Module:hrx-pron
Appearance
- The following documentation is located at Module:hrx-pron/documentation. [edit] Categories were auto-generated by Module:module categorization. [edit]
- Useful links: subpage list • links • transclusions • testcases • sandbox
This module is not to be directly used. It is used by Template:hrx-pr, see there for usage.
--Adapted from [[Module:es-pronunc]] by [[User:Benwing2]]
--[==[
FIXME:
1. Add unstressed words.
2. Clean up the code.
3. <aai> = /ɔːɪ̯/
4. Intervocalic <d> = /ɾ/?
8. Assimilation of nasals before stops
]==]
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 -- for testing
local lang = require("Module:languages").getByCode("hrx")
local u = mw.ustring.char
local rfind = mw.ustring.find
local rsubn = mw.ustring.gsub
local rmatch = mw.ustring.match
local rsplit = mw.text.split
local ulower = mw.ustring.lower
local uupper = mw.ustring.upper
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 DIA = u(0x0308) -- diaeresis = ̈
local BREVE = u(0x0306) --breve
local SYLDIV = u(0xFFF0) -- used to represent a user-specific syllable divider (.) so we won't change it
local vowel = "aeəiouàáéíóőúɔɛäëïÄËÏAIOU" -- vowels
local V = "[" .. vowel .. "]" -- vowel class
local Vo = "[" .. vowel .. u(0xFFF1) .. u(0xFFF2) .. u(0xFFF3) .. u(0xFFF4) .. u(0xFFF5) .. u(0xFFF6) ..
u(0xFFF9) .. u(0xFFFA) .. u(0xFFFB) .. u(0xFFFC) .. u(0x0400) .. u(0x0401) .. u(0x0402) .. u(0x0403) .. "]" --the extra characters stand for the long vowels in the hyphenation function
local accent = AC .. GR .. CFLEX
local accent_c = "[" .. accent .. "]"
local stress = AC .. GR
local stress_c = "[" .. AC .. GR .. "]"
local ipa_stress = "ˈˌ"
local ipa_stress_c = "[" .. ipa_stress .. "]"
local sylsep = "%-." .. SYLDIV -- 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 C_OR_WORDSEP = "[^" .. vowel .. separator_not_wordsep .. "]" -- consonant class, or word separator
local unstressed_words = require("Module:table").listToSet({
"ab", "de",
})
-- 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, remove_gr)
-- decompose everything but ä, ë, ï
text = unfd(text)
text = rsub(text, ".[" .. DIA .. BREVE .. "]", {
["a" .. DIA] = "ä",
["A" .. DIA] = "Ä",
["e" .. DIA] = "ë",
["E" .. DIA] = "Ë",
["i" .. DIA] = "ï",
["I" .. DIA] = "Ï",
["u" .. BREVE] = "ŭ",
["U" .. BREVE] = "Ŭ",
["i" .. BREVE] = "ĭ",
["I" .. BREVE] = "Ĭ",
})
if remove_gr then
text = rsub(text, "a" .. GR, "ä")
text = rsub(text, "ĭ", "i")
text = rsub(text, "ŭ", "u")
end
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
-- Main syllable-division algorithm
local function syllabify_from_spelling_or_pronun(text, is_spelling)
text = rsub(text, "##", "@@")
text = rsub(text, "# #", "@ @")
text = rsub(text, "#", ".")
-- Part 1: Divide before the last consonant in a cluster of consonants between vowels
text = rsub_repeatedly(text, "(" .. Vo .. accent_c .. "*)(" .. C .. Vo .. ")", "%1.%2")
text = rsub_repeatedly(text, "(" .. Vo .. accent_c .. "*" .. C .. "+)(" .. C .. Vo .. ")", "%1.%2")
-- Don't divide Cl or Cr where C is a stop or fricative, except for dl.
