The AFTOVolio phonetic phenomenon representation.

Type synonym to be used for clarity and better code readability.

The Array Int must be sorted in the ascending order to be used in the module correctly.

The String of converted phonetic language representation Char data is converted to this type to apply syllable segmentation or other transformations.

We can think of SegmentationPredFunction in terms of f (SI fN pN) ks [x_{1},x_{2},...,x_{i},...,x_{fN}]. Comparing with divCnsnts from the Syllable we can postulate that it consists of the following logical terms in the symbolic form:

1) phoneType x_{i} `elem` (X{...} = map P [Int8])

2) notEqC ks x_{i} x_{j} (j /= i)

combined with the standard logic Boolean operations of (&&), (||) and not. Further, the not can be transformed into the positive (affirmative) form using the notion of the universal set for the task. This transformation needs that the similar phonetic phenomenae (e. g. the double sounds -- the prolonged ones) belong to the one syllable and not to the different ones (so they are not related to different syllables, but just to the one and the same). Since such assumption has been used, we can further represent the function by the following data type and operations with it, see SegmentationPredFData.

List of the SegmentationRules1 sorted in the descending order by the fieldN SegmentationInfo1 data and where the length of all the SegmentationPredFunction lists of PhoneticPhenomenonRep are equal to the fieldN SegmentationInfo1 data by definition.

Basic group of delimiters in the String that are not converted to the phonemes or not silent sounds. Silent sounds or their absense, e. g. word gaps.

Is used for additional delimiters that are not included in the BasicSpaces group for some reason.

Function sndGroups converts a word being a list of PhoneticPhenomenonRep to the list of phonetically similar (consonants grouped with consonants and each vowel separately) sounds representations in PhoneticPhenomenonRep format.

Function divCnsnts is used to divide groups of consonants into two-elements lists that later are made belonging to different neighbour syllables if the group is between two vowels in a word. The group must be not empty, but this is not checked. The example phonetical information for the proper performance in Ukrainian can be found from the: https:/msn.khnu.km.uapluginfile.php302375mod_resourcecontent1/%D0%9B.3.%D0%86%D0%86.%20%D0%A1%D0%BA%D0%BB%D0%B0%D0%B4.%D0%9D%D0%B0%D0%B3%D0%BE%D0%BB%D0%BE%D1%81.pdf The example of the divCnsnts can be found at: https://hackage.haskell.org/package/aftovolio-0.6.2.0/docs/Aftovolio-Ukrainian-Syllable.html#v:divCnsnts

The function actually creates syllables using the provided data. Each resulting inner-most list is a phonetic language representation of the syllable according to the rules provided.

Is somewhat rewritten from the gBF3 function (not exported) from the mmsyn2-array package.

Function-predicate createsSyllable checks whether its argument is a phoneme representation that every time being presented in the text leads to the creation of the new syllable (in the PhoneticPhenomenonRep format). Usually it is a vowel, but in some languages there can be syllabic phonemes that are not considered to be vowels.

Function-predicate isSonorous1 checks whether its argument is a sonorous consonant representation in the PhoneticPhenomenonRep format.

Function-predicate isVoicedC1 checks whether its argument is a voiced consonant representation in the PhoneticPhenomenonRep format.

Function-predicate isVoiceless1 checks whether its argument is a voiceless consonant representation in the PhoneticPhenomenonRep format.

Binary function-predicate notCreatesSyllable2 checks whether its arguments are both consonant representations in the PhoneticPhenomenonRep format.

Binary function-predicate notEqC checks whether its arguments are not the same consonant sound representations (not taking palatalization into account).