Module:kilta-pron: Difference between revisions
Jump to navigation
Jump to search
No edit summary Tag: Manual revert |
No edit summary |
||
(133 intermediate revisions by the same user not shown) | |||
Line 5: | Line 5: | ||
local match = mw.ustring.match | local match = mw.ustring.match | ||
local u = mw.ustring.char | local u = mw.ustring.char | ||
local gsplit = mw.text.gsplit | |||
local split = mw.text.split | local split = mw.text.split | ||
local A = u(0x0301) -- COMBINING ACUTE | local A = u(0x0301) -- COMBINING ACUTE | ||
Line 14: | Line 14: | ||
local m_IPA = require("Module:IPA") | local m_IPA = require("Module:IPA") | ||
local consonants = "[ | local consonants = "[pβmtsnɾlʧkxqhyʤɡbvdƕ]" | ||
local vowels = "[ | local vowels = "[aeiouáéíóúəïüëæ]" | ||
local export = {} | local export = {} | ||
Line 25: | Line 25: | ||
end | end | ||
local | local phonemic_rules = { | ||
{"%-$", ""}, {"%-", " "}, | |||
{"hw", "ƕ"}, {"kw", "q"}, {"ch", "ʧ"}, {"au", "ü"}, {"ai", "ï"}, {"h", "x"}, | |||
{"v", "β"}, {"r", "ɾ"}, | |||
{"ë", "ə"}, {"əə+", "əː"}, | |||
return | {"kq", "qː"}, {"cʧ", "ʧː"}, {"(" .. consonants .. ")(" .. consonants .. ")", | ||
function(c1,c2) return same(c1,c2) and c1 .. "ː" or c1 .. c2 end}, | |||
} | |||
local | local phonetic_rules = { | ||
{" | {"a(·?ˈ?)([nm])", "æ%1%2"}, {"á(·?ˈ?)([nm])", "æː%1%2"}, | ||
{"n(ː?·?ˈ?)([kxqƕ])", "ŋ%1%2"}, | |||
{"([nŋm]·?)k", "%1ɡ"}, {"([nŋm]·?)q", "%1y"}, {"([nŋm]·?)p", "%1b"}, {"([nŋm]·?)ʧ", "%1ʤ"}, | |||
{" | {"([nŋm]·?)t", "%1d"}, | ||
{" | {"([nŋm])(ː?·?ˈ?)β", "%1%2b"}, | ||
{"β([iíeé])", "v%1"}, | |||
{"^(ˈ?)(" .. vowels .. ")", "%1ʔ%2"}, | |||
{"ü", "aʊ̯"}, | |||
{"ï", "aɪ̯"}, | |||
} | } | ||
local | local deacuter = { | ||
["á"] = "a", ["é"] = "e", ["í"] = "i", ["ó"] = "o", ["ú"] = "u", | |||
} | } | ||
local last_rules = { | local last_rules = { | ||
{" | {"q", "kʷ"}, {"ƕ", "xʷ"}, {"y", "ɡʷ"}, | ||
{"ʧ", "t͡ʃ"}, | {"ʧ", "t͡ʃ"}, | ||
{"ʤ", "d͡ʒ"}, | |||
{"ü", "au̯"}, | |||
{"ï", "ai̯"}, | |||
{"[·%.]ˈ", "ˈ"}, {"·", "."}, {"([áéíóú])", function(v) return deacuter[v] .. "ː" end}, | |||
} | } | ||
local sandhi_rules = { | |||
{"(" .. consonants .. ")u%sˈ?ʔ?(" .. vowels .. ")", "%1w%2"}, | |||
{"(" .. consonants .. ")i%sˈ?ʔ?(" .. vowels .. ")", "%1j%2"}, | |||
{"([mnŋ])(%sˈ?)βə", "%1%2bə"}, | |||
{"[mnŋ]%s(ˈ?[mpbβ])", "m %1"}, {"[mnŋ]%s(ˈ?v)", "ɱ %1"}, | |||
{"[mnŋ]%s(ˈ?[tdɾlsn])", "n %1"}, {"[mnŋ]%s(ˈ?[ʧʤ])", "n̠ %1"}, | |||
{"[mnŋ]%s(ˈ?[ŋkɡqƕy])", "ŋ %1"}, | |||
} | |||
local function syllabify(term) | |||
local first = term == "áxəpːi" | |||
local last = (term == "kolán" or term == "türá") | |||
local syllable = "(" .. consonants .. "*)(" .. vowels .. "ː?)(" .. consonants .. "-)" | |||
term = gsub(term, syllable, "·%1%2%3·") | |||
term = gsub(term, "^·", "") | |||
term = gsub(term, "·$", "") | |||
term = gsub(term, "··+", "·") | |||
term = gsub(term, "·(" .. consonants .. ")ː·", "%1·%1") | |||
term = gsub(term, "q·q", "k·q") | |||
term = gsub(term, "·(" .. consonants .. ")·", "%1·") | |||
term = gsub(term, "·(" .. consonants .. ")(" .. consonants .. ")", "%1·%2") | |||
term = gsub(term, "·([nslɾmt])$", "%1") | |||
term = gsub(term, "·(" .. consonants .. ")ː·", "·%1") | |||
local syllables = split(term, "·") | |||
if first then | |||
