Module:he-translit/old

local export = {}
local U = require("Module:string/char")
local gsub = mw.ustring.gsub

--[[
-- Uncomment this to redefine gsub so that it prints to the Lua log
-- the names of the code points in the replacements it's making.
local function print_code_point_names(text)
	if not text then return "" end
	local names = require "Module:array"()
	for cp in mw.ustring.gcodepoint(text) do
		names:insert(require "Module:Unicode data".lookup_name(cp))
	end
	return names:concat ", "
end

local actual_gsub = mw.ustring.gsub
local gsub = function(...)
	local old, pattern, repl = ...
	local new, count = actual_gsub(...)
	if old ~= new then
		mw.log(table.concat({
			print_code_point_names(old),
			print_code_point_names(new),
			pattern,
			tostring(repl)
		}, "\n") .. "\n")
	end
	return new, count
end
--]]

local sheva = U(0x05B0)
local hataf_segol = U(0x05B1)
local hataf_patah = U(0x05B2)
local hataf_qamats = U(0x05B3)
local hiriq = U(0x05B4)
local tsere = U(0x05B5)
local segol = U(0x05B6)
local patah = U(0x05B7)
local qamats = U(0x05B8)
local qamats_qatan = U(0x05C7)
local holam = U(0x05B9)
local holam_haser_for_waw = U(0x05BA)
local qubuts = U(0x05BB)
local dagesh_mappiq = U(0x05BC)
local shin_dot = U(0x05C1)
local sin_dot = U(0x05C2)

local macron_above = U(0x0304)
local macron_below = U(0x0331)
local macron = ""

local alef = "א"
local he = "ה"
local waw = "ו"
local yod = "י"
local vowel_letters = alef .. he .. waw .. yod
local vowel_letter = ""

-- '0' represents silent sheva
local vowel_points = (
	sheva .. hataf_segol .. hataf_patah .. hataf_qamats .. hiriq .. tsere ..
	segol .. patah .. qamats .. qamats_qatan .. holam .. qubuts .. '0' ..
	holam_haser_for_waw
)
local vowel_point = ""
local short_vowels = segol .. patah .. hiriq .. qubuts .. qamats_qatan
local short_vowel = ""

local shuruq = waw .. dagesh_mappiq
local holam_male = waw .. holam

-- use dummies characters that do not match as punctuation
-- the dummy letter stands in for final silent alef or he, or for the hiatus before a furtive patah,
-- or comes before a pre-transliterated waw to aid in matching
local dummy_letter = U(0x0627) -- ARABIC LETTER ALEF
local dummy_geresh = U(0x064E) -- ARABIC FATHA
local dummy_gershayim = U(0x064B) -- ARABIC FATHATAN
local real_geresh = '׳'
local real_gershayim = '״'
local letter_modifier = "??"
local letters = "אבגדהוזחטיכךלמםנןסעפףצץקרשת"
local letter = "" .. letter_modifier
local letter_not_waw = "" .. letter_modifier
local gutturals = "אהחע"
local guttural = ""

local vowel_letter_or_geresh = ""

-- note, the geresh and gershayim are included in this, which is why dummies are used in their place
local word_break_chars = "%s%p"
local word_break = ""
local word_start = "%f" -- matches the boundary but not the actual word break characters
local word_end = "%f" -- matches the boundary but not the actual word break characters

local tr_vowels = "aeiouāēīōūəăĕŏ0"

local biblical_to_modern = {
	 = '\'',
	 = 'v',
	 = 'g',
	 = 'd',
	 = 'v',
	 = 'zh',
	 = 'kh',
	 = 't',
	 = 'kh',
	 = '\'',
	 = 'f',
	 = 'ts',
	 = 'ch',
	 = 'k',
	 = 'sh',
	 = 's',
	 = 't',

	 = '\'',
	 = 'e',
	 = 'a',
	 = 'o',
	 = 'i',
	 = 'e',
	 = 'a',
	 = 'o',
	 = 'u',
}

-- helper function to remove vowel letters but keep gereshes
local function gereshes(str)
	return gsub(str, vowel_letter, '')
end

local biblical = {
	{
		-- replace geresh and gershayim with their dummy equivalents so that they won't match as word boundaries
		 = dummy_geresh,
		 = dummy_gershayim,
	},

	{
		-- The default order is: consonant, vowel point, dagesh or mappiq, shin or sin dot.
		-- The desired order is: consonant, shin or sin dot, dagesh or mappiq, vowel point.
		-- Also, move geresh and gershayim closer to the letter for easier handling (will be moved back later if not actually a modifier)
		)(" .. vowel_point .. "*)(" .. dagesh_mappiq .. "*)(*)(*)"] = "%1%4%5%3%2",
	},

	{
		-- special case: change qamats in כל to qamats qatan
		-- the problem is that כל might be preceded by prefixed clitics, which maybe be chained indefinitely,
		-- while other unrelated words might happen to end in כל with a qamats gadol; therefore, match either
		-- the entire word or only when preceded by a precisely recognized prefix
		 = "%1" .. qamats_qatan .. "%2",
		" .. dagesh_mappiq .. "?" .. patah .. "כ" .. dagesh_mappiq .. ")" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2",
		 = "%1" .. qamats_qatan .. "%2",
		כ" .. dagesh_mappiq .. ")" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2", -- patah is very archaic
		" .. dagesh_mappiq .. "?" .. sheva .. "כ)" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2",
	},

	{
		-- remove final alef and he, but only when preceded by a vowel
		" .. word_end] = "%1" .. dummy_letter,
		" .. word_end] = "%1" .. dummy_letter,
	},

