Module:Nihongo

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

require('strict');

--[[--------------------------< E R R O R _ M E S S A G E S >--------------------------------------------------

error messaging; keys to this table are the template base names:
	'nihongo', 'nihongo3', 'nihongo krt', 'nihongo foot' → 'nihongo' etc

]]

local err_msg = {
	['nihongo'] = 'Japanese or romaji text required',
	}

local err_cat = {
	['nihongo'] = '[[Category:Nihongo template errors]]',
	}


--[[--------------------------< C O N F I G U R A T I O N >----------------------------------------------------

configuration setting for the various templates.  keys to this table are the template names without capitalization

]]

local cfg = {
	['nihongo'] = {
		tag = 'ja',
		system = 'hepburn',
		system_link = '[[Hepburn romanization|Hepburn]]',
		err_msg = err_msg.nihongo,
		err_cat = err_cat.nihongo,
		},
	['nihongo3'] = {
		tag = 'ja',
		system = 'hepburn',
		err_msg = err_msg.nihongo,
		err_cat = err_cat.nihongo,
		},
	['nihongo krt'] = {
		tag = 'ja',
		system = 'hepburn',
		err_msg = err_msg.nihongo,
		err_cat = err_cat.nihongo,
		},
	['nihongo foot'] = {
		tag = 'ja',
		system = 'hepburn',
		system_link = '[[Hepburn romanization|Hepburn]]',
		err_msg = err_msg.nihongo,
		err_cat = err_cat.nihongo,
		},
	}


--[[--------------------------< E R R O R _ M E S S A G E >----------------------------------------------------

Creates an error message for {{nihongo}}, {{nihongo3}}, {{nihongo krt}}, and {{nihongo foot}} when these template are missing <japanese>
or <romaji> inputs; names the offending template, links to template page, and adds article to Category:Nihongo template errors

]]

local function error_message (template)
	local msg = {'<span class="error" style="font-size:100%">error: {{'};
	table.insert (msg, template);
	table.insert (msg, '}}: ');
	table.insert (msg, cfg[template].err_msg);
	table.insert (msg, ' ([[Template:');
	table.insert (msg, template);
	table.insert (msg, '|help]])</span>');
	if 0 == mw.title.getCurrentTitle().namespace then
		table.insert (msg, cfg[template].err_cat);
	end

	return table.concat (msg);	
end


--[[--------------------------< R E N D E R E R >--------------------------------------------------------------

Shared support function for nihingo(), nihongo3(), and nihongo_foot().  Calculates an index into formatting{}
from set/unset parameters:
	args[1] (english text) has a value of 8 (set) or 0 (unset)
	args[2] (native text) has a value of 4
	args[3] (romanized text) has a value of 2
	args[4] (extra) has a value of 1
index, the sum of these values, gets the appropriate format string from formatting{} table with associated values
from the formatting[index][2] table

]]

local function renderer (args, formatting, extra2)
	local output;
	local index = 0;															-- index into formatting{}
	local param_weight = {8, 4, 2, 1};											-- binary parameter weights: [1] = english (8), [2] = japanese (4), [3] = romaji (2), [4] = extra (1)

	for i=1, 5 do																-- spin through args[1] – args[4]
		index = index + (args[i] and param_weight[i] or 0);						-- calculate an index into formatting{}
	end

	output = (0 ~= index) and string.format (formatting[index][1] and formatting[index][1], formatting[index][2][1], formatting[index][2][2], formatting[index][2][3], formatting[index][2][4]) or nil;

	if extra2 then																-- always just attached to the end (if there is an end) so not part of formatting{}
		output = output and (output .. ' ' .. extra2) or '<5p4n>' .. extra2;	-- <5p4n> and </5p4n>: place holders for font-weight style spans; akin to stripmarkers, to be replaced
	end																			-- (nihongo and nihongo3) or removed (nihongo foot)

	return output and (output .. '</5p4n>') or '';								-- where there is output, add secret tag close
end


--[[--------------------------< R O M A N I Z E D _ K E R N >--------------------------------------------------

Add kerning when first or last character of romanized text contacts adjacent opening or closing parenthesis

