adapt case-conversion commands for Greek final sigma char

[Alexey-T]

from #6028 by @Doryforos:

I wonder whether it would be equally not complicated to gracefully deal with final sigma issues in case conversions.

Example:
Θουκυδίδης -> ΘΟΥΚΥΔΊΔΗΣ[ -> Θουκυδίδης]

[Alexey-T]

@Doryforos code of Greek_remove_accent plugin is VERY simple so I will embed its code to app. on case-conv commands.

it is ok logic for case-conv commands?

• first we do old code of case-convert
• next, we replace 'usual lowercase sigma' at word ends with 'final lowercase sigma'
• next, we do what that plugin does. loot at its file __init__.py. it replaces accented uppercase chars with not accented chars.

def do_format(s):

    s = s.replace('Ά' , 'Α')
    s = s.replace('Έ' , 'Ε')
    s = s.replace('Ή' , 'Η')
    s = s.replace('Ί' , 'Ι')
    s = s.replace('Ϊ' , 'Ι')
    s = s.replace('Ϊ́' , 'Ι')
    s = s.replace('Ό' , 'Ο')
    s = s.replace('Ύ' , 'Υ')
    s = s.replace('Ώ' , 'Ω')
    s = s.replace('ά' , 'α')
    s = s.replace('έ' , 'ε')
    s = s.replace('ή' , 'η')
    s = s.replace('ί' , 'ι')
    s = s.replace('ϊ' , 'ι')
    s = s.replace('ΐ' , 'ι')
    s = s.replace('ό' , 'ο')
    s = s.replace('ύ' , 'υ')
    s = s.replace('ώ' , 'ω')
    return s                     

• question: do we need to replace LOWERCASE accented chars to not accented? (plugin does it too, why?)

[Doryforos]

I am not certain about:

• s = s.replace('Ϊ́' , 'Ι'), we shall see it in action.
• 2nd Step: replace 'usual lowercase sigma' at word ends with 'final lowercase sigma'
  assuming this is done when converting to lowercase, is this step needed? [Why not convert directly to final sigma?]

Sometimes, being able to also remove accents from lowercase Greek may come in handy [dumb Linux/Unix terminals or terminal emulators, for one] — it is good to have this option available.

[Alexey-T]

I made the change, tested it a little. works as I wrote above in my plan.

beta win64
cudatext.zip

[Alexey-T]

note:
I corrected default value of opt like here:

  "nonword_chars": "-+*=/\\()[]{}<>\"'.,:;~?!@#$%^&|`…«»",

at the and of value I added 2 quote chars. do it too in your user.json. To avoid case-convert-problems with sample Greek text which has such quotes.

[Doryforos]

Thanks, however, with this EXE, final sigma is not converted to capital.

[Alexey-T]

wait, I confirm this bug - on win10 (on linux all is OK).

[Alexey-T]

new fix made, now it seems all OK:

cudatext.exe.zip

[Doryforos]

Yes — and I have done nothing regarding non-word characters [will keep using cudatext with this configuration, until I encounter some problem].

Thanks!

[Doryforos]

ΐ [U+0390: Greek small iota with tonos and dialytika (or diaeresis)], should convert to capital iota with dialytika only.

see #6030 (comment)

[Alexey-T]

we need which of these changes? ->

1st. U+0390 must convert to U+03AA

2nd. remove this code (adapted from Greek_...... plugin)

  S:= UnicodeStringReplace(S, 'Ϊ' , 'Ι', [rfReplaceAll]);

3nd. remove this code (adapted from Greek_...... plugin)

  S:= UnicodeStringReplace(S, 'Ϊ́' , 'Ι', [rfReplaceAll]);

i guess: 1st must be added, 2nd must be left, 3rd must be removed?

[Alexey-T]

After my guess (see above 'i guess.....') new beta:

cudatext.exe.zip