モジュール:ja

このモジュールについての説明文ページを モジュール:ja/doc に作成できます

local export = {}

local titleObj = mw.title.getCurrentTitle()
local pagename = titleObj.text
local namespace = titleObj.nsText

local str_gsub = string.gsub
local find = mw.ustring.find
local length = mw.ustring.len
local trim = mw.text.trim
local split = mw.text.split
local sub, gsub = mw.ustring.sub, mw.ustring.gsub
local match, gmatch = mw.ustring.match, mw.ustring.gmatch
local to_cp, to_char = mw.ustring.codepoint, mw.ustring.char

local Jpan = require("Module:scripts").getByCode("Jpan")
local lang = require("Module:languages").getByCode("ja")

-- note that arrays loaded by mw.loadData cannot be directly used by gsub
local data = mw.loadData("Module:ja/data")

-- Unicode normalization often converts these to the corresponding CJK Unified Ideographs characters
local compat_ideo = mw.ustring.char(0xF900) .. "-" .. mw.ustring.char(0xFAD9)

export.data = {
	joyo_kanji = data.joyo_kanji,
	jinmeiyo_kanji = data.jinmeiyo_kanji,
	grade1 = data.grade1,
	grade2 = data.grade2,
	grade3 = data.grade3,
	grade4 = data.grade4,
	grade5 = data.grade5,
	grade6 = data.grade6
}

local function track(code)
	if type(code) ~= "string" then
		error("The track function requires a string as argument.")
	end
	require("Module:debug").track("ja/" .. code)
end

local function change_codepoint(added_value)
	return function(char)
		return to_char(to_cp(char) + added_value)
	end
end

function export.enlarge(text)
	return  text
end

function export.hira_to_kata(text)
	if type(text) == "table" then text = text.args[1] end

	return (gsub(text, '[ぁ-ゖ]', change_codepoint(96)))
end

function export.kata_to_hira(text)
	if type(text) == "table" then text = text.args[1] end

	return (gsub(text, '[ァ-ヶ]', change_codepoint(-96)))
end

function export.fullwidth_to_halfwidth(text)
	if type(text) == "table" then text = text.args[1] end

	text = gsub(text, '　', ' ')
	return (gsub(text, '[！-～]', change_codepoint(-65248)))
end

function export.kana_to_romaji(text, options)
	-- options: no_diacritics, keep_period, hist, phonetic
	local str_find = string.find

	if type(text) == "table" then
		text = text.args[1]
	end

	if not options then options = {} end

	local tracking_has_percent = str_find(text, '%%')
	local text_old = ''
	if not options.hist then
		text_old = trim(require('Module:ja/k2r-old').kana_to_romaji(text, options.no_diacritics, options.keep_period))
	end

	-- conversions
	if not options.phonetic then
		text = gsub(text, '(%-)([はハ])$', '%1㊟㈛㊟%2') -- は as suffix (派 "-ha", etc.) and appearing at the end of string
		text = gsub(text, '(%-)([はハ]) ', '%1㊟㈛㊟%2 ') -- は as suffix and appearing mid-sentence
	end
	
	text = str_gsub(text, '％', '㊟㌫㊟') -- at [[見込む]], for example; avoid collision with % used in our ruby syntax
	text = str_gsub(text, '\'\'\'', '㊟⒝㊟')
	text = str_gsub(text, '<u>', '㊟㋑⒰㊟')
	text = str_gsub(text, '</u>', '㊟㋺⒰㊟')
	
	local text_styling = "㊟[㋑㋺⒝⒰]+㊟"
	
	-- avoid tampering with existing latin text: store it away
	local escape = {}
	local id = 0
	for latin in string.gmatch(text, "[a-z]+") do
		escape[id] = latin
		text = str_gsub(text, latin, "㊟㊕㊕㊟" .. id .. "㊟㊕㊕㊟")
		id = id + 1
	end

	-- special preformatting
	text = str_gsub(text, 'ヶげつ', 'かげつ')
	text = gsub(text, 'ヶ(' .. text_styling .. ')げつ', 'か%1げつ') -- 「'''ヶ'''げつ」
	text = str_gsub(text, 'ヶ', 'が')
	text = str_gsub(text, 'ヵ', 'か')
	text = gsub(text, '(.)[ゝヽ]', '%1%1')
	text = gsub(text, '(.)[ゞヾ]', function(char) return mw.ustring.toNFC(char .. char .. '゙') end) -- unicode hax

