Module:Ko-translit/clean

local p = {}
local m_utils = require('Module:Ko-utils')
local find = mw.ustring.find
local gsub_iterate = m_utils.gsub_iterate

-- Initial input validations
function p.validate_input(text)
	if not m_utils.contains_hangul(text) then
		error("Input must contain Hangul")
	elseif m_utils.contains_reference(text) then
		error("Input cannot contain references")
	-- Note: If Yale support is added, remove below line and make overall romanization "N/A"
	elseif find(text, "[ᄓ-ᅠᅶ-ᆧᇃ-ᇿ〮〯ㅤ-ㆎꥠ-꥿ힰ-퟿]") then
		error("Input contains Hangul not supported by RR and MR")
	elseif find(text, "[ᄀ-ᄒ]") or find(text, "[ᅡ-ᅵᆨ-ᇂ]") then
		error("Do not input conjoining Hangul jamo directly")
	elseif find(text, "`%*") then
		error("Use *` instead of `*")
	elseif find(text, "@%*") then
		error("Use *@ instead of @*")
	elseif find(text, "%^[^가-힣]") then
		error("^ must be immediately followed by Hangul syllabic block")
	elseif find(text, "[^%*0-9A-Za-z]`") or find(text, "[^0-9A-Za-z]%*`") or find(text, "`[^가-깋다-딯바-빟자-짛]") then
		error("Found invalid sequence containing `")
	elseif find(text, "[^%*ㄹ가-힣]@") or find(text, "[^가-힣]%*@") or find(text, "%*@[^가-깋다-딯바-빟자-짛]") or find(text, "ㄹ@[^가-깋다-딯바-빟사-싷자-짛]") or find(text, "@[^가-깋다-딯라-맇바-빟사-싷야-얳여-옅옇-옣요-욯유-윶윸-윻이-잍잏자-짛하-힣]") or find(text, "[받붙]잇@[가-깋]") or find(text, "앞엣@것") or find(text, "한솥엣@밥") then
		error("Found invalid sequence containing @")
	elseif find(text, "[^가-힣]%$") or find(text, "%$[^아어에엔엘여요으은을음읍의이인일임입]") then
		error("Found invalid sequence containing $")
	elseif find(text, "%%$") then
		error("Remove final %")
	elseif find(text, "[ _][ _]") then
		error("No two or more consecutive space characters")
	elseif find(text, "^[%$%*@_`]") or find(text, "^%%[^_가-힣]") or find(text, "[ _]%*") or find(text, "%*[ %*%-_]") or find(text, "%-%*") or find(text, "[﷐-﷒]") or find(text, "%%_$") or find(text, "[%$%*@%^`]$") then
		error("Invalid input")
	end
end

-- Validation after processing personal names and removing links & markups (before decomposing Hangul)
-- some invalid sequences can only be detected after doing those
function p.validate_composed(text)
	if find(text, "[ _][ _]") then
		error("No two or more consecutive space characters")
	elseif find(text, "^[%$%*@_`]") or find(text, "[ _]%*") or find(text, "%*[ %*%-_]") or find(text, "%-%*") or find(text, "[%$%*@%^_`]$") then
		error("Invalid input")
	end
end

-- Validation after Hangul syllable decomposition
-- some invalid sequences are difficult to detect with precomposed Hangul syllables
function p.validate_decomposed(text)
	if find(text, "[ᆨ-ᆪᆬ-ᆮᆴ-ᆶᆸᆹᆻᆽ-ᇂ]%*?﷐?@﷐?[ᄀᄃᄇᄉᄌ]") or find(text, "ᆰ%*?﷐?@﷐?[ᄀ-ᄊᄌ-ᄑ]") or find(text, "ᆲ﷐?@﷐?[ᄀ-ᄊᄌ-ᄑ]") or find(text, "ᆺ%*@[ᄀᄇ]") or find(text, "ᆺ%*?﷐?@﷐?[ᄁ-ᄆᄈ-ᄊᄌ-ᄑ]") or find(text, "[ᅡ-ᅵᆨ-ᆪᆬ-ᇂ]﷐?@﷐?ᄅ") or find(text, "[ᅡ-ᅵ]﷐?@﷐?ᄋ") or find(text, "[ᅡ-ᅵᆫ-ᆭᆯᆱ-ᆷᆼ]﷐?@﷐?ᄒ") then
		error("Found invalid sequence containing @")
	elseif find(text, "[ᅡ-ᅵᆨᆫᆭ-ᆯᆶ-ᆸᆼ]﷐?%$") then
		error("Found invalid sequence containing $")
	end
end