In this example, without kerning, the romanized characters 'j' and 'V' are italicized so will contact the parentheses
	(<i lang=\"ja-Latn\" title=\"Hepburn transliteration\">jV</i>)

<ret_string> is the formatted template output (except that the magic string '<5p4n>' has not yet been replaced)
<romanized> is the return from lang_module._xlit() so is not wrapped in parentheses

]]

local function romanized_kern (ret_string, romanized)
	if not romanized or ('' == romanized) then									-- if romanized not set
		return ret_string;														-- then we're done
	end

	local romanized_text = romanized:gsub ('%b<>', ''):gsub ('\'\'+', ''):gsub ('%[%[', ''):gsub ('%]%]', '');	-- strip HTML tags

	romanized = romanized:gsub ('([%(%)%.%%%+%-%*%?%[%^%$%]])', '%%%1');				-- escape lua pattern characters

	local romanized_has_leading_paren = ret_string:match ('%(' .. romanized);			-- has a value if (<romanized>; nil else
	local romanized_has_trailing_paren = ret_string:match (romanized .. '%)');		-- has a value if <romanized>); nil else

	local kern_lead_pattern = '^[jpy]';											-- list of characters that when italicized contact unitalicized leading parenthesis
	local kern_tail_pattern = '[dfijkltCEFHIJKMNPR-Z\'"%?!%]]$';				-- list of characters that when italicized contact unitalicized trailing parenthesis

	local kern_right = '<span style="margin-right:.09em">(</span>%1';			-- %1 is <romanized> capture
	local kern_left = '%1<span style="margin-left:.09em">)</span>';				-- %1 is <romanized> capture

	if romanized_has_leading_paren and romanized_text:match (kern_lead_pattern) then
		ret_string = ret_string:gsub ('%((' .. romanized .. ')', kern_right);		-- replace plain '(' with kerned '('; <romanized> included here to ensure that the correct '(' is kerned
	end
	if romanized_has_trailing_paren and romanized_text:match (kern_tail_pattern) then
		ret_string = ret_string:gsub ('(' .. romanized .. ')%)', kern_left);		-- replace plain ')' with kerned ')'; <romanized> included here to ensure that the correct ')' is kerned
	end

	return ret_string;															-- done
end


--[[--------------------------< C O M M O N >------------------------------------------------------------------

Common support for {{nihongo}}

render order: is translated (English), native, romanized

<template> is used to select the appropriate cfg table

]]

local function common (frame, template)
	local lang_module = require ('Module:Lang' .. (frame:getTitle():match ('/sandbox') or ''));	-- if this module is the sandbox, use Module:lang/sandbox; Module:Lang else

	local args = require ('Module:Arguments').getArgs (frame);
	
	local english, native, romanized, extra, extra2 = args[1], args[2], args[3], args.extra or args[4], args.extra2 or args[5];	-- meaningful names
	args[4] = extra or args[4];													-- ensure that extra is 'positional' for use by renderer()

	local lead = require ('Module:yesno')(args.lead);							-- make boolean

	if not (native or romanized) then											-- not present, return an error message
		return error_message (template);
	end
	if native then
		native = lead and lang_module._langx ({['code']=cfg[template].tag, ['text']=native, ['template']=template}) or
			lang_module._lang ({cfg[template].tag, native, ['template']=template});	-- add ja script with/without language prefix
	end
	if romanized then
		romanized = (lead and english and (cfg[template].system_link .. ': ') or '') .. lang_module._xlit ({'ja', cfg[template].system, romanized, ['template']=template}) or nil;
	end

