#!/usr/bin/env python
# -*- coding: UTF-8 -*-

# Licensed under the terms of the GNU GPLv3 or later
# Proprietary Software Must Perish!

import os
import re
import sys
from optparse import OptionParser

USAGE = "%prog [OPTION]... PATHNAME..."
VERSION = "%prog 0.2"
DESCRIPTION = """%prog runs through a list of given PATHNAMEs, trying to rename them in a 
way that makes sense for heavy users of the bourne shell
(e. g. by stripping whitespace and replacing problematic characters)."""

op = OptionParser(version = VERSION, usage = USAGE, description = DESCRIPTION )
op.add_option("-v", "--verbose", action = "store_true", dest = "verbose",
    help = "be verbose about what is going on")
op.add_option("-n", "--no-dirs", action = "store_true", dest = "nodirs",
    help = "do not rename directories")
op.add_option("-d", "--dry-run", action = "store_true", dest = "dryrun",
    help = "do not modify the filesystem/actually rename inodes")
op.add_option("-r", "--recursive", action = "store_true", dest = "recur",
    help = "recursively process directories")
op.add_option("-s", "--subst", action = "store", dest = "subst_str",
    default = "_", help = "substitution string for problematic characters (Default: _)")
op.add_option("-p", "--preserve-case", action = "store_true", dest = "preserve_case",
    default = False, help = "do not convert names to lowercase")
op.add_option("-c", "--collapse-underscores", action = "store_true", dest = "collapse_underscores",
    default = False, help = "collapse multiple occurences of _ to a single one")
(opts, args) = op.parse_args()

#hardcoded substitution table; annoying chars will be replaced by
#opts.subst_str

subst_list = [
(r'ä', 'ae'),
(r'ö', 'oe'),
(r'ü', 'ue'),
(r'ß', 'sz'),
(r'[]})({[ \t\n!"§$%&?`\'\\:*~°<>|^@]', opts.subst_str),
]
if opts.collapse_underscores == True:
    subst_list.append((r'__*', '_'))

#yuck! some day, this'll have to deal with unicode properly, I guess...
if opts.preserve_case == True:
    subst_additional_pairs = [(r'Ä', 'Ae'), (r'Ö', 'Oe'), (r'Ü', 'Ue')]
else:
    subst_additional_pairs = [(r'Ä', 'ae'), (r'Ö', 'oe'), (r'Ü', 'ue')]
for subst_pair in subst_additional_pairs:
    subst_list.append(subst_pair)

replace_list = []
for replace_seq in subst_list:
    replace_list.append((re.compile(replace_seq[0], (re.U)), replace_seq[1]))

#file-counter for verbose modes
f_seen = 0
f_renamed = 0


def sanitize_name(current_path):
    global f_seen
    global f_renamed
    f_seen += 1

    if (opts.nodirs and os.path.isdir(current_path)):
        if opts.verbose:
            print >> sys.stderr, current_path, "is a directory, skipping."
        return True

    (current_dir, current_file) = os.path.split(current_path)

    if opts.preserve_case == False:
        clean_file = current_file.lower()
    else:
        clean_file = current_file
    for (old, new) in replace_list:
        clean_file = re.sub(old, new, clean_file, 0)

    if clean_file != current_file:
        #avoid overwriting existing files, in a way similar to wget(1)
        while os.path.exists(os.path.join(current_dir, clean_file)):
            clean_file = clean_file + ".1"
        if opts.verbose:
            print >> sys.stderr, current_path, "->", os.path.join(current_dir, clean_file)
        if not opts.dryrun:
            try:
                os.rename(os.path.join(current_dir, current_file), os.path.join(current_dir, clean_file))
            except Exception, e:
                print >> sys.stderr, "An error occured while trying to rename", current_path, e
                return False
        f_renamed += 1
    return True

#main processing loop
if opts.recur == True:
    for pathname in args:
        if(os.path.isdir(pathname)):
            for root, dirs, files in os.walk(pathname, topdown = False):
                for name in files:
                    sanitize_name(os.path.join(root, name))
                for name in dirs:
                    sanitize_name(os.path.join(root, name))
        else:
            sanitize_name(os.path.join(root, pathname))
else:
    for pathname in args:
        sanitize_name(pathname)
if opts.verbose: 
    print >> sys.stderr, f_renamed, "of", f_seen, "files renamed."
if f_seen == 0: 
    print >> sys.stderr, "No pathnames given.\n"
    op.print_help()
if opts.dryrun: 
    print >> sys.stderr, "Dry-run requested; no changes were made."