Module:String utilities

Frae Wikipedia, the free beuk o knawledge
local module_name = "string_utilities"
local export = {}

local rfind = mw.ustring.find

local format_escapes = {
    ["op"] = "{",
    ["cl"] = "}",
}

function export.format(str, tbl)
    return (string.gsub(str, "{(\\?)((\\?)[^{}]*)}", function (p1, name, p2)
        if #p1 + #p2 == 1 then
            return format_escapes[name] or error(module_name .. ".format: unrecognized escape sequence '{\\" .. name .. "}'")
        else
            return tbl[name] or error(module_name .. ".format: '" .. name .. "' not found in table")
        end
    end))
end

-- Reimplementation of mw.ustring.split() that includes any capturing
-- groups in the splitting pattern. This works like Python's re.split()
-- function, except that it has Lua's behavior when the split pattern
-- is empty (i.e. advancing by one character at a time; Python returns the
-- whole remainder of the string).
function export.capturing_split(str, pattern)
    local ret = {}
    -- (.-) corresponds to (.*?) in Python or Perl; () captures the
    -- current position after matching.
    pattern = "(.-)" .. pattern .. "()"
    local start = 1
    while true do
        -- Did we reach the end of the string?
        if start > #str then
            table.insert(ret, "")
            return ret
        end
        -- match() returns all captures as multiple return values;
        -- we need to insert into a table to get them all.
        local captures = {mw.ustring.match(str, pattern, start)}
        -- If no match, add the remainder of the string.
        if #captures == 0 then
            table.insert(ret, mw.ustring.sub(str, start))
            return ret
        end
        local newstart = table.remove(captures)
        -- Special case: If we don't advance by any characters, then advance
        -- by one character; this avoids an infinite loop, and makes splitting
        -- by an empty string work the way mw.ustring.split() does. If we
        -- reach the end of the string this way, return immediately, so we
        -- don't get a final empty string.
        if newstart == start then
            table.insert(ret, mw.ustring.sub(str, start, start))
            table.remove(captures, 1)
            start = start + 1
            if start > #str then
            	return ret
            end
        else
            table.insert(ret, table.remove(captures, 1))
            start = newstart
        end
        -- Insert any captures from the splitting pattern.
        for _, x in ipairs(captures) do
            table.insert(ret, x)
        end
    end
end

local function uclcfirst(text, dolower)
	local function douclcfirst(text)
		-- Actual function to re-case of the first letter.
		local first_letter = mw.ustring.sub(text, 1, 1)
		first_letter = dolower and mw.ustring.lower(first_letter) or mw.ustring.upper(first_letter)
		return first_letter .. mw.ustring.sub(text, 2)
	end
	-- If there's a link at the beginning, re-case the first letter of the
	-- link text. This pattern matches both piped and unpiped links.
	-- If the link is not piped, the second capture (linktext) will be empty.
	local link, linktext, remainder = mw.ustring.match(text, "^%[%[([^|%]]+)%|?(.-)%]%](.*)$")
	if link then
		return "[[" .. link .. "|" .. douclcfirst(linktext ~= "" and linktext or link) .. "]]" .. remainder
	end
	return douclcfirst(text)
end

function export.ucfirst(text)
	return uclcfirst(text, false)
end

function export.lcfirst(text)
	return uclcfirst(text, true)
end

function export.add_indefinite_article(text, uppercase)
	local is_vowel = false
	-- If there's a link at the beginning, examine the first letter of the
	-- link text. This pattern matches both piped and unpiped links.
	-- If the link is not piped, the second capture (linktext) will be empty.
	local link, linktext, remainder = mw.ustring.match(text, "^%[%[([^|%]]+)%|?(.-)%]%](.*)$")
	if link then
		is_vowel = rfind(linktext ~= "" and linktext or link, "^[AEIOUaeiou]")
	else
		is_vowel = rfind(text, "^[AEIOUaeiou]")
	end
	return (is_vowel and (uppercase and "An " or "an ") or (uppercase and "A " or "a ")) .. text
end

return export