	-- [[Wiktionary:Grease_pit/2017/May#Formatting_for_individual_Japanese_readings]]
	if options.hist then
		text = gsub(text, '[やゆよわゐゑを]', '㊟⒳㊟%0')
		text = gsub(text, '.',
			{
				['つ'] = 'tu',
				['ち'] = 'ti',
				['づ'] = 'du',
				['ぢ'] = 'di',
				['し'] = 'si',
				['じ'] = 'zi',
				['ゐ'] = 'wi',
				['を'] = 'wo',
			}
		)
	end

	text = export.hira_to_kata(text)
	text = gsub(text, '.', data.kr)
	text = export.fullwidth_to_halfwidth(text)

	if options.hist then
		text = str_gsub(text, 'oo', 'o.o')
		text = str_gsub(text, 'ou', 'o.u')
		text = str_gsub(text, 'h', 'f')
		
		local old = text
		text = str_gsub(text, 'i㊟⒳㊟y', 'y') -- くゐやう kwyau
		text = str_gsub(text, '([kg])u㊟⒳㊟w', '%1w') 
		if old ~= text then
			--[=[
				There may be cases in which i or u is deleted incorrectly, and a
				period should be inserted.
				"Syncope" isn't quite accurate, as there wasn't a sound change.
				It's just an orthographic convention.
				[[Special:WhatLinksHere/Template:tracking/ja/mora syncope]]
			]=]
			mw.log(str_gsub(old, '㊟⒳㊟', '')  .. ' → ' .. str_gsub(text, '㊟⒳㊟', '') )
			track('mora syncope')
		end
		
		text = str_gsub(text, '㊟⒳㊟', '') -- ゑつ wetsu
	end

	-- markup
	text = str_gsub(text, '%%', '.') -- ruby "percent sign" syntax
	text = gsub(text, '([ッ¤])%.', '%1') -- 「し を ぼっ.す」; 「るい%じん%えん」→「rui.jin¤.en¤」

	-- 「テェェェ」→「テェーー」 (avoid funky romaji effected by the "(テュ→)teユ→tyu" line below)
	local kogaki_vowels = {'ァ','ィ','ゥ','ェ','ォ'}
	for _, char in ipairs(kogaki_vowels) do
		text = gsub(text, '('..char..')('..char..'+)', function(a,b) return a .. mw.ustring.rep('ー', length(b)) end)
	end

	-- (ゲェ→)geェ→gee (note that this causes things like ウゥ→ū and ギィ→gī)
	text = gsub(text, '[aiueo][ァィゥェォ]', {['aァ']='aa',['iィ']='ii',['uゥ']='uu',['eェ']='ee',['oォ']='oo',})

	-- (クヮ→)kuヮ→kwa
	text = gsub(text, '[u]([ヮ])', {['ヮ']='wa',})

	-- (クァ→)kuァ→kwa, (トァ→)toァ→twa, (ウィ→)uィ→wi
	text = gsub(text, '[uo]([ァィェォ])', {['ァ']='wa',['ィ']='wi',['ェ']='we',['ォ']='wo',})
	if not options.hist then
		-- (ツァ→)cwa→ca
		text = str_gsub(text, '([fvcsz])w', '%1')
	end

	-- (テュ→)teユ→tyu, (ギェ→)giェ→gye
	text = gsub(text, '[aiueo]([ャュェョ])', {['ャ']='ya',['ュ']='yu',['ェ']='ye',['ョ']='yo',})
	-- (ジュ→)jyu→ju
	text = gsub(text, '([xjq])y', '%1')

	-- (ティ→)teィ→ti (essentially forget about the vowel in between)
	text = gsub(text, '[aiueo]([ァィゥェォ])', {['ァ']='a',['ィ']='i',['ゥ']='u',['ェ']='e',['ォ']='o',})

	-- chouonpu and sokuon
	while str_find(text, '[aiueo]ー') or str_find(text, 'ッ *[bcdfghjklmnpqrstvwxyz]') or find(text, 'ッ' .. text_styling .. '[bcdfghjklmnpqrstvwxyz]') do
		text = str_gsub(text, '([aiueo])ー', '%1%1')
		text = str_gsub(text, 'ッ( *)([bcdfghjklmnpqrstvwxyz])', '%2%1%2')
		text = gsub(text, 'ッ(' .. text_styling .. ')([bcdfghjklmnpqrstvwxyz])', '%2%1%2')
	end
	-- deal with leftover sokuon not used as geminate
	text = str_gsub(text, 'ッ', '&apos;')

