local export = {}
local m_IPA = require("Module:IPA")
local lang = require("Module:languages").getByCode("pga")
local u = mw.ustring.char
local rfind = mw.ustring.find
local rsubn = mw.ustring.gsub
local rsplit = mw.text.split
local ulower = mw.ustring.lower
local AC = u(0x0301) -- acute = ́
local CFLEX = u(0x0302) -- circumflex = ̂
local vowel = "aɛɪɔʊ"
local V = "[" .. vowel .. "]"
local accent = AC .. CFLEX
local accent_c = "[" .. accent .. "]"
local ipa_stress = "ˈˌ"
local ipa_stress_c = "[" .. ipa_stress .. "]"
local separator = accent .. ipa_stress .. "# ."
local C = "[^" .. vowel .. separator .. "]" -- consonant
local unstressed_words = require("Module:table").listToSet({ }) --add here unstressed words
-- 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
-- version of rsubn() that returns a 2nd argument boolean indicating whether
-- a substitution was made.
local function rsubb(term, foo, bar)
local retval, nsubs = rsubn(term, foo, bar)
return retval, nsubs > 0
end
-- apply rsub() repeatedly until no change
local function rsub_repeatedly(term, foo, bar)
while true do
local new_term = rsub(term, foo, bar)
if new_term == term then
return term
end
term = new_term
end
end
--
function export.IPA(text, phonetic)
text = ulower(text or mw.title.getCurrentTitle().text)
-- convert commas and en/en dashes to IPA foot boundaries
text = rsub(text, "%s*[,–—]%s*", " | ")
-- question mark or exclamation point in the middle of a sentence -> IPA foot boundary
text = rsub(text, "([^%s])%s*[¡!¿?]%s*([^%s])", "%1 | %2")
-- canonicalize multiple spaces and remove leading and trailing spaces
local function canon_spaces(text)
text = rsub(text, "%s+", " ")
text = rsub(text, "^ ", "")
text = rsub(text, " $", "")
return text
end
text = canon_spaces(text)
-- Make certain monosyllabic words unstressed (not implemented yet)
local words = rsplit(text, " ")
for i, word in ipairs(words) do
if rfind(word, "%-$") and not rfind(word, accent_c) or unstressed_words[word] then
-- add CFLEX to the last vowel not the first one, or we will mess up 'que' by
-- adding the CFLEX after the 'u'
words[i] = rsub(word, "^(.*" .. V .. ")", "%1" .. CFLEX)
end
end
text = table.concat(words, " ")
-- Convert hyphens to spaces
text = rsub(text, "%-", " ")
-- canonicalize multiple spaces again, which may have been introduced by hyphens
text = canon_spaces(text)
-- now eliminate punctuation
text = rsub(text, "[¡!¿?']", "")
-- put # at word beginning and end and double ## at text/foot boundary beginning/end
text = rsub(text, " | ", "# | #")
text = "##" .. rsub(text, " ", "# #") .. "##"
--vowels
text = rsub(text, "aa", "a" .. AC)
text = rsub(text, "ee", "e" .. AC)
text = rsub(text, "ii", "i" .. AC)
text = rsub(text, "oo", "o" .. AC)
text = rsub(text, "uu", "u" .. AC)
--transcription
text = rsub(text, "sh", "ɕ")
text = rsub(text, "ng", "ŋ") --these two sounds do
text = rsub(text, "ny", "ɲ")
text = rsub(text, "[jg'fvyreiou]",
{ ["j"] = "ɟ", ["g"] = "ɡ", ["'"] = "ʔ", ["f"] = "ɸ", ["v"] = "β", ["y"] = "j", ["r"] = "ɾ", ["e"] = "ɛ", ["i"] = "ɪ", ["o"] = "ɔ", ["u"] = "ʊ"})
-- /e/,/i/,/o/,/u/ are open except in stressed syllables when not followed by /ɾ/.
--syllable division (rudimentary)
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C .. "?" .. V .. ")", "%1.%2")
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*" .. C .. ")(" .. C .. V .. ")", "%1.%2")
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*" .. C .. "+)(" .. C .. C .. V .. ")", "%1.%2")
text = rsub_repeatedly(text, "(" .. C .. ")%.s(" .. C .. ")", "%1s.%2")
local accent_to_stress_mark = { [AC] = "ˈ", [CFLEX] = "" }
local function accent_word(word, syllables)
if rfind(word, accent_c) then
for i = 1, #syllables do
syllables[i] = rsub(syllables[i], "^(.*)(" .. accent_c .. ")(.*)$",
function(pre, accent, post)
return accent_to_stress_mark[accent] .. pre .. post
end
)
end
else
syllables[1] = "ˈ" .. syllables[1]
end
end
local words = rsplit(text, " ")
for j, word in ipairs(words) do
-- accentuation
local syllables = rsplit(word, "%.")
accent_word(word, syllables)
--mark stressed syllables with an accent (to be used later), no need to mark stressed /a/
if phonetic then
for i = 1, #syllables do
if rfind(syllables[i], "ˈ") then
syllables[i] = rsub(syllables[i], "[ɛɪɔʊ]", {["ɛ"] = "é", ["ɪ"] = "í", ["ɔ"] = "ó", ["ʊ"] = "ú"})
end
end
end
-- Reconstruct the word.
if phonetic then
words[j] = table.concat(syllables, ".")
else
words[j] = table.concat(syllables, "")
end
end
text = table.concat(words, " ")
-- suppress syllable mark before IPA stress indicator
text = rsub(text, "%.(" .. ipa_stress_c .. ")", "%1")
--The vowels /e/, /i/, /o/, /u/ are open in unstressed syllables and stressed syllables when followed by /ɾ/. Only use <aeiou> in phonemic transcription.
if phonetic then
--FIXME: the following code works but could be more compact
text = rsub(text, "éɾ", "ɛɾ")
text = rsub(text, "é%.ɾ", "ɛ.ɾ")
text = rsub(text, "íɾ", "ɪɾ")
text = rsub(text, "í%.ɾ", "ɪ.ɾ")
text = rsub(text, "óɾ", "ɔɾ")
text = rsub(text, "ó%.ɾ", "ɔ.ɾ")
text = rsub(text, "úɾ", "ʊɾ")
text = rsub(text, "ú%.ɾ", "ʊ.ɾ")
text = rsub(text, "[éíóú]", {["é"] = "e", ["í"] = "i", ["ó"] = "o", ["ú"] = "u"})
else
text = rsub(text, "[ɛɪɔʊ]", {["ɛ"] = "e", ["ɪ"] = "i", ["ɔ"] = "o", ["ʊ"] = "u"})
end
--mark /h/ as optional
text = rsub(text, "h", "(h)")
-- remove # symbols at word and text boundaries
text = rsub(text, "#", "")
return mw.ustring.toNFC(text)
end
function export.show(frame)
local params = {
[1] = {},
["pre"] = {},
}
local parargs = frame:getParent().args
local args = require("Module:parameters").process(parargs, params)
local results = {}
local text = args[1] or mw.title.getCurrentTitle().text
table.insert(results, { pron = "/" .. export.IPA(text, false) .. "/" })
table.insert(results, { pron = "[" .. export.IPA(text, true) .. "]" })
local pre = args.pre and args.pre .. " " or ""
return "* " .. pre .. m_IPA.format_IPA_full(lang, results)
end
return export
