Cyrillic to Latin File Changer

cyrillic2latin_file_renamer.py

#!/usr/bin/python
# -*- coding: utf-8 -*-

from os import walk, rename, unlink, mkdir, remove
from os.path import isdir, exists
from sys import argv, exit, getfilesystemencoding
from shutil import copyfile
import shutil

conversion = {
	u'А': u'A',
        u'Б': u'B',
        u'В': u'V',
        u'Г': u'G',
        u'Д': u'D',
        u'Е': u'E',
        u'Ё': u'E',
        u'Ж': u'Zh',
        u'З': u'Z',
        u'И': u'I',
        u'Й': u'Y',
        u'К': u'K',
        u'Л': u'L',
        u'М': u'M',
        u'Н': u'N',
        u'О': u'O',
        u'П': u'P',
        u'Р': u'R',
        u'С': u'S',
        u'Т': u'T',
        u'У': u'U',
        u'Ф': u'F',
        u'Х': u'H',
        u'Ц': u'Ts',
        u'Ч': u'Ch',
        u'Ш': u'Sh',
        u'Щ': u'Sch',
        u'Ъ': u'',
        u'Ы': u'Y',
        u'Ь': u'',
        u'Э': u'E',
	u' ' : '_',
        }
def cyr2lat(s):
    retval = ""
    for c in s:
	if ord(c) > 128: 
		try: 
			c = conversion[c.upper()].lower() 
		except KeyError: 
			c=''
	elif c == ' ':
		c = '_'
        retval += c
    return retval
    
if len(argv) == 1:
    print "Usage: %s <dirs>" % argv[0]
    exit(-1)

processed = []

def recursive_walk(dir):
    # See http://docs.activestate.com/activepython/2.5/whatsnew/2.3/node6.html
    found = []
    dir = unicode(dir)
    for finfo in walk(dir, True):
        dirnames = finfo[1]
        fnames = finfo[2]
        for subdir in dirnames:
            subdir = "%s/%s" % (dir, subdir)
            if subdir in processed:
                continue
            for yield_val in recursive_walk(subdir):
                yield yield_val
        for fname in fnames:
            yield '%s/%s' % (dir, fname)
    raise StopIteration

if __name__ == "__main__":
    fs_enc = getfilesystemencoding()
    for dir in argv[1:]:
        for fpath in recursive_walk(dir):
            new_fpath = cyr2lat(fpath).lower()
	    print 'new path %s' % new_fpath
            print fpath.encode('utf-8')
            # First make dirs
            path_elts = new_fpath.split('/')
            for idx in range(len(path_elts))[1:]:
                subpath = '/'.join(path_elts[:idx])
                while True:
                    i = 0
                    if exists(subpath):
                        if not isdir(subpath):
                            print '%s exists but is not a directory, will try again' % subpath
                            subpath += str(i)
                            continue
                        else:
                            path_elts[idx - 1] = subpath.split('/')[-1]
                            break
                    else:
                        print 'Creating directory: %s' % subpath
                        mkdir(subpath)
                        break
            print 'Copying %s to %s' % (fpath, new_fpath)
            shutil.copyfile(fpath, new_fpath)
	    remove(fpath)

hi,
do you want bug report?

prerequisites:
wget https://gist.githubusercontent.com/braman/a8504c15ca537ea49c6a/raw/cyrillic2latin_file_renamer.py
chmod a+x ./cyrillic2latin_file_renamer.py
ls

.
..
243
cyrillic2latin_file_renamer.py

./cyrillic2latin_file_renamer.py .

result:

new path ./243
./243
Copying ./243 to ./243
Traceback (most recent call last):
  File "./cyrillic2latin_file_renamer.py", line 106, in <module>
    shutil.copyfile(fpath, new_fpath)
  File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/shutil.py", line 69, in copyfile
    raise Error("`%s` and `%s` are the same file" % (src, dst))
shutil.Error: `./243` and `./243` are the same file

---

it is because file after and before renaming is same.

Hi,
how to change the script to rename the file only with new name ?
Because the names are duplicated:
adriana - znaya, znaya adriana - znaq, znaq.mp3

Hi,
I've fixed errors in the script. You can get the modified version here.

Various bugs fixed in this version:

https://gist.github.com/tigran123/aafee6f58295ce6e6c06af9eb0aab0fb

First of all, the case should be preserved (no need to lowercase). Second, you forgot two letters: Ю and Я. Third, you should not copy the file as this is extremely slow (on a multi-terabyte library) and will fail if the amount of free filesystem space is less than the size of the file. Instead, just use rename().