	-- (ん→)n¤
	text = str_gsub(text, '¤([aiueoy])', "'%1")
	text = str_gsub(text, '¤', '')

	-- は, へ
	if not options.phonetic and str_find(text, "h[ae]") then
		for i, v in ipairs{
			{ "ha", "wa" },
			{ "he", "e" }
		} do
			local thingy = '[^a-z.㊟-]' -- not sure what this should be named
			text = gsub(text, "(" .. thingy .. ")" .. v[1] .. "(" .. thingy .. ")", "%1" .. v[2] .. "%2")
			text = gsub(text, "(" .. thingy .. ")" .. v[1] .. "$", "%1" .. v[2])
			text = gsub(text, "^" .. v[1] .. "(" .. thingy .. ")", v[2] .. "%1")
			if find(text, text_styling) then
				text = gsub(text, "(" .. thingy .. ")" .. v[1] .. "(" .. text_styling .. thingy ..")", "%1" .. v[2] .. "%2")
				text = gsub(text, "(" .. thingy .. ")" .. v[1] .. "(" .. text_styling .. ")$", "%1" .. v[2] .. "%2")
				text = gsub(text, "(" .. thingy .. text_styling .. ")" .. v[1] .. "(" .. text_styling .. thingy ..")", "%1" .. v[2] .. "%2")
				text = gsub(text, "(" .. thingy .. text_styling .. ")" .. v[1] .. "(" .. text_styling .. ")$", "%1" .. v[2] .. "%2")
			end
		end
	end
	-- change only when
	--   ① not flanked by a-z or a period ("^sore wa nani$", "^hyappou no .he hitotsu$")
	--   ② at the end of the string and not preceded by a-z or a period ("^are wa$")
	--   ③ at the beginning of the string and not followed by a-z or a period ("^he ikou$") [not sure this is actually necessary, but I suppose it is consistent with ②]
	-- this also means that "^ha$" becomes "ha"
	-- period can be used next to the kana (either side) to force the "dumb" romanization (i.e. "ha", "he")

	-- fix sh, ch, ts
	local function handle_digraphs(geminate, intervening, main, following)
		--「めちゃ」→「mecha」
		--「めっちゃ」→「metcha」
		--「めっっちゃ」→「mettcha」
		local corresp_geminate_form = {['x']='s',['q']='t',['c']='t'}
		local corresp_main = {['x']='sh',['q']='ch',['c']='ts'}
		
		local geminate_repl, main_repl
		
		-- So as not to convert ch to tsh.
		if not following or main .. following ~= "ch" then
			main_repl = corresp_main[main]
		end
		
		if geminate ~= "" then
			geminate_repl = string.rep(corresp_geminate_form[main], #geminate)
		end
		
		return (geminate_repl or geminate) .. (intervening or "") .. (main_repl or main) .. (following or "")
	end
	
	local function handle_digraphs2(geminate, main, following)
		return handle_digraphs(geminate, nil, main, following)
	end
	
	text = gsub(text, '([xqc]+)(' .. text_styling .. ')([xqc])', handle_digraphs)
	text = gsub(text, '([xqc]*)([xqc])(.?)', handle_digraphs2)
	

	-- macrons
	-- Will cause problems if combined vowel-macron characters are used below.
	if not options.no_diacritics then
		if not options.phonetic then
			text = str_gsub(text, 'ou', 'ō')
		end
		local macron = to_char(0x304)
		text = str_gsub(
			text,
			'([oaeui])%1',
			'%1' .. macron
		)
	end

	-- remove markup and convert real periods
	if not options.keep_period then
		text = str_gsub(text, '%.', '')
		text = str_gsub(text, '。', '◆.◇')
	end

	-- 
	text = str_gsub(text, '◇◆', '')
	text = str_gsub(text, '◆◇', '')
	text = str_gsub(text, ' *◆ *', '')
	text = str_gsub(text, ' *◇ *', ' ')

	-- restore latin text
	text = str_gsub(text, "㊟㊕㊕㊟(%d+)㊟㊕㊕㊟", function(id) return escape[tonumber(id)] end)

	-- clean up spaces
	text = trim(text)
	text = str_gsub(text, ' +', ' ')
	
