[BUG] Normalizers always tranform input to lowercase

[umoqnier]

Describe the bug
Normalizers always tranform input to lowercase and information can be missing for other NLP tasks

To Reproduce

Steps to reproduce the behavior:

from elotl.otomi.orthography import Normalizer
N = Normalizer("ots")
input_str = "Ra ngaho ra tso̱ y'o̱te:"
print(N.normalize(input_str))
>>> "ra ngajo ra tso̱ y'o̱te:"

Expected behavior

from elotl.otomi.orthography import Normalizer
N = Normalizer("ots")
input_str = "Ra ngaho ra tso̱ y'o̱te:"
print(N.normalize(input_str))
>>> "Ra ngajo ra tso̱ y'o̱te:"

Desktop (please complete the following information):

• OS: Any
• Version: 0.0.1.16

[Lguyogiro]

We could solve this easily for the cases where the whole word is upper case or is in title case. For cases of arbitrary capitalization it is a bit trickier since we would need to align them. The other solution is to update all of the normalizers to include a mayuscula-2-mayuscula version (essentially copying the existing lexc files and uppercasing them (assuming none of the supported orthographies use capitalization to distinguish between phonemes).

I am happy to do either one.

[umoqnier]

Ok, I think that if we solve the first scenario, it will be enough by the moment. For the second one (arbitrary caps) maybe we can do it in a pair programming session online. That can be helpful for my to understand the inside machinery of the package :)

[ximenina]

In my opinion, the mayuscula-2-mayuscula would be a better option for the long term.