Module:User:OblivionKhorasan/pnb-testing

This module sandbox lacks a documentation subpage. Please create it.
Useful links: root page • root page’s subpages • links • transclusions • testcases • sandbox
local m_str_utils = require("Module:string utilities")

local U = m_str_utils.char
local gsub = m_str_utils.gsub

local export = {}

local zabar = U(0x64E)
local zer = U(0x650)
local pesh = U(0x64F)
local tashdid = U(0x651) -- also called shadda
local jazm = "ْ"
local he = "ہ"

local punctuation = "%-:%(%)%*&٫؛؟،ـ«\".\'!»٪؉۔"
local numbers = "۱۲۳۴۵۶۷۸۹۰"

local ain = 'ع'
local alif = 'ا'
local ye = 'ی'
local ye2 = 'ے'
local ye3 = 'ئ'
local vaw = "و"
local nasal = 'ں'
local nunghunna = 'ن٘'
local aspirate = 'ھ'

local consonants = "بٻپتٹثجڄچحخدڊڈڋݙذرڑزژسشصضطظعغفقکݢگڳلࣇمنݨوہھئی"
local nonhe = "بٻپتٹثجڄچحخدڊڈڋݙذرڑزژسشصضطظعغفقکݢگڳلࣇمنݨوھئی"
local nonye = "بٻپتٹثجڄچحخدڊڈڋݙذرڑزژسشصضطظعغفقکݢگڳلࣇمنݨوہھئ"
local nonvaw = "ٻپتٹثجڄچحخدڊڈڋݙذرڑزژسشصضطظعغفقکݢگڳلࣇمنݨہھئی"
local vowels = "ایئےۓوؤ"
local indvowels = "آایےوؤ"
local semivowel = "یو"
local hes = "ہح"
local diacritics = "ًَُِّْٰ"
local ZZP = "َُِ"

local mapping = {
	 = 'ā',  = 'b',  = 'b̤',  = 'p',  = 't',  = 'ṭ',  = 's̱',
	 = 'j',  = 'j̈',  = 'c',  = 'ḥ',  = 'x', 
	 = 'd',  = 'ḍ',  = 'd̤',  = 'ẕ',  = 'r',  = "ṛ",  = 'z',  = 'ž',
	 = 's',  = 'ś',  = 'ṣ',  = 'ẓ', 
	 = 't̤',  = 'z̤',  = 'ʻ',  = 'ġ',  = 'f',  = 'q',
	 = 'k',  = 'g',  = 'g̈',  = 'ŋ',  = 'ṇ',  = 'ḷ',
	 = 'l',  = 'm',  = 'n',  = 'v',  = 'h',  = 'y',  = ".",  = 'ṉ',

	 = 'H',
	 = '',
	
	-- diacritics
	 = "a",
	 = "i",
	 = "u",
	 = "", -- also sukun - no vowel
	 = "-", -- ZWNJ (zero-width non-joiner)
	
	-- ligatures
	 = "lā",
	 = "allāh",
	
	-- kashida
	 = "-", -- kashida, no sound
	
	-- numerals
	 = "1",  = "2",  = "3",  = "4",  = "5",
	 = "6",  = "7",  = "8",  = "9",  = "0",
	
	-- punctuation (leave on separate lines)
	 = "?", -- question mark
	 = ",", -- comma
	 = ";", -- semicolon
	 = '“', -- quotation mark
	 = '”', -- quotation mark
	 = "%", -- percent
	 = "‰", -- per mille
	 = ".", -- decimals
	 = ",", -- thousand
}

function export.tr(text, lang, sc)
	
	--define the "end" of a word
	text = gsub(text, "#", "HASHTAG")
	text = gsub(text, " | ", "# | #")
	text = gsub(text, "\n" , "#".."\n" .. "#")
	text = gsub(text, "()" , "#".."%1" .. "#")
	text = "##" .. gsub(text, " ", "# #") .. "##"
	-- hastags now mark the beginning and end of a word
	
	-- EXCEPTIONS - leave as they are, unless they have been sorted out elsewhere

    text = gsub(text, '()' .. nunghunna, "%1ṉ")
    text = gsub(text, "الله", "allāh")
    
  	-- diacritics
  	text = gsub(text, '()' .. '()' .. aspirate, "%1h%2")
	text = gsub(text, pesh .. vaw .. jazm .. "", "ū")
	text = gsub(text, zabar .. vaw .. jazm .. "", "au")
	text = gsub(text, zer .. ye .. '()', "ī%1")
	text = gsub(text, zer .. ye .. "#", "ī")
	text = gsub(text, zer .. ye .. tashdid .. alif, "iyyā")
	text = gsub(text, zabar .. ye .. '()', "ai")
    text = gsub(text, jazm .. alif, "ā")
    
