Wrong os_sorted sorting with special character in filename

[toroConverter]

Describe the bug It can happen that the os_sorted functionality does not sort correctly files with special character inside

Expected behavior File sorting equal to Windows Explorer

Environment (please complete the following information):

• Python Version: 3.9
• Natsort Version: 8.0.2
• OS Windows
• If the bug involves LOCALE or humansorted:
  • Is PyICU installed? No
  • Do you have a locale set? If so, to what? Italian

To Reproduce

from natsort import os_sorted

file_list = ['Try.Me.Bug - 09 - One.Two.Three.[text].mkv',
             'Try.Me.Bug - 07 - One.Two.5.[text].mkv',
             'Try.Me.Bug - 08 - One.Two.Three[text].mkv']

file_list2 = ['TryMe - 02 - One Two Three [text].mkv',
              'TryMe_-_03_-_One_Two_Three_[text].mkv',
              'TryMe_-_01_-_One_Two_Three_[text].mkv']

for file in os_sorted(file_list):
    print(file)

for file in os_sorted(file_list2):
    print(file)

Expected sorting:

file_list Try.Me.Bug - 07 - One.Two.5.[text].mkv Try.Me.Bug - 08 - One.Two.Three[text].mkv Try.Me.Bug - 09 - One.Two.Three.[text].mkv

filelist2 TryMe_-_01_-_One_Two_Three_[text].mkv TryMe - 02 - One Two Three [text].mkv TryMe_-_03-_One_Two_Three_[text].mkv

Actual sorting

file_list Try.Me.Bug - 08 - One.Two.Three[text].mkv Try.Me.Bug - 09 - One.Two.Three.[text].mkv Try.Me.Bug - 07 - One.Two.5.[text].mkv

filelist2 TryMe - 02 - One Two Three [text].mkv TryMe_-_01_-_One_Two_Three_[text].mkv TryMe_-_03_-_One_Two_Three\[text].mkv

---

At the moment the only way to overcome this issue is to use os_sorted along with key=lambda x: re.sub(r'[^a-zA-Z0-9]+', ' ', x)

[SethMMorton]

I agree that the Try.Me.Bug - 07 - One.Two.5.[text].mkv case was a bug (at least undesired behavior) and I have a fix proposed in #146.

However, the TryMe_-_01_-_One_Two_Three_[text].mkv case is not a bug, and you would need to use something like key=lambda x: x.replace("_", " ") to sort it properly. This is because the code has no way to tell that you want to treat "_" and " " the same unless you tell it to.

[toroConverter]

@SethMMorton ok thanks a lot for your feedback and your work!