text = rsub(text, "([pbfvkctgɡ])%.([lr])", ".%1%2")
-- Don't divide ch
text = rsub(text, "c%.h", ".ch")
-- Part 2: Divide hiatuses (handling ai, ei, au, eu)
text = rsub(text,"([aeAE])([ui])", "%1&%2")
text = rsub_repeatedly(text, "(" .. Vo .. accent_c .. "*)(" .. Vo .. ")", "%1.%2")
text = rsub(text, "&", "")
--other fixes
text = rsub(text, "@@", "##")
text = rsub(text, "@ @", "# #")
return text
end
local function syllabify_from_spelling(text)
text = decompose(text)
-- start at FFF1 because FFF0 is used for SYLDIV
-- Temporary replacements for characters we want treated as default consonants. The C and related consonant regexes
-- treat all unknown characters as consonants (vowels have been added manually to Vo)
local TEMP_aa = u(0xFFF1)
local TEMP_Aa = u(0xFFF2)
local TEMP_ee = u(0xFFF4)
local TEMP_Ee = u(0xFFF5)
local TEMP_ie = u(0xFFF3)
local TEMP_Ie = u(0xFFF6)
local TEMP_Sch = u(0xFFF7)
local TEMP_sch = u(0xFFF8)
local TEMP_Tsch = u(0x0404)
local TEMP_tsch = u(0x0405)
local TEMP_AEae = u(0xFFF9)
local TEMP_aeae = u(0xFFFA)
local TEMP_Ii = u(0xFFFB)
local TEMP_ii = u(0xFFFC)
local TEMP_Oo = u(0x0400)
local TEMP_oo = u(0x0401)
local TEMP_Uu = u(0x0402)
local TEMP_uu = u(0x0403)
--vowel+h (by default, VhV will be interpreted as long vowel+vowel)
text = rsub(text, "([aeiouäAEIOUÄ])h([aeiouä])", "%1h.%2")
--substitutions
text = rsub(text, "aa", TEMP_aa)
text = rsub(text, "Aa", TEMP_Aa)
text = rsub(text, "ee", TEMP_ee)
text = rsub(text, "Ee", TEMP_Ee)
text = rsub(text, "ii", TEMP_ii)
text = rsub(text, "Ii", TEMP_Ii)
text = rsub(text, "oo", TEMP_oo)
text = rsub(text, "Oo", TEMP_Oo)
text = rsub(text, "uu", TEMP_uu)
text = rsub(text, "Uu", TEMP_Uu)
text = rsub(text, "Ää", TEMP_AEae)
text = rsub(text, "ää", TEMP_aeae)
text = rsub(text, "tsch", TEMP_tsch)
text = rsub(text, "Tsch", TEMP_Tsch)
text = rsub(text, "sch", TEMP_sch)
text = rsub(text, "Sch", TEMP_Sch)
--diphthongs (ai, ei, au, eu + falling dipthongs respelled with <ĭŭ>)
text = rsub(text, "([ĬĭŬŭ])(" .. Vo .. ")", "%1.%2")
text = rsub(text,"([aeAE])([ui])(" .. Vo .. ")", "%1%2.%3")
text = rsub(text, "ie", TEMP_ie)
text = rsub(text, "Ie", TEMP_Ie)
--the u in qu isn't a vowel
text = rsub(text, "([Qq])u", "%1w")
text = text:gsub("%.", SYLDIV)
text = text:gsub("#", SYLDIV)
text = syllabify_from_spelling_or_pronun(text, "is spelling")
-- Change user-specified . into SYLDIV so we don't shuffle it around when dividing into syllables.
text = text:gsub(SYLDIV, ".")
