Module:gl-pron

--Based on Module:es-pronunc by Benwing2

local export = {}

local m_IPA = require("Module:IPA") local m_table = require("Module:table") local audio_module = "Module:audio" local put_module = "Module:parse utilities"

local force_cat = false

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

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 usub = mw.ustring.sub local ulen = mw.ustring.len local unfd = mw.ustring.toNFD local unfc = mw.ustring.toNFC

local AC = u(0x0301) -- acute = ́ local GR = u(0x0300) -- grave = ̀ local CFLEX = u(0x0302) -- circumflex = ^ local CARO = u(0x030C) -- caron local TILDE = u(0x0303) -- tilde = ̃ local DIA = u(0x0308) -- diaeresis = ̈ local DOTA = u(0x0307) -- dot above local DOTB = u(0x0323) --dot below local BREVE = u(0x0306) --breve

local vowel = "aeiouüɛɔɐɪʊɑAEIOUÜ" -- vowel local V = "[" .. vowel .. "]" -- vowel class local accent = AC .. GR .. CARO local accent_c = "[" .. accent .. "]" local stress = AC local stress_c = "[" .. AC .. "]" local ipa_stress = "ˈˌ" local ipa_stress_c = "[" .. ipa_stress .. "]" local sylsep = "%-." -- hyphen included for syllabifying from spelling local sylsep_c = "[" .. sylsep .. "]" local wordsep = "# " local separator_not_wordsep = accent .. ipa_stress .. sylsep local separator = separator_not_wordsep .. wordsep local separator_c = "[" .. separator .. "]" local C = "[^" .. vowel .. separator .. "]" -- consonant class

local unstressed_words = m_table.listToSet({ --all words in this list containing non-final  should be listed later (look for "handle vowels"). FIXME add more unstressed words	"a", "as", "o", "os", -- definite articles	"con", "de", "en", "por", "sen", "so", --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

local function decompose(text) -- decompose everything but ñ and ü text = unfd(text) text = rsub(text, ".[" .. TILDE .. DIA .. "]", {		["n" .. TILDE] = "ñ",		["N" .. TILDE] = "Ñ",		["u" .. DIA] = "ü",		["U" .. DIA] = "Ü",	}) return text end

local function split_on_comma(term) if term:find(",%s") then return require(put_module).split_on_comma(term) else return rsplit(term, ",") end end

-- Remove any HTML from the formatted text and resolve links, since the extra characters don't contribute to the -- displayed length. local function convert_to_raw_text(text) text = rsub(text, "<.->", "") if text:find("%[%[") then text = require("Module:links").remove_links(text) end return text end

-- Return the approximate displayed length in characters. local function textual_len(text) return ulen(convert_to_raw_text(text)) end

local function construct_default_differences(dialect) if dialect == "Standard" then return { need_seseo = false, gheada_different = false, }	end return nil end

-- Main syllable-division algorithm local function syllabify_from_spelling_or_pronun(text) -- Part 1: Divide before the last consonant in a cluster of consonants between vowels. -- Then move the syllable division marker leftwards over clusters that can form onsets. text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C .. V .. ")", "%1.%2")	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*" .. C .. "+)(" .. C .. V .. ")", "%1.%2")	local cluster_r = is_spelling and "rɾ" or "ɾ" -- Don't divide Cl or Cr where C is a stop or fricative, except for dl. text = rsub(text, "([pbfvkctg])%.([l" .. cluster_r .. "])", ".%1%2")	text = text:gsub("d%.([" .. cluster_r .. "])", ".d%1") -- Don't divide ll or rr. text = rsub(text, "([lr])%.%1", ".%1%1")

-- Part 2: Divide hiatuses. Any aeo, or stressed iuüy, should be syllabically divided from a following aeo or -- stressed iuüy. Also divide ii and uu sequences text = rsub_repeatedly(text, "([aeoɛɔAEO]" .. accent_c .. "*)(h?[aeoɛɔ])", "%1.%2") text = rsub_repeatedly(text, "([aeoɛɔAEO]" .. accent_c .. "*)(h?" .. V .. stress_c .. ")", "%1.%2")	text = rsub(text, "([iuüyIUÜY]" .. stress_c .. ")(h?[aeoɛɔ])", "%1.%2") text = rsub_repeatedly(text, "([iuüyIUÜY]" .. stress_c .. ")(h?" .. V .. stress_c .. ")", "%1.%2")	text = rsub_repeatedly(text, "([iI]" .. accent_c .. "*)(h?i)", "%1.%2") text = rsub_repeatedly(text, "([uU]" .. accent_c .. "*)(h?u)", "%1.%2")

return text end

-- Generate the IPA of a given respelling.

function export.IPA(text, dialect, phonetic) local distincion = dialect == "Standard" or dialect == "Gheada" local gheismo = dialect == "Gheada" or dialect == "Gheada-seseo" local gheada_different = false local need_seseo = false

text = ulower(text or mw.title.getCurrentTitle.text) -- decompose everything but ñ and ü text = decompose(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 prefixes unstressed unless they have an explicit stress marker; also make certain monosyllabic words 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 CARO to the last vowel not the first one, or we will mess up 'que' by			-- adding the CARO after the 'u' words[i] = rsub(word, "^(.*" .. V .. ")", "%1" .. CARO) 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, " ", "# #") .. "##"

