Module:an-pron
- The following documentation is located at Module:an-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:an-pr, see there for usage.
--Adapted from [[Module:es-pronunc]] by [[User:Benwing2]]
--TODO
--This code is adapted from [[Module:pronunc]], which can output multiple dialectal pronounciations from a single spelling. This module will
--return a single pronunciation per spelling in all cases, so there are parts of the code which can be simplified.
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("an")
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 TILDE = u(0x0303) -- tilde = ̃
local DIA = u(0x0308) -- diaeresis = ̈
local SYLDIV = u(0xFFF0) -- used to represent a user-specific syllable divider (.) so we won't change it
local vowel = "aeiouüyAEIOUÜY" -- vowel; include y so we get single-word y correct and for syllabifying from spelling
local V = "[" .. vowel .. "]" -- vowel class
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 including h
local C_NOT_H = "[^" .. vowel .. separator .. "h]" -- consonant class not including h
local C_OR_WORDSEP = "[^" .. vowel .. separator_not_wordsep .. "]" -- consonant class including h, or word separator
local T = "[^" .. vowel .. "lrɾjw" .. separator .. "]" -- obstruent or nasal
local unstressed_words = m_table.listToSet({ -- TODO add more
"o", "ro",-- definite articles
"y", --conjunction
})
-- version of rsubn() that discards all but the first return value
local function rsub(term, foo, bar)
local retval = rsubn(term, foo, bar)
return retval
end
-- version of rsubn() that returns a 2nd argument boolean indicating whether
-- a substitution was made.
local function rsubb(term, foo, bar)
local retval, nsubs = rsubn(term, foo, bar)
return retval, nsubs > 0
end
-- apply rsub() repeatedly until no change
local function rsub_repeatedly(term, foo, bar)
while true do
local new_term = rsub(term, foo, bar)
if new_term == term then
return term
end
term = new_term
end
end
local function decompose(text)
-- decompose everything but ñ and ü
text = unfd(text)
text = rsub(text, ".[" .. TILDE .. DIA .. "]", {
["n" .. TILDE] = "ñ",
["N" .. TILDE] = "Ñ",
["u" .. DIA] = "ü",
["U" .. DIA] = "Ü",
})
return text
end
local function split_on_comma(term)
if term:find(",%s") then
return require(put_module).split_on_comma(term)
else
return rsplit(term, ",")
end
end
-- Remove any HTML from the formatted text and resolve links, since the extra characters don't contribute to the
-- displayed length.
local function convert_to_raw_text(text)
text = rsub(text, "<.->", "")
if text:find("%[%[") then
text = require("Module:links").remove_links(text)
end
return text
end
-- Return the approximate displayed length in characters.
local function textual_len(text)
return ulen(convert_to_raw_text(text))
end
-- Main syllable-division algorithm. Can be called either directly on spelling (when hyphenating) or after
-- non-trivial processing of respelling in the direction of pronunciation (when generating pronunciation).
local function syllabify_from_spelling_or_pronun(text, is_spelling)
-- Part 1: Divide before the last consonant in a cluster of consonants between vowels (but don't divide a VhV
-- sequence; [[prohibir]] should be prohi.bir). Then move the syllable division marker leftwards over clusters that
-- can form onsets.
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C_NOT_H .. V .. ")", "%1.%2")
text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*" .. C .. "+)(" .. C .. V .. ")", "%1.%2")
-- NOTE: When run on pronun, we have already eliminated c and v, but not when run on spelling.
