
def get_free_space() : #{
return int(os.popen("df -B 1 | grep `mount | grep " + dev_path + \
" | awk '{print $1}'`" + " | awk '{print $4}'") \
.read().replace("\n", ""))
#}

# returns the # of bytes of a given file name
def get_file_size(file_name) : #{
return int(os.popen("ls -l \"" + file_name + "\" | awk '{print $5}'") \
.read().replace("\n", ""))
#}

if __name__ == "__main__" : #{
if (not player_is_mounted()) : #{
print "Your mp3 player needs to be mounted on " + dev_path + \
" before you can write to it."
sys.exit()
#}
os.popen("mkdir -p " + dest_path)

# find our entire music collection, if it's very large this could be bad
# it works ok over here
file_list = os.popen("find -L \"" + src_path + \
"\" -type f -iname \"*.mp3\" -o -iname \"*.ogg\"") \
.read().split("\n")[0:-1]

# loop until there are no more files to copy
# if the device gets close to full we will stop too
while (len(file_list) > 0) : #{
# see how much the player can fit
free_space = get_free_space()
if (free_space < player_is_almost_full) : #{
break # can't really fit anything else
#}

# choose a random file number
index = random.randint(0, len(file_list) - 1)

# quotes in the file name will F everything up
file_list[index].replace("\"", "\\\"")

# see if it will fit on the player and copy it if so
if (free_space > get_file_size(file_list[index])) : #{
os.popen("cp \"" + file_list[index] + "\" " + dest_path)
#}

# put the last file in our list where the one we just copied was
# if it's the last file, we'll just toss it
if (index != len(file_list) - 1) : #{
file_list[index] = file_list.pop(len(file_list) - 1)
#}
else : #{
file_list.pop(len(file_list) - 1)
#}
#}
#}
let me know how it works if you try it.

[awi]

I had a similar idea some time ago, but never the time to do it (not yet I hope :) ). Not only a random selection, but also one that holds your preferences giving more probability to your favorites songs, but still making it less pickable after n1 times, and after n1 times you could be sure of having all of your songs transfered, at least one time. Letting you choose the time period for this event. Kind of cool for an office too, don't you think? Or to have music all the time at home.
Doing a file random choose and running every time you are going to copy it, kills the all random concept. A real random selection, should give you, in average, an uniform selection on all of your tracks, as the number of copies increases.
If you use your script independently, the chances that you get all the songs, shrinks.

[junior aspirin]

hi, thanks for this, but i cannot seem to get it to get eather to work :-k

the bash script returns "Have 0 tracks in the collection"
where as the python one, does nothing

could it be to do with simlinks? my music folder is symlinked to "/mnt/media/my music"

also on a posibly related note:
when i unmount a device ie my media player i get this error:
unmount error
unable to eject media
>eject: unable to eject, last error: Invalid argument

any ideas?

[jerome bettis]

lol it does nothing??

"by god! .. do you know what this means?? .... it means that this damn thing doesn't work at all!!!"

are you sure the folders are set right? i can throw in some print statements to show what's going on if those are set right. it works flawlessly for me over here. it should just start copying stuff to it and stop when it's full or you run out of tracks, i'm kind of upset it does nothing lol.

the best reason i can come up with for it doing nothing right now, is that the file list has 0 elements. that would mean that the find command isn't working, which probably means that the dir is set wrong or there's a slight problem somewhere.

[junior aspirin]

ok, i think its to do with the symlink. i pointed it to "/mnt/media/My Music" and it complains about the space ie it looks for "/mnt/media/My" and "music". when i point it to a file with no spaces the treminal outputs:
Traceback (most recent call last):
File "/home/paul/random.py", line 58, in ?
index = random.randint(0, len(file_list) - 1)
AttributeError: 'module' object has no attribute 'randint'


any ideas on how to fix both these?

as a note this is how my music files are organised

/mnt/media/My Music/%artist% - %album%/&track Number% - %artist% - %track% - %album%.%mp3 or ogg%

%blah% is a variable
dont know if that helps as you can see the My Music folder is left over from windowz xp, as i dual boot, occationaly.

[jerome bettis]

ok i put quotes around the source path in the find command, so that space won't screw it up. i also turned the option to follow symlinks on, try it now.

not sure why randint doesn't exist, it's a standard python thing, try this first and we'll go from there.

[junior aspirin]

now it just skips to the error.

do you think i need to install something from synaptic? if so what? :-k

[jerome bettis]

hah you named the script random.py didn't you?

rename it and it will work. :cool:

[junior aspirin]

ooooo groovy, thanks. why does random, not work? or was it the .py on the end?

thanks anyway that is great!

[psychicdragon]

It was trying to import randint from itself instead of the python module random because you named the script random.py.

[wsmith]

Fill my MP3 player one-liner:

find ./ -iname '*.mp3' | sort -R | while read fn ; do if ! cp -v "$fn" /media/disk ; then rm "$fn" && break ; fi ; done


Find all .mp3 files in the current directory
Sort them randomly
Try to copy them to the media
If the copy fails [i.e. disk full], abort.

[FatCapybara]

WSmith - now that's what I call a solution. One simple, not showily obfuscated, statement.

I've got it running now and have saved it for regular use.

Thank you.

[ghostdog74]

# returns whether or not dev_path is mounted
def player_is_mounted() : #{
return (os.popen("mount | awk '{print $3}' | grep " + \
dev_path).read().replace("\n", "") == dev_path)
#}

# returns the # of bytes avail on the player
def get_free_space() : #{
return int(os.popen("df -B 1 | grep `mount | grep " + dev_path + \
" | awk '{print $1}'`" + " | awk '{print $4}'") \
.read().replace("\n", ""))
#}

# returns the # of bytes of a given file name
def get_file_size(file_name) : #{
return int(os.popen("ls -l \"" + file_name + "\" | awk '{print $5}'") \
.read().replace("\n", ""))
#}

On top of being non portable, if you have no choice to call system commands, call it and manipulate the return results inside Python. Don't pipe processes like that, its ugly and creates extra overheads.
The functionality of awk/grep etc can be done with Python itself

Also, why bother with ls -l, when Python provides these functionalities through the os module?? os.listdir() does just this, along with os.path.isdir, os.path.isfile, etc etc