ข้ามไปเนื้อหา

มอดูล:fax-IPA

จาก วิกิพจนานุกรม พจนานุกรมเสรี

-- Based on [[Module:es-pronunc]] by: Benwing
-- Adapted by Santi2222

local export = {}

local m_IPA = require("Module:IPA")

local lang = require("Module:languages").getByCode("fax")

local u = require("Module:string/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 GR = u(0x0300) -- grave =  ̀
local CFLEX = u(0x0302) -- circumflex =  ̂
local TILDE = u(0x0303) -- tilde =  ̃
local DIA = u(0x0308) -- diaeresis =  ̈

local vowel = "aeiou"
local V = "[" .. vowel .. "]"
local W = "[jw]" -- glide
local accent = AC .. GR .. CFLEX
local accent_c = "[" .. accent .. "]"
local stress_c = "[" .. AC .. GR .. "]"
local ipa_stress = "ˈˌ"
local ipa_stress_c = "[" .. ipa_stress .. "]"
local separator = accent .. ipa_stress .. "# ."
local separator_c = "[" .. separator .. "]"
local C = "[^" .. vowel .. separator .. "]" -- consonant

local unstressed_words = require("Module:table").listToSet({ --feel free to add more unstressed words
	"o", "u", "us", "os", "a", "as", "da", "das", "do", "dos", "du", "dus", "no", "nos", "nu", "nus", "na", "nas", "un", -- articles and preposition+article contractions
	"me", "mus", "nus", "te", "vus", "le", "se", "mo", "mu", "to", "tu",-- unstressed pronouns and contraction
	"e", "i", "mais", "si", "que",  -- conjunctions
	"con", "cun", "de", "in", "en", "pa", --prepositions
})

-- 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

-- ɟ and ĉ are used internally to represent [ʝ⁓ɟ͡ʝ] and [t͡ʃ]
--
function export.IPA(text)
	local debug = {}
	------------------------------------

	text = ulower(text or mw.title.getCurrentTitle().text)
	-- decompose everything but ñ and ü
	text = mw.ustring.toNFD(text)
	text = rsub(text, ".[" .. TILDE .. DIA .. "]", {
		["n" .. TILDE] = "ñ",
		["u" .. DIA] = "ü",
	})
	-- 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 prefixes unstressed unless they have an explicit stress marker; also make certain
	-- monosyllabic words (e.g. [[el]], [[la]], [[de]], [[en]], etc.) without stress marks be
	-- unstressed.
	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, "-y", "y") --intercept the pronoun "-y"
	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, " ", "# #") .. "##"

	table.insert(debug, text)

	--c, g, q
	text = rsub(text, "c([ie])", "θ%1")
	text = rsub(text, "gü([ie])", "ɡw%1")
	text = rsub(text, "gu([ie])", "ɡ%1") -- special IPA g, not normal g
	text = rsub(text, "g([ie])", "j%1") -- <ge,gi> is pronounced as <je,ji>, to be handled later
	text = rsub(text, "ng([^aeiouüwhlr])", "n%1") -- [[Bangkok]], [[ángstrom]], [[branding]]
	text = rsub(text, "qu([ie])", "k%1")
	text = rsub(text, "z", "θ")

	table.insert(debug, text)

	--digraphs
	text = rsub(text, "ch", "č") --not the real symbol
	text = rsub(text, "dx", "đ") --not the real symbol
	text = rsubb(text, "ll", "ʎ")
	text = rsubb(text, "nh", "ŋ")
	text = rsubb(text, "xh", "ʒ")
	text = rsubb(text, "sh", "z")

	text = rsub(text, "#ps", "#s") -- [[psicología]]
	
	text = rsub(text, "[cgjñrvxy]",
			{ ["c"] = "k", ["g"] = "ɡ", ["j"] = "x", ["ñ"] = "ɲ", ["r"] = "ɾ", ["v"] = "b", ["x"] = "ʃ", ["y"] = "j"})

	-- trill in #r, lr, nr, sr, rr
	text = rsub(text, "ɾɾ", "r")
	text = rsub(text, "([#lns])ɾ", "%1r")

	--assimilation of nasals and sibilants
	text = rsub(text, "n([# .]*[bpm])", "m%1")
	text = rsub(text, "n([# .]*[kɡ])", "ŋ%1")
	text = rsub(text, "s([# .]*[nml])", "z%1")

	--double uu and ii
	text = rsubb(text, "uu", "wu")
	text = rsubb(text, "ii", "ji")

	table.insert(debug, text)