--handle vowels (internally, we use an accute accent to denote stress) text = rsub(text, "a" .. BREVE, "ɐ") text = rsub(text, "e" .. BREVE, "ɪ") text = rsub(text, "o" .. BREVE, "ʊ") text = rsub(text, "e" .. GR, "ɛ" .. AC) text = rsub(text, "o" .. GR, "ɔ" .. AC) text = rsub(text, "e" .. CFLEX, "E" .. AC) --temporary symbol text = rsub(text, "o" .. CFLEX, "O" .. AC) --temporary symbol text = rsub(text, "o".. DOTB, "ɔ") text = rsub(text, "e" .. DOTB, "ɛ") text = rsub(text, "e" .. DOTA, "E") --temporary symbols. we will use for undetermined mid vowels. The template will return an error text = rsub(text, "o" .. DOTA, "O") --if they remain in pre-tonic or accented syllables --handle reintegrationist norm forms, if any -- "nh" and word-final M stay unchanged so they won't affect standard forms. text = rsub(text, "mh", "nh") text = rsub(text, "lh", "ll") text = rsub(text, "ç", "z") text = rsub(text, "ss", "s") text = rsub(text, "g([ieɛE])", "x%1") text = rsub(text, "j", "x") --determine the values of  automatically (in some cases only) text = rsub(text, "ol#", "ɔl#") text = rsub(text, "ols#", "ɔls#") text = rsub(text, "el#", "ɛl#") text = rsub(text, "els#", "ɛls#") text = rsub(text, "ou", "Ou") text = rsub(text, "edo#", "Edo#") text = rsub(text, "edos#", "Edos#") text = rsub(text, "ello#", "Ello#") text = rsub(text, "ellos#", "Ellos#") text = rsub(text, "eza#", "Eza#") text = rsub(text, "ezas#", "Ezas#") text = rsub(text, "o" .. AC .. "n#", "O" .. AC .. "n#") text = rsub(text, "o" .. AC .. "ns#", "O" .. AC .. "ns#") text = rsub(text, "mente#", "mEnte#") --the -mente ending in adverbs will be handled automatically --some prepositions/function words will be handled automatically (no need to add "de") text = rsub(text, "#en#", "#En") text = rsub(text, "#co" .. CARO .. "n#", "#cʊ" .. CARO .. "n#") text = rsub(text, "#po" .. CARO .. "r#", "#pʊ" .. CARO .. "r#") text = rsub(text, "#si" .. CARO .. "n#", "#sɪn#")

--FIXME add automatic suport for prefixes such as termo-, ... -- handle certain combinations text = rsub(text, "ch", "ĉ") --not the real sound text = rsub(text, "nh", "ŋ") text = rsub(text, "ll", "ɟ") text = rsub(text, "#p([st])", "#%1") -- psicoloxía, pterodáctilo

--c, g, q	text = rsub(text, "c([ieɛE])", (distincion and "θ" or "z") .. "%1") -- not the real seseo sound text = rsub(text, "gu([ieɛE])", "g%1") text = rsub(text, "gü([ieɛE])", "gu%1") text = rsub(text, "qu([ieɛE])", "k%1") text = rsub(text, "z", distincion and "θ" or "z") -- not the real seseo sound if rfind(text, "[θz]") then need_seseo = true end -- map various consonants to their phoneme equivalent (we will switch to the correct <ɡ> later) text = rsub(text, "[vgñrxc]", {["v"]="b", ["ñ"]="ɲ", ["r"]="ɾ", ["x"]="ʃ", ["c"] = "k" })

-- trill/tap text = rsub(text, "ɾɾ", "r") text = rsub(text, "([#lnszθ])ɾ", "%1r")

--assimilation of nasals text = rsub(text, "n([# .]*[bp])", "m%1") text = rsub(text, "n([# .]*[kg])", "ŋ%1") text = rsub(text, "ns#", "ŋs#") text = rsub(text, "mn", "ŋn") text = rsub(text, "nm", "ŋm") text = rsub(text, "(" .. V .. ")gn(" .. V .. ")", "%1ŋn%2") text = rsub(text, "n##", "ŋ##") text = rsub(text, "n# #(" .. V .. ")", "ŋ# #%1")

-- remove silent h before syllable division text = rsub(text, "h", "")

-- convert i/u between vowels to glide local vowel_to_glide = { ["i"] = "j", ["u"] = "w" } -- i and u between vowels -> consonant-like substitutions: paranoia... text = rsub_repeatedly(text, "(.*" .. V .. accent_c .. "*h?)([iu])(" .. V .. ")",		function (v1, iu, v2) return v1 .. vowel_to_glide[iu] .. v2 end	) --some special cases involving the sequences  text = rsub(text, "iu#", "iw#") text = rsub(text, "iu" .. AC, "i.u" .. AC)

--syllable division text = syllabify_from_spelling_or_pronun(text)

--diphthongs text = rsub(text, "i([aeɛoɔuEO])", "j%1") text = rsub(text, "u([aeɛoɔuiEO])", "w%1")

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

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, "[^" .. vowel .. "nŋs#]#") or rfind(word, C .. "[nŋs]#")) or #syllables == 1 and rfind(word, V) 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%.te#") then local mente_syllables -- Words ends in -mente (converted above to -mênte); add a stress to the preceding portion 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 --if the aperture of the stressed vowel is not given, throw an error for i = 1, #syllables do			if string.find(syllables[i], "ˈ") and (string.find(syllables[i], "e") or string.find(syllables[i], "o")) then error("Please specify whether the stressed vowels are close-mid or open-mid (see documentation for details).") end end

