Sware: 1 revision imported

2026-04-21T12:44:14Z

1 revision imported

← Older revision	Revision as of 12:44, 21 April 2026
(No difference)

wikt>Theknightwho: Use faster implementation of mw.ustring.char.

2024-04-05T03:06:37Z

Use faster implementation of mw.ustring.char.

New page

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 = "[" .. macron_above .. macron_below .. "]"

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

-- '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 = "[" .. vowel_points .. "]"
local short_vowels = segol .. patah .. hiriq .. qubuts .. qamats_qatan
local short_vowel = "[" .. short_vowels .. "]"

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 = "[" .. shin_dot .. sin_dot .. "]?[" .. dummy_geresh .. dummy_gershayim .. "]?"
local letters = "אבגדהוזחטיכךלמםנןסעפףצץקרשת"
local letter = "[" .. letters .. dummy_letter .. "]" .. letter_modifier
local letter_not_waw = "[אבגדהזחטיכךלמםנןסעפףצץקרשת" .. dummy_letter .. "]" .. letter_modifier
local gutturals = "אהחע"
local guttural = "[" .. gutturals .. "]"

local vowel_letter_or_geresh = "[" .. vowel_letters .. dummy_geresh .. dummy_gershayim .. "]"

-- 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 = "[" .. word_break_chars .. "]"
local word_start = "%f[^" .. word_break_chars .. "]" -- matches the boundary but not the actual word break characters
local word_end = "%f[" .. word_break_chars .. "]" -- matches the boundary but not the actual word break characters

local tr_vowels = "aeiouāēīōūəăĕŏ0"

local biblical_to_modern = {
['ʾ'] = '\'',
['b' .. macron_below] = 'v',
['g' .. macron_above] = 'g',
['d' .. macron_below] = 'd',
['w'] = 'v',
['ž'] = 'zh',
['ḥ'] = 'kh',
['ṭ'] = 't',
['k' .. macron_below] = 'kh',
['ʿ'] = '\'',
['p' .. macron_above] = 'f',
['ṣ'] = 'ts',
['č'] = 'ch',
['q'] = 'k',
['š'] = 'sh',
['ś'] = 's',
['t' .. macron_below] = 't',

['ə'] = '\'',
['ĕ'] = 'e',
['ă'] = 'a',
['ŏ'] = 'o',
['ī'] = 'i',
['ē'] = 'e',
['ā'] = 'a',
['ō'] = 'o',
['ū'] = '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
[real_geresh] = dummy_geresh,
[real_gershayim] = 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)
["([" .. letters .. "])(" .. vowel_point .. "*)(" .. dagesh_mappiq .. "*)([" .. shin_dot .. sin_dot .. "]*)([" .. dummy_geresh .. dummy_gershayim .. "]*)"] = "%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
[word_start .. "(כ" .. dagesh_mappiq .. "?)" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2",
["([הבכל]" .. dagesh_mappiq .. "?" .. patah .. "כ" .. dagesh_mappiq .. ")" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2",
["(מ" .. dagesh_mappiq .. "?" .. hiriq .. "כ" .. dagesh_mappiq .. ")" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2",
["(ש" .. shin_dot .. dagesh_mappiq .. "?[" .. segol .. patah .. "]כ" .. dagesh_mappiq .. ")" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2", -- patah is very archaic
["([ובכלד]" .. dagesh_mappiq .. "?" .. sheva .. "כ)" .. qamats .. "(ל)" .. word_end] = "%1" .. qamats_qatan .. "%2",
},

{
-- remove final alef and he, but only when preceded by a vowel
["(" .. vowel_point .. vowel_letter_or_geresh .. "*)[" .. alef .. he .. "]" .. word_end] = "%1" .. dummy_letter,
["(" .. shuruq .. vowel_letter_or_geresh .. "*)[" .. alef .. he .. "]" .. word_end] = "%1" .. dummy_letter,
},

{
-- these are the cases, other than the above, where a final letter should be ignored
[hiriq .. vowel_letter_or_geresh .. "-[" .. yod .. dummy_letter .. "]" .. word_end] = "ī",
["([" .. tsere .. segol .. "])" .. vowel_letter_or_geresh .. "-[" .. yod .. "]" .. word_end] = "%1",
["([" .. holam .. qubuts .. "])" .. vowel_letter_or_geresh .. "-[" .. waw .. "]" .. word_end] = "%1",
},

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

["(" .. vowel_point .. ")" .. shuruq] = "%1" .. dummy_letter .. "ww", -- when waw + dagesh is not a shuruq
["(" .. vowel_point .. vowel_letter_or_geresh .. "-)" .. shuruq .. "(" .. vowel_letter_or_geresh .. "-" .. vowel_point .. ")"] = "%1" .. dummy_letter .. "ww%2", -- when waw + dagesh is not a shuruq
["(" .. vowel_point .. ")" .. holam_male] = "%1" .. dummy_letter .. "w" .. holam, -- when waw + holam is not a holam male