	-- remove double ampersands used in ruby
	text = str_gsub(text, '&&(.-)&&', '%1')

	-- uppercase markup
	text = str_gsub(text, "(%^)(㊟⒝㊟)", "%2%1") -- move ^ to an effective position if placed before bold markup
	text = str_gsub(text, "(%^)( )", "%2%1") -- same but with spaces
	text = gsub(text, '%^(.)', mw.ustring.upper) -- uppercase conversion

	-- clean up spaces again
	text = str_gsub(text, ' +', ' ')

	-- conversions
	text = str_gsub(text, '㊟⒝㊟', '\'\'\'')
	text = str_gsub(text, '㊟㋑⒰㊟', '<u>')
	text = str_gsub(text, '㊟㋺⒰㊟', '</u>')
	text = str_gsub(text, '㊟㈛㊟', '')
	text = str_gsub(text, '㊟㌫㊟', '%%')

	-- unicode NFC
	text = mw.ustring.toNFC(text)

	-- comparison with old "to romaji" code
	if not options.hist then
		-- comparison with old kana_to_romaji() code
		if text ~= text_old then
			if str_find(text_old, '%(ba%)') then -- adjective conjugation table :/
				track('k2r diff ba')
			elseif mw.ustring.lower(text) == mw.ustring.lower(text_old) then
				track('k2r diff caps')
			elseif str_find(text_old, 'ッ') then
				track('k2r diff w xtu')
			elseif tracking_has_percent then
				track('k2r diff pc')
			elseif gsub(text, '%s*([“”!?])%s*', '%1') == gsub(text_old, '%s*([”!?])%s*', '%1') then
				track('k2r diff punc spacing')
			else
				track('k2r diff')
			end
			mw.log('new]' .. text .. '[')
			mw.log('old]' .. text_old .. '[')
		end
	end

	if find(text, '[ぁ-ー]') then
		track('k2r failure')
	end
	
	return text
end

-- removes spaces and hyphens from input
-- intended to be used when checking manual romaji to allow the
-- insertion of spaces or hyphens in manual romaji without appearing "wrong"
function export.rm_spaces_hyphens(f)
	local text = type(f) == 'table' and f.args[1] or f
	text = str_gsub(text, '.', { [' '] = '', ['-'] = '', ['.'] = '', ['\''] = '' })
	text = str_gsub(text, '&nbsp;', '')
	return text
end

function export.romaji_to_kata(f)
	local text = type(f) == 'table' and f.args[1] or f
	text = gsub(text, '.', data.rd)
	text = str_gsub(text, '(.)%1', {
		k = 'ッk', s = 'ッs', t = 'ッt', p = 'ッp',
		b = 'ッb', d = 'ッd', g = 'ッg', j = 'ッj'
	})
	text = str_gsub(text, 'tc', 'ッc')
	text = str_gsub(text, 'tsyu', 'ツュ')
	text = str_gsub(text, 'ts[uoiea]', {['tsu']='ツ',['tso']='ツォ',['tsi']='ツィ',['tse']='ツェ',['tsa']='ツァ'})
	text = str_gsub(text, 'sh[uoiea]', {['shu']='シュ',['sho']='ショ',['shi']='シ',['she']='シェ',['sha']='シャ'})
	text = str_gsub(text, 'ch[uoiea]', {['chu']='チュ',['cho']='チョ',['chi']='チ',['che']='チェ',['cha']='チャ'})
	text = str_gsub(text, "n[uoiea']?", {['nu']='ヌ',['no']='ノ',['ni']='ニ',['ne']='ネ',['na']='ナ'})
	text = str_gsub(text, '[wvtrpsnmlkjhgfdbzy][yw]?[uoiea]', data.rk)
	text = str_gsub(text, "n'?", 'ン')
	text = str_gsub(text, '[aeiou]', {
		u = 'ウ', o = 'オ', i = 'イ', e = 'エ', a = 'ア'
	})
	return text
end

