local export = {}
local lang = require("Module:languages").getByCode("zlw-ocs")
local m_links = require("Module:links")
local m_table = require("Module:table")
local m_string_utilities = require("Module:string utilities")
local u = mw.ustring.char
local rsplit = mw.text.split
local rfind = mw.ustring.find
local rmatch = mw.ustring.match
local rsubn = mw.ustring.gsub
local ulen = mw.ustring.len
local uupper = mw.ustring.upper
-- 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
export.TEMP_CH = u(0xFFF0) -- used to substitute ch temporarily in the default-reducible code
export.TEMP_SOFT_LABIAL = u(0xFFF1)
local lc_vowel = "aeiouyáéíóúýě"
local uc_vowel = uupper(lc_vowel)
export.vowel = lc_vowel .. uc_vowel
export.vowel_c = "[" .. export.vowel .. "]"
export.non_vowel_c = "[^" .. export.vowel .. "]"
-- Consonants that can never form a syllabic nucleus.
local lc_non_syllabic_cons = "bcdfghjkmnpqstvwxzčňšžďťľṕḿźś" .. export.TEMP_CH .. "b́" .. "v́" .. "f́"
local uc_non_syllabic_cons = uupper(lc_non_syllabic_cons)
export.non_syllabic_cons = lc_non_syllabic_cons .. uc_non_syllabic_cons
export.non_syllabic_cons_c = "[" .. export.non_syllabic_cons .. "]"
local lc_syllabic_cons = "lrř"
local uc_syllabic_cons = uupper(lc_syllabic_cons)
local lc_cons = lc_non_syllabic_cons .. lc_syllabic_cons
local uc_cons = uupper(lc_cons)
export.cons = lc_cons .. uc_cons
export.cons_c = "[" .. export.cons .. "]"
export.lowercase = lc_vowel .. lc_cons
export.lowercase_c = "[" .. export.lowercase .. "]"
export.uppercase = uc_vowel .. uc_cons
export.uppercase_c = "[" .. export.uppercase .. "]"
local lc_velar = "kgh"
local uc_velar = uupper(lc_velar)
export.velar = lc_velar .. uc_velar
export.velar_c = "[" .. export.velar .. "]"
local lc_plain_labial = ""
local lc_labial = lc_plain_labial .. export.TEMP_SOFT_LABIAL
local uc_plain_labial = uupper(lc_plain_labial)
local uc_labial = uupper(lc_labial)
export.plain_labial = lc_plain_labial .. uc_plain_labial
export.labial = lc_labial .. uc_labial
export.labial_c = "[" .. export.labial .. "]"
local lc_paired_palatal = "ňďťľṕḿźś"
local uc_paired_palatal = uupper(lc_paired_palatal)
export.paired_palatal = lc_paired_palatal .. uc_paired_palatal
local lc_paired_plain = "ndtlbpmvfzs"
local uc_paired_plain = uupper(lc_paired_plain)
export.paired_plain = lc_paired_plain .. uc_paired_plain
export.paired_palatal_to_plain = {
["b́"] = "b",
["B́"] = "B",
["ṕ"] = "p",
["Ṕ"] = "P",
["ḿ"] = "m",
["Ḿ"] = "M",
["v́"] = "v",
["V́"] = "V",
["f́"] = "f",
["F́"] = "F",
["ź"] = "z",
["Ź"] = "Z",
["ś"] = "s",
["Ś"] = "S",
["ň"] = "n",
["Ň"] = "N",
["ť"] = "t",
["Ť"] = "T",
["ď"] = "d",
["Ď"] = "D",
["ľ"] = "l",
["Ľ"] = "L",
}
export.paired_plain_to_palatal = {}
for k, v in pairs(export.paired_palatal_to_plain) do
export.paired_plain_to_palatal[v] = k
end
export.paired_palatal_to_plain[export.TEMP_SOFT_LABIAL] = ""
for labial in mw.ustring.gmatch(export.plain_labial, ".") do
export.paired_plain_to_palatal[labial] = labial .. export.TEMP_SOFT_LABIAL
end
local lc_followable_by_e_hacek = lc_paired_plain
local uc_followable_by_e_hacek = uc_paired_plain
export.followable_by_e_hacek = lc_followable_by_e_hacek .. uc_followable_by_e_hacek
export.followable_by_e_hacek_c = "[" .. export.followable_by_e_hacek .. "]"
local lc_inherently_soft = "žščřcjďťň"
local uc_inherently_soft = uupper(lc_inherently_soft)
export.inherently_soft = lc_inherently_soft .. uc_inherently_soft
export.inherently_soft_c = "[" .. export.inherently_soft .. "]"
function export.is_monosyllabic(word)
-- Treat ou as a single vowel.
word = word:gsub("ou", "ů")
word = word:gsub("ay$", "aj")
-- Convert all vowels to 'e'.
word = rsub(word, export.vowel_c, "e")
-- All consonants next to a vowel are non-syllabic; convert to 't'.
word = rsub(word, export.cons_c .. "e", "te")
word = rsub(word, "e" .. export.cons_c, "et")
-- Convert all remaining non-syllabic consonants to 't'.
word = rsub(word, export.non_syllabic_cons_c, "t")
-- At this point, what remains is 't', 'e', or a syllabic consonant. Count the latter two types.