-- When run on pronun, don't include r, which at this point represents the trill.
local cluster_r = is_spelling and "rɾ" or "ɾ"
-- Don't divide Cl or Cr where C is a stop or fricative, except for dl.
text = rsub(text, "([pbfvkctg])%.([l" .. cluster_r .. "])", ".%1%2")
text = text:gsub("d%.([" .. cluster_r .. "])", ".d%1")
-- Don't divide ch, sh, ph, th, dh, fh, kh or gh. Do allow bh to be divided ([[subhumano]], [[subhúmedo]], etc.).
text = rsub(text, "([csptdfkg])%.h", ".%1h")
-- Don't divide rr.
text = rsub(text, "r.r", ".rr")
-- Don't divide tz
text = rsub(text, "t%.z", ".tz")
-- Part 2: Divide hiatuses. Any aeo, or stressed iuüy, should be syllabically divided from a following aeo or
-- stressed iuüy. Also divide ii and uu sequences ([[antiincendios]], [[shiita]], [[vacuum]]). Note that words with
-- ii or uu next to a vowel (e.g. [[hawaiiano]]) will not make it to this point unchanged; the i or u adjacent to
-- a vowel (or the second one if both are adjacent to vowels) will get converted to a consonant symbol (temporarily
-- when syllabifying spelling).
text = rsub_repeatedly(text, "([aeoAEO]" .. accent_c .. "*)(h?[aeo])", "%1.%2")
text = rsub_repeatedly(text, "([aeoAEO]" .. accent_c .. "*)(h?" .. V .. stress_c .. ")", "%1.%2")
text = rsub(text, "([iuüyIUÜY]" .. stress_c .. ")(h?[aeo])", "%1.%2")
text = rsub_repeatedly(text, "([iuüyIUÜY]" .. stress_c .. ")(h?" .. V .. stress_c .. ")", "%1.%2")
text = rsub_repeatedly(text, "([iI]" .. accent_c .. "*)(h?i)", "%1.%2")
text = rsub_repeatedly(text, "([uU]" .. accent_c .. "*)(h?u)", "%1.%2")
return text
end
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.
local TEMP_I = u(0xFFF1)
local TEMP_U = u(0xFFF2)
local TEMP_Y_CONS = u(0xFFF3)
local TEMP_QU = u(0xFFF4)
local TEMP_QU_CAPS = u(0xFFF5)
local TEMP_GU = u(0xFFF6)
local TEMP_GU_CAPS = u(0xFFF7)
local TEMP_H = u(0xFFF8)
local TEMP_NY = u(0xFFF9)
local TEMP_NY_CAPS = u(0xFFFA)
local TEMP_LL = u(0xFFFB)
local TEMP_LL_CAPS = u(0xFFFC)
-- Change user-specified . into SYLDIV so we don't shuffle it around when dividing into syllables. TODO fix hyphenation of words like paral·lel
text = rsub(text, "ny", TEMP_NY)
text = rsub(text, "Ny", TEMP_NY_CAPS)
text = rsub(text, "ll", TEMP_LL)
text = rsub(text, "Ll", TEMP_LL_CAPS)
text = text:gsub("%.", SYLDIV)
text = rsub(text, "y(" .. V .. ")", TEMP_Y_CONS .. "%1")
-- We don't want to break -sh- except in desh-, e.g. [[deshuesar]], [[deshonra]], [[deshecho]]. Normally, -sh- is
-- automatically preserved, so we replace the h with a temporary symbol to avoid this.
text = text:gsub("^([Dd]es)h", "%1" .. TEMP_H)
text = text:gsub("([ %-][Dd]es)h", "%1" .. TEMP_H)
-- qu mostly handled correctly automatically, but not in quietud
text = rsub(text, "qu(" .. V .. ")", TEMP_QU .. "%1")
text = rsub(text, "Qu(" .. V .. ")", TEMP_QU_CAPS .. "%1")
text = rsub(text, "gu(" .. V .. ")", TEMP_GU .. "%1")
text = rsub(text, "Gu(" .. V .. ")", TEMP_GU_CAPS .. "%1")
local vowel_to_glide = { ["i"] = TEMP_I, ["u"] = TEMP_U }
-- i and u between vowels -> consonant-like substitutions: [[paranoia]], [[baiano]], [[abreuense]], [[alauita]],
-- [[Malaui]], etc.; also with h, as in [[marihuana]], [[parihuela]], [[antihielo]], [[pelluhuano]], [[náhuatl]],
-- etc. When we do this we need to help the syllabification particularly of words with -hiV- and -huV- in them,
-- otherwise we get e.g. 'an.tih.ie.lo' because we converted the i following the h to a consonant. Add .* at the
-- beginning so we go right-to-left, in the case of [[hawaiiano]] -> ha.wai.iano.
