TKinter User Input into OS Walk - Get user input to pull File Name

Question

This is difficult to summarize what I am trying to do, I doubt the Title made it any easier to follow but...

I am using tkinter to build an options dialog box, allowing the user to enter their own file name breakdown structure. Os.Walk would read and setup the folder structure based on the users input.

Config Parser/User Input

[Alpha] aoth = file[:1], file[:3]

The issue is that I can breakdown the structure based on user input (I.E Read 'aoth' in Alpha structure, break down by two folders). However Os.Walk/Python considers this as the actual input. Broke down the code to only show the relevant information:

for root, subFolders, files in os.walk(Entry.get(self.sourceE)):
    for file in files:
        if not file.startswith('.'):
            if file[0].isalpha():
                alpfol = '%s' % self.alpha_fol.get()
                alset = self.alphaa_radio.get()
                if alset == 1:
                    file_strut = '%s' % file[:1]
                elif alset == 2:
                    file_strut = '%s/%s' % (file[:1], file[:2])
                elif alset == 3:
                    files_strip = [st.strip() for st in (self.radio_aBoxV.get()).split(',')]
                    files_count = len(files_strip)
                    ###
                    file_count_check = '%s/'*files_count
                    file_strut = file_count_check % tuple(files_strip)
                    ###
                subFolder = os.path.join(Entry.get(self.destE), alpfol, file_strut)
            checkFile = os.path.join(subFolder, file)
                ........

I know it's not the most elegant code, but alset 1/2 work flawlessly whereas alset 3 considers file[:1]/file[:3] as literal input. How would I make the users input read the actual file name and break down accordingly? I'm sure its something very simple I'm missing.

Thanks for reading this!

Answer 1

Well, like you suggested it's a pretty simple and obvious error:

"file[:1],file[:2]" is not the same as file[:1],file[:2].

Your first two options are hardcoded, so they're taking the reference file and applying slice notation to it. The last one takes a String and just throws it around; the program has no idea that you mean for it to convert the String "file" to the Reference file and the String "[:1]" and convert it to slice notation [:1].

The solution to the problem is to make that conversion, probably by going into the String and pulling the necessary information out of it. You do this to get the number of partitions, but never convert the reference and slice notation. You could take those partitions you have in your code and do some finangling using split or slice notations to "cut" the important bits out, but I'm going to suggest Regex (re module in Python) as it's a more flexible solution.

Python RE HowTo

Below is a sample that I have mostly working; I have a comment at the one point that may cause it to give imperfect results (both negative start and step at same time), but it's a really obscure case and you can get the code up to 100% just by removing the option to use the step portion of slice notation. You'll also note that simply taking slice notations blindly from the user can lead to malformed directory names.

(Note- I don't actually program in Python 2, but as far as I know this is 2.7 safe; hopefully only necessary changes are minor syntax adjustments)

import os, os.path
import re

PATTERNRE=re.compile('''(?P<parts>[Ff]ile\[(?P<start>-?\d*):+(?P<end>-?\d*):?(?P<step>-?\d*)\])''')

def formatindices(start,end,step,file): ## Slice Notation only accepts Ints,
                                        ## and int('{empty string}') raises an error,
                                        ## so we're programming around it
    step = int(step) if step else 1
    if not start:
        if step>0: start=0
        else: start=-1
    else:
        start=int(start)
    end = int(end) if end else len(file) ## Having both negative Start and Step will
                                         ## break this on the else statement, couldn't
                                         ## find an integer to represent "index before 0"
    return start,end,step

files=['hello.py','world.pyc','helloworld.tar.gz.','foo.pyo','bar.pyo','foobar.whl']
patterns=['fl<;>','file[:]','file[:1],file[:2]','file[1:2]','file[:2],file[2:4]','file[:-3:2],file[:-2:-1],file[-1:]']

for i,(file,pattern) in enumerate(zip(files,patterns),start=1):
    try:
        if not file.startswith('.') and file[0].isalpha():
            print "Trial %d: %s=>%s" % (i,file,pattern)

            matches=PATTERNRE.finditer(pattern) ## Use Re to find all "file[x:y:z]"
            ## Convert to  list of tuples
            indices=[(match.group('start'),match.group('end'),match.group('step')) for match in matches]
            ## Using IOError because a lot that can otherwise go wrong comes up as
            ## ValueError (Which is the most-logical ErrorName)
            if not indices: raise IOError('Bad Formatting')

            ## Standardize indexes
            indices=[formatindices(start,end,step,file) for start,end,step in indices]
            print 'Indices: '+'|'.join(','.join(str(item) for item in startendstep) for startendstep in indices)

            ## Build File Structure
            file_strut='\\'.join(file[start:end:step] for start,end,step in indices)
            print 'Folder Structure: '+file_strut

            ## Put together
            subFolder = os.path.join('directoryfrom self.destE', 'somefolder alpfol', file_strut)
            print'Output: '+subFolder+'\n'
    except IOError:
        print 'No Pattern, Skipping\n'