-- Based on ] by Benwing2 and ] by Rex Aurorum.
-- Original module by TagaSanPedroAko.
-- Partly rewritten by TagaSanPedroAko, merging code from ] back into this module; {{tl-pr}} restructured
-- to take inline modifiers, like {{es-pr}}.
-- Modified for {{ms-pr}}.
local export = {}
-- Specifically, the following accents are recognized:
-- "Johor-Selangor" = Standard Malaysian Malay pronunciation, based on the Malay of the Sultanate of Johor. Pronounces ''i'' and ''u'' in closed final syllable as /e/ and /o/, word-final ''a'' as schwa, and syllable-final R as silent.
-- "Northern Malay Peninsula" = Similar to Johor-Selangor, except it pronounces ''u'' and ''i'' in closed final syllable the same where following syllable-final consonant is N or NG, word-final A is retained and syllable-final R is always pronounced.
-- "Bahasa Baku" = standard Singapore Malay accent, also the same accent used in Indonesian, i.e. ''i'' and ''u'' in closed final syllable are always retained, syllabification also different
-- A "style spec" (indicated by the style= parameter to {{ms-pr}}) restricts the output to certain styles.
-- A style spec can be one of the following:
-- 1. An accent, e.g. "Johor", "Northern Peninsula"; if specified, only styles containing this isogloss are output.
-- 2. A negated accent, e.g. "-Bahasa Baku".
local force_cat = false -- enable for testing
local m_IPA = require("Module:IPA")
local m_str_utils = require("Module:string utilities")
local m_table = require("Module:table")
local accent_qualifier_module = "Module:accent qualifier"
local audio_module = "Module:audio"
local headword_data_module = "Module:headword/data"
local hyphenation_module = "Module:hyphenation"
local labels_module = "Module:labels"
local parse_utilities_module = "Module:parse utilities"
local rhymes_module = "Module:rhymes"
local set_utilities_module = "Module:set utilities"
local lang = require("Module:languages").getByCode("ms")
local maxn = table.maxn
local rfind = m_str_utils.find
local rsubn = m_str_utils.gsub
local rsplit = m_str_utils.split
local toNFC = mw.ustring.toNFC
local toNFD = mw.ustring.toNFD
local trim = mw.text.trim
local u = m_str_utils.char
local ulen = m_str_utils.len
local ulower = m_str_utils.lower
local AC = u(0x0301) -- acute = ́
local GR = u(0x0300) -- grave = ̀
local CFLEX = u(0x0302) -- circumflex = ̂
local MAC = u(0x0304) -- macron
local BR = u(0x0306) -- breve = ˘
local vowel = "aeéèioòuəɛɔ" -- vowel
local V = ""
local W = "" -- glide
local NV = ""
local accent = AC .. GR .. MAC .. BR
local accent_c = ""
local stress_c = ""
local ipa_stress = "ˈ"
local ipa_stress_c = ""
local separator = accent .. ipa_stress .. "# .-"
local C = "" -- consonant
local unstressed_words = m_table.listToSet {
--prepositions
"di", "ké",
-- letter names. Excluding ones that also occur as regular words
"é", "bi", "si", "i", "éf", "ji", "hec", "ay", "jé", "él", "ém", "én", "o", "pi",
"kiu", "ar", "es", "ti", "yu", "vi", "dabel yu", "éks", "way", "zéd",
"dan" -- conjunction
}
local unstressed_affixes = m_table.listToSet {
-- NOTE: prefixes here aren't currently used with prefixes themselves because they are all assumed unstressed
-- in the absence of an explicit accent marker..
"-an", "-in", "ber-", "di-", "meng-", "me-", "peng-", "pe-", "per-", "-nya", "-mu", "-ku"
}
local function track(page)
require("Module:debug/track")("ms-pron/" .. page)
return true
end
-- 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
-- Combine two sets of qualifiers, either of which may be nil or a list of qualifiers. Remove duplicate qualifiers.
-- Return value is nil or a list of qualifiers.
local function combine_qualifiers(qual1, qual2)
if not qual1 then
return qual2
end
if not qual2 then
return qual1
end
local qualifiers = m_table.deepcopy(qual1)
for _, qual in ipairs(qual2) do
m_table.insertIfNot(qualifiers, qual)
end
return qualifiers
end
local function decompose(text)
-- decompose everything but é
text = toNFD(text)
text = rsub(text, "." .. "", {
= "é",
= "É",
})
return text
end
local function remove_accents(str)
str = decompose(str)
str = rsub(str, "(.)" .. accent_c, "%1")
return str
end
local function split_on_comma(term)
if term:find(",%s") then
return require(parse_utilities_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(accent)
if accent == "Johor-Selangor" then
return {
bahasa_baku_different = false,
}
end
return nil
end
-- ĵ, ɟ and ć are used internally to represent , and
--
function export.IPA(text)
local piawai = dialect == "Johor-Selangor"
local northern = dialect == "Northern Peninsula"
local baku = dialect == "Bahasa Baku"
local need_piawai = false
local need_northern = false
local need_baku = false
local debug = {}
text = ulower(text)
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*()", "%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, " $", "")
return text
end
text = canon_spaces(text)
-- Make prefixes unstressed unless they have an explicit stress marker; also make certain
-- monosyllabic words (e.g. ], ], ], ], etc.) without stress marks be
-- unstressed.