text = rsub_repeatedly(text, "(.*" .. V .. accent_c .. "*)(h?)([iu])(" .. V .. ")",
function (v1, h, iu, v2) return v1 .. "." .. h .. vowel_to_glide[iu] .. v2 end
)
text = syllabify_from_spelling_or_pronun(text, "is spelling")
text = text:gsub(SYLDIV, ".")
text = text:gsub(TEMP_I, "i")
text = text:gsub(TEMP_U, "u")
text = text:gsub(TEMP_Y_CONS, "y")
text = text:gsub(TEMP_QU, "qu")
text = text:gsub(TEMP_QU_CAPS, "Qu")
text = text:gsub(TEMP_GU, "gu")
text = text:gsub(TEMP_GU_CAPS, "Gu")
text = text:gsub(TEMP_H, "h")
text = text:gsub(TEMP_NY, "ny")
text = text:gsub(TEMP_NY_CAPS, "Ny")
text = text:gsub(TEMP_LL, "ll")
text = text:gsub(TEMP_LL_CAPS, "Ll")
text = unfc(text)
return text
end
-- Generate the IPA of a given respelling
function export.IPA(text)
-- start at FFF1 because FFF0 is used for SYLDIV
local TEMP_Y = u(0xFFF1)
local TEMP_W = u(0xFFF2)
text = ulower(text or mw.title.getCurrentTitle().text)
-- decompose everything but ñ and ü
text = decompose(text)
-- convert commas and en/en dashes to IPA foot boundaries
text = rsub(text, "%s*[,–—]%s*", " | ")
-- question mark or exclamation point in the middle of a sentence -> IPA foot boundary
text = rsub(text, "([^%s])%s*[¡!¿?]%s*([^%s])", "%1 | %2")
-- canonicalize multiple spaces and remove leading and trailing spaces
local function canon_spaces(text)
text = rsub(text, "%s+", " ")
text = rsub(text, "^ ", "")
text = rsub(text, " $", "")
return text
end
text = canon_spaces(text)
-- Make prefixes unstressed unless they have an explicit stress marker; also make certain
-- monosyllabic words (e.g. [[el]], [[la]], [[de]], [[en]], etc.) without stress marks be
-- unstressed.
local words = rsplit(text, " ")
for i, word in ipairs(words) do
if rfind(word, "%-$") and not rfind(word, accent_c) or unstressed_words[word] then
-- add CFLEX to the last vowel not the first one, or we will mess up 'que' by
-- adding the CFLEX after the 'u'
words[i] = rsub(word, "^(.*" .. V .. ")", "%1" .. CFLEX)
end
end
text = table.concat(words, " ")
-- Convert hyphens to spaces, to handle [[Austria-Hungría]], [[franco-italiano]], etc; with the exception of verbal clitics (with the temporary symbol @)
text = rsub(text, "-se", "@se")
text = rsub(text, "-te", "@te")
text = rsub(text, "-me", "@me")
text = rsub(text, "-tos", "@tos")
text = rsub(text, "-vos", "@vos")
text = rsub(text, "-nos", "@nos")
text = rsub(text, "-mos", "@mos")
text = rsub(text, "-lo", "@lo")
text = rsub(text, "-el", "@el")
text = rsub(text, "-los", "@los")
text = rsub(text, "-es", "@es")
text = rsub(text, "-la", "@la")
text = rsub(text, "-las", "@las")
text = rsub(text, "-le", "@le")
text = rsub(text, "-li", "@li")
text = rsub(text, "-les", "@les")
text = rsub(text, "-lis", "@lis")
text = rsub(text, "-en", "@en")
text = rsub(text, "-ne", "@ne")
text = rsub(text, "-i", "@i")
text = rsub(text, "-bi", "@bi")
text = rsub(text, "%-", " ")
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, " ", "# #") .. "##"
--glides
text = rsub(text, "#y" .. CFLEX, "#i" .. CFLEX) -- conjunction
text = rsub(text, "ny", "ɲ")
text = rsub(text, "y", "ʝ")
text = rsub(text, "hi([aeou])", "ʝ%1")
-- handle certain combinations; sh handling needs to go before x handling to avoid issues with [[exhausto]]
text = rsub(text, "ch", "ĉ") --not the real sound
text = rsub(text, "ll", "ʎ")
text = rsub(text, "#ps", "#s") -- [[psicolochía]]
text = rsub(text, "x", "ʃ")
--c, g, q
text = rsub(text, "c([ie])", "θ%1")
text = rsub(text, "z", "θ")
text = rsub(text, "g([ie])", "x%1")
text = rsub(text, "gu([ie])", "g%1")
text = rsub(text, "gü([ie])", "gu%1")
text = rsub(text, "qü([ie])", "ku%1")
text = rsub(text, "ng([^aeiouüwhlr])", "n%1") -- TODO is this really needed?
