Module:Deranize: Difference between revisions

Module:Deranize (view source)

Revision as of 02:20, 25 March 2024

711 bytes added , 25 March

m

clean :s

Laqme

Bureaucrats, Interface administrators, Suppressors, Administrators

568

edits

@@ Line 1: / Line 1: @@
--- neither mw.ustring.toNFD nor mw.ustring.match doesn't work properly, so resorting to this for the time being (lol)
+-- Converts Latin text to Derani.
-local detone = {
+-- Supports: syllable splitting, tone, prefixes, hiatus/diphthong marks, punctuation.
-  ["a"] = {"a", 1},
+-- For cartouches and null-raı, try: deranize("Ruaq sá [poq] sá* da.")
-  ["u"] = {"u", 1},
-  ["ı"] = {"ı", 1},
+u = mw.ustring
-  ["o"] = {"o", 1},
-  ["e"] = {"e", 1},
+function get_tone(word)
-  ["á"] = {"a", 2},
+   local nfd = u.toNFD(word)
-   ["ú"] = {"u", 2},
+   if u.find(nfd, "́") then return 2 end
-   ["í"] = {"ı", 2},
+   if u.find(nfd, "̈") then return 3 end
-  ["ó"] = {"o", 2},
+   if u.find(nfd, "̂") then return 4 end
-   ["é"] = {"e", 2},
+   return 1
-  ["ä"] = {"a", 3},
+end
-   ["ü"] = {"u", 3},
-  ["ï"] = {"ı", 3},
+function strip_tone(word)
-   ["ö"] = {"o", 3},
+   return u.gsub(u.gsub(u.toNFD(word), "[́̈̂]", ""), "i", "ı")
-  ["ë"] = {"e", 3},
+end
-  ["â"] = {"a", 4},
-   ["û"] = {"u", 4},
-  ["î"] = {"ı", 4},
-  ["ô"] = {"o", 4},
-  ["ê"] = {"e", 4},
-}
 local deranimap = {
@@ Line 55: / Line 49: @@
    ["e"]  = "󱚴", -- DERANI LETTER FOFUAQ
    -- tone marks
+  [1]    = "",
    [2]    = "󱛊", -- DERANI COMBINING RISING TONE
    [3]    = "󱛋", -- DERANI COMBINING LOW GLOTTAL TONE
    [4]    = "󱛌", -- DERANI COMBINING RISING-FALLING TONE
-   -- TODO: the rest of the owl
+   -- prefix
+  ["-"] = "󱛒",
+  ["hiatus"] = "󱛍",
+  ["diphthong"] = "󱛎",
+  [""] = "",
 }
-function deranize(frame)
+local derani_punctuation = {
-   assert(frame.args[1] ~= nil and frame.args[2] == nil, "This function requires exactly one argument")
+   ["\""] = "󱛓",
-   local normalized = mw.ustring.gsub(mw.ustring.lower(frame.args[1]), "i", "ı")
+  [":"] = "󱛓",
-   mw.log(normalized)
+  [","] = " 󱛔",
+  ["."] = " 󱛕",
+  ["!"] = " 󱛖",
+   ["?"] = " 󱛗",
+  ["["] = "󱛘",
+  ["]"] = "󱛙",
+  ["*"] = " 󱛚",
+   ["_"] = "󱛛",
+}
+function deranize_word(word)
    local res = {}
    local ix = 1
-   local len = mw.ustring.len(normalized)
+  local tone = get_tone(word)
+   local len = mw.ustring.len(word)
    -- NB. this is not PCRE regex, see https://www.mediawiki.org/wiki/Extension:Scribunto/Lua_reference_manual#Ustring_patterns
-  --              ↓ initpos (these ones eval to an index)                     precoda ↓        ↓ postcoda
+   local toaqre = "()([mbpfndtzcsrljꝡgk']?h?)([auıoe])([̣]?)([́̈̂̀]?)([auıoe]?[auıoe]?)()([qm]?)([-·]?)()([auıoe]?)"
-  --                ↓ initial               ↓ medial                ↓ final             ↓ coda   ↓ lingeringvowel
+  local is_first_syllable = true
-   local toaqre = "()([mbpfndtzcsrljꝡgk']?h?)([auıoeáúíóéäüïöëâûîôê])([auıoe]?[auıoe]?)()([qm]?)()([auıoe]?)"
+   while ix <= len do
-   while ix < len do