	local formatting = {														-- <5p4n> and </5p4n>: place holders for font-weight style spans; akin to stripmarkers, replaced  before function returns
		{'<5p4n>(%s)', {extra}}, 												-- 1 - (extra)
		{'%s<5p4n>', {romanized}},												-- 2 - romanized
		{'%s<5p4n> (%s)', {romanized, extra}},									-- 3 - romanized (extra)
		{'<5p4n>(%s)', {native}},												-- 4 - native
		{'<5p4n>(%s, %s)', {native, extra}},									-- 5 - (native, extra)
		{'%s<5p4n> (%s)', {romanized, native}},									-- 6 - romanized (native)
		{'%s<5p4n> (%s, %s)', {romanized, native, extra}},						-- 7 - romanized (native, extra)
		{'%s<5p4n>', {english}},												-- 8 - english
		{'%s<5p4n> (%s)', {english, extra}},									-- 9 - english (extra)
		{'%s<5p4n> (%s)', {english, romanized}},								-- 10 - english (romanized)
		{'%s<5p4n> (%s, %s)', {english, romanized, extra}},						-- 11 - english (romanized, extra)
		{'%s<5p4n> (%s)', {english, native}},									-- 12 - english (native)
		{'%s<5p4n> (%s, %s)', {english, native, extra}},						-- 13 - english (native, extra)
		{'%s<5p4n> (%s, %s)', {english, native, romanized}},					-- 14 - english (native, romanized)
		{'%s<5p4n> (%s, %s, %s)', {english, native, romanized, extra}},			-- 15 - english (native, romanized, extra)
		}

	local ret_string = renderer (args, formatting, extra2)
	ret_string = romanized_kern (ret_string, romanized);						-- kern romanized text when appropriate
	ret_string = ret_string:gsub ('<5p4n>', '<span style="font-weight: normal">'):gsub ('</5p4n>', '</span>');	-- replace 'secret' tags with proper tags

	return ret_string;															-- because gsub returns the number of replacements made as second return value
end


--[[--------------------------< C O M M O N _ R O M A N I Z E D _ N A T I V E _ T R A N S L A T E D >----------

Common support for {{nihongo3}}

render order: is romanized, native, translated (English)

<template> is used to select the appropriate cfg table

]]

local function common_romanized_native_translated (frame, template)
	local lang_module = require ('Module:Lang' .. (frame:getTitle():match ('/sandbox') or ''));	-- if this module is the sandbox, use Module:lang/sandbox; Module:Lang else
	local args = require ('Module:Arguments').getArgs (frame);
	
	local english, native, romanized, extra, extra2 = args[1], args[2], args[3], args.extra or args[4], args.extra2 or args[5];	-- meaningful names
	args[4] = extra or args[4];													-- ensure that extra is 'positional' for use by renderer()

	if not (native or romanized) then											-- not present, return an error message
		return error_message (template);
	end
	native = native and lang_module._lang ({cfg[template].tag, native, ['template']=template}) or nil;
	romanized = romanized and lang_module._xlit ({cfg[template].tag, cfg[template].system, romanized, ['template']=template}) or nil;
	
	local formatting = {														-- <5p4n> and </5p4n>: place holders for font-weight style spans; akin to stripmarkers, replaced  before function returns
		{'<5p4n>(%s)', {extra}}, 												-- 1 - (extra)
		{'%s<5p4n>', {romanized}},												-- 2 - romanized
		{'%s<5p4n> (%s)', {romanized, extra}},									-- 3 - romanized (extra)
		{'<5p4n>(%s)', {native}},												-- 4 - native
		{'<5p4n>(%s, %s)', {native, extra}},									-- 5 - (native, extra)
		{'%s<5p4n> (%s)', {romanized, native}},									-- 6 - romanized (native)
		{'%s<5p4n> (%s, %s)', {romanized, native, extra}},						-- 7 - romanized (native, extra)
		{'%s<5p4n>', {english}},												-- 8 - english
		{'%s<5p4n> (%s)', {english, extra}},									-- 9 - english (extra)
		{'%s<5p4n> (%s)', {romanized, english}},								-- 10 - romanized (english)
		{'%s<5p4n> (%s, %s)', {romanized, english, extra}},						-- 11 - romanized (english, extra)
		{'%s<5p4n> (%s)', {english, native}},									-- 12 - english (native)
		{'%s<5p4n> (%s, %s)', {english, native, extra}},						-- 13 - english (native, extra)
		{'%s<5p4n> (%s, %s)', {romanized, native, english}},					-- 14 - romanized (native, english)
		{'%s<5p4n> (%s, %s, %s)', {romanized, native, english, extra}},			-- 15 - romanized (native, english, extra)
		}

	local ret_string = renderer (args, formatting, extra2)
	ret_string = ret_string:gsub ('<5p4n>', '<span style="font-weight: normal">'):gsub ('</5p4n>', '</span>');	-- replace 'secret' tags with proper tags
	return ret_string;															-- because gsub returns the number of replacements made as second return value
end