syllables[1] = "ˈ" .. syllables[1] | |||
elseif last then | |||
syllables[#syllables] = "ˈ" .. syllables[#syllables] | |||
elseif #syllables > 1 then | |||
syllables[#syllables - 1] = "ˈ" .. syllables[#syllables - 1] -- penultimate stress | |||
end | |||
return table.concat(syllables, "·") | |||
end | |||
function export.crux(term) | function export.crux(term) | ||
term = mw.ustring.lower | term = mw.ustring.lower(term) | ||
local phonemic | local phonemic = term | ||
for _, rule in ipairs(phonemic_rules) do | for _, rule in ipairs(phonemic_rules) do | ||
phonemic = gsub( | phonemic = gsub(phonemic, rule[1], rule[2]) | ||
end | end | ||
phonemic = syllabify(phonemic) | |||
local phonetic = phonemic | |||
for _, rule in ipairs(phonetic_rules) do | for _, rule in ipairs(phonetic_rules) do | ||
phonetic = gsub(phonetic, rule[1], rule[2]) | phonetic = gsub(phonetic, rule[1], rule[2]) | ||
end | end | ||
Line 92: | Line 127: | ||
end | end | ||
local phonemicAll = table.concat(phonemic, " ") | |||
local phoneticAll = table.concat(phonetic, " ") | |||
for _, rule in ipairs(sandhi_rules) do | |||
phoneticAll = gsub(phoneticAll, rule[1], rule[2]) | |||
end | |||
for _, rule in ipairs(last_rules) do | |||
phonemicAll = gsub(phonemicAll, rule[1], rule[2]) | |||
phoneticAll = gsub(phoneticAll, rule[1], rule[2]) | |||
end | |||
return phonemicAll, phoneticAll | |||
end | end | ||
Line 103: | Line 150: | ||
local term = args[1] | local term = args[1] | ||
local phonemic, phonetic = separate_word(term) | local phonemic, phonetic = separate_word(term) | ||
local IPA_args = {{pron = '[' .. phonetic .. ']'}} | |||
if phonemic ~= phonetic then table.insert(IPA_args, 1, {pron = '/' .. phonemic .. '/'}) end | |||
return "* " .. m_IPA.format_IPA_full(lang, IPA_args) | return "* " .. m_IPA.format_IPA_full({lang = lang, items = IPA_args}) | ||
end | end | ||
return export | return export |
Latest revision as of 18:07, 12 August 2024
Documentation for this module may be created at Module:kilta-pron/doc
local sub = mw.ustring.sub
local find = mw.ustring.find
local gmatch = mw.ustring.gmatch
local gsub = mw.ustring.gsub
local match = mw.ustring.match
local u = mw.ustring.char
local gsplit = mw.text.gsplit
local split = mw.text.split
local A = u(0x0301) -- COMBINING ACUTE
local D = u(0x0308) -- COMBINING DIAERESIS
local lang = require("Module:languages").getByCode("kilta")
local m_IPA = require("Module:IPA")
local consonants = "[pβmtsnɾlʧkxqhyʤɡbvdƕ]"
local vowels = "[aeiouáéíóúəïüëæ]"
local export = {}
local function same(foo, bar)
foo, bar = mw.ustring.toNFD(foo), mw.ustring.toNFD(bar) -- decompose diacritics
foo, bar = match(foo, "^."), match(bar, "^.") -- sort out the letter
return foo == bar and true or false
end
local phonemic_rules = {
{"%-$", ""}, {"%-", " "},
{"hw", "ƕ"}, {"kw", "q"}, {"ch", "ʧ"}, {"au", "ü"}, {"ai", "ï"}, {"h", "x"},
{"v", "β"}, {"r", "ɾ"},
{"ë", "ə"}, {"əə+", "əː"},
{"kq", "qː"}, {"cʧ", "ʧː"}, {"(" .. consonants .. ")(" .. consonants .. ")",
function(c1,c2) return same(c1,c2) and c1 .. "ː" or c1 .. c2 end},
}
local phonetic_rules = {
{"a(·?ˈ?)([nm])", "æ%1%2"}, {"á(·?ˈ?)([nm])", "æː%1%2"},
{"n(ː?·?ˈ?)([kxqƕ])", "ŋ%1%2"},
{"([nŋm]·?)k", "%1ɡ"}, {"([nŋm]·?)q", "%1y"}, {"([nŋm]·?)p", "%1b"}, {"([nŋm]·?)ʧ", "%1ʤ"},
