モジュール:uk-pronunciation
ナビゲーションに移動
検索に移動
このモジュールについての説明文ページを モジュール:uk-pronunciation/doc に作成できます
local export = {}
local lang = require("モジュール:languages").getByCode("uk")
local m_IPA = require("モジュール:IPA")
local m_table = require("モジュール:table")
local com = require("モジュール:uk-common")
local u = mw.ustring.char
local rfind = mw.ustring.find
local rsplit = mw.text.split
local rsubn = mw.ustring.gsub
local AC = u(0x301)
local GR = u(0x300)
local vowel_no_i = "ɑɛɪuɔɐoʊe"
local vowel = vowel_no_i .. "i"
local vowel_c = "[" .. vowel .. "]"
local consonant_no_w = "bdzʒᵊɡɦmnlrpftskxʃᵊj"
local consonant_no_w_c = "[" .. consonant_no_w .. "]"
local consonant = consonant_no_w .. "ʋβ̞wʍ"
local consonant_c = "[" .. consonant .. "]"
local palatalizable = "tdsznlrbpʋfɡmkɦxʃᵊʒᵊ"
local palatalizable_c = "[" .. palatalizable .. "]"
-- version of rsubn() that discards all but the first return value
local function rsub(term, foo, bar)
local retval = rsubn(term, foo, bar)
return retval
end
function export.remove_pron_notations(text, remove_grave)
-- Remove grave accents from annotations but maybe not from phonetic respelling
if remove_grave then
text = mw.ustring.toNFC(rsub(mw.ustring.toNFD(text), GR, ""))
end
return text
end
local perm_syl_onset = m_table.listToSet({
'spr', 'str', 'skr', 'spl', 'skl',
'sp', 'st', 'sk', 'sf', 'sx',
'pr', 'br', 'tr', 'dr', 'kr', 'gr', 'ɦr', 'fr', 'xr',
'pl', 'bl', 'kl', 'gl', 'ɦl', 'fl', 'xl',
})
function export.pronunciation(text, allow_unstressed, output, ann)
if type(text) == "table" then
local iparams = {
["output"] = {},
}
local params = {
[1] = {},
["allow_unstressed"] = {type = "boolean"},
["ann"] = {},
["output"] = {},
}
local iargs = require("モジュール:parameters").process(text.args, iparams)
local args = require("モジュール:parameters").process(text:getParent().args, params)
text, allow_unstressed, output, ann = args[1], args.allow_unstressed, iargs.output, args.ann
end
if not text then
text = mw.title.getCurrentTitle().text
end
-- Returns an error if the text contains alphabetic characters that are not Cyrillic.
require("モジュール:script utilities").checkScript(text, "Cyrl")
local origterm = text
-- Lowercase and decompose ѐ and ѝ into letter + accent char
text = mw.ustring.lower(com.decompose_grave(text))
if not allow_unstressed and com.needs_accents(text) then
error("Multisyllabic words that are not prefixes or suffixes must have an acute accent marking the stress, unless allow_unstressed=1 is given: " .. text)
end
-- convert commas and en/en dashes to IPA foot boundaries
text = rsub(text, '%s*[,–—]%s*', ' | ')
-- canonicalize multiple spaces
text = rsub(text, '%s+', ' ')
local phonetic_chars_map = {
-- single characters that map to IPA sounds; these are processed last
[3] = {
["а"] = "ɑ", ["б"] = "b", ["в"] = "ʋ", ["г"] = "ɦ", ["ґ"] = "ɡ",
["д"] = "d", ["е"] = "ɛ", ["є"] = "jɛ", ["ж"] = "ʒᵊ", ["з"] = "z",
["и"] = "ɪ", ["і"] = "i", ["ї"] = "ji", ["й"] = "j", ["к"] = "k",
["л"] = "l", ["м"] = "m", ["н"] = "n", ["о"] = "ɔ", ["п"] = "p",
["р"] = "r", ["с"] = "s", ["т"] = "t", ["у"] = "u", ["ф"] = "f",
["х"] = "x", ["ц"] = "t͡s", ["ч"] = "t͡ʃᵊ", ["ш"] = "ʃᵊ", ["щ"] = "ʃᵊt͡ʃᵊ",
["ь"] = "ʲ", ["ю"] = "ju", ["я"] = "jɑ", ["’"] = "j",
-- accented vowels
[AC] = "ˈ", [GR] = "ˌ",
},
-- character sequences of two that map to IPA sounds
[2] = {
["дж"] = "d͡ʒᵊ", ["дз"] = "d͡z",
-- Dental plosives assimilate to following hissing/hushing consonants, which is not noted in the spelling.