--[[--------------------------< C O M M O N _ N A T I V E _ R O M A N I Z E D _ T R A N S L A T E D >----------

Common support for {{nihongo krt}}

render order: is native, romanized, translated (English)

<template> is used to select the appropriate cfg table

]]

local function common_native_romanized_translated (frame, template)
	local lang_module = require ('Module:Lang' .. (frame:getTitle():match ('/sandbox') or ''));	-- if this module is the sandbox, use Module:lang/sandbox; Module:Lang else
	local args = require ('Module:Arguments').getArgs (frame);
	
	local english, native, romanized, extra, extra2 = args[1], args[2], args[3], args.extra or args[4], args.extra2 or args[5];	-- meaningful names
	args[4] = extra or args[4];													-- ensure that extra is 'positional' for use by renderer()

	if not (native or romanized) then											-- not present, return an error message
		return error_message (template);
	end
	native = native and lang_module._lang ({cfg[template].tag, native, ['template']=template}) or nil;
	romanized = romanized and lang_module._xlit ({cfg[template].tag, cfg[template].system, romanized, ['template']=template}) or nil;
	
	local formatting = {														-- <5p4n> and </5p4n>: place holders for font-weight style spans; akin to stripmarkers, replaced  before function returns
		{'<5p4n>(%s)', {extra}}, 												-- 1 - (extra)
		{'%s<5p4n>', {romanized}},												-- 2 - romanized
		{'%s<5p4n> (%s)', {romanized, extra}},									-- 3 - romanized (extra)
		{'<5p4n>%s', {native}},													-- 4 - native
		{'<5p4n>%s (%s)', {native, extra}},										-- 5 - native (extra)
		{'<5p4n>%s (%s)', {native, romanized}},									-- 6 - native (romanized)
		{'<5p4n>%s (%s, %s)', {native, romanized, extra}},						-- 7 - native (romanized, extra)
		{'%s<5p4n>', {english}},												-- 8 - english
		{'%s<5p4n> (%s)', {english, extra}},									-- 9 - english (extra)
		{'%s<5p4n> (%s)', {romanized, english}},								-- 10 - romanized (english)
		{'%s<5p4n> (%s, %s)', {romanized, english, extra}},						-- 11 - romanized (english, extra)
		{'<5p4n>%s (%s)', {native, english}},									-- 12 - native (english)
		{'<5p4n>%s (%s, %s)', {native, english, extra}},						-- 13 - native (english, extra)
		{'<5p4n>%s (%s, %s)', {native, romanized, english}},						-- 14 - native (romanized, english)
		{'<5p4n>%s (%s, %s, %s)', {native, romanized, english, extra}},			-- 15 - native (romanized, english, extra)
		}

	local ret_string = renderer (args, formatting, extra2)
	ret_string = romanized_kern (ret_string, romanized);						-- kern romanized text when appropriate
	ret_string = ret_string:gsub ('<5p4n>', '<span style="font-weight: normal">'):gsub ('</5p4n>', '</span>');	-- replace 'secret' tags with proper tags
	return ret_string;															-- because gsub returns the number of replacements made as second return value
end