text = rsub(text, "qu([ie])", "k%1")
text = rsub(text, "q", "k") -- [[quan]]
text = rsub(text, "([aeo])iʃ", "%1Iʃ")
-- map various consonants to their phoneme equivalent
text = rsub(text, "[cjñrv]", {["c"]="k", ["j"]="x", ["ñ"]="ɲ", ["r"]="ɾ", ["v"]="b"})
--pronunciation of <r>/<rr>
text = rsub(text, "ɾɾ", "r")
text = rsub(text, "([#lnsθ])ɾ", "%1r")
text = rsub(text, "ɾ#", "R#")
text = rsub(text, "ɾ%-", "R-")
-- voiceless stop to voiced before obstruent or nasal (see Basic Grammar of Aragonese)
local voice_stop = { ["p"] = "b", ["t"] = "d", ["k"] = "g" }
text = rsub(text, "([ptk])(" .. separator_c .. "*" .. T .. ")",
function(stop, after) return voice_stop[stop] .. after end)
text = rsub(text, "!", "t")
text = rsub(text, "n([# .]*[bpm])", "m%1")
-- remove silent h and · before syllable division
text = rsub(text, "h", "")
text = rsub(text, "·", ".")
-- convert i/u between vowels to glide
local vowel_to_glide = { ["i"] = "j", ["u"] = "w" }
text = rsub_repeatedly(text, "(.*" .. V .. accent_c .. "*h?)([iu])(" .. V .. ")",
function (v1, iu, v2) return v1 .. vowel_to_glide[iu] .. v2 end
)
--syllable division
text = syllabify_from_spelling_or_pronun(text, false)
--diphthongs; do not include TEMP_Y here
text = rsub(text, "i([aeou])", "j%1")
text = rsub(text, "u([aeio])", "w%1")
local accent_to_stress_mark = { [AC] = "ˈ", [GR] = "ˌ", [CFLEX] = "" }
local function accent_word(word, syllables)
-- Now stress the word. If any accent exists in the word (including ^ indicating an unaccented word),
-- put the stress mark(s) at the beginning of the indicated syllable(s). Otherwise, apply the default
-- stress rule.
if rfind(word, accent_c) then
for i = 1, #syllables do
syllables[i] = rsub(syllables[i], "^(.*)(" .. accent_c .. ")(.*)$",
function(pre, accent, post) return accent_to_stress_mark[accent] .. pre .. post end
)
end
else
-- Default stress rule. Words without vowels (e.g. IPA foot boundaries) don't get stress.
if #syllables > 1 and (rfind(word, "[^" .. vowel .. "ns#]#") or rfind(word, C .. "[ns]#")) or #syllables == 1 and rfind(word, V) then
syllables[#syllables] = "ˈ" .. syllables[#syllables]
elseif #syllables > 1 and rfind(word, "[aeo][iu][ns#]") then
syllables[#syllables] = "ˈ" .. syllables[#syllables]
elseif #syllables > 1 then
syllables[#syllables - 1] = "ˈ" .. syllables[#syllables - 1]
end
end
end
local words = rsplit(text, " ")
for j, word in ipairs(words) do
-- accentuation
local syllables = rsplit(word, "%.")
