ls files --> list packer

[kpp9c]

I would like to use the power of Python to build some list structures
for me.

Namely i have organized a bunch of folders that have soundfiles in them
and would like Python to slurp up all the .aif/.aiff (or .wav whatever)
files in a given set of directories. My friend hacked up this is perl:

$files = `ls /snd/Public/*.aiff`;

@snd_filelist = split('\n',$files);

$i = 0;
while ($file = @snd_filelist[$i]) {
print "file $i = @snd_filelist[$i]\n";
$i++;
}


The only catch with the above code (besides its hideousness hee hee) is
if you have a directory w/i the structure, but in general it works and
with this i can just put gobs of files into separate dirs pack them
into a list and then send them to my script that scrambles them and
plays them.

I would like something similar, that works with python that is more
elegant and maybe even more robust.

but i am failing miserably and my perl friends mock me.

cheers,
kp8

 

[kpp9c]

and one example of a slightly fancier version would be a variation that
looks recursively into subdirectories and makes separate lists for each
subdirectory encountered.

so if i had a directory called "~/snd/"

and in "~/snd/" i had:

"~/snd/one/"
"~/snd/two/"
"~/snd/three/"

each with soundfiles in it.

I could get those packed in three separate lists named after the
directory or some such thing....

This would be so awesome because my carefully organizing your
directory, you would be carefully
organizing your data, change your dir structure or add/delete some
files and you would get a new structure in your script... prolly would
work with scripting your iTunes music folder too...

gosh ... sorry ... just thinking out-loud here and getting kind of
giddy! reaching for the python book ...

 

[I V]

Namely i have organized a bunch of folders that have soundfiles in them
and would like Python to slurp up all the .aif/.aiff (or .wav whatever)
files in a given set of directories. My friend hacked up this is perl:

$files = `ls /snd/Public/*.aiff`;

You could use posix.popen to duplicate the perl hack:

files = posix.popen('ls /snd/Public/*.aiff').read().strip()

@snd_filelist = split('\n',$files);

snd_filelist = files.split('\n')

I would like something similar, that works with python that is more
elegant and maybe even more robust.

Lucklily, python lets you avoid this kind of horrible hack. Try
os.listdir:

snd_filelist = [f for f in os.listdir('/snd/Public/') if
f.endswith('.aiff')]

I think it's more elegant, and it's certainly more robust.

but i am failing miserably and my perl friends mock me.

Now you get to mock your perl friends!

 

[kpp9c]

cool i just tried:

import os
snd_filelist = [f for f in os.listdir('/Users/foo/snd') if f.endswith('.aif')]

and it worked! and will take a huge bite out of my big script ... which
i make by doing an ls
in the terminal and editing (boo hoo)

one one lc and one import!

cool..

that other sillyness i mentioned is not strickly required ... just
dreaming but i know involves some kind of os walk type thing prolly ...
meanwhile this is so exciting!

Thank you!!!!

 

[kpp9c]

gosh i could even use other string methods like startswith to take all
the files in a given directory which i have organized with a prefix and
have them stuffed in different lists ... i think ...

snd_filelist = [f for f in os.listdir('/Users/foo/snd') if
f.endswith('.aif') & f.startswith('r')]

\m/ (>.<) \m/

yeah!

runnin' to the interpreta now...

 

[Jeremy Sanders]

snd_filelist = [f for f in os.listdir('/snd/Public/') if
f.endswith('.aiff')]

Or even

from glob import glob

snd_filelist = glob('/snd/Public/*.aiff')

Jeremy

 

[Singletoned]

Try using The Path module:
http://www.jorendorff.com/articles/python/path/.

I wrote a little script to traverse a directory structure which you
could use. (You just pass a function to it and it runs it on each file
in the directory. You want it to run on each directory instead, so
I've changed it a little for you).

import path

def traverse(directory, function, depth=0, onfiles=True, ondirs=False):
thedir = path.path(directory)
if onfiles == True:
for item in thedir.files():
function(item, depth)
if ondirs == True:
for item in thedir.dirs():
function(item, depth)
for item in thedir.dirs():
traverse(item, function, depth+1, onfiles, ondirs)

You can use it like so:

def printaifs(thedir, depth):
print thedir.name
for item in thedir.files("*.aif'"):
print "\t" + item

traverse(r"/Users/foo/snd", printaifs, onfiles=False, ondirs=True)

NB: I've quickly adapted this whilst away from an installation of
Python, so it is untested, but should mostly work, unless there's a
typo.