word = word:gsub("t", "")
return ulen(word) <= 1
end
function export.apply_vowel_alternation(alt, stem)
local modstem, origvowel
if alt == "quant" then
-- [[kráva]] "cow", gen pl. [[krav]]
-- [[húba]] "mushroom", gen pl. [[hub]]
-- [[hól]] "cane", gen sg. [[holi]]
modstem = rsub(stem, "(.)([áéíóúý])(" .. export.cons_c .. "*)$",
function(pre, vowel, post)
origvowel = vowel
if vowel == "á" then
return pre .. "a" .. post
elseif vowel == "é" then
return pre .. "e" .. post
elseif vowel == "í" then
return pre .. "i" .. post
elseif vowel == "ó" then
return pre .. "o" .. post
elseif vowel == "ý" then
return pre .. "y" .. post
else
return pre .. "u" .. post
end
end
)
-- [[snieh]] "snow", gen sg. [[sněhu]]
modstem = rsub(modstem, "ie(" .. export.cons_c .. "*)$", "ě%1")
else
return stem, nil
end
return modstem, origvowel
end
local function make_try(word)
return function(from, to)
local stem = rmatch(word, "^(.*)" .. from .. "$")
if stem then
return stem .. to
end
return nil
end
end
function export.iotate(word)
local try = make_try(word)
return
try("st", "šč") or
try("št", "šč") or
try("zd", "žď") or
try("žd", "žď") or
try("sl", "šľ") or
try("zl", "žľ") or
try("stn", "šťň") or
try("zdn", "žďň") or
try("sn", "šň") or
try("zn", "žň") or
try("n", "ň") or
try("l", "ľ") or
try("t", "c") or
try("d", "z") or
try("s", "š") or
try("z", "ž") or
try("r", "ř") or
try("ch", "š") or
try("[kc]", "č") or
try("[hg]", "ž") or
try("p", "ṕ") or
try("m", "ḿ") or
try("z", "ź") or
try("s", "ś") or
try("b", "b́") or
try("v", "v́") or
try("f", "f́") or
word
end
function export.onlyndt(word)
local try = make_try(word)
return
try("n", "ň") or
try("d", "ď") or
try("t", "ť") or
word
end
function export.apply_first_palatalization(word, is_soft)
local try = make_try(word)
return
try("r", "ř") or
try("ch", "š") or
try("[hg]", "ž") or
try("sk", "šč") or
try("ck", "čšč") or
try("[kc]", "č") or
is_soft and try("z", "ž") or
word
end
function export.apply_second_palatalization(word, is_adjective)
local try = make_try(word)
return
try("ch", "š") or
try("[hg]", "z") or
try("r", "ř") or
is_adjective and try("sk", "šč") or
is_adjective and try("ck", "čšč") or
try("k", "c") or
word
end
function export.reduce(word)
local pre, letter, vowel, post = rmatch(word, "^(.*)([" .. export.cons .. "y%-])([e])(" .. export.cons_c .. "+)$")
if not pre then
return nil
end
return pre .. letter .. post
end
function export.dereduce(base, stem)
local pre, letter, post = rmatch(stem, "^(.*)(" .. export.cons_c .. ")(" .. export.cons_c .. ")$")
if not pre then
return nil
end
local epvowel
if rfind(letter, "[" .. export.paired_palatal .. "]") then
letter = export.paired_palatal_to_plain[letter]
epvowel = "ě"
else
epvowel = "e"
end
return pre .. letter .. epvowel .. post
end
function export.convert_paired_plain_to_palatal(stem, ending)
if ending and not rfind(ending, "^[ěií]") then
return stem
end
local stembegin, lastchar = rmatch(stem, "^(.*)([" .. export.followable_by_e_hacek .. "])$")
if lastchar then
return stembegin .. export.paired_plain_to_palatal[lastchar]
else
return stem
end
end
function export.convert_paired_palatal_to_plain(stem, ending)
-- For stems that alternate between n/t/d and ň/ť/ď, we always maintain the stem in the latter format and convert
-- to the corresponding plain as needed, with e -> ě.
if ending and not rfind(ending, "^[Eěií]") then
stem = stem:gsub(export.TEMP_SOFT_LABIAL, "")
return stem, ending
end
local stembegin, lastchar = rmatch(stem, "^(.*)([" .. export.paired_palatal .. "])$")
if lastchar then
ending = ending and rsub(ending, "^e", "ě") or nil
stem = stembegin .. export.paired_palatal_to_plain[lastchar]
end
-- 'E' has served its purpose of preventing the e -> ě conversion after a paired palatal (i.e. it depalatalizes
-- paired palatals).
ending = ending and rsub(ending, "^E", "e") or nil
return stem, ending
end
function export.combine_stem_ending(base, slot, stem, ending)
if stem == "?" then
return "?"
else
-- Convert ňe and ňě -> ně. Convert nE and ňE -> ne. Convert ňi and ni -> ni.
stem, ending = export.convert_paired_palatal_to_plain(stem, ending)
-- We specify endings with -e/ě using ě, but some consonants cannot be followed by ě; convert to plain e.
if base.all_uppercase then
stem = uupper(stem)
end
return stem .. ending
end
end
return export