--Reduction of word-final vowels if not string.find(syllables[#syllables], "ˈ") then syllables[#syllables] = rsub(syllables[#syllables], "a#", "ɐ#") syllables[#syllables] = rsub(syllables[#syllables], "ɛ#", "ɪ#") syllables[#syllables] = rsub(syllables[#syllables], "e#", "ɪ#") syllables[#syllables] = rsub(syllables[#syllables], "ɔ#", "ʊ#") syllables[#syllables] = rsub(syllables[#syllables], "o#", "ʊ#") end -- Reconstruct the word. words[j] = table.concat(syllables, ".") words[j] = rsub_repeatedly(words[j], "ˈ(.+)ˈ", "ˌ%1ˈ")--make all primary stresses but the last one be secondary words[j] = rsub(words[j], "%.(" .. ipa_stress_c .. ")", "%1") -- suppress syllable mark before IPA stress indicator --determine whether there's an undetermined vowel in the pre-tonic part (exclude monosyllabic unaccented words) if string.find(words[j], "ˈ") then local extracted = string.match(words[j], "(.*)ˈ") -- take everything before the stressed syllable if string.find(extracted, "e") or string.find(extracted, "o") then error("Please specify whether the pre-tonic vowels are close-mid or open-mid (see documentation for details).") end end end

text = table.concat(words, " ") text = rsub(text, "E", "e") --no need for the fake symbols anymore text = rsub(text, "O", "o") --diphthongs text = rsub(text, "([aeɛoɔ])i", "%1j") text = rsub(text, "([aeɛo])u", "%1w") --we will use /wi/ after velar sounds and /uj/ elsewhere text = rsub(text, "wi", "uj") text = rsub(text, "([kgħ])uj", "%1wi") text = rsub(text, "ju", "iw")

--real sound of seseo Z	text = rsub(text, "z", "s")

--phonetic transcription if phonetic then --allophones of /a/ text = rsub(text, "al", "ɑl") --syllable ending in /l/ text = rsub(text, "aŋ", "ɑŋ") --preceded or followed by /ŋ/ in the same syllable -- text = rsub(text, "a" .. TILDE .. "ŋ", "ɑ" .. TILDE .. "ŋ") text = rsub(text, "ŋa", "ŋɑ") text = rsub(text, "aw", "ɑw") --in the falling dipthong /aw/ text = rsub(text, "aˈu", "ɑˈu") --hiatus with /aˈu/ text = rsub(text, "a%.([oʊ])", "ɑ.%1") --followed by /o/~/ʊ/ text = rsub(text, "([gk])a", "%1ɑ") --in contact with /k/ or /g/ text = rsub(text, "a([gk])", "ɑ%1") text = rsub(text, "a%.([gk])", "ɑ.%1") text = rsub(text, "aˈ([gk])", "ɑˈ%1") text = rsub(text, "a([ɟʃĉj])", "a̠%1") --before palatal consonants/in a the falling dipthong /aj/ text = rsub(text, "a%.([ɟʃĉj])", "a̠.%1") --nasalization between nasal consonants text = rsub(text, "([ŋmnɲ])(" .. V .. ")%.([ŋmnɲ])", "%1%2" .. TILDE .. ".%3") text = rsub(text, "([ŋmnɲ])(" .. V .. ")ˈ([ŋmnɲ])", "%1%2" .. TILDE .. "ˈ%3") text = rsub(text, "([ŋmnɲ])(" .. V .. ")([ŋmnɲ])", "%1%2" .. TILDE .. "%3") -- θ, s before voiced consonants local voiced = "mnɲbdɟgʎ" local r = "ɾr" local tovoiced = { ["θ"] = "θ̬", ["s"] = "z", }		local function voice(sound, following) return tovoiced[sound] .. following end text = rsub(text, "([θs])(" .. separator_c .. "*[" .. voiced .. r .. "])", voice)

--allophones of /s/ text = rsub(text, "s([ĉʃ])", "S%1") --fake symbol text = rsub(text, "s%.([ĉʃ])", "S.%1") text = rsub(text, "z([ɲɟ])", "Z%1") --fake symbol text = rsub(text, "z%.([ɲɟ])", "Z.%1") -- fricative vs. stop allophones; first convert stops to fricatives, then back to stops after nasals and sometimes after l		local stop_to_fricative = {["b"] = "β", ["d"] = "ð", ["g"] = "ɣ"} local fricative_to_stop = {["β"] = "b", ["ð"] = "d", ["ɣ"] = "g"} text = rsub(text, "[bdg]", stop_to_fricative) text = rsub(text, "([mnɲŋ]" .. separator_c .. "*)([βɣ])",			function(nasal, fricative) return nasal .. fricative_to_stop[fricative] end		) text = rsub(text, "([lʎmnɲŋ]" .. separator_c .. "*)ð",			function(nasal_l, fricative) return nasal_l .. "d" end		) text = rsub(text, "(##" .. ipa_stress_c .. "*)([βɣð])",			function(stress, fricative) return stress .. fricative_to_stop[fricative] end		) --plosives before consonants text = rsub(text, "([βɣð])%.([smtʃ])", function(fricative, consonant) return fricative_to_stop[fricative] .. "." .. consonant end )

--/t,d/ are dental text = rsub(text, "[td]", {["t"] = "t̪", ["d"] = "d̪"})

--/s/ and its voiced allophone [z] are usually apico-alveolar in non-seseo dialects if distincion then text = rsub(text, "[sz]", {["s"] = "s̺", ["z"] = "z̺"}) end -- nasal assimilation before consonants local labiodental, dental, palatalized = "ɱ", "n̪", "nʲ" local nasal_assimilation = { ["f"] = labiodental, ["t"] = dental, ["d"] = dental, ["θ"] = dental, ["ĉ"] = palatalized, ["ʃ"] = palatalized, }		text = rsub(text, "n(" .. separator_c .. "*)(.)",			function(stress, following) return (nasal_assimilation[following] or "n") .. stress .. following end		)