Hope this helps at least a little...

Ed

PS The Python Tutor list tends to be a much better place to discuss
this kind of stuff. If you haven't yet encountered Kent and Alan's
help, then you have a joyous experience ahead of you.

 

[Magnus Lycka]

that other sillyness i mentioned is not strickly required ... just
dreaming but i know involves some kind of os walk type thing prolly ...

os.walk isn't exactly rocket science... Something similar to this?
.... txt_files = [x for x in files if x.endswith('.txt')]
.... if txt_files:
.... print dir, txt_files

 

[kpp9c]

that is nice.... but the little further wrinkle, which i have no idea
how to do, would be to have the contents of each directory packed into
a different list.... since you have no idea before hand how many lists
you will need (how many subdirs you will enounter) ... well that is
where the hairy part comes in...

-kp--

 

[kpp9c]

os.listdir works great ... just one problem, it packs the filenames
only into a list... i need the full path and seach as i might i se NO
documentation on python.org for os.listdir()

how do i either grab the full path or append it later ...

 

[Kent Johnson]

os.listdir works great ... just one problem, it packs the filenames
only into a list... i need the full path and seach as i might i se NO
documentation on python.org for os.listdir()

Docs for os.listdir() are here:
http://docs.python.org/lib/os-file-dir.html

When all else fails try the index:
http://docs.python.org/lib/genindex.html

how do i either grab the full path or append it later ...

Use os.path.join():

base = '/some/useful/path'
for f in os.listdir(base):
full_path_to_file = os.path.join(base, f)

Kent

 

[kpp9c]

Thank you... i was looking in the wrong place cause all i found was
this relatively useless doc:

http://docs.python.org/lib/module-os.html

which says almost nothing.

 

[kpp9c]

nice! two little lines that do a boatload of work! hee hee

 

[Scott David Daniels]

Thank you... i was looking in the wrong place cause all i found was
this relatively useless doc:
http://docs.python.org/lib/module-os.html
which says almost nothing.

In one of its subsections, cleverly named "Files and Directories",
I see a nice description of listdir.

http://docs.python.org/lib/os-file-dir.html

You also might want to read about os.walk in the same page.
In the os.path module you can see more path name manipulations.
If you intend to know a language, you should read its manuals
fast; what you want is a vague impression where information
lives and what information is there. Maybe half a year later
do it again. After that every couple of years often suffices.
At the very least, go through the full tutorial, read docs on
the "obvious" modules for everyone and for your particular
area of endeavor, and then on a snacking basis get yourself
through the rest unless you decide that you never want to deal
with, for example, unix-specific services or internet protocols.

Don't expect to acquire _any_ language with "just in time" (JIT)
techniques. Perhaps JIT works for magic. When you acquire a
language with JIT, you miss the subtleties of the language.
You will be doomed to writing the same kinds of programs in every
language you touch ("writing Fortran in Algol" is what we used to
call it). I've worked on code that was Java-in-Python, and it was
frustratingly hard to understand.

--Scott David Daniels
(e-mail address removed)

 

[Magnus Lycka]

that is nice.... but the little further wrinkle, which i have no idea
how to do, would be to have the contents of each directory packed into
a different list.... since you have no idea before hand how many lists
you will need (how many subdirs you will enounter) ... well that is
where the hairy part comes in...

What's the problem? If you'll get an unknown bundle of objects in a
program, you just put them in a container. A list or a dict will do
fine. Have a look at the Python tutorial.

You get a file list for each directory from os.walk. You either keep
a list called "directories" and for each turn in the loop, you do
"directories.append((dir, sound_files))", or you have a dict called
"directories", and do "directories[dir] = sound_files" in the loop.

Something like this untested code:

def isSoundFile(x):
# home work

dirs = {}
root = raw_input('start directory')
for dir, dummy, files in os.walk(root):
dirs[dir] = [x for x in files if isSoundFile(x)]

for dir in sorted(dirs):
print dir
for fn in dirs[dir]:
print "\t", fn
print