	--syllable division
	local break_iu = { ["i"] = "i.", ["u"] = "u." }
	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)([iu])(" .. V .. ")",
			function(v1, iu, v2)
				return v1 .. break_iu[iu] .. v2
			end
	)
	text = rsub(text, "i([aeou])", "j%1")
	text = rsub(text, "u([aeo])", "w%1")

	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C .. W .. "?" .. 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(text, "([pbktdɡ])%.([lɾ])", ".%1%2")
	text = rsub_repeatedly(text, "(" .. C .. ")%.s(" .. C .. ")", "%1s.%2")
	-- Any aeo, or stressed iu, should be syllabically divided from a following aeo or stressed iu.
	text = rsub_repeatedly(text, "([aeo]" .. accent_c .. "*)([aeo])", "%1.%2")
	text = rsub_repeatedly(text, "([aeo]" .. accent_c .. "*)(" .. V .. stress_c .. ")", "%1.%2")
	text = rsub(text, "([iu]" .. stress_c .. ")([aeo])", "%1.%2")
	text = rsub_repeatedly(text, "([iu]" .. stress_c .. ")(" .. V .. stress_c .. ")", "%1.%2")
	text = rsub_repeatedly(text, "i(" .. accent_c .. "*)i", "i%1.i")
	text = rsub_repeatedly(text, "u(" .. accent_c .. "*)u", "u%1.u")

	table.insert(debug, text)

	--diphthongs
	text = rsub(text, "([aeou])i", "%1ɪ")
	text = rsub(text, "([aeio])u", "%1ʊ")

	table.insert(debug, text)

	local accent_to_stress_mark = { [AC] = "ˈ", [GR] = "ˌ", [CFLEX] = "" }

	local function accent_word(word, syllables)
		-- Now stress the word. If any accent exists in the word (including ^ indicating an unaccented word),
		-- put the stress mark(s) at the beginning of the indicated syllable(s). Otherwise, apply the default
		-- stress rule.
		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
			-- Default stress rule. Words without vowels (e.g. IPA foot boundaries) don't get stress.
			if #syllables > 1 and rfind(word, "[^aeiouns#]#") or #syllables == 1 and rfind(word, "[aeiou]") then
				syllables[#syllables] = "ˈ" .. syllables[#syllables]
			elseif #syllables > 1 and rfind(word, "[ɪʊ][ns#]") then
				syllables[#syllables] = "ˈ" .. syllables[#syllables]
			elseif #syllables > 1 then
				syllables[#syllables - 1] = "ˈ" .. syllables[#syllables - 1]
			end
		end
	end

	local words = rsplit(text, " ")
	for j, word in ipairs(words) do
		-- accentuation
		local syllables = rsplit(word, "%.")

		if rfind(word, "men%.ti#") then
			local mente_syllables
			-- Words ends in -mente (converted above to ménte); add a stress to the preceding portion
			-- (e.g. [[agriamente]] -> 'ágriaménte') unless already stressed (e.g. [[rápidamente]]).
			-- It will be converted to secondary stress further below. Essentially, we rip the word apart
			-- into two words ('mente' and the preceding portion) and stress each one independently.
			mente_syllables = {}
			mente_syllables[2] = table.remove(syllables)
			mente_syllables[1] = table.remove(syllables)
			accent_word(table.concat(syllables, "."), syllables)
			accent_word(table.concat(mente_syllables, "."), mente_syllables)
			table.insert(syllables, mente_syllables[1])
			table.insert(syllables, mente_syllables[2])
		else
			accent_word(word, syllables)
		end
		
		-- Reconstruct the word.
		words[j] = table.concat(syllables, "")
	end

	text = table.concat(words, " ")

	--make all primary stresses but the last one be secondary
	text = rsub_repeatedly(text, "ˈ(.+)ˈ", "ˌ%1ˈ")

	table.insert(debug, text)

	-- remove silent "h" and convert fake symbols to real ones
	local final_conversions = {
		["h"] = "",
		["č"] = "t͡ʃ", 
		["đ"] = "d͡ʒ",
		["ɪ"] = "i̯",
		["ʊ"] = "u̯",
	}

	text = rsub(text, "[hčđɪʊ]", final_conversions)

	-- 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) .. "/" })

	local pre = args.pre and args.pre .. " " or ""

	return "* " .. pre .. m_IPA.format_IPA_full { lang = lang, items = results }
end

return export