local words = rsplit(text, " ")
for i, word in ipairs(words) do
if rfind(word, "%-$") and not rfind(word, accent_c) or unstressed_words then
-- add BR to the last vowel not the first one
-- adding the BR after the 'u'
words = rsub(word, "^(.*" .. V .. ")", "%1" .. BR)
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, " ", "# #") .. "##"
table.insert(debug, text)
--"i" or "u" to glide (as part of a diphthong)
text = rsub(text, "(" .. V .. ")i()", "%1ɟ%2")
text = rsub(text, "(" ..V.. ")u()", "%1w%2")
-- syllable-initial X (e.g. in ], ], ])
text = rsub(text, "x("..V..")", "z%1")
-- handle certain combinations; kh, ng, ny and sy handling needs to go first, so is Johor-Selangor word-final R and A
text = rsub(text, "kh", "x")
text = rsub(text, "ng", "ŋ")
text = rsub(text, "ny", "ɲ")
text = rsub(text, "sy", "ʃ")
text = rsub(text, "bh", "b") -- in terms from Sanskrit e.g. ]
text = rsub(text, "gh", "ɣ") -- in Arabic-derived terms e.g. ]
text = rsub(text, "dh", "ð") -- in Arabic-derived terms e.g. ]
-- Syllable-final R is silent and lengthens nearby vowel in Johor accent
text = rsub(text, "(".. V ..")r", "%1" .. (piawai and "ː" or "r"))
-- Syllable-final A is schwa in Johor accent
text = rsub(text, "a()", (piawai and "ə" or "a") .. "%1")
if rfind(text, "()") or rfind(text, "()")
then
need_piawai = true
end
table.insert(debug, text)
--alphabet-to-phoneme
text = rsub(text, "",
--="ɡ": U+0067 LATIN SMALL LETTER G → U+0261 LATIN SMALL LETTER SCRIPT G
{ = "ć", = "ə", = "e", = "ɡ", = "ĵ", = "ʔ", = "j" })
-- reduce any remaining double consonants (e.g. ], ]);
text = rsub(text, "(" .. C .. ")%1", "%1")
-- I or U in closed final syllable
text = rsub(text, "i()()", (baku and "i" or "e") .. "%1%2")
text = rsub(text, "u()()", (baku and "u" or "o") .. "%1%2")
text = rsub(text, "i()()", (piawai and "e" or "i") .. "%1%2")
text = rsub(text, "u()()", (piawai and "o" or "u") .. "%1%2")
if rfind(text, "()()") or rfind(text, "()()") then
need_baku = true
end
if rfind(text, "()()") or rfind(text, "()()") then
need_piawai = true
end
--syllable division
local vowel_to_glide = { = "j", = "w", = "w" }
-- i, o and u between vowels -> j and u e.g. ])
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)()(" .. V .. ")",
function(v1, iou, v2)
return v1 .. vowel_to_glide .. v2
end
)
table.insert(debug, text)
-- "mb", "mp", "nd", "nk", "nt" combinations
text = rsub_repeatedly(text, "(m)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "(n)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "(n)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "(ŋ)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "()()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "()()(?)()(" .. V .. ")", "%1%2%3.%4%5")
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C .. W .. "?" .. V .. ")", "%1.%2")
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*" .. C .. "+)(" .. 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")
text = rsub_repeatedly(text, "(" .. accent_c .. "*)()", "%1.%2")
table.insert(debug, text)
--diphthongs (Johor and Northern Peninsula only)
text = rsub(text, "i(".. V .. ")", baku and "i.%1" or "j%1")
text = rsub(text, "u(".. V .. ")", baku and "u.%1" or "w%1")
if rfind(text, "(?)(".. V .. ")") or rfind(text, "(?)(".. V .. ")") then
need_baku = true
end
table.insert(debug, text)
local accent_to_stress_mark = { = "ˈ", = "" }
local function accent_word(word, syllables)
-- Now stress the word. If any accent exists in the word (including macron 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 = rsub(syllables, "^(.*)(" .. accent_c .. ")(.*)$",
function(pre, accent, post)
return accent_to_stress_mark .. 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, V) then
syllables = "ˈ" ..syllables -- Monosyllabic words wity schwa don't get stress
elseif #syllables == 2 and rfind(word, "()(".. C .."?)()") then
syllables = "ˈ" .. syllables
elseif #syllables == 2 and rfind(word, "()(".. C .."?)()") then
syllables = "ˈ" .. syllables
elseif #syllables >= 3 and rfind(word, "()(".. C .."?)()(".. C .."?)()(".. C .. C .."?)(".. V .."?)(".. C .."?)()") then
syllables = "ˈ" .. syllables
elseif #syllables > 1 then
syllables = "ˈ" .. syllables
end
end
end
local words = rsplit(text, " ")
for j, word in ipairs(words) do
-- accentuation
local syllables = rsplit(word, "%.")
accent_word(word, syllables)
-- Reconstruct the word.
words = table.concat(syllables, ".")
end
text = table.concat(words, " ")
-- suppress syllable mark before IPA stress indicator
text = rsub(text, "%.(" .. ipa_stress_c .. ")", "%1")
--make all primary stresses but the last one be secondary
text = rsub_repeatedly(text, "ˈ(.+)ˈ", "ˌ%1ˈ")
table.insert(debug,text)
--correct final consonant placement before Y or W
text = rsub(text,"()(".. C ..")()(" .. V .. ")","%2%1%3%4")
table.insert(debug,text)
text = rsub_repeatedly(text, "^(*)()(*)", "%1%3%2") -- Move stress inside word boundary fix at start
text = rsub_repeatedly(text, "()(*)()(*)", "%1%2%4%3") -- Move stress inside word boundary fix at start
local ms_IPA_table = {
= text,
= text,
= text
}
for key, value in pairs(ms_IPA_table) do
text = ms_IPA_table
--phonetic transcription
if key == "phonetic" then
table.insert(debug, text)
--phonemic diphthongs
text = rsub(text, "(".. V ..")()()", "%1i̯%3")
text = rsub(text, "(".. V ..")w()", "%1u̯%2")
table.insert(debug, text)
--change i, u in closed final syllables (standard Malaysian only)
text = rsub(text, "(".. C .."?)e(".. C ..")()","%1ɛ%2%3")
text = rsub(text, "(".. C .."?)o(".. C ..")()","%1ɔ%2%3")
table.insert(debug, text)
--devoice final B, D, G and final K
text = rsub(text, "b()","p̚%1")
text = rsub(text, "d()","t̚%1")
text = rsub(text, "ɡ()","k̚%1")
text = rsub(text, "k()","k̚%1")
--/n/ and /ŋ/ sandhi
text = rsub(text,"()(*)", "m%2")
text = rsub(text,"()(*)","n%2")
text = rsub(text,"()(*)","ɲ%2")
--T is dental in Malay
text = rsub(text, "t","t̪")
mw.log(text)
end
table.insert(debug, text)
-- convert fake symbols to real ones
local final_conversions = {
= "t͡ʃ", -- fake "c" to real "c"
= "j", -- fake "i" to real "i"
= "d͡ʒ" -- fake "j" to real "j"
}
text = rsub(text, "", final_conversions)
-- Do not have multiple syllable break consecutively
text = rsub_repeatedly(text, "(+)", ".")
text = rsub_repeatedly(text, "(?)(‿)(?)", "%2")
-- remove # symbols at word and text boundaries
text = rsub_repeatedly(text, "(?)#(?)", "")
-- resuppress syllable mark before IPA stress indicator
text = rsub(text, "%.(" .. ipa_stress_c .. ")", "%1")
text = rsub_repeatedly(text, "(?)(" .. ipa_stress_c .. ")(?)", "%2")
ms_IPA_table = canon_spaces(toNFC(text))
end
return ms_IPA_table
end
function export.show(frame)
local params = {
= {},
= {},
= {type = "number", default = 1},
}
local parargs = frame:getParent().args
local args = require("Module:parameters").process(parargs, params)
local results = {}
local text = args or mw.title.getCurrentTitle().text
local IPA_result = export.IPA(text)
table.insert(results, { pron = "/" .. IPA_result .. "/" })
table.insert(results, { pron = " .. "]" })
local pre = args.pre and args.pre .. " " or ""
local bullet = (args.bullets ~= 0) and "* " or ""
return bullet .. pre .. m_IPA.format_IPA_full { lang = lang, items = results }
end
local function parse_gloss(arg)
local poses, gloss
if arg:find("%^") then
poses, gloss = arg:match("^(.-)%^(.*)$")
if gloss == "" then
gloss = nil
end
else
gloss = arg
end
if poses then
poses = split_on_comma(poses)
local m_headword_data = mw.loadData(headword_data_module)
for i, pos in ipairs(poses) do
poses = m_headword_data.pos_aliases or pos
end
end
return {
poses = poses,
gloss = gloss,
}
end
-- Parse a raw accent spec, which is one or more comma-separated accent qualifiers.