	-- Initial alif
	text = gsub(text, alif .. zabar, "a")
	text = gsub(text, alif .. zer .. ye .. jazm .. "", "ī")
	text = gsub(text, alif .. ye .. alif .. "", "eyā")

	text = gsub(text, alif .. zer, "i")
	text = gsub(text, alif .. vaw .. jazm .. "", "o")
	text = gsub(text, alif .. ye .. '()', "e%1")
	text = gsub(text, alif .. ye .. jazm .. "", "e")

	text = gsub(text, alif .. pesh .. vaw .. jazm .. "", "u")	
	text = gsub(text, alif .. pesh, "u")

	-- Tashdeed
	text = gsub(text, '()' .. tashdid, "%1%1")
	text = gsub(text, '()' .. tashdid .. '()', "%1%1%2")
	text = gsub(text, '()' .. ye .. '()' .. tashdid, "%1yy%2")
	text = gsub(text, '()' .. vaw .. '()' .. tashdid, "%1ww%2")
	-- For some reason the tashdeed gets pushed after the other diacritics, so this line is necessary for tashdeed to work with other diacritics
	text = gsub(text, '()' .. '()' .. tashdid, "%1%1%2")
	
	-- tanween diacritic / no need to mess about
	text = gsub(text, '()' .. 'ً' .. alif, "%1an")
	text = gsub(text, alif .. 'ً', "an")
	text = gsub(text, '()' .. 'ً', "%1an")

	-- tall zabar -- / no need to mess about
	text = gsub(text, '()' .. 'ٰ', "á")
	text = gsub(text, '()' .. 'ٰ' .. '()', "%1á")

	-- ‘ain
	text = gsub(text, alif .. ain , "ā‘") 
	text = gsub(text, ain .. alif  .. '()', "ʻā%1") 
	text = gsub(text, '()' .. '(?)' .. ain, "%1%2ʻ")
	text = gsub(text, ain .. zer  .. '()', "ʻi%1")
	text = gsub(text, ain .. pesh  .. '()', "ʻu%1")
	text = gsub(text, ain .. zer .. ye .. '()', "ʻī%1")
    text = gsub(text, ain .. pesh .. vaw .. '()', "ʻū%1")
    
    ---  alif
    text = gsub(text, "" .. zabar .. alif, "%1ā")
	text = gsub(text, '()' .. alif, "%1ā")
	text = gsub(text, '()' .. tashdid .. alif, "%1%1ā")
	text = gsub(text, "#" .. alif .. ye, "ī")
	text = gsub(text, "#" .. alif .. vaw, "o")
	text = gsub(text, "#" .. alif .. ye2, "e")
  
	-- vaw
	text = gsub(text, vaw .. '()', "v%1")
	text = gsub(text, '()' .. vaw .. jazm, "%1o")

	-- medial/final consonants

    --- (e) -- works
	text = gsub(text, '()' .. ye .. jazm .. '()', "%1e%2")
	text = gsub(text, '()' .. ye .. '()', "%1e%2")
	
    --- he 
	text = gsub(text, zabar .. he .. '()', "ah%1")
	text = gsub(text, '()' .. zabar .. he, "%1ah")
	text = gsub(text, '()' .. he .. "#", "%1ā")
	text = gsub(text, '()' .. he .. he .. "#", "%1hā")
	text = gsub(text, '()' .. jazm .. he .. he .. "#", "%1hā")
	text = gsub(text, '()' .. vaw .. he .. "#", "%1oh")
	text = gsub(text, '()' .. he .. "#", "%1ā")

    --- vaw

	text = gsub(text, '()' .. vaw .. jazm, "%1o")
	text = gsub(text, '()' .. tashdid .. vaw, "%1%1o")
	
	--- ye
	text = gsub(text, '()' .. ye .. '()', "%1ī%2")
	text = gsub(text, '()' .. ye .. "#", "%1ī")
	text = gsub(text, '()' .. ye2, "%1e")
	text = gsub(text, '()' .. zabar .. ye2, "%1ai")
	text = gsub(text, '()' .. ye .. '()', "%1e%2")
	text = gsub(text, '()' .. ye .. alif, "%1eyā")
	text = gsub(text, '()' .. ye .. alif, "%1yā")
	text = gsub(text, '()' .. ye .. alif, "%1yā")
	text = gsub(text, "#" .. ye .. alif, "yā")

    -- get rid of hashtags (not needed)
    text = gsub(text, "#", "")
    text = gsub(text, "HASHTAG", "#")

	text = gsub(text, '.', mapping)

	
	text = gsub(text, 'ا', "ā")
	text = gsub(text, 'īā', "iyā")
	text = gsub(text, 'ehh', "eh")
	text = gsub(text, 'īhh', "īh")
	text = gsub(text, 'n٘', "ṉ")
	text = gsub(text, 'ṉhh', "ṉh")
	text = gsub(text, 'yṉ', "eṉ")
	text = gsub(text, 'H', "'")
	--

	return text
end
return export
Module:User:OblivionKhorasan/pnb-testing

Wikious

Boobota

Sagapedia