-- expects: any mix of kanji and kana
-- determines the script types used
-- e.g. given イギリス人, it returns Kana+Hani
function export.script(f)
	text, script = type(f) == 'table' and f.args[1] or f, {}

	if find(text, '[ぁ-ゖ]') then
		table.insert(script, 'Hira')
	end
	-- TODO: there are two kanas.  This should insert Kata.
	if find(text, '[ァ-ヺー]') then
		table.insert(script, 'Kana')
	end
	-- 一 is unicode 4e00, previously used 丁 is 4e01
	if find(text, '[㐀-䶵一-鿌' .. compat_ideo .. '𠀀-𯨟]') then
		table.insert(script, 'Hani')
	end
	-- matching %a should have worked but matched the end of every string
	if find(text, '[a-zA-ZāēīōūĀĒĪŌŪａ-ｚＡ-Ｚ]') then
		table.insert(script, 'Romaji')
	end
	if find(text, '[0-9０-９]') then
		table.insert(script, 'Number')
	end
	if find(text, '[〆々]') then
		table.insert(script, 'Abbreviation')
	end

	return table.concat(script, '+')
end

-- when counting morae, most small hiragana belong to the previous mora,
-- so for purposes of counting them, they can be removed and the characters
-- can be counted to get the number of morae.  The exception is small tsu,
-- so data.nonmora_to_empty maps all small hiragana except small tsu.
function export.count_morae(text)
	if type(text) == "table" then
		text = text.args[1]
	end
	-- convert kata to hira (hira is untouched)
	text = export.kata_to_hira(text)
	-- remove all of the small hiragana such as ょ except small tsu
	text = gsub(text,'.',data.nonmora_to_empty)
	-- remove zero-width spaces
	text = gsub(text, '‎', '')
	-- return number of characters, which should be the number of morae
	return length(text)
end

-- accepts: any mix of kana
-- returns: a hiragana sort key designed for WMF software
-- this is like sort() but doesn't return |sort=sortkey,
-- just the sort key itself, but unlike sort(), this
-- replaces the long vowel mark with its vowel
function export.jsort(text)
	if text == nil then
		return nil
	end
	
	if type(text) == "table" then
		text = text.args[1]
	end
	local sort2 = text
	
	-- remove western spaces, hyphens, and periods
	-- diff=41967612: also remove caret
	sort2 = gsub(sort2, '[ %-%.%^・]', '')

	sort2 = export.kata_to_hira(sort2)

	-- replace the long vowel mark with the vowel that it stands for
	for key,value in pairs(data.longvowels) do
		sort2 = gsub(sort2,key,value)
	end

	-- 濁音・半濁音・小さい文字を基本となる字に変換
	local sort1 = sort2
	for key,value in pairs(data.basechars) do
		sort1 = gsub(sort1,key,value)
	end

	if sort1 == sort2 then
		return sort1
	else
		return sort1 .. ' ' .. sort2
	end
end

-- returns a sort key with |sort= in front, e.g.
-- |sort=はつくん ばつぐん if given ばつぐん
function export.sort(f)
	local text = type(f) == 'table' and f.args[1] or f
	local sort = export.jsort(text)
	return ('|' .. 'sort' .. '=' .. sort)
end

-- returns the "stem" of a verb or -i adjective, that is the term minus the final character
function export.definal(f)
	return sub(f.args[1],1,(length(f.args[1])-1))
end

function export.remove_ruby_markup(text)
	return string.gsub(text, "[%^%-%. %%]", "")
end

-- see also Template:JAruby
-- meant to be called from another module
function export.add_ruby_backend(term, kana, from_ja_link)
	if term == kana then
		return export.enlarge(term)
	end
	local pattern = ""
	-- holds the whole segments of markup enclosed in <ruby>...</ruby>
	local ruby_markup = {}
	-- range of kana: '[ぁ-ゖァ-ヺ]'
	-- nonkana: [^ぁ-ゖァ-ヺ]
	local kanji_pattern = "[A-Za-z0-9々㐀-䶵一-鿌" .. compat_ideo .. "𠀀-𯨟０-９Ａ-Ｚａ-ｚ]"
	
	local str_find = string.find
	local str_gsub = string.gsub
	
	-- If term and kana aren't identical, then term should contain stuff that can
	-- have kana put above it.
	if not find(term, kanji_pattern) then
		require("Module:debug").track("ja/ruby/no kanji")
		-- error("No kanji or other ruby-annotatable characters in first argument of add_ruby_backend.")
		return export.enlarge(term)
	end
	
	local orig_term = term
	
	-- Escape decimal numeric entities (such as &#32; representing a space) and HTML tags, which do not need ruby annotations.
	local function escape(t, i)
		return function(text, patt)
			return str_gsub(text, patt,
				function(capture)
					i = i + 1
					t[i] = capture
					return "㊟" .. string.rep("&", i) .. "㊟"
				end)
		end
	end
	