-- lateral assimilation before consonants text = rsub(text, "l(" .. separator_c .. "*)(.)",			function(stress, following)				local l = "l"				if following == "t" or following == "d" or following == "θ" then -- dentialveolar					l = "l̪"				elseif following == "ĉ" or following == "ʃ" then -- alveolopalatal					l = "lʲ"				elseif following == "k" or following == "ɣ" then -- velarized					l = "ɫ"				end				return l .. stress .. following			end) text = rsub(text, "l#", "ɫ#") end --dialectal gheada text = rsub(text, "g", (gheismo and "ħ" or "g")) text = rsub(text, "ɣ", (gheismo and "ħ" or "ɣ")) if rfind(text, "[gɣħ]") then gheada_different = true end

-- convert fake symbols to real ones local final_conversions = { ["ĉ"] = "t͡ʃ", -- fake "ch" to real "ch" ["S"] = "sʲ", ["Z"] = "zʲ", ["g"] = "ɡ", --non-IPA to IPA }	text = rsub(text, "[ĉSZg]", final_conversions) text = rsub(text, "([βðɣ])", "%1̞")   -- voiced fricatives are actually approximants --don't show vowel reduction in phonemic transcriptions if not phonetic then text = rsub(text, "ʊ", "o") text = rsub(text, "ɪ", "e") text = rsub(text, "ɐ", "a") end -- remove # symbols at word and text boundaries text = rsub(text, "#", "") text = unfc(text)

local differences = nil if dialect == "Standard" then differences = { gheada_different = gheada_different, need_seseo = need_seseo, }	end local ret = { text = text, differences = differences, }	return ret end

function export.IPA_string(frame) local iparams = { [1] = {},		["style"] = {required = true}, ["phonetic"] = {type = "boolean"}, }	local iargs = require("Module:parameters").process(frame.args, iparams) local retval = export.IPA(iargs[1], iargs.style, iargs.phonetic) return retval.text end

-- Generate all relevant dialect pronunciations and group into styles. See Module:es-pronunc for details. local function express_all_styles(style_spec, dodialect) local ret = { pronun = {}, expressed_styles = {}, }

local function express_style(hidden_tag, tag, representative_dialect, matching_styles) matching_styles = matching_styles or representative_dialect

-- If style specified, make sure it matches the requested style. local style_matches if not style_spec then style_matches = true else local style_parts = rsplit(matching_styles, "%-") local or_styles = rsplit(style_spec, "%s*,%s*") for _, or_style in ipairs(or_styles) do				local and_styles = rsplit(or_style, "%s*%+%s*") local and_matches = true for _, and_style in ipairs(and_styles) do					local negate if and_style:find("^%-") then and_style = and_style:gsub("^%-", "") negate = true end local this_style_matches = false for _, part in ipairs(style_parts) do						if part == and_style then this_style_matches = true break end end if negate then this_style_matches = not this_style_matches end if not this_style_matches then and_matches = false end end if and_matches then style_matches = true break end end end if not style_matches then return end

-- Fetch the representative dialect's pronunciation if not already present. if not ret.pronun[representative_dialect] then dodialect(ret, representative_dialect) end -- Insert the new style into the style group, creating the group if necessary. local new_style = { tag = tag, pronun = ret.pronun[representative_dialect], }		for _, hidden_tag_style in ipairs(ret.expressed_styles) do			if hidden_tag_style.tag == hidden_tag then table.insert(hidden_tag_style.styles, new_style) return end end table.insert(ret.expressed_styles, {			tag = hidden_tag,			styles = {new_style},		}) end

-- For each type of difference, figure out if the difference exists in any of the given respellings. dodialect(ret, "Standard") local differences = {} for _, difftype in ipairs { "need_seseo", "gheada_different" } do		for _, pronun in ipairs(ret.pronun["Standard"]) do			if pronun.differences[difftype] then differences[difftype] = true end end end local need_seseo = differences.need_seseo local gheada_different = differences.gheada_different

-- Now, based on the observed differences, figure out how to combine the individual dialects into styles and -- style groups. if not gheada_different then if not need_seseo then express_style(false, false, "Standard") else express_style(false, "standard", "Standard") express_style(false, "seseo", "Seseo") end else if not need_seseo then express_style(false, "standard", "Standard") express_style(false, "gheada", "Gheada") else express_style(false, "standard", "Standard") express_style(false, "gheada", "Gheada") express_style(false, "gheada and seseo", "Gheada-seseo") end end

return ret end

local function format_all_styles(expressed_styles, format_style, width) for i, style_group in ipairs(expressed_styles) do		if #style_group.styles == 1 then style_group.formatted, style_group.formatted_len = format_style(style_group.styles[1].tag, style_group.styles[1], i == 1) else style_group.formatted, style_group.formatted_len = format_style(style_group.tag, style_group.styles[1], i == 1) for j, style in ipairs(style_group.styles) do				style.formatted, style.formatted_len = format_style(style.tag, style, i == 1 and j == 1) end end end

local maxlen = 0 for i, style_group in ipairs(expressed_styles) do		local this_len = style_group.formatted_len if #style_group.styles > 1 then for _, style in ipairs(style_group.styles) do				this_len = math.max(this_len, style.formatted_len) end end maxlen = math.max(maxlen, this_len) end

local lines = {}

local need_major_hack = false for i, style_group in ipairs(expressed_styles) do		if #style_group.styles == 1 then table.insert(lines, style_group.formatted) need_major_hack = false else local inline = '\n \n' .. style_group.formatted .. " "			local full_prons = {} for _, style in ipairs(style_group.styles) do				table.insert(full_prons, style.formatted) end local full = '\n \n' .. table.concat(full_prons, "\n") .. " "			local em_length = math.floor(maxlen * width) --this allows us to have a different coefficient in IPA transcriptions and rhymes table.insert(lines, ' ' .. inline .. full .. " ") need_major_hack = true end end

