How to check if it is a file or folder for an archive in python?

Question

I have an archive which I do not want to extract but check for each of its contents whether it is a file or a directory.

os.path.isdir and os.path.isfile do not work because I am working on archive. The archive can be anyone of tar,bz2,zip or tar.gz(so I cannot use their specific libraries). Plus, the code should work on any platform like linux or windows. Can anybody help me how to do it?

Solution

You've stated that you need to support "tar, bz2, zip or tar.gz". Python's tarfile module will automatically handle gz and bz2 compressed tar files, so there is really only 2 types of archive that you need to support: tar and zip. (bz2 by itself is not an archive format, it's just compression).

You can determine whether a given file is a tar file with tarfile.is_tarfile(). This will also work on tar files compressed with gzip or bzip2 compression. Within a tar file you can determine whether a file is a directory using TarInfo.isdir() or a file with TarInfo.isfile().

Similarly you can determine whether a file is a zip file using zipfile.is_zipfile(). With zipfile there is no method to distinguish directories from normal file, but files that end with / are directories.

So, given a file name, you can do this:

import zipfile
import tarfile

filename = 'test.tgz'

if tarfile.is_tarfile(filename):
    f = tarfile.open(filename)
    for info in f:
        if info.isdir():
            file_type = 'directory'
        elif info.isfile():
            file_type = 'file'
        else:
            file_type = 'unknown'
        print('{} is a {}'.format(info.name, file_type))

elif zipfile.is_zipfile(filename):
    f = zipfile.ZipFile(filename)
    for name in f.namelist():
         print('{} is a {}'.format(name, 'directory' if name.endswith('/') else 'file'))

else:
    print('{} is not an accepted archive file'.format(filename))

When run on a tar file with this structure:

(py2)[mhawke@localhost tmp]$ tar tvfz /tmp/test.tgz
drwxrwxr-x mhawke/mhawke     0 2016-02-29 12:38 x/
lrwxrwxrwx mhawke/mhawke     0 2016-02-29 12:38 x/4 -> 3
drwxrwxr-x mhawke/mhawke     0 2016-02-28 21:14 x/3/
drwxrwxr-x mhawke/mhawke     0 2016-02-28 21:14 x/3/4/
-rw-rw-r-- mhawke/mhawke     0 2016-02-28 21:14 x/3/4/zzz
drwxrwxr-x mhawke/mhawke     0 2016-02-28 21:13 x/2/
-rw-rw-r-- mhawke/mhawke     0 2016-02-28 21:13 x/2/aa
drwxrwxr-x mhawke/mhawke     0 2016-02-28 21:13 x/1/
-rw-rw-r-- mhawke/mhawke     0 2016-02-28 21:13 x/1/abc
-rw-rw-r-- mhawke/mhawke     0 2016-02-28 21:13 x/1/ab
-rw-rw-r-- mhawke/mhawke     0 2016-02-28 21:13 x/1/a

The output is:

x is a directory
x/4 is a unknown
x/3 is a directory
x/3/4 is a directory
x/3/4/zzz is a file
x/2 is a directory
x/2/aa is a file
x/1 is a directory
x/1/abc is a file
x/1/ab is a file
x/1/a is a file

Notice that x/4 is "unknown" because it is a symbolic link.

There is no easy way, with zipfile, to distinguish a symlink (or other file types) from a directory or normal file. The information is there in the ZipInfo.external_attr attribute, but it's messy to get it back out:

import stat

linked_file = f.filelist[1]
is_symlink = stat.S_ISLNK(linked_file.external_attr >> 16L)