	local escaped1 = {}
	local escape1 = escape(escaped1, 0)
	
	term = escape1(term, "(&&.-&&)")
	term = escape1(term, "(&#x?%d+;)")
	term = escape1(term, "(&[A-Za-z]+;)")
	term = escape1(term, "(<[^>]+>)")
	
	local escaped2 = {}
	local escape2 = escape(escaped2, 0)
	
	kana = escape2(kana, "(&&.-&&)")
	kana = escape2(kana, "(&#%d+;)")
	kana = escape2(kana, "(&[A-Za-z]+;)")
	kana = escape2(kana, "(<[^>]+>)")
	
	if escaped1[1] or escaped2[1] then
		-- [[Special:WhatLinksHere/Template:tracking/ja/ruby/escaped]]
		require("Module:debug").track("ja/ruby/escaped")
	end
	
	if #escaped2 ~= #escaped1 then
		local relationship
		if #escaped1 > #escaped2 then
			relationship = "more"
		else
			relationship = "fewer"
		end
		
		mw.logObject(escaped1)
		mw.logObject(escaped2)
		error("There are " .. relationship .. " escaped things in the text to be annotated than in the ruby text.")
	end
	
	-- links without pipes will fail
	term = str_gsub(term, '%[%[([^|%]]+)%]%]', '[[%1|%1]]')

	-- remove links from kana
	kana = str_gsub(kana, '%[%[([^|%]]+)%]%]', '%1')
	kana = str_gsub(kana, '%[%[[^%]]+|([^%]]+)%]%]', '%1')

	-- build up pattern
	-- escape the magic characters in the term
	pattern = str_gsub(term, '%[%[[^%]]+|([^%]]+)%]%]', '%1')
	pattern = require("Module:string").pattern_escape(pattern)

	pattern = str_gsub(pattern, "[%[%]]+", " *")
	kana = str_gsub(kana, "[%[%]]+", '')
	pattern = str_gsub(pattern, " *('+) *", "%1")
	kana = str_gsub(kana, " *('+) *", "%1")
	pattern = str_gsub(pattern, " +", " ")
	kana = str_gsub(kana, " +", " ")

	-- remove periods and caret signs and hyphens
	pattern = str_gsub(pattern, '%%[.^-]', '')
	kana = str_gsub(kana, '[.^-]', '')

	-- in order to make a pattern that will find the ruby,
	-- replace every unbroken string of kanji with a sub-pattern
	
	-- mw.log("before adding (..-):", pattern)
	pattern = gsub(pattern, kanji_pattern .. '+', '(..-)')
	-- get a pattern like
	-- (.+)ばか(.+)ばか(.+)ばかばかばああか(.+) when given 超ばか猿超ばか猿超ばかばかばああか猿
	-- it turns out we need to keep the spaces sometimes
	-- so that kana don't "leak" in ambiguous cases like 捨すてて撤退 where it's not clear if it's
	-- す, てったい or すて, ったい.  only solution now is to put spaces in the "term" param
	-- if they fall between kana

	-- build up term (e.g. [[歌う|歌った]])
	local replaced = {}
	local count = 0
	term = str_gsub(term, '%]', '%]') -- escape the "]" character so that it cannot appear, example becomes [[歌う|歌った%]%]
	term = gsub(
		str_gsub(term, "%-", ""),
		kanji_pattern .. '+',
		function(text)
			count = count + 1
			-- remove spaces
			text = str_gsub(text, ' ', '')
			table.insert(replaced, text)
			return '[' .. count .. ']'
		end
	) -- example becomes [[[1]う|[2]った%]%]

	
	while str_find(term, '%[%[[^|]*%[%d+%][^|]*|') do
		term = str_gsub(
			term,
			'(%[%[[^|]*)%[(%d+)%]([^|]*|)',
			function(a,b,c)
				return a .. replaced[tonumber(b)] .. c
			end
		)
	end
	 -- example becomes [[歌う|[2]った%]%]
	
	-- anchor pattern at ends of string, because quantifiers are non-greedy
	pattern = "^" .. pattern .. "$"
	