      local shouldbreak = false
-     local initpos, initial, medial, final, precoda, coda, postcoda, lingeringvowel = mw.ustring.match(normalized, toaqre, ix)
+     local pos_init, initial, medial, underdot, diacritic, final, pos_precoda, coda, hyphen, pos_postcoda, lingering_vowel = mw.ustring.match(word, toaqre, ix)
-     if initpos == nil then
+     if pos_init == nil then
-       shouldbreak = true
+       break
-      initpos = nil
      end
-    mw.log(table.concat({ix, initpos, initial, medial, final, precoda, coda, postcoda, lingeringvowel}, ", "))
+     res[#res+1] = u.toNFC(u.sub(word, ix, pos_init - 1))
-     res[#res+1] = mw.ustring.toNFC(mw.ustring.sub(normalized, ix, initpos - 1))
+     if coda == "m" and lingering_vowel ~= "" then
-    if shouldbreak then break end
-     if coda == "m" and lingeringvowel ~= "" then
        coda = ""
-       ix = precoda
+       ix = pos_precoda
      else
-       ix = postcoda
+       ix = pos_postcoda
+    end
+    local glyphs = {initial, tone, medial}
+    if initial == "" and is_first_syllable then glyphs[1] = "'" end
+    if not is_first_syllable then glyphs[2] = "" end
+    if glyphs[1] == "'" or glyphs[1] == "ꝡ" then
+      -- Move tone onto the first vowel
+      glyphs[2], glyphs[3] = glyphs[3], glyphs[2]
+    end
+    local v = medial..final
+    if v == "aı" or v == "ao" or v == "eı" or v == "oı" then
+      glyphs[#glyphs+1] = "diphthong"
+    elseif final ~= "" then
+      glyphs[#glyphs+1] = "hiatus"
+    end
+    for _, fin in ipairs(mw.text.split(final, "")) do
+      glyphs[#glyphs+1] = fin
      end
-    local nucleus, tone = unpack(detone[medial])
-    local glyphs = {initial, nucleus, tone}
      if coda ~= "" then
        glyphs[#glyphs+1] = coda .. "_"
      end
-     for _, fin in ipairs(mw.text.split(final, "")) do
+     if underdot ~= "" or hyphen ~= "" then
-       glyphs[#glyphs+1] = fin
+       glyphs[#glyphs+1] = "-"
      end
      for _, glyph in ipairs(glyphs) do
-       local mapped = deranimap[glyph]
+       res[#res+1] = deranimap[glyph]
-      if mapped then
-      	res[#res+1] = mapped
-      end
      end
-     -- TODO: actually implement derani hiatus symbol, o'aomo, etc. insertion logic here 🤪
+     is_first_syllable = false
    end
+  res[#res+1] = u.toNFC(u.sub(word, ix, len))
    return table.concat(res)
 end
-return {deranize = deranize}
+function deranize(frame)
+  assert(frame.args[1] ~= nil and frame.args[2] == nil, "This function requires exactly one argument")
+  local text = u.gsub(u.lower(u.toNFD(frame.args[1])), "i", "ı")
+  local converted = u.gsub(text, "(%S+)", deranize_word)
+  for k, v in pairs(derani_punctuation) do
+    converted = u.gsub(converted, "%" .. k, v)
+  end
+  return converted
+end
+function clean(frame)
+  assert(frame.args[1] ~= nil and frame.args[2] == nil, "This function requires exactly one argument")
+  local cleaned = u.gsub(frame.args[1], "[:%[%]*]", "")
+  return cleaned
+end
+return {deranize = deranize, clean = clean}