text = text:gsub("qw", "qu")
text = text:gsub("Qw", "Qu")
text = text:gsub(TEMP_aa, "aa")
text = text:gsub(TEMP_Aa, "Aa")
text = text:gsub(TEMP_ee, "ee")
text = text:gsub(TEMP_Ee, "Ee")
text = text:gsub(TEMP_ie, "ie")
text = text:gsub(TEMP_Ie, "Ie")
text = text:gsub(TEMP_ii, "ii")
text = text:gsub(TEMP_Ii, "Ii")
text = text:gsub(TEMP_oo, "oo")
text = text:gsub(TEMP_Oo, "Oo")
text = text:gsub(TEMP_uu, "uu")
text = text:gsub(TEMP_Uu, "Uu")
text = text:gsub(TEMP_AEae, "Ää")
text = text:gsub(TEMP_aeae, "ää")
text = text:gsub(TEMP_sch, "sch")
text = text:gsub(TEMP_Sch, "Sch")
text = text:gsub(TEMP_tsch, "tsch")
text = text:gsub(TEMP_Tsch, "Tsch")
text = unfc(text)
return text
end
-- Generate the IPA of a given respelling
function export.IPA(text)
local debug = {}
text = ulower(text or mw.title.getCurrentTitle().text)
-- decompose everything but ä, ë, ï
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, " $", "")
return text
end
text = canon_spaces(text)
-- Make prefixes unstressed unless they have an explicit stress marker; also make certain
-- monosyllabic words without stress marks be unstressed.
local words = rsplit(text, " ")
for i, word in ipairs(words) do
if rfind(word, "%-$") and not rfind(word, accent_c) or unstressed_words[word] then
-- add CFLEX to the last vowel not the first one, or we will mess up 'que' by
-- adding the CFLEX after the 'u'
words[i] = rsub(word, "^(.*" .. V .. ")", "%1" .. CFLEX)
end
end
text = table.concat(words, " ")
-- Convert hyphens to spaces
text = rsub(text, "%-", " ")
-- 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, " ", "# #") .. "##"
--x/ç
text = rsub(text, "x", "X") --the character <x>
text = rsub(text, "auch", "auxx")
text = rsub(text, "([aou])(" .. accent_c .. "?)ch", "%1%2xx")
text = rsub(text, "eich", "eiçç")
text = rsub(text, "euch", "euçç")
text = rsub(text, "äich", "eiçç")
text = rsub(text, "([äei])(" .. accent_c .. "?)ch", "%1%2çç")
text = rsub(text, "augh", "aux")
text = rsub(text, "([aou])(" .. accent_c .. "?)gh", "%1%2x")
text = rsub(text, "eigh", "eiç")
text = rsub(text, "eugh", "euç")
text = rsub(text, "äigh", "eiç")
text = rsub(text, "([äëei])(" .. accent_c .. "?)gh", "%1%2ç")
--c, ng, qu
text = rsub(text, "tsch", "čč")
text = rsub(text, "sch", "ʃʃ")
text = rsub(text, "#ch", "ʃ")
text = rsub(text, "ck", "kk")
text = rsub(text, "c([ei])", "s%1")
text = rsub(text, "c", "k")
text = rsub(text, "ng", "ŋŋ")
text = rsub(text, "qu", "kW") --fake symbol used to distinguish this sound from the usual /v/ written as <w>
--diphthongs/n-phthongs
text = rsub(text, "ieh", "í" .. AC)
text = rsub(text, "ih", "í" .. AC)
text = rsub(text, "eu", "ɔ" .. AC .. "ɪ̯")
text = rsub(text, "ai", "a" .. AC .. "ɪ̯")
text = rsub(text, "ei", "a" .. AC .. "ɪ̯")
text = rsub(text, "aai", "ő" .. AC .. "ɪ̯")
text = rsub(text, "au", "a" .. AC .. "ʊ̯")
text = rsub(text, "oi", "o" .. AC .. "ɪ̯")
text = rsub(text, "ĭ", "ɪ̯")
text = rsub(text, "ŭ", "ʊ̯")
-- double vowels (we will handle them with a single character)
text = rsub(text, "aa", "ő" .. AC)
text = rsub(text, "ää", "à" .. AC)
text = rsub(text, "äh", "à" .. AC)
text = rsub(text, "a" .. GR, "ä" .. AC) --note that <à> is used internally for long ä=/ɛː/