--[[--------------------------< C O M M O N _ F O O T >--------------------------------------------------------

Common support for {{nihongo foot}}

render order: is English<ref>native, romanized</ref>

<template> is used to select the appropriate cfg table

]]

local function common_foot (frame, template)
	local lang_module = require ('Module:Lang' .. (frame:getTitle():match ('/sandbox') or ''));	-- if this module is the sandbox, use Module:lang/sandbox; Module:Lang else
	local args = require ('Module:Arguments').getArgs (frame);
	
	local english, native, romanized, extra, extra2 = args[1], args[2], args[3], args.extra or args[4], args.extra2 or args[5];	-- meaningful names
	args[4] = extra or args[4];													-- ensure that extra is 'positional' for use by renderer()
	local post = args[6] or args.post;
	local group = args.group;
	local ref_name = args.ref_name
	local lead = require ('Module:yesno')(args.lead);							-- make boolean

	if not (native or romanized) then											-- not present, return an error message
		return error_message (template);
	end
	if native then
		native = lead and lang_module._langx ({['code']=cfg[template].tag, ['text']=native, ['template']=template}) or
			lang_module._lang ({cfg[template].tag, native, ['template']=template});	-- add ja script with/without language prefix
	end
	if romanized then
		romanized = (lead and (cfg[template].system_link .. ': ') or '') .. lang_module._xlit ({'ja', cfg[template].system, romanized, ['template']=template}) or nil;
	end
	
	local formatting = {
		{'%s', {extra}}, 														-- 1 - extra
		{'%s', {romanized}},													-- 2 - romanized
		{'%s, %s', {romanized, extra}},											-- 3 - romanized, extra
		{'%s', {native}},														-- 4 - native
		{'%s, %s', {native, extra}},											-- 5 - native, extra
		{'%s, %s', {native, romanized}},										-- 6 - native romanized
		{'%s, %s, %s', {native, romanized, extra}},								-- 7 - native romanized, extra
																				-- from here english is used in the mapping but not rendered by renderer so not included in the table
		{'', {''}},																-- 8 - english
		{'%s', {extra}},														-- 9 - extra
		{'%s', {romanized}},													-- 10 - romanized
		{'%s, %s', {romanized, extra}},											-- 11 - romanized, extra
		{'%s', {native}},														-- 12 - native
		{'%s, %s', {native, extra}},											-- 13 - native, extra
		{'%s, %s', {native, romanized}},										-- 14 - native romanized
		{'%s, %s, %s', {native, romanized, extra}},								-- 15 - native romanized, extra
		}

	if english and post then													-- rewrite english to include |post=
		english = english .. post;												-- if english has a value append post else just post
	elseif post then
		english = post;															-- english not set, use post
	elseif not english then														-- neither are set
		english = '';															-- make english an empty string for concatenation
	end

	if native or romanized or extra or extra2 then								-- no ref tag when none of these are set (it would be empty)
		local content = renderer (args, formatting, extra2);
		content = content:gsub ('<5p4n>', ''):gsub ('</5p4n>$', '', 1);			-- strip secret <5p4n> and </5p4n> tags added by renderer(); spans not used by this template

		if english:match ('\'\'+$') then										-- if <english> is italic, bold, or both
			local text = english:gsub ('%b<>', ''):gsub ('\'\'+', ''):gsub ('%[%[', ''):gsub ('%]%]', '');	-- strip markup
			if text:match ('[dfijkltCEFHIJKMNPR-Z\'"%?!%]]$') then				-- when <english> ends with one of these characters
				english =  '<span style="margin-right:.09em">' .. english .. '</span>';	-- add kerning
			end
		end

		return english .. frame:extensionTag ({name='ref', args={group=group, name=ref_name}, content=content});	-- english with attached reference tag
	else
		return english;															-- nothing to be inside ref tag so just return english
	end
end


--[=[-------------------------< N I H O N G O >----------------------------------------------------------------

Implement {{nihongo}} using Module:Lang for language and transliteration markup

{{Nihongo|<English>|<japanese>|<romaji>|<extra>|<extra2>|lead=yes}}

<English>, <japanese>, and <romaji> are positional parameters
	<English>: rendered as presented; purports to be English translation of <kanji/kana>
	<japanese>: Japanese language text using Japanese script; TODO: require?
	<romaji>: Hepburn romanization (transliteration); TODO: in Module:Lang/data change tooltip text to 'Hepburn romanization'?
<extra> and <extra2> are positional or named: |extra= and |extra2=; mixing can be problematic
	<extra> is rendered as presented preceeded with <comma><space>
	<extra2> is rendered as presented preceeded with <space>
|lead=: takes one value 'yes'; renders language name same as {{langx|ja}} but also adds [[Hepburn romanization|Hepburn]]:<space> ahead of the romanization; TODO: in Module:Lang, turnoff tooltip for transl when |lead=yes