["дс"] = "d͡zs", ["дш"] = "d͡ʒᵊʃᵊ", ["дч"] = "d͡ʒᵊt͡ʃᵊ", ["дц"] = "d͡zt͡s",
["тс"] = "t͡s", ["тш"] = "t͡ʃᵊʃᵊ", ["тч"] = "t͡ʃᵊː", ["тц"] = "t͡sː",
},
-- character sequences of three that map to IPA sounds
[1] = {
["дзь"] = "d͡zʲ",
-- Dental plosives assimilate to following hissing/hushing consonants, which is not noted in the spelling.
["тьс"] = "t͡sʲː"
},
}
local pronuns = {}
-- FIXME, not completely correct, we need to treat hyphens at beginning and end of
-- a word as indicating unstressed pronunciation.
for _, phonetic in ipairs(rsplit(text, "[%s%-]+")) do
phonetic = "#" .. phonetic .. "#"
local orthographic_replacements = {
-- first apply consonant cluster simplifications that always occur orthographically
["нтськ" ] = "ньськ",
["стськ" ] = "ськ",
["нтст" ] = "нст",
["стч" ] = "шч",
["стд" ] = "зд",
["стс" ] = "сː",
["стськ" ] = "ськ",
["#зш" ] = "#шː",
["зш" ] = "жш",
["#зч" ] = "#шч",
["зч" ] = "жч",
-- then long consonants that are orthographically geminated.
["([бвгґд])%1" ] = "%1ː",
["([^д]+)жж" ] = "%1жː", -- джж sequence encode diphonemic дж
["([^д]+)зз" ] = "%1зː", -- дзз sequence encode diphonemic дз
["([йклмнпрстфхцчшщ])%1"] = "%1ː",
["дждж" ] = "джː",
["дздз" ] = "дзː",
}
for regex, replacement in pairs(orthographic_replacements) do
phonetic = rsub(phonetic, regex, replacement)
end
-- remap apostrophe to '!' so that it doesn't conflict with IPA stress mark
phonetic = rsub(phonetic, "'", "!")
-- replace multiple letter sequences
for _, replacements in ipairs(phonetic_chars_map) do
for key, replacement in pairs(replacements) do
phonetic = rsub(phonetic, key, replacement)
end
end
-- move stress mark, added by phonetic_chars_map, before vowel
phonetic = rsub(phonetic, "([ɑɛiɪuɔ])([ˈˌ])", "%2%1")
-- add accent if the word is monosyllabic and not allow_unstressed,
-- so that monosyllabic words without explicit stress marks get stressed
-- vowel allophones; we use a different character from the regular
-- primary stress mark so we can later remove it without affecting
-- explicitly user-added accents on monosyllabic words, as in нема́ за́ що.
local _, numberOfVowels = rsubn(phonetic, "[ɑɛiɪuɔ]", "")
if (numberOfVowels == 1) and not allow_unstressed then
phonetic = rsub(phonetic, "([ɑɛiɪuɔ])", "⁀%1")
end
-- palatalizable consonants before /i/ or /j/ become palatalized
phonetic = rsub(phonetic, "(" .. palatalizable_c .. ")([ː]?)([ˈˌ⁀]?)i", "%1ʲ%2%3i")
phonetic = rsub(phonetic, "(" .. palatalizable_c .. ")([ː]?)j", "%1ʲ%2")
-- eliminate garbage sequences of [ʲːj] resulting from -тьс- cluster followed by [j]
phonetic = rsub(phonetic, "ʲːj", "ʲː")
-- consonant simplification: ст + ц' → [с'ц']. We do it here because of palatalization.
