Add to_list_by_hyphen_space processor

[marukaz]

close #871

I ran python -m unittest tests/library/test_postprocessors.py it passed.

[elronbandel]

Not sure about it but i think its easier to have all the logic defined in the prepare file.

[elronbandel]

Suggested change

	ToListByHyphenSpace(field="prediction", process_every_value=False),
	SplitStrip(delimiter="- ", strip_every_element=True, process_every_value=False),

I think it it is cleaner to have it like this rather then define a new class.

[marukaz]

Thank you for the review! Yes, I agree, and in fact, when I personally used this processor, I implemented it in the same file. However, since the implementation of to_list_by_comma defined a class and separated the files, I followed that approach. I don't have a strong preference, so we could either consolidate this processor into one file, refactor to_list_by_comma, or keep it as is following the current implementation of to_list_by_comma. Which option do you prefer?

[elronbandel]

Also, wouldn't be safer to split by "\n-"? What do you think @marukaz?

[marukaz]

@elronbandel I think the leading hyphen will remain when splitting by '\n-'

[elronbandel]

So just to be on the safe side how about:

from unitxt.string_operators import RegexSplit

RegexSplit(by="(?<=^|\n)- ")

Which will split by either "- " if it is at the beginning of the string or if its after "\n"
@marukaz if you do want to go there we can close merge as it is.

[marukaz]

That sounds better! I have updated the code accordingly. Also, I followed your comment and removed the class, consolidating everything into the same file.

[elronbandel]

Neat!