	-- apply that pattern to the kana to collect the rubies
	-- if this fails, try it without spaces
	if find(kana, pattern) == nil then
		kana = str_gsub(kana, ' ', '')
		if not find(kana, pattern) then
			mw.log("failed match:", "\n", kana, #kana, "\n", pattern, #pattern, "\n", orig_term, #orig_term)
			error("The pattern did not match the kana.")
		end
	end
	
	local ruby = { match(kana, pattern) }
	if require("Module:fun").some(function(kana) return kana:find("^ +$") end, ruby) then
		-- [[Special:WhatLinksHere/Template:tracking/ja/ruby spaces]]
		track('ruby spaces')
		mw.log('One of the kana in the ruby table derived from ' .. kana .. ' contains nothing but spaces.')
	end
	-- local ruby = {}
	-- for c in gmatch(kana, pattern) do table.insert(ruby, c) end

	-- find the kanji strings again and combine them with their ruby to make the <ruby> markup
	local kanji_segments = {}
	for c in string.gmatch(term, '%[(%d+)%]') do
		table.insert(kanji_segments, replaced[tonumber(c)])
	end
	
	if #kanji_segments ~= #ruby then
		mw.logObject(kanji_segments)
		mw.logObject(ruby)
		error("There are " .. #kanji_segments .. " kanji segments but only " .. #ruby .. " ruby segments. There should be an equal number of each.")
	end
	
	for i, kanji_segment in pairs(kanji_segments) do
		if not ruby[i] then
			error('No ruby for kanji segment "' .. kanji_segment .. '".')
		end
		
		local kanji, kana
		if mw.ustring.len(kanji_segment) > 1 and mw.ustring.len(kanji_segment) == mw.ustring.len(ruby[i]) then
			-- Split kanji and kana into individual characters.
			kanji = mw.text.split(kanji_segment, "")
			kana = mw.text.split(ruby[i], "")
		else
			kanji, kana = { kanji_segment }, { ruby[i] }
		end
		
		--[[
			Apportions kana to kanji based on phonological rules.
			Moraic nasal and sokuon are grouped with the previous kana,
			then sequences containing long vowels are grouped together if
			necessary.
			This captures many of the easier cases.
			漢字(かんじ)	京都(きょうと)		一気(いっき)
				↓				↓					↓
			漢(かん)字(じ)	京(きょう)都(と)	一(いっ)気(き)
		]]
		if length(kanji_segment) > 1 and not find(ruby[i], "[ァ-ヶ]") then
			local value1, value2 = require("Module:User:Erutuon/ja").divideKana(kanji_segment, ruby[i])
			if value1 and value2 and #value1 == #value2 then
				kanji, kana = value1, value2
			end
		end
		
		local ruby_string = {}
		for i = 1, #kanji do
			-- To prevent, for instance, "stop" being annotated with "stop" in [[言う]].
			if kanji[i] == kana[i] then
				table.insert(ruby_string, kanji[i])
			else
				table.insert(ruby_string, "<ruby>" .. kanji[i] .. "<rp>&nbsp;(</rp><rt>" .. kana[i] .. "</rt><rp>) </rp></ruby>")
			end
		end
		table.insert(ruby_markup, table.concat(ruby_string))
	end

	count = 0
	term = str_gsub(term, '%[%d+%]', function()
		count = count + 1
		return ruby_markup[count]
	end)

	term = str_gsub(term, '%%%]', ']')
	term = str_gsub(term, '%%', '')
	term = str_gsub(term, ' ', '')
	
	term = str_gsub(term,
		"㊟(&+)㊟",
		function(ampersands)
			return escaped1[#ampersands]
		end)
	
	term = str_gsub(term, "&&(.-)&&", "%1")
	
	--done
	return export.enlarge(term)
end

-- do the work of Template:ja-kanji
function export.kanji(frame)
	local pagename = mw.title.getCurrentTitle().text
	-- only do this if this entry is a kanji page and not some user's page
	if find(pagename, "[㐀-䶵一-鿯" .. compat_ideo .. "𠀀-𯨟]") then
		local params = {
			grade = {},
			rs = {},
			shin = {},
			kyu = {},
			head = {},
		}
		local args = require("Module:parameters").process(frame:getParent().args, params)
		
		local rs = args["rs"] or require("Module:zh-sortkey").makeSortKey(pagename, "ja")
		local shin = args["shin"]
		local kyu = args["kyu"]
		local head = args["head"]
		
		local grade_replacements = {
			c = "7",
			n = "8",
			uc = "9",
			r = "0",
		}
		local grade = tonumber(args["grade"])
		grade = grade_replacements[grade] or grade

		local wikitext = {}
		local categories = {}

		local catsort = rs or pagename

		-- display the kanji itself at the top at 275% size
		table.insert(wikitext, '<div><span lang="ja" class="Jpan" style="font-size:275%; line-height:1;">' .. (args["head"] or pagename) .. '</span></div>')