-- Due to the т +ц → [ц:] rule length is present. According to Орфоепскі словник p. 13,
-- both forms are proper, without length in normal (colloquial) speech and with length
-- in slow speech, so we parenthesize the length as optional.
phonetic = rsub(phonetic, "st͡sʲ([ː]?)", "sʲt͡sʲ(%1)")
-- assimilation: voiceless + voiced = voiced + voiced
-- should /ʋ/ be included as voiced? Орфоепічний словник doesn't voice initial cluster of шв (p. 116)
local voiced_obstruent = "[bdzʒᵊɡɦ]"
local voicing = {
["p"] = "b",
["f"] = "v",
["t"] = "d",
["tʲ"] = "dʲ",
["s"] = "z",
["sʲ"] = "zʲ",
["ʃᵊ"] = "ʒᵊ",
["k"] = "ɡ",
["x"] = "ɣ",
["t͡s"] = "d͡z",
["t͡sʲ"] = "d͡zʲ",
["t͡ʃᵊ"] = "d͡ʒᵊ",
["ʃᵊt͡ʃᵊ"] = "ʒᵊd͡ʒᵊ",
}
for voiceless, voiced in pairs(voicing) do
phonetic = rsub(phonetic, voiceless .. "(" .. voiced_obstruent .. "+)", voiced .. "%1")
end
-- In the sequence of two consonants, of which the second is soft, the first is pronounced soft too
-- unless the first consonant is a labial, namely б, п, в, ф, м.
phonetic = rsub(phonetic, "([tdsznl])(.)ʲ", "%1ʲ%2ʲ")
phonetic = rsub(phonetic, "([tdsznl])t͡sʲ", "%1ʲt͡sʲ")
phonetic = rsub(phonetic, "([tdsznl])d͡zʲ", "%1ʲd͡zʲ")
phonetic = rsub(phonetic, "t͡s(.)ʲ", "t͡sʲ%1ʲ")
phonetic = rsub(phonetic, "d͡z(.)ʲ", "d͡zʲ%1ʲ")
phonetic = rsub(phonetic, "d͡zt͡sʲ", "d͡zʲt͡sʲ")
phonetic = rsub(phonetic, "t͡sd͡zʲ", "t͡sʲd͡zʲ")
-- Hushing consonants ж, ч, ш assimilate to the following hissing consonants, giving a long hissing consonant:
-- [ʒᵊ] + [t͡sʲ] → [zʲt͡sʲ], [t͡ʃᵊ] + [t͡sʲ] → [t͡sʲː], [ʃᵊ] + [t͡sʲ] → [sʲt͡sʲ], [ʃᵊ] + [sʲ] → [sʲː]
phonetic = rsub(phonetic, "ʒᵊt͡sʲ", "zʲt͡sʲ")
phonetic = rsub(phonetic, "t͡ʃᵊt͡sʲ", "t͡sʲː")
phonetic = rsub(phonetic, "ʃᵊt͡sʲ", "sʲt͡sʲ")
phonetic = rsub(phonetic, "ʃᵊsʲ", "sʲː")
-- Hissing consonants before hushing consonants within a word assimilate - on зш and зч word-initially and
-- word-medially see above.
-- [s] + [ʃᵊ] → [ʃᵊː], [z] + [ʃᵊ] → [ʒʃᵊ], [z] + [t͡s] → [ʒt͡s]
-- [z] + [d͡ʒᵊ] → [ʒd͡ʒᵊ]
phonetic = rsub(phonetic, "zʒᵊ", "ʒᵊː")
phonetic = rsub(phonetic, "sʃᵊ", "ʃᵊː")
phonetic = rsub(phonetic, "zt͡s", "ʒt͡s")
phonetic = rsub(phonetic, "zd͡ʒᵊ", "ʒd͡ʒᵊ")
-- cleanup: excessive palatalization: CʲCʲCʲ → CCʲCʲ
phonetic = rsub(phonetic, "([^ɑɛiɪuɔ]+)ʲ([^ɑɛiɪuɔ]+)ʲ([^ɑɛiɪuɔ]+)ʲ", "%1%2ʲ%3ʲ")
-- unstressed /ɑ/ has an allophone [ɐ]
phonetic = rsub(phonetic, "([^ˈˌ⁀])ɑ", "%1ɐ")
-- unstressed /u/ has an allophone [ʊ]
phonetic = rsub(phonetic, "([^ˈˌ⁀])u", "%1ʊ")
-- unstressed /ɔ/ has by assimilation an allophone [o] before a stressed syllable with /u/ or /i/
phonetic = rsub(phonetic, "ɔ([bdzʒᵊɡɦmnlrpftskxʲʃ͡ᵊ]+)([ˈˌ⁀][uiʊ])", "o%1%2")
-- one allophone [e] covers unstressed /ɛ/ and /ɪ/
phonetic = rsub(phonetic, "([^ˈˌ⁀])[ɛɪ]", "%1e")
-- Remove the monosyllabic stress we auto-added to ensure that vowels in
-- monosyllabic words get stressed allophones. Do this before vocalizing
-- /ʋ/ and /j/. NOTE: Nothing below should depend on stress marks being
-- present.