text = rsub(text, "e", "ə")
text = rsub(text, "əh", "é" .. AC)
text = rsub(text, "əər#", "é" .. AC .. "a#")
text = rsub(text, "əə", "é" .. AC)
text = rsub(text, "ëə", "é" .. AC .. "ə")
text = rsub(text, "ër", "ɚ")
text = rsub(text, "ë", "e")
text = rsub(text, "iə", "í" .. AC)
text = rsub(text, "ïə", "í" .. AC .. "ə")
text = rsub(text, "ə" .. AC, "e" .. AC)
text = rsub(text, "ii", "í" .. AC) --not really used, probably unnecesary
text = rsub(text, "oo", "ó" .. AC)
text = rsub(text, "oh", "ó" .. AC)
text = rsub(text, "uu", "ú" .. AC)
text = rsub(text, "uh", "ú" .. AC)
text = rsub(text, "ï", "i")
--consonants
text = rsub(text, "w", "vv")
text = rsub(text, "W", "w")
text = rsub(text, "tz", "zz")
text = rsub(text, "khə#", "çə#") --the suffix -che (note that c has been changed to k in a previous step)
text = rsub(text, "s([ptk])", "ʃ%1") --palatalize s before (orthographic) fortis stops (this doesn't affect borrowings with <x>)
text = rsub(text, "X", "ks")
--<ptk> represent plain consonants word-initially only when followed by a vowel
text = rsub(text, "#p(" .. V .. ")", "#P%1")
text = rsub(text, "#t(" .. V .. ")", "#T%1")
text = rsub(text, "#k(" .. V .. ")", "#K%1")
fortis_lenis_conv = {["P"] = "p", ["T"] = "t", ["K"] = "k", ["p"] = "b", ["t"] = "d", ["k"] = "g"}
text = rsub(text, "[PTKptk]", fortis_lenis_conv)
--word finally, <v> devoices to /f/
text = rsub(text, "v#", "f#")
--Sequence äer
text = rsub(text, "äər", "à" .. AC .. "a")
-- accent all words (if no accent is found, it is added after the first vowel)
local words = rsplit(text, " ")
for j, word in ipairs(words) do
if not rfind(word, accent_c) then
words[j] = rsub(word, "^(.-" .. V .. ")", "%1" .. AC)
end
end
text = table.concat(words, " ")
--Ending -er when pronounced /-ˈɛa/ and -for when pronounced /ˈfɔa/
text = rsub(text, "[eə]" .. AC .. "r#", "ɛ" .. AC .. "a#")
text = rsub(text, "#fo" .. AC .. "r#", "#fɔ" .. AC .. "a#")
--Orthographically short and phonologically long vowels, convert stressed /ə/ into /e/
local short_long_v = {{"a", "á"}, {"ä", "à"}, {"e", "é"}, {"ə", "é"}, {"i", "í"}, {"o", "ó"}, {"u", "ú"}}
for _, pair in ipairs(short_long_v) do
text = rsub(text, pair[1] .. AC .. "(" .. C .. "?)" .. "#", pair[2] .. AC .. "%1#")
text = rsub(text, pair[1] .. AC .. "(" .. C .. "?)" .. "(" .. V .. ")", pair[2] .. AC .. "%1%2")
end
text = rsub(text, "ə" .. AC, "e" .. AC)
--Rhotics
text = rsub(text, "o" .. AC .. "ər#", "ó" .. AC .. "a#")
text = rsub(text, "([eə])" .. AC .. "([mnŋr])" .. "(" .. C .. ")", "ä" .. AC .. "%2%3")
text = rsub(text, "o" .. AC .. "([mnŋr])" .. "(" .. C .. ")", "ɔ" .. AC .. "%1%2")
text = rsub(text, "ər#", "a#")
text = rsub(text, "ő" .. AC .. "r#", "ő" .. AC)
text = rsub(text, "rr#", "r#")
text = rsub(text, "r([dlnsʃčtz])", "%1%1")
text = rsub(text, "ɚ", "er")
--n -> ŋ before velar stops
text = rsub(text, "n(%.*[gk])", "ŋ%1")
--syllable division
text = syllabify_from_spelling_or_pronun(text, false)
--stress
local accent_to_stress_mark = { [AC] = "ˈ", [GR] = "ˌ", [CFLEX] = "" }
local words = rsplit(text, " ")
for j, word in ipairs(words) do
local syllables = rsplit(word, "%.")
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
words[j] = table.concat(syllables, ".")