-- major hack to get bullets working on the next line after a div box return table.concat(lines, "\n") .. (need_major_hack and "\n " or "") end

local function dodialect_pronun(args, ret, dialect) ret.pronun[dialect] = {} for i, term in ipairs(args.terms) do		local phonemic, phonetic, differences if term.raw then phonemic = term.raw_phonemic phonetic = term.raw_phonetic differences = construct_default_differences(dialect) else phonemic = export.IPA(term.term, dialect, false) phonetic = export.IPA(term.term, dialect, true) differences = phonemic.differences phonemic = phonemic.text phonetic = phonetic.text end local refs if not term.ref then refs = nil else refs = {} for _, refspec in ipairs(term.ref) do				local this_refs = require("Module:references").parse_references(refspec) for _, this_ref in ipairs(this_refs) do					table.insert(refs, this_ref) end end end

ret.pronun[dialect][i] = { raw = term.raw, phonemic = phonemic, phonetic = phonetic, refs = refs, q = term.q,			qq = term.qq, a = term.a,			aa = term.aa, differences = differences, }	end end

local function generate_pronun(args) local function this_dodialect_pronun(ret, dialect) dodialect_pronun(args, ret, dialect) end

local ret = express_all_styles(args.style, this_dodialect_pronun)

local function format_style(tag, expressed_style, is_first) local pronunciations = {} local formatted_pronuns = {}

local function ins(formatted_part) table.insert(formatted_pronuns, formatted_part) end

-- Loop through each pronunciation. For each one, add the phonemic and phonetic versions to `pronunciations`, -- for formatting by Module:IPA, and also create an approximation of the formatted version so that we can -- compute the appropriate width of the HTML switcher div box that holds the different per-dialect variants. for j, pronun in ipairs(expressed_style.pronun) do			-- Add tag to left qualifiers if first one -- FIXME: Consider using accent qualifier for the tag instead. local qs = pronun.q			if j == 1 and tag then if qs then qs = m_table.deepcopy(qs) table.insert(qs, tag) else qs = {tag} end end

local first_pronun = #pronunciations + 1

if not pronun.phonemic and not pronun.phonetic then error("Internal error: Saw neither phonemic nor phonetic pronunciation") end

if pronun.phonemic then -- missing if 'raw:[...]' given -- don't display syllable division markers in phonemic local slash_pron = "/" .. pronun.phonemic:gsub("%.", "") .. "/"				table.insert(pronunciations, {					pron = slash_pron,				}) ins(slash_pron) end

if pronun.phonetic then -- missing if 'raw:/.../' given local bracket_pron = "[" .. pronun.phonetic .. "]"				table.insert(pronunciations, {					pron = bracket_pron,				}) ins(bracket_pron) end

local last_pronun = #pronunciations

if qs then pronunciations[first_pronun].q = qs			end if pronun.a then pronunciations[first_pronun].a = pronun.a			end if j > 1 then pronunciations[first_pronun].separator = ", " ins(", ") end if pronun.qq then pronunciations[last_pronun].qq = pronun.qq			end if pronun.aa then pronunciations[last_pronun].aa = pronun.aa			end if qs or pronun.qq or pronun.a or pronun.aa then -- Note: This inserts the actual formatted qualifier text, including HTML and such, but the later call -- to textual_len removes all HTML and reduces links. ins(require("Module:pron qualifier").format_qualifiers {					lang = lang,					text = "",					q = qs,					qq = pronun.qq,					a = pronun.a,					aa = pronun.aa,				}) end

if pronun.refs then pronunciations[last_pronun].refs = pronun.refs -- Approximate the reference using a footnote notation. This will be slightly inaccurate if there are -- more than nine references but that is rare. ins(string.rep("[1]", #pronun.refs)) end if first_pronun ~= last_pronun then pronunciations[last_pronun].separator = " " ins(" ") end end

local bullet = string.rep("*", args.bullets) .. " "		-- Here we construct the formatted line in `formatted`, and also try to construct the equivalent without HTML -- and wiki markup in `formatted_for_len`, so we can compute the approximate textual length for use in sizing -- the toggle box with the "more" button on the right. local pre = is_first and args.pre and args.pre .. " " or "" local post = is_first and args.post and " " .. args.post or "" local formatted = bullet .. pre .. m_IPA.format_IPA_full { lang = lang, items = pronunciations, separator = "" } .. post local formatted_for_len = bullet .. pre .. "IPA(key): " .. table.concat(formatted_pronuns) .. post return formatted, textual_len(formatted_for_len) end

ret.text = format_all_styles(ret.expressed_styles, format_style, 0.55)

return ret end

local function parse_respelling(respelling, pagename, parse_err) local raw_respelling = respelling:match("^raw:(.*)$") if raw_respelling then local raw_phonemic, raw_phonetic = raw_respelling:match("^/(.*)/ %[(.*)%]$") if not raw_phonemic then raw_phonemic = raw_respelling:match("^/(.*)/$") end if not raw_phonemic then raw_phonetic = raw_respelling:match("^%[(.*)%]$") end if not raw_phonemic and not raw_phonetic then parse_err(("Unable to parse raw respelling '%s', should be one of /.../, [...] or /.../ [...]")				:format(raw_respelling)) end return { raw = true, raw_phonemic = raw_phonemic, raw_phonetic = raw_phonetic, }	end if respelling == "+" then respelling = pagename end return {term = respelling} end

-- Return the number of syllables of a phonemic representation, which should have syllable dividers in it but no hyphens. local function get_num_syl_from_phonemic(phonemic) -- Maybe we should just count vowels instead of the below code. phonemic = rsub(phonemic, "|", " ") -- remove IPA foot boundaries local words = rsplit(phonemic, " +") for i, word in ipairs(words) do -- IPA stress marks are syllable divisions if between characters; otherwise just remove. word = rsub(word, "(.)[ˌˈ](.)", "%1.%2") word = rsub(word, "[ˌˈ]", "") words[i] = word end -- There should be a syllable boundary between words. phonemic = table.concat(words, ".") return ulen(rsub(phonemic, "[^.]", "")) + 1 end

