Lao wrong transliteration

[aleray]

Hi,

your project looks wonderful. I came across it while looking for Lao romanization. I think there is a mistake in the following example...

import epitran
epi = epitran.Epitran('lao-Laoo')
res = epi.transliterate(u'ຂ້ອຍມັກກາເຟ')
print(res)
'kʰɔːɲmakkaːfeː'

rom = epitran.reromanize.ReRomanizer('lao-Laoo', 'anglocentric')
print(rom.reromanize(res))
'koonymakkaafee'

the characters "ອ" and "ຍ" in "ຂ້ອຍ" are consonants, but together — and in this case — they form the (complex) vowel/diphtong "ອຍ". See: https://r12a.github.io/scripts/laoo/lo#diphthongsz.

I think the "ຍ" got treated here as the consonants "ny", but should not. I'm not a specialist so please forgive me if I'm wrong.

Thanks

[martinarisk]

Is kʰɔːj˥˥˨mak̚˧kaː˩˨feː˩˨[1] correct?

[1] IPA Phonemization at https://www.hashtron.cloud/

[aleray]

Hello,

actually I not familiar with IPA. Are "˥˥˨˧˩˨˩˨" tone signs, aren't they? If I remove them, "kʰɔːjmakkaːfeː" seems to be better/correct, as far as I can tell.

The consonnant "ຂ" is however more often romanized as "kh". I see it in the IPA ("kʰ") but not in the romanization, as shown below

In [4]: import epitran
   ...: epi = epitran.Epitran('lao-Laoo')
   ...: rom = epitran.reromanize.ReRomanizer('lao-Laoo', 'anglocentric')
   ...: print(rom.reromanize("kʰɔːjmakkaːfeː"))
kooymakkaafee

Thanks