-- suppress syllable mark before IPA stress indicator
words[j] = rsub(words[j], "%.(" .. ipa_stress_c .. ")", "%1")
--make all primary stresses but the last one be secondary
words[j] = rsub_repeatedly(words[j], "ˈ(.+)ˈ", "ˈ%1ˌ")
end
text = table.concat(words, " ")
--final conversions
final_conversions={
["g"] = "k", ["k"] = "kʰ",
["b"] = "p", ["p"] = "pʰ",
["d"] = "t", ["t"] = "tʰ",
["ő"] = "ɔː", ["á"] = "aː", ["ä"] = "ɛ", ["à"] = "ɛː", ["é"] = "eː", ["í"] = "iː", ["ó"]= "oː", ["ú"] = "uː"}
text = rsub(text, "[gkbpdtőáäàéíóú]",final_conversions)
text = rsub(text, "#", "")
text = rsub_repeatedly(text, "(" .. C .. ")%1", "%1") --remove double consonants
text = rsub_repeatedly(text, "(" .. C .. ")([%.ˈˌ])%1", "%2%1") --remove double consonants
text = rsub(text, "z", "t͡s")
text = rsub(text, "č", "t͡ʃ")
text = rsub(text, "ˌ" .. u(0x032F), u(0x032F) .. "ˌ")
text = unfc(text)
return {text = text}
end
local function express_all_styles(dodialect)
local ret = {
pronun = {},
expressed_styles = {},
}
dodialect(ret)
table.insert(ret.expressed_styles, {
styles = {{pronun = ret.pronun,}},
})
return ret
end
local function format_all_styles(expressed_styles, format_style)
for i, style_group in ipairs(expressed_styles) do
style_group.formatted, style_group.formatted_len =
format_style(style_group.styles[1].tag, style_group.styles[1], i == 1)
end
local lines = {}
for i, style_group in ipairs(expressed_styles) do
table.insert(lines, style_group.formatted)
end
return table.concat(lines, "\n")
end
local function dodialect_pronun(args, ret)
ret.pronun = {}
for i, term in ipairs(args.terms) do
local phonemic
if term.raw then
phonemic = term.raw_phonemic
else
phonemic = export.IPA(term.term)
phonemic = phonemic.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[i] = {
raw = term.raw,
phonemic = phonemic,
refs = refs,
q = term.q,
qq = term.qq,
a = term.a,
aa = term.aa,
}
end
end
local function generate_pronun(args)
local function this_dodialect_pronun(ret)
dodialect_pronun(args, ret)
end
local ret = express_all_styles(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 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 then
error("Internal error: Saw neither phonemic nor phonetic pronunciation")
end
if pronun.phonemic then
local slash_pron = "/" .. pronun.phonemic:gsub("%.", "") .. "/"
table.insert(pronunciations, {
pron = slash_pron,
})
ins(slash_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)
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_respelling:match("^/(.*)/ %[(.*)%]$")
if not raw_phonemic then
raw_phonemic = raw_respelling:match("^/(.*)/$")
end
if not raw_phonemic 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,
}
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("t͡s", "ts")
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, true), "")
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 == "." or ci == "#" then
table.insert(result, ci)
i = i + 1
elseif ci == AC or ci == CFLEX 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
local function generate_hyph_obj(term)
return {syllabification = term, hyph = split_syllabified_spelling(term)}
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
not words[1]:find("%-$") and -- no if word is a prefix
not (words[1]:find("^%-") and words[1]:find(CFLEX)) 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, hyphs, parsed_respellings, rhyme_ret)
rhyme_ret.pronun = {}
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 hyphenation.
if not num_syl then
num_syl = {}
for _, hyph in ipairs(hyphs) do
if should_generate_rhyme_from_respelling(hyph.syllabification) then
local this_num_syl = 1 + ulen(rsub(hyph.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 dialect, pronun in pairs(parsed.pronun.pronun) 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, {
rhyme = rhyme.rhyme,
num_syl = num_syl,
qualifiers = rhyme.qualifiers,
})
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_hyph(arg, parse_err)
-- None other than qualifiers
local param_mods = {}
return parse_pron_modifier(arg, parse_err, generate_hyph_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 <t:...> 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 <g:...> 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 {{fax-pr}}.