local function parse_accents(arg)
return require(labels_module).split_labels_on_comma(arg)
end
-- Return the number of syllables of a phonemic or phonetic representation, which should have syllable dividers in it
-- but no hyphens.
local function get_num_syl_from_ipa(pron)
-- Maybe we should just count vowels instead of the below code.
pron = rsub(pron, "|", " ") -- remove IPA foot boundaries
local words = rsplit(pron, " +")
for i, word in ipairs(words) do
-- IPA stress marks are syllable divisions if between characters; otherwise just remove.
word = rsub(word, "(.)(.)", "%1.%2")
word = rsub(word, "", "")
words = word
end
-- There should be a syllable boundary between words.
pron = table.concat(words, ".")
return ulen(rsub(pron, "", "")) + 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)
-- NOTE: This works because the phonemic vowels are just possibly with diacritics that are separate
-- Unicode chars.
phonemic = rsplit(phonemic, " ")
phonemic = phonemic
phonemic = rsub(rsub(phonemic, ".*", ""), "^" .. NV .. "*", ""):gsub("%.", "")
phonemic = rsub_repeatedly(phonemic, "^(" .. V .. ")", "%1" )
return phonemic
end
local function split_syllabified_spelling(spelling)
spelling = "#" .. spelling .. "#"
spelling = rsub_repeatedly(spelling, "%.()", "·%1")
spelling = rsub_repeatedly(spelling, "#", "")
spelling = rsplit(spelling, "%.")
for key, value in ipairs(spelling) do
spelling = rsub_repeatedly(value, "·", ".")
end
return spelling
end
-- "Align" syllabified respelling `syllab` to original spelling `spelling` by matching character-by-character, allowing
-- for extra syllable and accent markers in the syllabification and certain mismatches in the consonants. The goal is to
-- produce the appropriately syllabified version of the original spelling (the pagename) by matching characters in the
-- syllabified respelling to the original spelling, putting the syllable boundaries in the appropriate places in the
-- original spelling.
--
-- If we encounter an extra syllable marker (.), we allow and keep it. If we encounter an extra accent marker in thes
-- syllabification, we drop it. We allow for mismatches in capitalization and for certain other mismatches, e.g. extra
-- glottal stops (written q), respelling of unadapted loanwords, etc. If we can't match, we return nil
-- indicating the alignment failed.
local function align_syllabification_to_spelling(syllab, spelling)
local result = {}
local function concat_result()
-- Postprocess to remove dots (syllable boundaries) next to hyphens.
return (toNFC(table.concat(result)):gsub("%.%-", "-"):gsub("%-%.", "-"))
end
-- Replace glottal stop (q) from respelling with k, as found for original
-- spelling. (FIXME: Diacritics should be removed , but they're currently removed earlier, in
-- syllabify_from_spelling(). We should probably get rid of the removal there and put it here.)
syllab = decompose(syllab:gsub("ː", "")):gsub("7", "q")
local syll_chars = rsplit(ulower(syllab), "")
local spelling_chars = rsplit(spelling, "")
local i = 1
local j = 1
local function matches(uci, ucj)
-- Return true if a syllabified respelling character (uci) matches the corresponding spelling char (ucj).
-- Both uci and ucj should be lowercase.
-- Sound is at the key, values are the letters sound can match
local matching_chars = {
= {"e", "a"},
= {"k", "q"},
= {"j", "c", "x"}
}
return uci == ucj or (matching_chars and m_table.contains(matching_chars, ucj))
end
local function silent_spelling_letter(ucj)
return ucj == "h" or ucj == "'" or ucj == "-"
end
local function syll_at(pos)
return syll_chars or ""
end
local function spell_at(pos)
return spelling_chars or ""
end
local function uspell_at(pos)
local c = spelling_chars
return c and ulower(c) or ""
end
while i <= #syll_chars or j <= #spelling_chars do
local uci = syll_at(i)
local cj = spell_at(j)
local ucj = uspell_at(j)
if uci == "g" and syll_at(i - 1) == "n" and syll_at(i + 1) == "." and matches(syll_at(i + 2), ucj) and
not matches(syll_at(i + 2), uspell_at(j + 1)) then
-- As a special case, before checking whether the corresponding characters match, we have to skip an extra
-- g in an -ng- sequence in the syllabified respelling if the corresponding spelling character matches the
-- next respelling character (taking into account the syllable boundary).
i = i + 1
elseif uci == "g" and ucj == "g" and uspell_at(j + 1) == TILDE then
table.insert(result, cj)
table.insert(result, uspell_at(j + 1))
i = i + 1
j = j + 2
elseif matches(uci, ucj) then
table.insert(result, cj)
i = i + 1
j = j + 1
elseif ucj == uspell_at(j - 1) and uci == "." and ucj ~= syll_at(i + 1) then
-- See below. We want to allow for a doubled letter in spelling that is pronounced single, and preserve the
-- doubled letter. But it's tricky in the presence of syllable boundaries on both sides of the doubled
-- letter as well as doubled letters pronounced double. Specifically, there are three possibilities,
-- exemplified by:
-- (1) syll='Ma.ka.ssar', spelling='Makassar' -> 'Ma.ka.ssar';
-- (2) syll='Ta.llo', spelling='Tallo' -> 'Ta.llo'.
table.insert(result, cj)
j = j + 1
elseif silent_spelling_letter(ucj) and uci == "." and ucj ~= syll_at(i + 1) and
not rfind(uspell_at(j + 1), V) then
-- See below for apostrophe in spelling. This condition is parallel to the one directly above
-- for silent doubled letters in spelling and handles the case of syllab='Abdurrahmān', spelling='Abdurahman',
-- which should be syllabified 'Ab.du.rrah.man'. But we need a check to see that the next spelling character
-- isn't a vowel, because in that case we want the silent letter to go after the period, e.g.
-- syllab='Jumu.ah', spelling='Jumu'ah' -> 'Ju.mu.ah'.
table.insert(result, cj)
j = j + 1
elseif uci == "." then
table.insert(result, uci)
i = i + 1
elseif ucj == uspell_at(j - 1) then
-- A doubled letter in spelling that is pronounced single. Examples:
-- * syllab='Da.rus.sa.lam', spelling='Darussalam' -> 'Da.rus.sa.lam' (with s)
-- * syllab='Ka.a.ba', spelling='Kaaba' -> 'Kaa.ba' (with a)
table.insert(result, cj)
j = j + 1
elseif silent_spelling_letter(ucj) then
-- A silent h, apostrophe or hyphen in spelling. Examples:
-- * syllab='Ramadān', spelling='Ramadhan' -> 'Ra.ma.dhan'
table.insert(result, cj)
j = j + 1
elseif uci == AC or uci == GR or uci == MAC or uci == BR or
uci == "y" or uci == "w" then
-- skip character
i = i + 1
else
-- non-matching character
mw.log(("Syllabification alignment mismatch for pagename '%s' (position %s, character %s), syllabified respelling '%s' (position %s, character %s), aligned result so far '%s'"
):format(spelling, j, ucj, syllab, i, uci, concat_result()))
return nil
end
end
if i <= #syll_chars or j <= #spelling_chars then
-- left-over characters on one side or the other
mw.log(("Syllabification alignment mismatch for pagename '%s' (%s), syllabified respelling '%s' (%s), aligned result so far '%s'"
):format(
spelling, j > #spelling_chars and "end of string" or ("position %s, character %s"):format(j, uspell_at(j)),
syllab, i > #syll_chars and "end of string" or ("position %s, character %s"):format(i, syll_at(i)),
concat_result()))
return nil
end
return concat_result()
end
local function generate_syll_obj(term)
return {syllabification = term, hyph = split_syllabified_spelling(term)}
end
-- Word should already be decomposed.
