Module:Sandbox/HenryLi/Article

local p = {}

-- 數字
local digits = {"零","壹","貳","叁","肆","伍","陸","柒","捌","玖"}
local units = {"","拾","佰","仟"}
local bigUnits = {"","萬","億"}

-- 轉換個十百千
local function fourDigitToChinese(num)
    local str = ""
    local zero = false
    local n = tonumber(num)
    local numStr = tostring(n)
    local len = #numStr

    for i = 1, len do
        local digit = tonumber(numStr:sub(i,i))
        local pos = len - i + 1
        if digit == 0 then
            zero = true
        else
            if zero then
                str = str .. digits[1]
                zero = false
            end
            str = str .. digits[digit+1] .. units[pos]
        end
    end

    return str
end

-- 主轉換函數
local function toChineseFormal(num)
    if type(num) ~= "number" then
        num = tonumber(num) or 0
    end

    if num == 0 then return digits[1] end

    local str = ""
    local parts = {}
    local unitIndex = 1

    while num > 0 do
        local segment = num % 10000
        if segment ~= 0 then
            local segmentStr = fourDigitToChinese(segment) .. bigUnits[unitIndex]
            table.insert(parts,1,segmentStr)
        else
            table.insert(parts,1,"")
        end
        num = math.floor(num / 10000)
        unitIndex = unitIndex + 1
    end

    str = table.concat(parts)

    -- 清理多餘零
    str = mw.ustring.gsub(str,"零+","零")
    str = mw.ustring.gsub(str,"零$","")

    -- 特殊處理十位 10-19
    str = mw.ustring.gsub(str,"^壹拾","拾")

    return str
end

-- 讀參數
local function getArg(frame, name_or_number, default)
    local function clean(value)
        if not value then return nil end
        value = mw.text.trim(value)
        if value == "" then return nil end
        return value
    end

    local value = frame.args[name_or_number]
    value = clean(value)

    if not value and frame.getParent then
        local parent = frame:getParent()
        if parent and parent.args then
            value = clean(parent.args[name_or_number])
        end
    end

    if not value then
        return default
    end

    return value
end

-- 計字
-- 0x4E00  0x9FFF  本
-- 0x3400  0x4DBF  甲
-- 0x20000 0x2A6DF 乙
-- 0x2A700 0x2B73F 丙
-- 0x2B740 0x2B81F 丁
-- 0x2B820 0x2CEAF 戊
-- 0xF900  0xFAFF  容
local function countCharacter(text)
    local _, count = mw.ustring.gsub(text, "[一-龯㐀-䶿豈-﫿]", "")
    return count
end


-- 省非文計
local function cleanText(text)
    -- 記。唔現
    text = mw.ustring.gsub(text, "<!%-%-.-%-%->", "")
    -- 攷
    text = mw.ustring.gsub(text, "<ref[^>/]->.-</ref>", "")
    text = mw.ustring.gsub(text, "<ref[^>]*/>", "")
    -- 模
    text = mw.ustring.gsub(text, "{{[^{}]-}}", "")
    -- 表
    text = mw.ustring.gsub(text, "{|.-|}", "")
	-- 題
    text = mw.ustring.gsub(text, "\n=+[^\n]+=+\n", "\n")
    -- 圖。檔
    text = mw.ustring.gsub(text, "%[%[File:.-%]%]", "")
    text = mw.ustring.gsub(text, "%[%[Image:.-%]%]", "")
    -- 省內通 [[甲|乙]] → 乙
    text = mw.ustring.gsub(text, "%[%[[^|%]]-|", "")
    text = mw.ustring.gsub(text, "%[%[", "")
    text = mw.ustring.gsub(text, "%]%]", "")
	-- 省外通
    text = mw.ustring.gsub(text, "%[https?://.-%]", "")

    return text
end

-- 字數評級
local function grade(count)
	if count < 150 then -- 小學一二年班 百五字
		return "草稿"
	elseif count < 600 then -- 中學畢業 八千卦估算
        return "簡介"
    elseif count < 1500 then -- 夠料 一萬六千卦估算
        return "及格"
	elseif count < 10000 then
		return "長文"
    else
        return "超長"
    end
end

local function hasSource(text)
    if mw.ustring.find(text, "<ref") then
        return true
    else
        return false
    end
end

local function wikidata(title)
	local entityId = mw.wikibase.getEntityIdForTitle(title)
	if entityId then

	    local entity = mw.wikibase.getEntity(entityId)
	    if entity then
	        return entity
		else
			return nil
	    end
	else
		return nil
	end
end

local function commonscat(entity)
	local commons = nil
	if not entity then
		return nil
	end
    if entity.sitelinks and entity.sitelinks.commonswiki then
		local name = entity.sitelinks.commonswiki.title
		if mw.ustring.match(name, "^Category:") then
	        return entity.sitelinks.commonswiki.title .. ""
		end
    end

	local claims = entity.claims

    if claims then
        if claims["P373"] then
			return "Category:" .. claims["P373"][1].mainsnak.datavalue.value .. ""
		end
		if claims["P948"] then
			return "Category:" .. claims["P948"][1].mainsnak.datavalue.value .. ""
        end
    end
	return commons
end

local function hasImage(frame, title)
	local html = frame:preprocess("{{:" .. title.fullText .. "}}")

    if html and string.find(html, "<img") then
        return true
    else
        return false
    end
end

local function isDisambiguity(entity)
	if entity and entity.claims and entity.claims["P31"] then
        for _, claim in ipairs(entity.claims["P31"]) do
            local value = claim.mainsnak.datavalue.value.id
            if value == "Q4167410" then
                return true
            end
        end
	end
	return false
end

p.count = function(frame)

	local page = getArg(frame, 1)
	local title

    if page then
        title = mw.title.new(page)
    else
        title = mw.title.getCurrentTitle()
    end

	if not title or not title.exists then
        return "無此版題"
    end

	if title.isRedirect then
		return "跳轉"
	end

    -- 限文章
    -- if title.namespace ~= 0 then
    --     return "祇容文章"
    -- end

    local text = title:getContent()
    if not text then
        return "莫能讀"
    end

	local sourced = hasSource(text)
	local sourceText
	if sourced then
		sourceText = ""
	else
		sourceText = "。欠來源"
	end

	local wikidataConnected = wikidata(page)
	local wikidataText
	if wikidataConnected then
		wikidataText = ""
	else
		wikidataText = "。欠資料庫"
	end
	if isDisambiguity(wikidataConnected) then
		return "搞清楚"
	end
	local imageText
	if not hasImage(frame, title) then
		imageText = "無圖"
		if wikidataConnected then
			-- 若果無圖
			-- 要做啲嘢
			imageText = imageText .. "。[[commons:" .. commonscat(wikidataConnected) .. "]]"
		end	
	end


	text = cleanText(text)
	local count = countCharacter(text)
	local level = grade(count)
	local countChinese = toChineseFormal(count)


    return string.format("%s字。%s%s%s%s", countChinese, level, sourceText, wikidataText, imageText)
end

return p