if rfind(word, "ment#") then
local mente_syllables
-- Words ends in -ment (converted above to mént); add a stress to the preceding portion
mente_syllables = {}
mente_syllables[2] = table.remove(syllables)
mente_syllables[1] = table.remove(syllables)
accent_word(table.concat(syllables, "."), syllables)
accent_word(table.concat(mente_syllables, "."), mente_syllables)
table.insert(syllables, mente_syllables[1])
table.insert(syllables, mente_syllables[2])
else
accent_word(word, syllables)
end
-- Reconstruct the word.
words[j] = table.concat(syllables, ".")
end
text = table.concat(words, " ")
text = rsub(text, "%-", "")
-- 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")
--other conversions, silent letters
text = rsub(text, "R", "(ɾ)")
text = rsub(text, "tθ#", "θ#")
text = rsub(text, "([nsɾ])t#", "%1#")
text = rsub(text, "nd#", "n#")
text = rsub(text, "ɾs#", "s#")
--semivowel symbols
text = rsub(text, "([aeou][i])", "%1̯")
text = rsub(text, "([aeio][u])", "%1̯")
--value of conjunction <y>
text = rsub(text, "([aeiou][# .]*)#i#([ˈ# .]*[aeiou])", "%1#ʝ#‿%2")
text = rsub(text, "i̯([# .]*)#i#([ˈ# .]*[aeiou])", "i̯%1#ʝ#‿%2")
text = rsub(text, "u̯([# .]*)#i#([ˈ# .]*[aeiou])", "u̯%1#ʝ#‿%2")
text = rsub(text, "#i#([ˈ# .]*[aeiou])", "#j#‿%1")
text = rsub(text, "([aeio][# .]*)#i#", "%1‿#i̯#")
-- convert fake symbols to real ones
local final_conversions = {
["ĉ"] = "t͡ʃ", -- fake "ch" to real "ch"
["g"] = "ɡ", -- U+0067 LATIN SMALL LETTER G → U+0261 LATIN SMALL LETTER SCRIPT G
["I"] = "(i̯)",
}
text = rsub(text, "[ĉgI]", final_conversions)
-- remove # symbols at word and text boundaries
text = rsub(text, "#", "")
text = rsub(text, "‿ ", "‿")
text = rsub(text, " ‿", "‿")
text = rsub(text, "([jʝ])‿ˈ", "ˈ%1‿")
text = rsub(text, "([jʝ])‿ˌ", "ˌ%1‿")
text = unfc(text)
local ret = {
text = text,
}
return ret
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)
-- NOTE: This works because the phonemic vowels are just [aeiou] possibly with diacritics that are separate
-- Unicode chars. If we want to handle things like ɛ or ɔ we need to add them to `vowel`.
return rsub(rsub(phonemic, ".*[ˌˈ]", ""), "^[^" .. vowel .. "]*", ""):gsub("%.", ""):gsub("t͡ʃ", "tʃ")
end
local function split_syllabified_spelling(spelling)
return rsplit(spelling, "%.")
end
-- "Align" syllabification to original spelling by matching character-by-character, allowing for extra syllable and
-- accent markers in the syllabification. If we encounter an extra syllable marker (.), we allow and keep it. If we
-- encounter an extra accent marker in the syllabification, we drop it. In any other case, we return nil indicating
-- the alignment failed.
local function align_syllabification_to_spelling(syllab, spelling)
local result = {}
local syll_chars = rsplit(decompose(syllab), "")
local spelling_chars = rsplit(decompose(spelling), "")
local i = 1
local j = 1
while i <= #syll_chars or j <= #spelling_chars do
local ci = syll_chars[i]
local cj = spelling_chars[j]
if ci == cj then
table.insert(result, ci)
i = i + 1
j = j + 1
elseif ci == "." then
table.insert(result, ci)
i = i + 1
elseif ci == AC or ci == GR or ci == 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 (e.g. [[Austria-Hungría]])
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 {{es-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