local function word_has_vowels(word)
word = ulower(word)
return rfind(word, V) or word:find("y")
end
local function any_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_has_vowels(word) then
return true
end
end
return false
end
local function should_generate_rhyme_from_respelling(term)
local words = rsplit(decompose(term), " +")
local last_word = words
local should_generate_cat = #words == 1
local should_generate_rhyme =
not last_word:find("%-$") and -- no if word is a prefix
not (last_word:find("^%-") and last_word:find(MAC)) and -- no if word is an unstressed suffix
word_has_vowels(last_word) -- no if word has no vowels (e.g. a single letter)
return should_generate_rhyme, should_generate_cat
end
local function should_generate_rhyme_from_ipa(ipa)
local should_generate_cat = not ipa:find("%s")
local should_generate_rhyme = word_has_vowels(decompose(ipa))
return should_generate_rhyme, should_generate_cat
end
local function should_generate_rhyme_from_termobj(termobj)
if termobj.raw then
return should_generate_rhyme_from_ipa(termobj.raw_phonemic or termobj.raw_phonetic)
else
return should_generate_rhyme_from_respelling(termobj.term)
end
end
local function process_specified_rhymes(rhymes, sylls, parsed_respellings)
local rhyme_ret = {}
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, try to take it from
-- the syllabification.
if not num_syl then
num_syl = {}
for _, syll in ipairs(sylls) do
if should_generate_rhyme_from_respelling(syll.syllabification) then
local this_num_syl = 1 + ulen(rsub(syll.syllabification, "", ""))
m_table.insertIfNot(num_syl, this_num_syl)
else
no_num_syl = true
break
end
end
if no_num_syl or #num_syl == 0 then
num_syl = nil
end
end
-- If that fails and term is single-word, try to take it from the phonemic.
if not no_num_syl and not num_syl then
for _, parsed in ipairs(parsed_respellings) do
for _, pronun in ipairs(parsed.pronuns) do
-- Check that pronun.phonemic exists (it may not if raw phonetic-only pronun is given), and rhyme
-- isn't suppressed (which may happen if the term has a qualifier "colloquial", "obsolete" or the
-- like or is an auto-generated "glottal stop elision" pronunciation).
if pronun.phonemic and not pronun.no_rhyme 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_ipa(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
local rhymeobj = m_table.shallowcopy(rhyme)
rhymeobj.num_syl = num_syl
table.insert(rhyme_ret, rhymeobj)
end
end
-- Parse a pronunciation modifier in `arg`, the argument portion in an inline modifier (after the prefix), which
-- specifies a pronunciation property such as rhyme, syllabification, homophones or audio. The argument can itself have
-- inline modifiers, e.g. <audio:Foo.ogg<a:Jakarta>>. The allowed inline modifiers are specified by `param_mods` (of
-- the format expected by `parse_inline_modifiers()`); in addition to any modifiers specified there, the modifiers
-- <q:...>, <qq:...>, <a:...> and <aa:...> are always accepted (and can be repeated). `generate_obj` and `parse_err` are
-- like in `parse_inline_modifiers()` and specify respectively a function to generate the object into which modifier
-- properties are stored given the non-modifier part of the argument, and a function to generate an error message (given
-- the message). Normally, a comma-separated list of pronunciation properties is accepted and parsed, where each element
-- in the list can have its own inline modifiers and where no spaces are allowed next to the commas in order for them to
-- be recognized as separators. If `no_split_on_comma` is given, only a single pronunciation property is accepted. If
-- `has_outer_container` is given, the list of pronunciation properties is embedded in the `terms` property of an outer
-- container, into which other list-level modifiers can also be stored (by setting `overall = "true"` in the respective
-- spec in `param_mods`). The return value is a list if neither `no_split_on_comma` nor `has_outer_container` are given,
-- otherwise a container object (which, in the case of `has_outer_container`, will contain a list inside of it, in the
-- `terms` property).
local function parse_pron_modifier(arg, parse_err, generate_obj, param_mods, no_split_on_comma, has_outer_container)
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(parse_utilities_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,
outer_container = has_outer_container and {} or nil,
})
elseif no_split_on_comma then
return generate_obj(arg)
else
local retval = {}
for _, term in ipairs(split_on_comma(arg)) do
table.insert(retval, generate_obj(term))
end
if has_outer_container then
retval = {
terms = retval,
}
end
return retval
end
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",
type = "number",
sublist = true,
},
}
return parse_pron_modifier(arg, parse_err, generate_obj, param_mods)
end
local function parse_syll(arg, parse_err)
local param_mods = {
cap = { overall = true},
}
-- We need to pass in has_outer_container because we have an overall property <cap:...> (the caption, defaulting
-- to "Syllabification") applying to the whole set of syllabifications.
return parse_pron_modifier(arg, parse_err, generate_syll_obj, param_mods, nil, "has outer container")
end
local function parse_homophone(arg, parse_err)
local function generate_obj(term)
return {term = term}
end
local param_mods = {
t = {
-- ] (called from ]) expects the gloss in "gloss".
item_dest = "gloss",
},
gloss = {},
pos = {},
alt = {},
lit = {},
id = {},
g = {
-- ] (called from ]) expects the genders in "genders".