-- Get the rhyme by truncating everything up through the last stress mark + any following consonants, and remove -- syllable boundary markers. local function convert_phonemic_to_rhyme(phonemic) return rsub(rsub(phonemic, ".*[ˌˈ]", ""), "^[^" .. vowel .. "]*", ""):gsub("%.", ""):gsub("t͡ʃ", "tʃ") end

local function split_syllabified_spelling(spelling) return rsplit(spelling, "%.") end

-- "Align" syllabification to original spelling by matching character-by-character, allowing for extra syllable and -- accent markers in the syllabification. If we encounter an extra syllable marker (.), we allow and keep it. If we -- encounter an extra accent marker in the syllabification, we drop it. In any other case, we return nil indicating -- the alignment failed. local function align_syllabification_to_spelling(syllab, spelling) local result = {} local syll_chars = rsplit(decompose(syllab), "") local spelling_chars = rsplit(decompose(spelling), "") local i = 1 local j = 1 while i <= #syll_chars or j <= #spelling_chars do		local ci = syll_chars[i] local cj = spelling_chars[j] if ci == cj then table.insert(result, ci) i = i + 1 j = j + 1 elseif ci == "." then table.insert(result, ci) i = i + 1 elseif ci == AC or ci == GR or ci == CARO then -- skip character i = i + 1 else -- non-matching character return nil end end if i <= #syll_chars or j <= #spelling_chars then -- left-over characters on one side or the other return nil end return unfc(table.concat(result)) end

-- Word should already be decomposed. local function word_has_vowels(word) return rfind(word, V) end

local function all_words_have_vowels(term) local words = rsplit(decompose(term), "[ %-]") for i, word in ipairs(words) do -- Allow empty word; this occurs with prefixes and suffixes. if word ~= "" and not word_has_vowels(word) then return false end end return true end

local function should_generate_rhyme_from_respelling(term) local words = rsplit(decompose(term), " +") return #words == 1 and -- no if multiple words not words[1]:find(".%-.") and -- no if word is composed of hyphenated parts (e.g. Austria-Hungría) not words[1]:find("%-$") and -- no if word is a prefix not (words[1]:find("^%-") and words[1]:find(CARO)) and -- no if word is an unstressed suffix word_has_vowels(words[1]) -- no if word has no vowels (e.g. a single letter) end

local function should_generate_rhyme_from_ipa(ipa) return not ipa:find("%s") and word_has_vowels(decompose(ipa)) end

local function dodialect_specified_rhymes(rhymes, parsed_respellings, rhyme_ret, dialect) rhyme_ret.pronun[dialect] = {} for _, rhyme in ipairs(rhymes) do		local num_syl = rhyme.num_syl local no_num_syl = false

-- If user explicitly gave the rhyme but didn't explicitly specify the number of syllables and term is single-word, try to take the number of syllables from the phonemic. if not no_num_syl and not num_syl then for _, parsed in ipairs(parsed_respellings) do				for dialect, pronun in pairs(parsed.pronun.pronun[dialect]) do -- Check that pronun.phonemic exists (it may not if raw phonetic-only pronun is given). if pronun.phonemic then if not should_generate_rhyme_from_ipa(pronun.phonemic) then no_num_syl = true break end -- Count number of syllables by looking at syllable boundaries (including stress marks). local this_num_syl = get_num_syl_from_phonemic(pronun.phonemic) m_table.insertIfNot(num_syl, this_num_syl) end end if no_num_syl then break end end if no_num_syl or #num_syl == 0 then num_syl = nil end end

table.insert(rhyme_ret.pronun[dialect], {			rhyme = rhyme.rhyme,			num_syl = num_syl,			qualifiers = rhyme.qualifiers,			differences = construct_default_differences(dialect),		}) end end

local function parse_pron_modifier(arg, parse_err, generate_obj, param_mods, no_split_on_comma) local retval = {}

if arg:find("<") then local insert = { store = "insert" } param_mods.q = insert param_mods.qq = insert param_mods.a = insert param_mods.aa = insert return require(put_module).parse_inline_modifiers(arg, {			param_mods = param_mods,			generate_obj = generate_obj,			parse_err = parse_err,			splitchar = not no_split_on_comma and "," or nil,		}) elseif no_split_on_comma then table.insert(retval, generate_obj(arg)) else for _, term in ipairs(split_on_comma(arg)) do			table.insert(retval, generate_obj(term)) end end

return retval end

local function parse_rhyme(arg, parse_err) local function generate_obj(term) return {rhyme = term} end local param_mods = { s = { item_dest = "num_syl", convert = function(arg, parse_err) local nsyls = rsplit(arg, ",") for i, nsyl in ipairs(nsyls) do					if not nsyl:find("^[0-9]+$") then parse_err("Number of syllables '" .. nsyl .. "' should be numeric") end nsyls[i] = tonumber(nsyl) end return nsyls end, },	}

return parse_pron_modifier(arg, parse_err, generate_obj, param_mods) end

local function parse_homophone(arg, parse_err) local function generate_obj(term) return {term = term} end local param_mods = { t = { -- We need to store the  inline modifier into the "gloss" key of the parsed term, -- because that is what Module:links (called from Module:homophones) expects. item_dest = "gloss", },		gloss = {}, pos = {}, alt = {}, lit = {}, id = {}, g = { -- We need to store the  inline modifier into the "genders" key of the parsed term, -- because that is what Module:links (called from Module:homophones) expects. item_dest = "genders", convert = function(arg) return rsplit(arg, ",") end, },	}

return parse_pron_modifier(arg, parse_err, generate_obj, param_mods) end

local function generate_audio_obj(arg) local file, gloss if arg:find("#") then file, gloss = arg:match("^(.-)%s*#%s*(.*)$") else file, gloss = arg:match("^(.-)%s*;%s*(.*)$") end if not file then file = arg gloss = "Audio" end return {file = file, gloss = gloss} end

local function parse_audio(arg, parse_err) -- None other than qualifiers local param_mods = {}

