Module:Linguistic

From Wikispecies
Jump to navigation Jump to search

Documentation for this module may be created at Module:Linguistic/doc

--[[
  __  __           _       _        _     _                   _     _   _      
 |  \/  | ___   __| |_   _| | ___ _| |   (_)_ __   __ _ _   _(_)___| |_(_) ___ 
 | |\/| |/ _ \ / _` | | | | |/ _ (_) |   | | '_ \ / _` | | | | / __| __| |/ __|
 | |  | | (_) | (_| | |_| | |  __/_| |___| | | | | (_| | |_| | \__ \ |_| | (__ 
 |_|  |_|\___/ \__,_|\__,_|_|\___(_)_____|_|_| |_|\__, |\__,_|_|___/\__|_|\___|
                                                  |___/                        
Simple internationalization functions that can be called by other modules.
This Module was copied from Wikimedia Commons, so please request changes there.

Maintainers:
* Zolo - original version
* Jarekt

Dependencies:
* Module:I18n/or
* Module:Delink
* Module:Yesno
]]

-- ==================================================
-- === Internal functions ===========================
-- ==================================================

local function langSwitch(list,lang)
	local langList = mw.language.getFallbacksFor(lang)
	table.insert(langList,1,lang)
	for i,language in ipairs(langList) do
		if list[language] then
			return list[language]
		end
	end
	return nil
end

-- ==================================================
-- === External functions ===========================
-- ==================================================
local p = {}

function p.vowelfirst (str)
	if str then 
		local vowels = 'aeiouyąăẵằẳặȃắâẫấầẩậãäǟāáàȁǎảẚåǻḁạǡæǣǽĕȇêễếềểệḙẽḛëēḕéḗèȅěẻẹęȩḝǝĭȋîĩḭïḯīíìȉǐỉịįıŏȏôỗốồổộõṏṍöōṑóṓòȍǒỏọǫǭơỡớờởợøǿŭȗûṷũṻṹṵüǖǘǜǚṳūúùȕǔủůụųưữứừửựŷỹÿȳýỳỷẙỵ'
		str = mw.ustring.lower(mw.ustring.sub(str,1,1))
		return mw.ustring.find(vowels, str, 1, true ) 
	end
end

function p.inparentheses(str)
	if (not str) or (str == "") then
		return nil
	end
	return "(".. str .. ")" -- same in all languages  ?
end

function p.of(word, lang, raw, gender, number, determiner) -- rough translation of "of" in various languages
-- note that the cases when on "of" is employed varies a lot among languages, so it is more prudent to call this from lang specific function only
	if not raw or mw.text.trim(raw) == "" then 
		local args = {}
		local Delink = require('Module:Delink')._delink
		args[1] = word
		raw = mw.ustring.lower(Delink(args))
	end

	-- raw is the string without the Wikiformatting so that it correctly analyses the string that is [[:fr:Italie|Italie]] -> 'italie'
	-- any way to automate this ?
	
	if lang == 'fr' then 
		local yesno = require('Module:Yesno')
		determiner = yesno(determiner,false)

		if determiner then
			if string.sub(number or '',1,1)=='p' then -- number == 'plural'
				return 'des ' .. word
			elseif p.vowelfirst(raw) then
				return 'de l’' .. word
			elseif string.sub(gender or '',1,1)=='f' then -- gender == 'feminine'
				return 'de la ' .. word
			else
				return 'du ' .. word
			end
		else
			if p.vowelfirst(raw) then
				return 'd’' .. word
			else
				return 'de ' .. word
			end
		end
	elseif lang == 'ca' then
		-- implement [[Template:Of/ca]] or  https://ca.wikipedia.org/wiki/Plantilla:Deod%27/base for case where "{{{context}}}" is "en" (default on Commons)
		if ( p.vowelfirst(raw) and not mw.ustring.find( 'ia|ià|ie|io|iu|ua|ue|ui|uí|uï|uo|ya|ye|yi|yo|yu|', mw.ustring.sub(raw,1,2) .. '|')) then
			return 'd\'' .. word
		else
			return 'de ' .. word
		end
	end	

end

function p.offromwiki(frame)
	args = frame.args
	return p.of(args.word, args.lang, args.raw, args.gender, args.number, args.determiner)
end

function p.noungroup(noun, adj, lang)
	if not noun or noun == '' then 
		return nil -- not '' so that it is not counted as a string by mw.listToText
	end
	if not adj or adj == ''
		then return noun
	end
	
	local wordsep = mw.message.new( "Word-separator" ):inLanguage(lang):plain()
	local LUT = {ar=2, ca=2, cs=1, da=1, de=1, el=1, en=1, es=2, et=1, fi=1,
		fr=2, gl=2, he=1, hu=1, it=2, ja=1, la=2, nds=1, nl=1, no=1, pl=1, pt=2, ro=2,
		ru=1, scn=2, sk=1, sl=1, sr=1, sv=1, sw=1, tr=1, vec=1, vi=1, zh=1}
	-- source for many: language subpages of [[Template:Technique]]
	-- gl and pt aren't consistent in [[Template:Technique/gl]] and
	--     [[Template:Technique/pt]], but seem to be 2
	-- shouldn't he be 2? in [[Template:Technique/he]] it seems to be 1 though
	-- mk inconsistent in [[Template:Technique/mk]]
	local case = LUT[lang]
	if case==1 then     -- adjective before the noun
		return adj .. wordsep .. noun
	elseif case==2 then -- adjective after the noun
		return noun .. wordsep .. adj
	else                -- order unknown
		return noun .. ' (' .. adj .. ')'
	end
end

function p.conj(args, lang, conjtype)
	local comma    = mw.message.new( "comma-separator"):inLanguage(lang):plain()
	local wordsep  = mw.message.new( "Word-separator" ):inLanguage(lang):plain()
	local andtable = { ar=' و', he=' ו', ja='および', pl=' i '} -- languages with a problem with the MediaWiki:And
	if conjtype == 'comma' then
		return mw.text.listToText(args, comma, comma)
	elseif conjtype == 'or' then 
		local wordor = 	langSwitch(require('Module:I18n/or'), lang)
		return mw.text.listToText(args, comma, wordor  .. wordsep)
	elseif conjtype == 'explicit or' then -- adds "or" betwen all words when the context can be confusing
		local wordor = 	langSwitch(require('Module:I18n/or'), lang)
		return mw.text.listToText(args, wordor .. wordsep, wordor  .. wordsep)
	elseif conjtype and conjtype ~= 'and' and conjtype ~= '' then
		return mw.text.listToText(args, conjtype, conjtype)
	elseif andtable[lang] then 
		return mw.text.listToText(args, comma, andtable[lang])
	else
		local wordand = mw.message.new( "and" ):inLanguage(lang):plain()
		return mw.text.listToText(args, comma, wordand .. wordsep)
	end
end

function p.conjfromWiki(frame)
	args = frame.args
	if not args or not args[1] then
		args = frame:getParent().args
	end
	local lang = args.lang
	if not lang or mw.text.trim(lang) == '' then
		lang = frame:callParserFunction( "int", "lang" )
	end
	newargs = {}  -- transform args metatable into a table so it can be concetenated
	for i, j in pairs(args) do
			if type(i) == 'number' then
				j = mw.text.trim(j)
				if j ~= '' then
					table.insert(newargs, j)
				end
			else 
				if i ~= 'type' and i ~= 'lang' then 
					return 'error: bad parameter in template:Conj: ' .. i .. '[[Category:Pages with incorrect template usage/Conj|A]]'
				end
			end
	end
	return p.conj(newargs, lang, args.type)
end

return p