item_dest = "genders",
sublist = true,
},
}
return parse_pron_modifier(arg, parse_err, generate_obj, param_mods)
end
local function generate_audio_obj(arg)
local file, gloss = arg:match("^(.-)%s*#%s*(.*)$")
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
local function syllabify_from_spelling(text, pagename)
-- Auto syllabifications start --
local vowel = vowel .. "ẃý" -- vowel
local V = ""
local NV = ""
local C = "" -- consonant
text = decompose(text)
local origtext = remove_accents(text)
text = string.lower(text)
text = rsub(text, " ", "․ ")
text = rsub(text, "$", "․")
-- put # at word beginning and end and double ## at text/foot boundary beginning/end
text = rsub(text, " | ", "# | #")
text = "##" .. rsub(text, " ", "# #") .. "##"
text = rsub_repeatedly(text, "(?)#(?)", "#")
text = rsub(text, "bh", "β")
text = rsub(text, "eu", "ɨ")
text = rsub(text, "ng", "ŋ")
text = rsub(text, "ny", "ɲ")
text = rsub(text, "c", "ć")
text = rsub(text, "j", "ĵ")
text = rsub(text, "kh", "x")
text = rsub(text, "sy", "ʃ")
text = rsub(text, "ʔ()", "q%1")
text = rsub(text, "(" .. V .. ")(" .. accent_c .. "?)ẃ()()" ,"%1%2w%3%4")
text = rsub(text, "(" .. V .. ")(" .. accent_c .. "?)ẃ()()" ,"%1%2w%3%4")
text = rsub(text, "(" .. V .. ")(" .. accent_c .. "?)ý()()" ,"%1%2y%3%4")
text = rsub(text, "(" .. V .. ")(" .. accent_c .. "?)ý()()" ,"%1%2y%3%4")
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C .. V .. ")", "%1.%2")
-- "mb", "mp", "nd", "nk", "nt" combinations
text = rsub_repeatedly(text, "(m)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "(n)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "(n)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "(n)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "(ŋ)()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "()()()(" .. V .. ")", "%1%2.%3%4")
text = rsub_repeatedly(text, "()()(?)()(" .. V .. ")", "%1%2%3.%4%5")
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_repeatedly(text, "(" .. C .. ")%.s(" .. C .. ")", "%1s.%2")
-- Any aeéèo, or stressed iu, should be syllabically divided from a following aeéèo or stressed iu.
text = rsub_repeatedly(text, "(" .. accent_c .. "*)()", "%1.%2")
text = rsub_repeatedly(text, "(" .. accent_c .. "*)(" .. V .. accent_c .. ")", "%1.%2")
text = rsub(text, "(" .. accent_c .. ")()", "%1.%2")
text = rsub_repeatedly(text, "(" .. accent_c .. ")(" .. V .. accent_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")
text = rsub_repeatedly(text, "i()", "i.%1")
text = rsub_repeatedly(text, "u()", "u.%1")
--correct final consonant placement before any glides (e.g. in ], ] ])
text = rsub(text,"()(".. C .. ")(".. W ..")(" .. V .. ")","%2%1%3%4")
text = rsub(text, "β", "bh")
text = rsub(text, "ć", "c")
text = rsub(text, "ĵ", "j")
text = rsub(text, "x", "kh")
text = rsub(text, "ŋ", "ng")
text = rsub(text, "ɲ", "ny")
text = rsub(text, "ʃ", "sy")
text = rsub(text, "q", "k")
text = rsub(text, "ɨ", "eu")
text = rsub(text, "()", "e")
text = rsub(text, "ò", "o")
text = remove_accents(text)
text = rsub_repeatedly(text, "(+)", ".")
text = rsub(text, "?-?", "-")
text = rsub(text, "()", "|%1")
text = rsub(text, "()", "|%1")
text = rsub(text, "()+", "%1")
-- remove # symbols at word and text boundaries
text = rsub_repeatedly(text, "(?)#(?)", "")
text = rsub(text, "․", ".")
-- Fix Capitalization --
local syllbreak = 0
for i=1, #text do
if text:sub(i,i) == "|" and origtext:sub(i-syllbreak, i-syllbreak) ~= "." and origtext:sub(i-syllbreak, i-syllbreak) ~= "7" then
syllbreak = syllbreak + 1
elseif origtext:sub(i-syllbreak, i-syllbreak) == text:sub(i,i):upper() then
text = table.concat({text:sub(1, i-1), text:sub(i,i):upper(), text:sub(i+1)})
end
end
-- Fix hyphens --
-- FIXME!!! Why are we relying on looking at the pagename here? This should not be happening.
origtext = pagename
if (table.concat(rsplit(origtext, "-")) == table.concat(rsplit(table.concat(rsplit(text, "|")), "-"))) then
syllbreak = 0
for i=1, #text do
if text:sub(i,i) == "|" then
if origtext:sub(i-syllbreak, i-syllbreak) == "-" then
text = table.concat({text:sub(1, i-1), "-", text:sub(i+1)})
else
syllbreak = syllbreak + 1
end
end
end
end
-- FIXME! Hack -- up above we changed periods to vertical bars. The rest of the code expects periods so change
-- them back. We should clean up the code above to leave the periods alone.
return (text:gsub("|", "%."))
end
function export.syllabify_and_align(respelling, pagename)
local syllabification = syllabify_from_spelling(respelling, pagename)
return align_syllabification_to_spelling(syllabification, pagename)
end
local function css_wrap(text, classes)
return ('<span class="%s">%s</span>'):format(classes, text)
end
local function format_glosses(glosses)
if not glosses then
return ""
end
local formatted_glosses = {}
for _, glossobj in ipairs(glosses) do
local gloss_parts = {}
if glossobj.gloss then
table.insert(gloss_parts, css_wrap("“", "mention-gloss-double-quote") ..
css_wrap(glossobj.gloss, "mention-gloss") .. css_wrap("”", "mention-gloss-double-quote"))
end
if glossobj.poses then
for _, pos in ipairs(glossobj.poses) do
table.insert(gloss_parts, css_wrap(pos, "ann-pos"))
end
end
table.insert(formatted_glosses, table.concat(gloss_parts, css_wrap(",", "mention-gloss-comma") .. " "))
end
return " " .. css_wrap("(", "mention-gloss-paren annotation-paren") ..
table.concat(formatted_glosses, css_wrap(";", "mention-gloss-semicolon") .. " ") ..
css_wrap(")", "mention-gloss-paren annotation-paren")
end
local function format_pronuns(pronuns)
local pronunciations = {}
-- Loop through each pronunciation. For each one, add the phonemic and phonetic versions to `pronunciations`,
-- for formatting by ].
for j, pronun in ipairs(pronuns) do
local qs = pronun.q
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,
})
end
if pronun.phonetic then -- missing if 'raw:/.../' given
local bracket_pron = ""
table.insert(pronunciations, {
pron = bracket_pron,
})
end
local last_pronun = #pronunciations
if pronun.q then
pronunciations.q = pronun.q
end
if j > 1 then
pronunciations.separator = ", "
end
if pronun.qq then
pronunciations.qq = pronun.qq
end
if pronun.refs then
pronunciations.refs = pronun.refs
end
if first_pronun ~= last_pronun then
pronunciations.separator = " "
end
end
return m_IPA.format_IPA_full { lang = lang, items = pronunciations, separator = "" }
end
local function format_pronun_line(parsed)
local formatted_pronuns = format_pronuns(parsed.pronuns)
local pre = is_first and parsed.pre and parsed.pre .. " " or ""
local post = is_first and parsed.post and " " .. parsed.post or ""
return pre .. formatted_pronuns .. format_glosses(parsed.t) .. post
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
-- External entry point for {{id-pr}}.
function export.show_full(frame)
--------------------------------- 1. Parse the arguments. ------------------------------------
local params = {
= {list = true},
= {},
= {},
= {},
= {list = true},
= {},
}
local parargs = frame:getParent().args
local args = require("Module:parameters").process(parargs, params)
local pagename = args.pagename or mw.title.getCurrentTitle().subpageText
local respellings = #args > 0 and args 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_syll = args.syll and
parse_syll(args.syll, function(msg) overall_parse_err(msg, "syll", args.syll) 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 > 0 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)
table.insert(overall_audio, parsed_audio)
end
end
-- Parse each respelling. Individual arguments in 1=, 2=, etc. can consist of one or more comma-separated
-- respellings, each of which can have inline modifiers <q:...>, <qq:...>, <a:...>, <aa:...> or <ref:...>.