function export.show_pr(frame)
local params = {
[1] = {list = true},
["rhyme"] = {},
["hyph"] = {},
["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_hyph = args.hyph and
parse_hyph(args.hyph, function(msg) overall_parse_err(msg, "hyph", args.hyph) 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 },
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,
},
hyph = {
overall = true,
store = "insert-flattened",
convert = parse_hyph,
},
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 = {}, hyph = {}, 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 = {},
hyph = {},
hmp = {},
bullets = 1,
})
end
end
if overall_hyph then
local hyphs = {}
for _, hyph in ipairs(overall_hyph) do
if hyph.syllabification == "+" then
hyph.syllabification = syllabify_from_spelling(pagename)
hyph.hyph = split_syllabified_spelling(hyph.syllabification)
elseif hyph.syllabification == "-" then
overall_hyph = {}
break
end
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) then
no_auto_rhyme = true
break
end
elseif not should_generate_rhyme_from_respelling(term.term) then
no_auto_rhyme = true
break
end
end
if #parsed.hyph == 0 then
if not overall_hyph and all_words_have_vowels(pagename) then
for _, term in ipairs(parsed.terms) do
if not term.raw then
local syllabification = syllabify_from_spelling(term.term)
local aligned_syll = align_syllabification_to_spelling(syllabification, pagename)
if aligned_syll then
m_table.insertIfNot(parsed.hyph, generate_hyph_obj(aligned_syll))
end
end
end
end
else
for _, hyph in ipairs(parsed.hyph) do
if hyph.syllabification == "+" then
hyph.syllabification = syllabify_from_spelling(pagename)
hyph.hyph = split_syllabified_spelling(hyph.syllabification)
elseif hyph.syllabification == "-" then
parsed.hyph = {}
break
end
end
end
-- Generate the rhymes.
local function dodialect_rhymes_from_pronun(rhyme_ret)
rhyme_ret.pronun = {}
for _, pronun in ipairs(parsed.pronun.pronun) 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) 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
table.insert(rhyme_ret.pronun, {
rhyme = rhyme,
num_syl = {num_syl},
})
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(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.hyph, {parsed}, rhyme_ret)
end
parsed.rhyme = express_all_styles(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 all_hyphs
if overall_hyph then
all_hyphs = overall_hyph
else
all_hyphs = {}
for _, parsed in ipairs(parsed_respellings) do
for _, hyph in ipairs(parsed.hyph) do
m_table.insertIfNot(all_hyphs, hyph)
end
end
end
local function dodialect_overall_rhyme(rhyme_ret, dialect)
return dodialect_specified_rhymes(overall_rhyme, all_hyphs, parsed_respellings, rhyme_ret)
end
overall_rhyme = express_all_styles(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)
end
-- If all sets of pronunciations have the same hyphenations, display them only once at the bottom.
-- Otherwise, display hyphenations beneath each set, indented.
local first_hyphs
local all_hyph_sets_eq = true
for j, parsed in ipairs(parsed_respellings) do
if j == 1 then
first_hyphs = parsed.hyph
elseif not m_table.deepEquals(first_hyphs, parsed.hyph) then
all_hyph_sets_eq = false
break
end
end
local function format_hyphenations(hyphs, num_bullets)
local hyphtext = require("Module:hyphenation").format_hyphenations { lang = lang, hyphs = hyphs, caption = "Syllabification" }
return string.rep("*", num_bullets) .. " " .. hyphtext
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_hyph_sets_eq and #parsed.hyph > 0 then
table.insert(textparts, "\n")
table.insert(textparts, format_hyphenations(parsed.hyph, 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_hyph_sets_eq and #first_hyphs > 0 then
table.insert(textparts, "\n")
table.insert(textparts, format_hyphenations(first_hyphs, min_num_bullets))
end
if overall_hyph and #overall_hyph > 0 then
table.insert(textparts, "\n")
table.insert(textparts, format_hyphenations(overall_hyph, 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