-- Remove links and markup, which are unnecessary or interfere with assimilation
function p.remove_links_and_markup(text)
	local replacements = {
		-- Remove bold/italic wikitext syntax
		-- It is not impossible to allow bold/italic when it does not interfere with assimilation, but determining when to allow or disallow that adds complication for little practical gain
		{"'''", ""},
		{"''", ""},
		-- Remove HTML tags (except br)
		{"<[Bb][Rr] */?>", "&#10;"},
		{"</?[A-Za-z][^>]->", ""},
		{"&#10;", "<br>"},
		-- Remove wikilinks
		{"%[%[[^%|]+%|(..-)%]%]", "%1"},
		{"%[%[", ""},
		{"%]%]", ""}
	}
	text = gsub_iterate(text, replacements)
	text = mw.text.killMarkers(text)
	return text
end

-- Converting escaped special chars to HTML tags to preserve them
function p.escaped_to_html(text)
	local replacements = {
		{"\\%$", "&#36;"},
		{"\\%%", "&#37;"},
		{"\\%*", "&#42;"},
		{"\\@", "&#64;"},
		{"\\%^", "&#94;"},
		{"\\_", "&#95;"},
		{"\\`", "&#96;"}
	}
	return gsub_iterate(text, replacements)
end

-- Converting HTML tags back to unescaped chars
function p.html_to_ascii(text)
	local replacements = {
		{"&#36;", "$"},
		{"&#37;", "%%"},
		{"&#42;", "*"},
		{"&#64;", "@"},
		{"&#94;", "^"},
		{"&#95;", "_"},
		{"&#96;", "`"}
	}
	return gsub_iterate(text, replacements)
end

-- Unwrapping enclosed Hangul text
-- Actually not very necessary, but these are also classified as Hangul chars in Unicode
-- No distinction is made between parenthesized and circled chars
-- Needs to be executed before decomposing Hangul
function p.unwrap_enclosed_hangul(text)
	local replacements = {
		{"[㈀㉠]", "(기역)"},
		{"[㈁㉡]", "(니은)"},
		{"[㈂㉢]", "(디귿)"},
		{"[㈃㉣]", "(리을)"},
		{"[㈄㉤]", "(미음)"},
		{"[㈅㉥]", "(비읍)"},
		{"[㈆㉦]", "(시옷)"},
		{"[㈇㉧]", "(이응)"},
		{"[㈈㉨]", "(지읒)"},
		{"[㈉㉩]", "(치읓)"},
		{"[㈊㉪]", "(키읔)"},
		{"[㈋㉫]", "(티읕)"},
		{"[㈌㉬]", "(피읖)"},
		{"[㈍㉭]", "(히읗)"},
		{"[㈎㉮]", "(가)"},
		{"[㈏㉯]", "(나)"},
		{"[㈐㉰]", "(다)"},
		{"[㈑㉱]", "(라)"},
		{"[㈒㉲]", "(마)"},
		{"[㈓㉳]", "(바)"},
		{"[㈔㉴]", "(사)"},
		{"[㈕㉵]", "(아)"},
		{"[㈖㉶]", "(자)"},
		{"[㈗㉷]", "(차)"},
		{"[㈘㉸]", "(카)"},
		{"[㈙㉹]", "(타)"},
		{"[㈚㉺]", "(파)"},
		{"[㈛㉻]", "(하)"},
		{"㈜", "(주)"},
		{"㈝", "(오전)"},
		{"㈞", "(오후)"},
		{"㉼", "(참고)"},
		{"㉽", "(주의)"},
		{"㉾", "(우)"}
	}
	return gsub_iterate(text, replacements)
end

-- Final validation and postprocessing (both MR and RR)
function p.final_processing(text)
	-- result should not contain Hangul
	if m_utils.contains_hangul(text) then
		error("Result contains Hangul; debugging required")
	end

	text = p.html_to_ascii(text) -- convert HTML encodings back to ASCII

	-- if result is nothing (e.g. when input is just ㅇ)
	if text == "" then
		text = "—"
	end

	return text
end
return p