-- In addition, the respellings as a whole of a given argument can be followed by various inline modifiers,
-- such as <t:...>, <rhyme:...>, <syll:...>, etc. The result of parsing goes into `parsed_respellings`, which
-- is a list of objects (one per numbered argument), each of which is a table of the form
--
-- {
-- terms = {TERM, TERM, ...},
-- audio = {AUDIO, AUDIO, ...},
-- rhyme = {RHYME, RHYME, ...},
-- syll = {SYLL, SYLL, ...},
-- hmp = {HMP, HMP, ...},
-- t = {GLOSS, GLOSS, ...},
-- pre = "PRE-TEXT" or nil,
-- post = "POST-TEXT" or nil,
-- bullets = NUM_BULLETS,
-- accents = {"ACCENT", "ACCENT", ...},
-- }
--
-- In this structure, TERM is an object that usually has the form
--
-- {
-- term = "RESPELLING",
-- ref = {"REF-SPEC", "REF-SPEC", ...},
-- q = {"QUALIFIER", "QUALIFIER", ...},
-- qq = {"QUALIFIER", "QUALIFIER", ...},
-- }
--
-- Note that in this structure, "REF-SPEC" of the form parsable by parse_references() in ].
--
-- Alternatively, if phonemic or phonetic IPA is given in place of a respelling, TERM will have the form
--
-- {
-- raw = true,
-- phonemic = "PHONEMIC",
-- phonetic = "PHONETIC",
-- ref = {"REF-SPEC", "REF-SPEC", ...},
-- q = {"QUALIFIER", "QUALIFIER", ...},
-- qq = {"QUALIFIER", "QUALIFIER", ...},
-- }
--
-- AUDIO is a table of the form
--
-- {
-- file = "FILE",
-- gloss = "GLOSS",
-- q = {"QUALIFIER", "QUALIFIER", ...},
-- qq = {"QUALIFIER", "QUALIFIER", ...},
-- a = {"ACCENT-QUALIFIER", "ACCENT-QUALIFIER", ...},
-- aa = {"ACCENT-QUALIFIER", "ACCENT-QUALIFIER", ...},
-- }
--
-- RHYME is a table of the form
--
-- {
-- rhyme = "RHYME",
-- num_syl = {NUM_SYL, NUM_SYL, ...},
-- q, qq, a, aa = (as for AUDIO),
-- }
--
-- SYLL is a table of the form (where `hyph` is required to be named this way for ])
--
-- {
-- syllabification = "SYL.LAB.LES",
-- hyph = {"SYL", "LAB", "LES"},
-- q, qq, a, aa = (as for AUDIO),
-- }
--
-- HMP is a table of the form
--
-- {
-- term = "HOMOPHONE",
-- gloss = "GLOSS" or nil,
-- pos = "POS" or nil,
-- alt = "ALT" or nil,
-- lit = "LIT" or nil,
-- id = "ID" or nil,
-- g = {"G", "G", ...},
-- q, qq, a, aa = (as for AUDIO),
-- }
--
-- GLOSS is a table of the form
--
-- {
-- poses = {"POS", "POS", ...} or nil,
-- gloss = "GLOSS" or nil,
-- }
for i, respelling in ipairs(respellings) do
if respelling:find("<") then
local param_mods = {
pre = { overall = true },
post = { overall = true },
bullets = {
overall = true,
type = "number",
},
t = {
overall = true,
store = "insert",
convert = parse_gloss,
},
rhyme = {
overall = true,
store = "insert-flattened",
convert = parse_rhyme,
},
syll = {
overall = true,
-- Not `store = "insert-flattened"`. parse_syll() does not generates a list but a structure where
-- the syllabifications are in `terms` and there's an additional overall property `cap` for the
-- caption (defaulting to "Syllabification"). FIXME: Rethink whether we even want "insert-flattened"
-- or just "insert" for the remaining pronunciation properties.
convert = parse_syll,
},
hmp = {
overall = true,
store = "insert-flattened",
convert = parse_homophone,
},
audio = {
overall = true,
store = "insert", -- not "insert-flattened" because parse_audio returns a single object
convert = parse_audio,
},
ref = { store = "insert" },
q = { store = "insert" },
qq = { store = "insert" },
a = {
item_dest = "accents",
overall = true,
convert = parse_accents,
},
}
local parsed = require(parse_utilities_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,
pre_normalize_modifiers = function(data)
local modtext = data.modtext
modtext = modtext:match("^<(.*)>$")
if not modtext then
error(("Internal error: Passed-in modifier isn't surrounded by angle brackets: %s"):format(
data.modtext))
end
if modtext:find("%^") and not modtext:find("^t:") then
modtext = "t:" .. modtext
end
return "<" .. modtext .. ">"
end,
splitchar = ",",
outer_container = {},
})
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,
bullets = 1,
})
end
end
--------------------------------- 2. Generate IPA, rhymes and syllabification. ------------------------------------
-- Used for categorization below.
local syllabification_alignment_failed = false
-- Canonicalize syllabifications in `sylls` by convering '+' to the default syllabification of the pagename, '#' to
-- the pagename itself, and '-' to no syllabification (return `null_syll`). If '-' not seen, return `sylls`.
local function canonicalize_syllabification(sylls, null_syll)
for _, syll in ipairs(sylls.terms) do
if syll.syllabification == "+" then
syll.syllabification = syllabify_from_spelling(pagename, pagename)
syll.hyph = split_syllabified_spelling(syll.syllabification)
elseif syll.syllabification == "#" then
syll.syllabification = pagename
syll.hyph = {syll.syllabification}
elseif syll.syllabification == "-" then
return null_syll
end
end
return sylls
end
if overall_syll then
overall_syll = canonicalize_syllabification(overall_syll, {})
end
local function doesnt_count_for_rhyme(list)
if not list then
return false
end
local accent_no_count = {"colloquial", "obsolete", "nonstandard"}
for _, item in ipairs(list) do
for _, word_no_count in ipairs(accent_no_count) do
if item:find("%f" .. word_no_count .. "%f") then
return true
end
end
end
return false
end
-- Loop over individual respellings, processing each.
for _, parsed in ipairs(parsed_respellings) do
-- First, sort the specified accents and default to "Standard Indonesian".
local default_accent = "Standard Indonesian"
if not parsed.accents then
parsed.accents = {default_accent}
end
-- If more than one respelling given, then if any accent or qualifier has the words 'colloquial', 'obsolete' or
-- 'nonstandard' in them, don't generate a rhyme or a '#-syllable word' category.
