Python function to handle multiple file input types for an argument

If I understand you well, then you want to get a single argument, and depending on whether it is a directory, a file, or a mask, you want to have a function that returns filenames from the directory, the file, or the files got by expanding the mask.  Is that what you want?

If yes, I suggest to create a generator function that produces the filenames independently on what you pass.  Or it could be no file, or the single file, or the files got via expanded mask, or the files from inside the directory.  (I know I am repeating the same using the other words.)

Then if you get the argument myArg, the you will use it this way:

for fname in myFiles(myArg):
     process(fname)

Select allOpen in new window

Do you want something like this?

@pepr: how would this work with my current logic? I am only going to be using the options.filename and doing away with the options.directory argument.

    filenames_or_wildcards = []

# remove next line if you do not want allow to run the script without the -f -d
# option, but with arguments
    filenames_or_wildcards = args # take all filenames passed in the command line

# if -f was specified add them (in current working directory)
    if options.filename is not None:
        filenames_or_wildcards.append(options.filename)

# if -d was specified or nothing at all add files from dir
    if options.directory is not None:
        filenames_or_wildcards.append( os.path.join(options.directory, "*") )

# Now expand all wildcards
# glob.glob transforms a filename or a wildcard in a list of all matches
# a non existing filename will be 'dropped'
    all_files = []
    for filename_or_wildcard in filenames_or_wildcards:
        all_files.extend( glob.glob(filename_or_wildcard) )

Select allOpen in new window

Well, I was typing meanwhile. The myFiles() processes a single argument but it can easily be adjusted to proces or a single argument or a list of arguments.  Any single argument could be a file, a dir, a mask.

Thanks for the quick response!

Have a look at the modified function:

b.py

import glob
import os
import sys


def myFiles(arg):
    '''A generator function that diggs all files from the argument.'''
    
    if isinstance(arg, str):     # single string
        lst = [ arg ]            # convert to the list
    elif isinstance(arg, list):  # list is OK,
        lst = arg                # no conversion
    else:
        print 'Unexpected arg:', repr(arg)
        sys.exit(1)
    # Warning: the arg identifier reused here (not ideal).
    for arg in lst:
        if os.path.isfile(arg):                 # for a single file, return the file
            yield arg
            
        elif os.path.isdir(arg):                # for a directory, return all files
            for item in os.listdir(arg):        # produces bare names
                fname = os.path.join(arg, item) # path + bare name
                if os.path.isfile(fname):       # only the files are returned
                    yield fname  # no normalization of slashes/backslashes here
        else:
            for fname in glob.glob(arg):        # it could be a mask -- expand it
                if os.path.isfile(fname):       # only the files are returned
                    yield os.path.normpath(fname)  # normalize slashes/backslashes

print '-' * 70
arg = 'testDir/a.txt'      # existing file
print 'Test for', repr(arg)
print list(myFiles(arg))

print '-' * 70
arg = 'testDir/.'           # existing directory -- another form
print 'Test for', repr(arg)
print list(myFiles(arg))

print '-' * 70
arg = ['noDir/*.txt',       # mask expanding to no file
       'testDir/*',         # all files from the existing directory
       'testDir/?.txt',     # one-letter + extension
       'testDir/a*.txt',    # a-prefixed filenames
       'testDir/*a.txt',    # a-suffixed filenames
      ]
print 'Test for', repr(arg)
print list(myFiles(arg))

Select allOpen in new window

It prints:

C:\tmp\_Python\adamshields\Q_26901398>python b.py
----------------------------------------------------------------------
Test for 'testDir/a.txt'
['testDir/a.txt']
----------------------------------------------------------------------
Test for 'testDir/.'
['testDir/.\\a.txt', 'testDir/.\\aa.txt', 'testDir/.\\ax.txt', 'testDir/.\\ba.txt', 'testDir/.\\x.txt']
----------------------------------------------------------------------
Test for ['noDir/*.txt', 'testDir/*', 'testDir/?.txt', 'testDir/a*.txt', 'testDir/*a.txt']
['testDir\\a.txt', 'testDir\\aa.txt', 'testDir\\ax.txt', 'testDir\\ba.txt', 
'testDir\\x.txt', 'testDir\\a.txt', 'testDir\\x.txt', 'testDir\\a.txt', 
'testDir\\aa.txt', 'testDir\\ax.txt', 'testDir\\a.txt', 'testDir\\aa.txt', 
'testDir\\ba.txt']

Select allOpen in new window

Notice that the names are not uniqued in the case.

Great thanks, lastly, if I were using:

def main()

Where would def myFiles(arg): be located, before, after or in the main()?

Thanks

It does not matter in Python.  But it may be good to keep conventions from other languages and put it earlier than main().

Actually, Python interprets the .py file from beginning to the end.  However, the first and the only interpretation means compilation.  Any def, class or such is converted to the internal object, any other command is interpreted.

Okay thanks!