phonetic = rsub(phonetic, "⁀", "")
-- /ʋ/ has an allophone [u̯] in a syllable coda
phonetic = rsub(phonetic, "(" .. vowel_c .. ")ʋ([" .. consonant_no_w .. "#])", "%1u̯%2")
-- /ʋ/ has an allophone [w] before /ɔ, u/ and voiced consonants (not after a vowel; [ʋ] before vowel already converted)
phonetic = rsub(phonetic, "ʋ([ˈˌ]?[ɔuoʊbdzʒᵊɡɦmnlr])", "w%1")
-- /ʋ/ has an allophone [β̞] before remaining vowels besides /i/
-- Not sure whether this looks good.
-- phonetic = rsub(phonetic, "ʋ([ˈˌʲ]*[" .. vowel_no_i .. "])", "β̞%1")
-- /ʋ/ has an allophone [ʍ] before before voiceless consonants (not after a vowel; [ʋ] before vowel already converted)
phonetic = rsub(phonetic, "ʋ([pftskxʃᵊ])", "ʍ%1")
-- in a syllable-final position (i.e. the first position of a syllable coda) /j/ has an allophone [i̯]:
phonetic = rsub(phonetic, "(" .. vowel_c .. ")j([" .. consonant_no_w .. "#])", "%1i̯%2")
-- also at the beginning of a word before a consonant
phonetic = rsub(phonetic, "#j(" .. consonant_no_w_c .. ")", "#i̯%1")
-- remove old orthographic apostrophe
phonetic = rsub(phonetic, "!", "")
-- stress mark in correct place
-- (1) Put the stress mark before the final consonant of a cluster (if any).
phonetic = rsub(phonetic, "([^#" .. vowel .. "]?[ʲː]*)([ˈˌ])", "%2%1")
-- (2) Continue moving it over the rest of an affricate with a tie bar.
phonetic = rsub(phonetic, "([^#" .. vowel .. "]͡)([ˈˌ])", "%2%1")
-- (3) Continue moving it over any "permanent onset" clusters (e.g. st, skr, pl, also Cj).
phonetic = rsub(phonetic, "(.)(ʲ?)(" .. consonant_c .. ")(ʲ?)([ˈˌ])(" .. consonant_c .. ")",
function(a, aj, b, bj, stress, c)
if perm_syl_onset[a .. b .. c] then
return stress .. a .. aj .. b .. bj .. c
elseif perm_syl_onset[b .. c] or c == "j" then
return a .. aj .. stress .. b .. bj .. c
else
return a .. aj .. b .. bj .. stress .. c
end
end)
-- (4) If we're in the middle of an affricate with a tie bar, continue moving back
-- if the following consonant is /j/, else move forward.
phonetic = rsub(phonetic, "([^#" .. vowel .. "]͡)([ˈˌ])(.ʲ?j)", "%2%1%3")
phonetic = rsub(phonetic, "([^#" .. vowel .. "]͡)([ˈˌ])(.ʲ?)", "%1%3%2")
-- (5) Move back over any remaining consonants at the beginning of a word.
phonetic = rsub(phonetic, "#([^#" .. vowel .. "]+)([ˈˌ])", "#%2%1")
-- (6) Move back over u̯ or i̯ at the beginning of a word.
phonetic = rsub(phonetic, "#([ui]̯)([ˈˌ])", "#%2%1")
phonetic = rsub(phonetic, "ʲ?ːʲ", "ʲː")
-- use dark [ɫ] for non-palatal /l/
phonetic = rsub(phonetic, "l([^ʲ])", "ɫ%1")
table.insert(pronuns, phonetic)
end
phonetic = rsub(table.concat(pronuns, " "), "#", "")
if output == "template" then
local ipa = m_IPA.format_IPA_full(lang, { { pron = "[" .. phonetic .. "]" } } )
local anntext
if ann == "1" or ann == "y" then
-- remove secondary stress annotations
anntext = "<b class=\"Cyrl\" lang=\"uk\">" .. export.remove_pron_notations(origterm, true) .. "</b>: "
elseif ann then
anntext = "'''" .. ann .. "''': "
else
anntext = ""
end
return anntext .. ipa
else
return phonetic
end
end
return export