local more_than_one_respelling = #parsed.terms > 1 or #parsed_respellings > 1
local is_standard = m_table.contains(parsed.accents, default_accent)
local all_terms_no_rhyme = more_than_one_respelling and doesnt_count_for_rhyme(parsed.accents)
parsed.pronuns = {}
for i, term in ipairs(parsed.terms) do
local phonemic, phonetic
if term.raw then
phonemic = term.raw_phonemic
phonetic = term.raw_phonetic
else
local ret = export.IPA(term.term)
phonemic = ret.phonemic
phonetic = ret.phonetic
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
local no_rhyme, rhyme_with_cat
-- Same check as above for colloquial/obsolete/relaxed but check the qualifiers, which are attached to
-- individual respellings rather than a single-line set of respellings.
no_rhyme = all_terms_no_rhyme or more_than_one_respelling and (
doesnt_count_for_rhyme(term.q) or doesnt_count_for_rhyme(term.qq)
)
if not no_rhyme then
local should_generate_rhyme, should_generate_cat = should_generate_rhyme_from_termobj(term)
no_rhyme = not should_generate_rhyme
rhyme_with_cat = should_generate_cat
end
local pronobj = {
raw = term.raw,
phonemic = phonemic,
phonetic = phonetic,
refs = refs,
q = term.q,
qq = term.qq,
no_rhyme = no_rhyme,
rhyme_with_cat = rhyme_with_cat,
}
table.insert(parsed.pronuns, pronobj)
end
if not parsed.syll then
if not overall_syll and any_words_have_vowels(pagename) then
for _, term in ipairs(parsed.terms) do
if not term.raw then
local syllabification = syllabify_from_spelling(term.term, pagename)
local aligned_syll = align_syllabification_to_spelling(syllabification, pagename)
if aligned_syll then
if not parsed.syll then
parsed.syll = {terms = {}}
end
m_table.insertIfNot(parsed.syll.terms, generate_syll_obj(aligned_syll))
else
syllabification_alignment_failed = true
end
end
end
end
else
parsed.syll = canonicalize_syllabification(parsed.syll, nil)
end
if not parsed.rhyme then
if overall_rhyme then
parsed.rhyme = nil
else
-- Generate the rhymes.
for _, pronun in ipairs(parsed.pronuns) 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), and rhyme isn't suppressed (which may happen if the term has
-- a qualifier "colloquial", "obsolete" or the like ).
if pronun.phonemic and not pronun.no_rhyme then
-- Count number of syllables by looking at syllable boundaries (including stress marks).
local num_syl = get_num_syl_from_ipa(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)
-- Copying qualifiers to rhymes:
-- (1) If there's only one pronunciation, displaying any associated qualifier on the rhyme is
-- is redundant, so don't do it.
-- (2) If there are multiple pronunciations, then we generally do want to copy the qualifier(s)
-- from pronunciation to rhyme, but only if a given rhyme either derives from a single
-- pronunciation, or derives from multiple pronunciations all of which share the same
-- qualifier(s). We do NOT want to combine two different qualifiers from two different
-- pronunciations.
-- (3) If there are multiple pronunciations that map to a single rhyme, and all pronunciations
-- share qualifiers, then we might consider omitting the qualifiers as redundant; but this
-- case will rarely happen so it might not be worth worrying about.
-- (4) Similarly, if there are multiple pronunciations where some have the rhyme suppressed (see
-- above), and all pronunciations share qualifiers, then we might consider omitting the
-- qualifiers as redundant; but again, this case will rarely happen (especially since in
-- almost all cases the suppressed-rhyme pronunciation will have distinctive qualifiers) so
-- it probably isn't worth worrying about. Note that in the common case where the qualifiers
-- of the rhyme-suppressed pronunciation differ from those of the rhyme-included
-- pronunciation, we do want to include the qualifiers of the rhyme-included pronunciation
-- (imagine e.g. there are two pronunciations marked "standard" and "colloquial"; we want to
-- mark the rhyme as "standard").
-- (4) There are two different types of qualifiers (left and right); when comparing qualifiers,
-- we need to compare the entire set of both qualifiers and make sure they both match
-- (although it will be rare to have both left and right qualifiers on a single
-- pronunciation).
local saw_already = false
if not parsed.rhyme then
parsed.rhyme = {}
end
for _, existing in ipairs(parsed.rhyme) 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 'biologi' (4 syllables) and
-- 'biologi' (5 syllables), both of which have the same rhyme /ogi/.
m_table.insertIfNot(existing.num_syl, num_syl)
if not m_table.deepEquals(existing.q, pronun.q) or not
m_table.deepEquals(existing.qq, pronun.qq) then
existing.q = nil
existing.qq = nil
end
break
end
end
if not saw_already then
table.insert(parsed.rhyme, {
rhyme = rhyme,
num_syl = {num_syl},
q = #parsed.pronuns > 1 and pronun.q or nil,
qq = #parsed.pronuns > 1 and pronun.qq or nil,
nocat = not pronun.rhyme_with_cat,
})
end
end
end
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
parsed.rhyme = process_specified_rhymes(parsed.rhyme, parsed.syll and parsed.syll.terms or {}, {parsed})
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 all_sylls
if overall_syll then
all_sylls = overall_syll
else
all_sylls = {}
for _, parsed in ipairs(parsed_respellings) do
if parsed.syll then
for _, syll in ipairs(parsed.syll.terms) do
m_table.insertIfNot(all_sylls, syll)
end
end
end
end
overall_rhyme = process_specified_rhymes(overall_rhyme, all_sylls, parsed_respellings)
end
end
-- Determine whether all sets of pronunciations have the same value for a pronunciation property (rhymes,
-- syllabifications or homophones). If so, we display them them only once at the bottom, otherwise beneath each set,
-- indented. This function takes one argument, the name of a slot specifying the pronunciation property, and
-- returns two values, a boolean indicating whether all values are the same and the first value seen (which will
-- be the only value seen if all values are the same).