-- Don't split on comma because some filenames have embedded commas not followed by a space -- (typically followed by an underscore). return parse_pron_modifier(arg, parse_err, generate_audio_obj, param_mods, "no split on comma") end

-- External entry point for. function export.show_pr(frame) local params = { [1] = {list = true}, ["rhyme"] = {}, ["hmp"] = {}, ["audio"] = {list = true}, ["pagename"] = {}, }	local parargs = frame:getParent.args local args = require("Module:parameters").process(parargs, params) local pagename = args.pagename or mw.title.getCurrentTitle.subpageText

-- Parse the arguments. local respellings = #args[1] > 0 and args[1] or {"+"} local parsed_respellings = {} local function overall_parse_err(msg, arg, val) error(msg .. ": " .. arg .. "=" .. val) end local overall_rhyme = args.rhyme and parse_rhyme(args.rhyme, function(msg) overall_parse_err(msg, "rhyme", args.rhyme) end) or nil local overall_hmp = args.hmp and parse_homophone(args.hmp, function(msg) overall_parse_err(msg, "hmp", args.hmp) end) or nil local overall_audio if args.audio then overall_audio = {} for _, audio in ipairs(args.audio) do			local parsed_audio = parse_audio(audio, function(msg) overall_parse_err(msg, "audio", audio) end) if #parsed_audio > 1 then error("Internal error: Saw more than one object returned from parse_audio") end table.insert(overall_audio, parsed_audio[1]) end end for i, respelling in ipairs(respellings) do		if respelling:find("<") then local param_mods = { pre = { overall = true }, post = { overall = true }, style = { overall = true }, bullets = { overall = true, convert = function(arg, parse_err) if not arg:find("^[0-9]+$") then parse_err("Modifier 'bullets' should have a number as argument, but saw '" .. arg .. "'") end return tonumber(arg) end, },				rhyme = { overall = true, store = "insert-flattened", convert = parse_rhyme, },				hmp = { overall = true, store = "insert-flattened", convert = parse_homophone, },				audio = { overall = true, store = "insert-flattened", convert = parse_audio, },				ref = { store = "insert" }, q = { store = "insert" }, qq = { store = "insert" }, a = { store = "insert" }, aa = { store = "insert" }, }

local parsed = require(put_module).parse_inline_modifiers(respelling, {				paramname = i,				param_mods = param_mods,				generate_obj = function(term, parse_err)					return parse_respelling(term, pagename, parse_err)				end,				splitchar = ",",				outer_container = {					audio = {}, rhyme = {}, hmp = {}				}			}) if not parsed.bullets then parsed.bullets = 1 end table.insert(parsed_respellings, parsed) else local termobjs = {} local function parse_err(msg) error(msg .. ": " .. i .. "=" .. respelling) end for _, term in ipairs(split_on_comma(respelling)) do				table.insert(termobjs, parse_respelling(term, pagename, parse_err)) end table.insert(parsed_respellings, {				terms = termobjs,				audio = {},				rhyme = {},				hmp = {},				bullets = 1,			}) end end

-- Loop over individual respellings, processing each. for _, parsed in ipairs(parsed_respellings) do		parsed.pronun = generate_pronun(parsed) local no_auto_rhyme = false for _, term in ipairs(parsed.terms) do			if term.raw then if not should_generate_rhyme_from_ipa(term.raw_phonemic or term.raw_phonetic) then no_auto_rhyme = true break end elseif not should_generate_rhyme_from_respelling(term.term) then no_auto_rhyme = true break end end

-- Generate the rhymes. local function dodialect_rhymes_from_pronun(rhyme_ret, dialect) rhyme_ret.pronun[dialect] = {}

if not parsed.pronun.pronun[dialect] then dodialect_pronun(parsed, parsed.pronun, dialect) end for _, pronun in ipairs(parsed.pronun.pronun[dialect]) do				-- We should have already excluded multiword terms and terms without vowels from rhyme generation (see				-- `no_auto_rhyme` below). But make sure to check that pronun.phonemic exists (it may not if raw				-- phonetic-only pronun is given). if pronun.phonemic then -- Count number of syllables by looking at syllable boundaries (including stress marks). local num_syl = get_num_syl_from_phonemic(pronun.phonemic) -- Get the rhyme by truncating everything up through the last stress mark + any following -- consonants, and remove syllable boundary markers. local rhyme = convert_phonemic_to_rhyme(pronun.phonemic) local saw_already = false for _, existing in ipairs(rhyme_ret.pronun[dialect]) do						if existing.rhyme == rhyme then saw_already = true -- We already saw this rhyme but possibly with a different number of syllables, -- e.g. if the user specified two pronunciations 'biología' (4 syllables) and -- 'bi.ología' (5 syllables), both of which have the same rhyme /ia/. m_table.insertIfNot(existing.num_syl, num_syl) break end end if not saw_already then local rhyme_diffs = nil if dialect == "Standard" then rhyme_diffs = {} if rhyme:find("θ") then rhyme_diffs.need_seseo = true end if rhyme:find("ɡ") then rhyme_diffs.gheada_different = true end end table.insert(rhyme_ret.pronun[dialect], {							rhyme = rhyme,							num_syl = {num_syl},							differences = rhyme_diffs,						}) end end end end