		-- display information for the grade

		-- if grade was not specified, determine it now
		if not grade then
			grade = export.kanji_grade(pagename)
		end
		
		local in_parenthesis = {}
		local grade_links = {
			[1] = "[[w:教育漢字|第1学年の「教育」漢字]]",
			[2] = "[[w:教育漢字|第2学年の「教育」漢字]]",
			[3] = "[[w:教育漢字|第3学年の「教育」漢字]]",
			[4] = "[[w:教育漢字|第4学年の「教育」漢字]]",
			[5] = "[[w:教育漢字|第5学年の「教育」漢字]]",
			[6] = "[[w:教育漢字|第6学年の「教育」漢字]]",
			[7] = "[[w:常用漢字|「常用」漢字",
			[8] = "[[w:人名用漢字|「人名用」漢字",
			[9] = "[[w:表外漢字|「表外」漢字]]",
			[0] = "[[w:部首|部首]]",
		}
		if grade_links[grade] then
			table.insert(in_parenthesis, grade_links[grade])
		else
			table.insert(categories, "[[Category:日本語 漢字 級の欠落|" .. catsort .. "]]")
		end

		-- link to shinjitai if shinjitai was specified, and link to kyujitai if kyujitai was specified

		if kyu then
			table.insert(in_parenthesis, '[[新字体]]、[[旧字体]]：<span lang="ja" class="Jpan">[[' .. kyu .. '#日本語|' .. kyu .. ']]</span>')
		elseif shin then
			table.insert(in_parenthesis, '[[旧字体]]、[[新字体]]：<span lang="ja" class="Jpan">[[' .. shin .. '#日本語|' .. shin .. ']]</span>')
		end
		table.insert(wikitext, "''(" .. table.concat(in_parenthesis, ",&nbsp;") .. "'')")

		-- add categories
		table.insert(categories, "[[Category:日本語 漢字|" .. catsort .. "]]")
		local grade_categories = {
			[1] = "第1学年の漢字",
			[2] = "第2学年の漢字",
			[3] = "第3学年の漢字",
			[4] = "第4学年の漢字",
			[5] = "第5学年の漢字",
			[6] = "第6学年の漢字",
			[7] = "常用漢字",
			[8] = "人名用漢字",
			[9] = "表外漢字",
			[0] = "CJKV部首",
		}
		table.insert(categories, "[[Category:" .. (grade_categories[grade] or error("The grade " .. grade .. " is invalid.")) .. "|" .. (grade == "0" and " " or catsort) .. "]]")

		-- error category
		if not rs then
			table.insert(categories, "[[Category:日本語 漢字 部首と画数の欠落]]")
		end

		return table.concat(wikitext, "") .. table.concat(categories, "\n")
	end
end

local grade1_pattern = ('[' .. data.grade1 .. ']')
local grade2_pattern = ('[' .. data.grade2 .. ']')
local grade3_pattern = ('[' .. data.grade3 .. ']')
local grade4_pattern = ('[' .. data.grade4 .. ']')
local grade5_pattern = ('[' .. data.grade5 .. ']')
local grade6_pattern = ('[' .. data.grade6 .. ']')
local secondary_pattern = ('[' .. data.secondary .. ']')
local jinmeiyo_kanji_pattern = ('[' .. data.jinmeiyo_kanji .. ']')
local hyogaiji_pattern = ('[^' .. data.joyo_kanji .. data.jinmeiyo_kanji .. ']')

function export.kanji_grade(kanji)
	if type(kanji) == "table" then
		kanji = kanji.args[1]
	end

	if find(kanji, hyogaiji_pattern) then return 9
	elseif find(kanji, jinmeiyo_kanji_pattern) then return 8
	elseif find(kanji, secondary_pattern) then return 7
	elseif find(kanji, grade6_pattern) then return 6
	elseif find(kanji, grade5_pattern) then return 5
	elseif find(kanji, grade4_pattern) then return 4
	elseif find(kanji, grade3_pattern) then return 3
	elseif find(kanji, grade2_pattern) then return 2
	elseif find(kanji, grade1_pattern) then return 1
	end

	return false
end

return export