local function all_sets_equal(parsed_slot)
local first_set
local all_sets_eq = true
for j, parsed in ipairs(parsed_respellings) do
if j == 1 then
first_set = parsed
elseif not m_table.deepEquals(first_set, parsed) then
all_sets_eq = false
break
end
end
return all_sets_eq, first_set
end
local all_rhyme_sets_eq, first_rhyme_ret = all_sets_equal("rhyme")
local all_syll_sets_eq, first_sylls = all_sets_equal("syll")
local all_hmp_sets_eq, first_hmps = all_sets_equal("hmp")
------------------------------ 3. Insert categories as appropriate. ---------------------------------
local categories = {}
local function get_rhymes_categories(rhymes)
if not rhymes then
return
end
end
get_rhymes_categories(overall_rhyme)
for _, parsed in ipairs(parsed_respellings) do
get_rhymes_categories(parsed.rhyme)
end
local function get_syll_categories(sylls)
if not sylls then
return
end
for _, syll in ipairs(sylls.terms) do
local syll_no_dot = "#" .. syll.syllabification .. "#"
syll_no_dot = syll.syllabification:gsub("%.()", "%1"):gsub("#", "")
if syll_no_dot ~= pagename then
mw.log(("For page '%s', saw syllabification '%s' not matching pagename"):format(
pagename, syll.syllabification))
m_table.insertIfNot(categories, ("%s terms with syllabification not matching pagename"):format(
lang:getCanonicalName()))
end
end
end
get_syll_categories(overall_syll)
for _, parsed in ipairs(parsed_respellings) do
get_syll_categories(parsed.syll)
end
if syllabification_alignment_failed then
table.insert(categories, ("%s terms where syllabification alignment failed"):format(lang:getCanonicalName()))
end
---------------------------- 4. Format IPA, rhymes and syllabification for display. -------------------------------
local function bullet_prefix(num_bullets)
return string.rep("*", num_bullets) .. " "
end
local function format_rhyme(rhymes)
return require(rhymes_module).format_rhymes {
lang = lang,
rhymes = rhymes,
force_cat = force_cat,
}
end
local function format_syllabifications(syllobj)
return require(hyphenation_module).format_hyphenations {
lang = lang,
hyphs = syllobj.terms,
caption = syllobj.cap or "Syllabification"
}
end
local function format_homophones(hmps)
return require("Module:homophones").format_homophones { lang = lang, homophones = hmps }
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, bullet_prefix(num_bullets) .. text)
end
return table.concat(ret, "\n")
end
-- Implement grouping by accent. If there is a run of more than one consecutive set of pronunciations with the
-- same accent, the accent goes on its own line and the pronunciations with this accent go below with an extra
-- bullet.
local prev_accents
local num_seen_with_these_accents
for j, parsed in ipairs(parsed_respellings) do
if m_table.deepEquals(prev_accents, parsed.accents) then
parsed.of_several_accents = "continuation"
num_seen_with_these_accents = num_seen_with_these_accents + 1
if num_seen_with_these_accents == 2 then
parsed_respellings.of_several_accents = "first"
end
else
prev_accents = parsed.accents
num_seen_with_these_accents = 1
end
end
-- Pull out autogenerated pronunciations and move to the next line, indented.
for _, parsed in ipairs(parsed_respellings) do
local saw_next_line_pronuns = false
for _, pronun in ipairs(parsed.pronuns) do
if pronun.move_to_next_line then
saw_next_line_pronuns = true
break
end
end
if saw_next_line_pronuns then
local this_line_pronuns = {}
local next_line_pronuns = {}
for _, pronun in ipairs(parsed.pronuns) do
if pronun.move_to_next_line then
table.insert(next_line_pronuns, pronun)
else
table.insert(this_line_pronuns, pronun)
end
end
-- Now see if there are qualifiers shared among all elements of the next-line pronuns and deduplicate if so.
local function deduplicate_qualifiers(field, keepfirst)
local saw_nil = false
for _, pronun in ipairs(next_line_pronuns) do
if not pronun then
saw_nil = true
break
end
end
if not saw_nil then
local m_setutil = require(set_utilities_module)
local qualifiers = {}
for _, pronun in ipairs(next_line_pronuns) do
table.insert(qualifiers, m_setutil.list_to_set(pronun))
end
local all_shared = m_setutil.intersect(unpack(qualifiers))
if next(all_shared) then
local first_index, last_index
if keepfirst then
first_index = 2
last_index = #pronun
else
first_index = 1
last_index = #pronun - 1
end
for i = first_index, last_index do
local pronun = next_line_pronuns
local new_qualifiers = {}
for _, q in ipairs(pronun) do
if not all_shared then
table.insert(new_qualifiers, q)
end
end
pronun = new_qualifiers
end
end
end
end
parsed.pronuns = this_line_pronuns
parsed.next_line_pronuns = next_line_pronuns
end
end
-- Now actually format the pronunciations.
local textparts = {}
local first_line = true
local function ins_line(linetext, num_bullets)
if not first_line then
table.insert(textparts, "\n")
end
first_line = false
table.insert(textparts, bullet_prefix(num_bullets) .. linetext)
end
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
local accent_grouping_offset = 0
if parsed.of_several_accents == "first" then
ins_line(require(accent_qualifier_module).format_qualifiers(lang, parsed.accents), parsed.bullets)
end
local pronuns = format_pronun_line(parsed)
local accent_prefix
if not parsed.of_several_accents then
accent_prefix = require(accent_qualifier_module).format_qualifiers(lang, parsed.accents) .. " "
else
accent_prefix = ""
accent_grouping_offset = 1
end
ins_line(accent_prefix .. pronuns, parsed.bullets + accent_grouping_offset)
if parsed.next_line_pronuns then
ins_line(format_pronuns(parsed.next_line_pronuns), parsed.bullets + accent_grouping_offset + 1)
end
if parsed.audio then
-- format_audio() inserts multiple lines and handles bullets by itself.
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) + accent_grouping_offset))
end
if not all_rhyme_sets_eq and parsed.rhyme then
ins_line(format_rhyme(parsed.rhyme), parsed.bullets + 1 + accent_grouping_offset)
end
if not all_syll_sets_eq and parsed.syll then
ins_line(format_syllabifications(parsed.syll), parsed.bullets + 1 + accent_grouping_offset)
end
if not all_hmp_sets_eq and parsed.hmp then
ins_line(format_homophones(parsed.hmp), parsed.bullets + 1 + accent_grouping_offset)
end
end
if overall_audio then
-- format_audio() inserts multiple lines and handles bullets by itself.
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
ins_line(format_rhyme(first_rhyme_ret), min_num_bullets)
end
if overall_rhyme then
ins_line(format_rhyme(overall_rhyme), min_num_bullets)
end
if all_syll_sets_eq and first_sylls then
ins_line(format_syllabifications(first_sylls), min_num_bullets)
end
if overall_syll then
ins_line(format_syllabifications(overall_syll), min_num_bullets)
end
if all_hmp_sets_eq and first_hmps then
ins_line(format_homophones(first_hmps), min_num_bullets)
end
if overall_hmp then
ins_line(format_homophones(overall_hmp), min_num_bullets)
end
return table.concat(textparts) ..
require("Module:utilities").format_categories(categories, lang, nil, nil, force_cat)
end
-- Meant to be called from a bot.
function export.pron_json(frame)
local iparams = {
= {list = true, required = true},
= {required = true},
}
local iargs = require("Module:parameters").process(frame.args, iparams)
local data = {}
local syllabification_from_pagename = syllabify_from_spelling(iargs.pagename, iargs.pagename)
for _, respelling in ipairs(iargs) do
local pronun = export.IPA(respelling)
local syllabification = export.syllabify_and_align(respelling, iargs.pagename)
local num_syl = get_num_syl_from_ipa(pronun.phonemic)
local rhyme = convert_phonemic_to_rhyme(pronun.phonemic)
table.insert(data, {
respelling = respelling,
phonemic = pronun.phonemic,
phonetic = pronun.phonetic,
syllabification = syllabification,
num_syl = num_syl,
rhyme = rhyme,
})
end
local retval = {
pagename = iargs.pagename,
syllabification_from_pagename = syllabification_from_pagename,
data = data,
}
return require("Module:JSON").toJSON(retval)
end
return export