[Tutor] Lists of files

[Karl Pflästerer]

On 14 Mai 2005, william.ohiggins at utoronto.ca wrote:

> Here's the problem - I want a list (array) of the files in a directory,
> and then I want to iterate over the list testing for image-ness (with
> imghdr.what()) and put all the image filenames in a global list.
>
> What I've tried is this:
>
> files = glob.glob('*.*')
>
> for file in files:
>     global pics 
>     pics = []
>     if imghdr.what(file):
>         # so far so good - file is a list of files in the directory
>         pics.append(file)
>         # I this this is the problem - my list only has the last
>         # alphabetical entry in it

The problem here is that in every iteration you set the list pics to
`[]'.  If you wanted to solve the problem like above (not very nice; try
to avoid globals if possible) you had to define `pics' outside the loop.

Bernard gave you an answer how you could solve it with a simple list
comprehension (just filter the output from os.listdir); it can also be
written like that:
     filter(imghdr.what, os.listdir('.'))
Just jump before to the directory you want to get searched with:
     os.chdir(path)

If you want a more general solution which also allows to search
recursively through a directory tree you could use something like that:

def filter_files (root, filepred=lambda f: f, dirpred=lambda d: False):
    filtered= []
    jn = os.path.join
    for path, dirs, files in os.walk(root, topdown=True):
        for d in dirs:
            if not dirpred(jn(path, d)): dirs.remove(d)
        filtered.extend([jn(path,f) for f in files if filepred(jn(path, f))])
    return filtered

The above takes two callback functions: filepred and dirpred.  For every
file filepred returns a true value that file gets appended to the list
of returned files.  `dirpred' allows you to recurse only in directories
where that function returns a true value.

So to have the same as above you write:
   filter_files('.', filepred=imghdr.what)
and if you wanted to search a directory tree without e.g. the ./bin
directory you could write:
   filter_files('.', filepred=imghdr.what, dirpred=lambda d: not d.endswith('bin'))


HTH

   Karl
-- 
Please do *not* send copies of replies to me.
I read the list