	{
		-- these are the cases, other than the above, where a final letter should be ignored
		" .. word_end] = "ī",
		)" .. vowel_letter_or_geresh .. "-" .. word_end] = "%1",
		)" .. vowel_letter_or_geresh .. "-" .. word_end] = "%1",
	},

	{
		 = "0%1" .. sheva, -- two shevas in a row
		 = "%10", -- after a short vowel, assume(!) a silent sheva
		 = "%10", -- gutturals cannot have a vocal sheva

		 = "%1" .. dummy_letter .. "ww", -- when waw + dagesh is not a shuruq
		 = "%1" .. dummy_letter .. "ww%2", -- when waw + dagesh is not a shuruq
		 = "%1" .. dummy_letter .. "w" .. holam, -- when waw + holam is not a holam male

		)" .. dagesh_mappiq] = "%1", -- handle mappiq (very rarely occurs on an alef)
	},

	{
		 = shuruq .. "ww", -- another potential case when waw + dagesh is not a shuruq
		 = shuruq .. "w" .. holam, -- another potential case when waw + holam is not a holam male

		-- tentatively lengthen hiriqs with vowel letters
		 = function(vlg, l) return "ī" .. gereshes(vlg) .. l end,

		-- rearrange furtive patach (mappiq should already have been removed, but handle it just in case)
		 = dummy_letter .. "a%1",
	},

	{
		-- remove vowel letters
		 = function(l, vlg) return l .. gereshes(vlg) .. shuruq end,
		 = function(vlg, l) return shuruq .. gereshes(vlg) .. l end,
		)"] = function(vlg, l) return shuruq .. gereshes(vlg) .. l end,
		 = function(vp, vlg, l) return vp .. gereshes(vlg) .. l end,
		)"] = function(vp, vlg, l) return vp .. gereshes(vlg) .. l end,
	},

	{
		-- handle two-character combinations first
		 = 'j',
		 = 'ž',
		' .. dummy_geresh] = 'č',
		 = 'š',
		 = 'ś',
	},

	{
		 = 'ʾ',
		 = 'b' .. macron_below,
		 = 'g' .. macron_above,
		 = 'd' .. macron_below,
		 = 'h',
		 = 'z',
		 = 'ḥ',
		 = 'ṭ',
		 = 'y',
		'] = 'k' .. macron_below,
		 = 'l',
		'] = 'm',
		'] = 'n',
		 = 's',
		 = 'ʿ',
		'] = 'p' .. macron_above,
		'] = 'ṣ',
		 = 'q',
		 = 'r',
		 = 't' .. macron_below,
	},

	{
		)' .. macron .. '?' .. dagesh_mappiq] = '%1', -- assume(!) dagesh qal at the beginning of a word
		()' .. macron .. '?' .. dagesh_mappiq] = '0%1', -- dagesh qal after sheva, and assume(!) silent sheva
		 = '%1' .. sheva .. '%1', -- vocal sheva between identical consonants
		 = 'ū',
	},

	{
		-- restore geresh and gershayim order
		)(" .. dagesh_mappiq .. "*)(" .. vowel_point .. "*)"] = "%2%3%1",
	},

	{
		-- handle ירושלם
		 = "ayi", -- in this case, the vowels are reversed by Unicode normalization rules
		 = "ayi", -- just in case they're in the correct order
		 = "āyi", -- pausal form of above
		 = "āyi", -- as above
		-- handle ירושלמה
		" .. patah] = "ay", -- in this case, the vowels are reversed by Unicode normalization rules
		"] = "ay", -- just in case they're in the correct order
		" .. qamats] = "āy", -- pausal form of above
		"] = "āy", -- as above
	},

	{
		 = 'ə',
		 = 'ĕ',
		 = 'ă',
		 = 'ŏ',
		 = 'i',
		 = 'ē',
		 = 'e',
		 = 'a',
		 = 'ā',
		 = 'o',
		 = 'u',
		 = '',
		 = '',
		 = 'ō',
		 = 'wō',
	},

	{
		 = '%1%1', -- gemination
	},

	{
		(k' .. macron_below .. ')'] = '%1%2', -- special case for יששכר
	},

	{
		 = 'o%1', -- assume(!) qamats qatan before silent sheva

		 = 'ō',
		 = 'w',
		 = 'š', -- assume(!) shin if no shin or sin dot
	},

	{
		-- handle bgdkpt letters in unvocalized words (such as acronyms)
		-" .. macron .. "-)" .. word_end] = function(w) return gsub(w, "()" .. macron, "%1") end
	},

	{
		"] = "",

		-- short vowels in non-final closed syllables (this rule should be expanded)
		 = "u%1%1",
		 = "i%1%1",
	},

	{
		 = "", -- final sheva is always silent

		 = '′',
		 = '″',
		 = '.', -- sof pasuq
		 = '-', -- maqaf
	},
}

function export.tr(text, lang, sc)
	-- default to modern for Hebrew, but not for other languages, such as Aramaic
	local modern = lang == "he"
	return export.biblical(text, modern)
end

function export.biblical(text, modern)
	-- decompose
	text = mw.ustring.toNFD(text)

	-- wrap with spaces to make initial and final replacements easier
	text = ' ' .. text .. ' '

	for _, replacements in ipairs(biblical) do
		for regex, replacement in pairs(replacements) do
			text = gsub(text, regex, replacement)
		end
	end

	-- unwrap spaces
	text = mw.ustring.match(text, "^ (.*) $")
	if text == nil then error("Something went wrong, wrapped spaces were deleted.") end

	-- must happen before recomposition
	if modern then
		text = gsub(text, "()%1", "%1")
		text = gsub(text, "" .. macron .. "?", function(x) return biblical_to_modern or x end)
		text = gsub(text, "''", "'")
	end

	-- recompose
	text = mw.ustring.toNFC(text)

	return text
end

return export