{"([nŋm]·?)t", "%1d"},
{"([nŋm])(ː?·?ˈ?)β", "%1%2b"},
{"β([iíeé])", "v%1"},
{"^(ˈ?)(" .. vowels .. ")", "%1ʔ%2"},
{"ü", "aʊ̯"},
{"ï", "aɪ̯"},
}
local deacuter = {
["á"] = "a", ["é"] = "e", ["í"] = "i", ["ó"] = "o", ["ú"] = "u",
}
local last_rules = {
{"q", "kʷ"}, {"ƕ", "xʷ"}, {"y", "ɡʷ"},
{"ʧ", "t͡ʃ"},
{"ʤ", "d͡ʒ"},
{"ü", "au̯"},
{"ï", "ai̯"},
{"[·%.]ˈ", "ˈ"}, {"·", "."}, {"([áéíóú])", function(v) return deacuter[v] .. "ː" end},
}
local sandhi_rules = {
{"(" .. consonants .. ")u%sˈ?ʔ?(" .. vowels .. ")", "%1w%2"},
{"(" .. consonants .. ")i%sˈ?ʔ?(" .. vowels .. ")", "%1j%2"},
{"([mnŋ])(%sˈ?)βə", "%1%2bə"},
{"[mnŋ]%s(ˈ?[mpbβ])", "m %1"}, {"[mnŋ]%s(ˈ?v)", "ɱ %1"},
{"[mnŋ]%s(ˈ?[tdɾlsn])", "n %1"}, {"[mnŋ]%s(ˈ?[ʧʤ])", "n̠ %1"},
{"[mnŋ]%s(ˈ?[ŋkɡqƕy])", "ŋ %1"},
}
local function syllabify(term)
local first = term == "áxəpːi"
local last = (term == "kolán" or term == "türá")
local syllable = "(" .. consonants .. "*)(" .. vowels .. "ː?)(" .. consonants .. "-)"
term = gsub(term, syllable, "·%1%2%3·")
term = gsub(term, "^·", "")
term = gsub(term, "·$", "")
term = gsub(term, "··+", "·")
term = gsub(term, "·(" .. consonants .. ")ː·", "%1·%1")
term = gsub(term, "q·q", "k·q")
term = gsub(term, "·(" .. consonants .. ")·", "%1·")
term = gsub(term, "·(" .. consonants .. ")(" .. consonants .. ")", "%1·%2")
term = gsub(term, "·([nslɾmt])$", "%1")
term = gsub(term, "·(" .. consonants .. ")ː·", "·%1")
local syllables = split(term, "·")
if first then
syllables[1] = "ˈ" .. syllables[1]
elseif last then
syllables[#syllables] = "ˈ" .. syllables[#syllables]
elseif #syllables > 1 then
syllables[#syllables - 1] = "ˈ" .. syllables[#syllables - 1] -- penultimate stress
end
return table.concat(syllables, "·")
end
function export.crux(term)
term = mw.ustring.lower(term)
local phonemic = term
for _, rule in ipairs(phonemic_rules) do
phonemic = gsub(phonemic, rule[1], rule[2])
end
phonemic = syllabify(phonemic)
local phonetic = phonemic
for _, rule in ipairs(phonetic_rules) do
phonetic = gsub(phonetic, rule[1], rule[2])
end
return phonemic, phonetic
end
function separate_word(term)
local phonemic, phonetic = {}, {}
for word in gsplit(term, " ") do
local phonemicEach, phoneticEach = export.crux(word)
table.insert(phonemic, phonemicEach)
table.insert(phonetic, phoneticEach)
end
local phonemicAll = table.concat(phonemic, " ")
local phoneticAll = table.concat(phonetic, " ")
for _, rule in ipairs(sandhi_rules) do
phoneticAll = gsub(phoneticAll, rule[1], rule[2])
end
for _, rule in ipairs(last_rules) do
phonemicAll = gsub(phonemicAll, rule[1], rule[2])
phoneticAll = gsub(phoneticAll, rule[1], rule[2])
end
return phonemicAll, phoneticAll
end
function export.show(frame)
local parent_args = frame:getParent().args
local params = {
[1] = { default = mw.title.getCurrentTitle().nsText == 'Template' and "kílta" or mw.title.getCurrentTitle().text },
}
local args = require("Module:parameters").process(parent_args, params)
local term = args[1]
local phonemic, phonetic = separate_word(term)
local IPA_args = {{pron = '[' .. phonetic .. ']'}}
if phonemic ~= phonetic then table.insert(IPA_args, 1, {pron = '/' .. phonemic .. '/'}) end
return "* " .. m_IPA.format_IPA_full({lang = lang, items = IPA_args})
end
return export