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

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

-- tentatively lengthen hiriqs with vowel letters
[hiriq .. "(" .. vowel_letter_or_geresh .. "+)(" .. letter .. ")"] = function(vlg, l) return "ī" .. gereshes(vlg) .. l end,

-- rearrange furtive patach (mappiq should already have been removed, but handle it just in case)
["(" .. guttural .. dagesh_mappiq .. "?)" .. patah .. word_end] = dummy_letter .. "a%1",
},

{
-- remove vowel letters
["(" .. letter .. ")(" .. vowel_letter_or_geresh .. "+)" .. shuruq] = function(l, vlg) return l .. gereshes(vlg) .. shuruq end,
[shuruq .. "(" .. vowel_letter_or_geresh .. "+)" .. "(" .. letter_not_waw .. ")"] = function(vlg, l) return shuruq .. gereshes(vlg) .. l end,
[shuruq .. "(" .. vowel_letter_or_geresh .. "+)" .. "(" .. waw .. "[^" .. holam .. dagesh_mappiq .. "])"] = function(vlg, l) return shuruq .. gereshes(vlg) .. l end,
["(" .. vowel_point .. ")" .. "(" .. vowel_letter_or_geresh .. "+)" .. "(" .. letter_not_waw .. ")"] = function(vp, vlg, l) return vp .. gereshes(vlg) .. l end,
["(" .. vowel_point .. ")" .. "(" .. vowel_letter_or_geresh .. "+)" .. "(" .. waw .. "[^" .. holam .. dagesh_mappiq .. "])"] = function(vp, vlg, l) return vp .. gereshes(vlg) .. l end,
},

{
-- handle two-character combinations first
['ג' .. dummy_geresh] = 'j',
['ז' .. dummy_geresh] = 'ž',
['[צץ]' .. dummy_geresh] = 'č',
['ש' .. shin_dot] = 'š',
['ש' .. sin_dot] = 'ś',
},

{
['א'] = 'ʾ',
['ב'] = 'b' .. macron_below,
['ג'] = 'g' .. macron_above,
['ד'] = 'd' .. macron_below,
['ה'] = 'h',
['ז'] = 'z',
['ח'] = 'ḥ',
['ט'] = 'ṭ',
['י'] = 'y',
['[כך]'] = 'k' .. macron_below,
['ל'] = 'l',
['[מם]'] = 'm',
['[נן]'] = 'n',
['ס'] = 's',
['ע'] = 'ʿ',
['[פף]'] = 'p' .. macron_above,
['[צץ]'] = 'ṣ',
['ק'] = 'q',
['ר'] = 'r',
['ת'] = 't' .. macron_below,
},

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

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

{
-- handle ירושלם
[hiriq .. patah] = "ayi", -- in this case, the vowels are reversed by Unicode normalization rules
[patah .. hiriq] = "ayi", -- just in case they're in the correct order
[hiriq .. qamats] = "āyi", -- pausal form of above
[qamats .. hiriq] = "āyi", -- as above
-- handle ירושלמה
["[0" .. sheva .. "]" .. patah] = "ay", -- in this case, the vowels are reversed by Unicode normalization rules
[patah .. "[0" .. sheva .. "]"] = "ay", -- just in case they're in the correct order
["[0" .. sheva .. "]" .. qamats] = "āy", -- pausal form of above
[qamats .. "[0" .. sheva .. "]"] = "āy", -- as above
},

{
[sheva] = 'ə',
[hataf_segol] = 'ĕ',
[hataf_patah] = 'ă',
[hataf_qamats] = 'ŏ',
[hiriq] = 'i',
[tsere] = 'ē',
[segol] = 'e',
[patah] = 'a',
[qamats] = 'ā',
[qamats_qatan] = 'o',
[qubuts] = 'u',
[shin_dot] = '',
[sin_dot] = '',
[holam_male] = 'ō',
[waw .. holam_haser_for_waw] = 'wō',
},

{
['(.)' .. macron .. '?' .. dagesh_mappiq] = '%1%1', -- gemination
},

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

{
['ā(%l' .. macron .. '?0)'] = 'o%1', -- assume(!) qamats qatan before silent sheva

[holam] = 'ō',
['ו'] = 'w',
['ש'] = 'š', -- assume(!) shin if no shin or sin dot
},

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

{
["[0" .. dummy_letter .. "]"] = "",

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

{
['ə' .. word_end] = "", -- final sheva is always silent

[dummy_geresh] = '′',
[dummy_gershayim] = '″',
['׃'] = '.', -- 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, "([%lʾʿ])%1", "%1")
text = gsub(text, "[%lʾʿ]" .. macron .. "?", function(x) return biblical_to_modern[x] or x end)
text = gsub(text, "''", "'")
end

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

return text
end

return export

Module:he-translit/old - Revision history

Sware: 1 revision imported

wikt>Theknightwho: Use faster implementation of mw.ustring.char.