if #parsed.rhyme == 0 then if overall_rhyme or no_auto_rhyme then parsed.rhyme = nil else parsed.rhyme = express_all_styles(parsed.style, dodialect_rhymes_from_pronun) end else local no_rhyme = false for _, rhyme in ipairs(parsed.rhyme) do				if rhyme.rhyme == "-" then no_rhyme = true break end end if no_rhyme then parsed.rhyme = nil else local function this_dodialect(rhyme_ret, dialect) return dodialect_specified_rhymes(parsed.rhyme, {parsed}, rhyme_ret, dialect) end parsed.rhyme = express_all_styles(parsed.style, this_dodialect) end end end

if overall_rhyme then local no_overall_rhyme = false for _, orhyme in ipairs(overall_rhyme) do			if orhyme.rhyme == "-" then no_overall_rhyme = true break end end if no_overall_rhyme then overall_rhyme = nil else local function dodialect_overall_rhyme(rhyme_ret, dialect) return dodialect_specified_rhymes(overall_rhyme, parsed_respellings, rhyme_ret, dialect) end overall_rhyme = express_all_styles(parsed.style, dodialect_overall_rhyme) end end

-- If all sets of pronunciations have the same rhymes, display them only once at the bottom. -- Otherwise, display rhymes beneath each set, indented. local first_rhyme_ret local all_rhyme_sets_eq = true for j, parsed in ipairs(parsed_respellings) do		if j == 1 then first_rhyme_ret = parsed.rhyme elseif not m_table.deepEquals(first_rhyme_ret, parsed.rhyme) then all_rhyme_sets_eq = false break end end

local function format_rhyme(rhyme_ret, num_bullets) local function format_rhyme_style(tag, expressed_style, is_first) local pronunciations = {} local rhymes = {} for _, pronun in ipairs(expressed_style.pronun) do				table.insert(rhymes, pronun) end local data = { lang = lang, rhymes = rhymes, qualifiers = tag and {tag} or nil, force_cat = force_cat, }			local bullet = string.rep("*", num_bullets) .. " "			local formatted = bullet .. require("Module:rhymes").format_rhymes(data) local formatted_for_len_parts = {} table.insert(formatted_for_len_parts, bullet .. "Rhymes: " .. (tag and "(" .. tag .. ") " or "")) for j, pronun in ipairs(expressed_style.pronun) do				if j > 1 then table.insert(formatted_for_len_parts, ", ") end if pronun.qualifiers then table.insert(formatted_for_len_parts, "(" .. table.concat(pronun.qualifiers, ", ") .. ") ")				end table.insert(formatted_for_len_parts, "-" .. pronun.rhyme) end return formatted, textual_len(table.concat(formatted_for_len_parts)) end

return format_all_styles(rhyme_ret.expressed_styles, format_rhyme_style, 0.75) end

-- If all sets of pronunciations have the same homophones, display them only once at the bottom. -- Otherwise, display homophones beneath each set, indented. local first_hmps local all_hmp_sets_eq = true for j, parsed in ipairs(parsed_respellings) do		if j == 1 then first_hmps = parsed.hmp elseif not m_table.deepEquals(first_hmps, parsed.hmp) then all_hmp_sets_eq = false break end end

local function format_homophones(hmps, num_bullets) local hmptext = require("Module:homophones").format_homophones { lang = lang, homophones = hmps } return string.rep("*", num_bullets) .. " " .. hmptext end

local function format_audio(audios, num_bullets) local ret = {} for i, audio in ipairs(audios) do			local text = require(audio_module).format_audio { lang = lang, file = audio.file, caption = audio.gloss, q = audio.q,				qq = audio.qq, a = audio.a,				aa = audio.aa, }			table.insert(ret, string.rep("*", num_bullets) .. " " .. text) end return table.concat(ret, "\n") end

local textparts = {} local min_num_bullets = 9999 for j, parsed in ipairs(parsed_respellings) do		if parsed.bullets < min_num_bullets then min_num_bullets = parsed.bullets end if j > 1 then table.insert(textparts, "\n") end table.insert(textparts, parsed.pronun.text) if #parsed.audio > 0 then table.insert(textparts, "\n") -- If only one pronunciation set, add the audio with the same number of bullets, otherwise -- indent audio by one more bullet. table.insert(textparts, format_audio(parsed.audio, #parsed_respellings == 1 and parsed.bullets or parsed.bullets + 1)) end if not all_rhyme_sets_eq and parsed.rhyme then table.insert(textparts, "\n") table.insert(textparts, format_rhyme(parsed.rhyme, parsed.bullets + 1)) end if not all_hmp_sets_eq and #parsed.hmp > 0 then table.insert(textparts, "\n") table.insert(textparts, format_homophones(parsed.hmp, parsed.bullets + 1)) end end if overall_audio and #overall_audio > 0 then table.insert(textparts, "\n") table.insert(textparts, format_audio(overall_audio, min_num_bullets)) end if all_rhyme_sets_eq and first_rhyme_ret then table.insert(textparts, "\n") table.insert(textparts, format_rhyme(first_rhyme_ret, min_num_bullets)) end if overall_rhyme then table.insert(textparts, "\n") table.insert(textparts, format_rhyme(overall_rhyme, min_num_bullets)) end if all_hmp_sets_eq and #first_hmps > 0 then table.insert(textparts, "\n") table.insert(textparts, format_homophones(first_hmps, min_num_bullets)) end if overall_hmp and #overall_hmp > 0 then table.insert(textparts, "\n") table.insert(textparts, format_homophones(overall_hmp, min_num_bullets)) end

return table.concat(textparts) end

return export