]=]

local function nihongo (frame)
	return common (frame, 'nihongo')
end


--[=[-------------------------< N I H O N G O 3 >--------------------------------------------------------------

Implement {{nihongo3}} using Module:Lang for language and transliteration markup

Similar to {{nihongo}} but changes rendered order and does not support |lead=

{{Nihongo3|<English>|<japanese>|<romaji>|<extra>|<extra2>}}

<English>, <japanese>, and <romaji> are positional parameters
	<English>: rendered as presented; purports to be English translation of <kanji/kana>
	<japanese>: Japanese language text using Japanese script; TODO: require?
	<romaji>: Hepburn romanization (transliteration); TODO: in Module:Lang/data change tooltip text to 'Hepburn romanization'?
<extra> and <extra2> are positional or named: |extra= and |extra2=; mixing can be problematic
	<extra> is rendered as presented preceded with <comma><space>
	<extra2> is rendered as presented preceded with <space>

]=]

local function nihongo3 (frame)
	return common_romanized_native_translated (frame, 'nihongo3')
end


--[=[-------------------------< N I H O N G O _ K R T >--------------------------------------------------------

Implement {{nihongo krt}} using Module:Lang for language and transliteration markup

Similar to {{nihongo}} but changes rendered order and does not support |lead=

{{Nihongo krt|<English>|<japanese>|<romaji>|<extra>|<extra2>}}

<English>, <japanese>, and <romaji> are positional parameters
	<English>: rendered as presented; purports to be English translation of <kanji/kana>
	<japanese>: Japanese language text using Japanese script; TODO: require?
	<romaji>: Hepburn romanization (transliteration); TODO: in Module:Lang/data change tooltip text to 'Hepburn romanization'?
	<extra> and <extra2> are positional or named: |extra= and |extra2=; mixing can be problematic
	<extra> is rendered as presented preceded with <comma><space>
	<extra2> is rendered as presented preceded with <space>

]=]

local function nihongo_krt (frame)
	return common_native_romanized_translated (frame, 'nihongo krt')
end


--[=[-------------------------< N I H O N G O _ F O O T >------------------------------------------------------

Implement {{nihongo_foot}} using Module:Lang for language and transliteration markup

{{Nihongo foot|<English>|<japanese>|<romaji>|<extra>|<extra2>|<post>|lead=yes|group}}

<English>, <japanese>, and <romaji> are positional parameters
	<English>: rendered as presented; purports to be English translation of <kanji/kana>
	<japanese>: Japanese language text using Japanese script; TODO: require?
	<romaji>: Hepburn romanization (transliteration); TODO: in Module:Lang/data change tooltip text to 'Hepburn romanization'?
<extra> and <extra2> are positional or named: |extra= and |extra2=; mixing can be problematic
	<extra> is rendered as presented preceded with <comma><space>
	<extra2> is rendered as presented preceded with <space>
<post> is positional or named: |post= is a postscript character preceding the <ref>..</ref> tag (after <English>)
|lead=: takes one value 'yes'; renders language name same as {{langx|ja}} but also adds [[Hepburn romanization|Hepburn]]:<space> ahead of the romanization;
	TODO: in Module:Lang, turnoff tooltip for transl when |lead=yes
	in the live template |lead= also adds the Help:Installing Japanese character sets link; this is not supported in this code (nihongo nor nihongo3 have this support)
|group=: the group attribute in <ref group="..."> and in {{reflist}}

]=]

local function nihongo_foot (frame)
	return common_foot (frame, 'nihongo foot')
end


--[[--------------------------< E X P O R T E D   F U N C T I O N S >------------------------------------------
]]

return {
	nihongo = nihongo,
	nihongo3 = nihongo3,
	nihongo_krt = nihongo_krt,
	nihongo_foot = nihongo_foot,
	}