Module:Text

Text – Module containing methods for the manipulation of text, wikimarkup and some HTML.

Functions for templates

All methods have an unnamed parameter containing the text.

The return value is an empty string if the parameter does not meet the conditions. When the condition is matched or some result is successfully found, strings of at least one character are returned.

char

Creates a string from a list of character codes.

1

Space-separated list of character codes

*

Number of repetitions of the list in parameter 1; (Default 1).

errors

0 – Silence errors

concatParams

Combine any number of elements into a list, like table.concat() in Lua.

From a template:

1

First element; missing and empty elements are ignored.

2 3 4 5 6 …

Further list elements

From Lua

args

table (sequence) of the elements

apply

Separator between elements; defaults to |

adapt

optional formatting, which will be applied to each element; must contain %s.

containsCJK

Returns whether the input string contains any CJK characters

• Returns nothing if there are no CJK characters

removeDelimited

Remove all text between delimiters, including the delimiters themselves.

getPlain

Remove wikimarkup (except templates): comments, tags, bold, italic, nbsp

isLatinRange

Returns some content, unless the string contains a character that would not normally be found in Latin text.

• Returns nothing if there is a non-Latin string.

isQuote

Returns some content if the parameter passed is a single character, and that character is a quote, such as '.

• Returns nothing for multiple characters, or if the character passed is not a quote.

listToText

Formats list elements analogously to mw.text.listToText().

The elements are separated by a comma and space ; the word "and" appears between the first and last.

Unnamed parameters become the list items.

Optional parameters for #invoke:

• format – Every list element will first be formatted with this format string; see here for how to construct this string. The string must contain at least one %s sequence.
• template=1 – List elements should be taken from the calling template.

Returns the resulting string.

quote

Wrap the string in quotes; quotes can be chosen for a specific language.

1

Input text (will be automatically trimmed); may be empty.

2

(optional) the ISO 639 language code for the quote marks; should be one of the supported languages (in German)

3

(optional) 2 for second level quotes. This means the single quote marks in a statement such as: Jack said, “Jill said ‘fish’ last Tuesday.”

quoteUnquoted

Wrap the string in quotes; quotes can be chosen for a specific language. Will not quote an empty string, and will not quote if there is a quote at the start or end of the (trimmed) string.

1

Input text (will be automatically trimmed); may be empty.

2

(optional) the ISO 639 language code for the quote marks; should be one of the supported languages (in German)

3

(optional) 2 for second level quotes. This means the single quote marks in a statement such as: Jack said, “Jill said ‘fish’ last Tuesday.”

removeDiacritics

Removes all diacritical marks from the input.

1

Input text

sentenceTerminated

Is this sentence terminated? Should work with CJK, and allows quotation marks to follow.

• Returns nothing if the sentence is unterminated.

ucfirstAll

The first letter of every recognized word is converted to upper case. This contrasts with the parser function {{ucfirst:}} which changes only the first character of the whole string passed.

A few common HTML entities are protected; the implementation of this may mean that numerical entities passed (e.g. &) are converted to & form

uprightNonlatin

Takes a string. Italicized non-Latin characters are un-italicized, unless they are a single Greek letter.

zip

Combines a tuple of lists by convolution. This is easiest to explain by example: given two lists, list1 = "a b c" and list2 = "1 2 3", then
zip(liste1, liste2, sep = " ", isep = "-", osep = "/")
outputs
a-1/b-2/c-3

• 1, 2, 3, … – Lists to be combined
• sep – A separator (in Lua regex form) used to split the lists. If empty, the lists are split into individual characters.
• sep1, sep2, sep3, … – Allows a different separator to be used for each list.
• isep – Output separator; placed between elements which were at the same index in their lists.
• osep – Output separator; placed between elements which had different original indices; i.e. between the groups joined with isep

split

Splits a string into chunks at the specified delimiter, and returns the first (or user-specified) chunk. This is non-Unicode-aware implementation of mw.text.split which, for ASCII-only text, can be up to 60 times faster.

• 1 (or text) – the text to be split
• 2 (or pattern) – the pattern to use when splitting the text. By default, this can be a string library pattern.
• 3 (or plain) – if set to "true", pattern will be interpreted as plain text, not a pattern.
• 4 (or index) – The chunk to return. If omitted, the first chunk will be returned. Can be set to a negative number to count from the end (e.g. -1 will return the last chunk).

Examples and test page

There are tests available (in German) to illustrate this in practice.

Use in another Lua module

All of the above functions can be called from other Lua modules. Use require(); the below code checks for errors loading it:

local lucky, Text = pcall( require, "Module:Text" )
if type( Text ) == "table" then
    Text = Text.Text()
else
    -- In the event of errors, Text is an error message.
    return "<span class=\"error\">" .. Text .. "</span>"
end

You may then call:

• Text.char( apply, again, accept )
• Text.concatParams( args, separator, format )
• Text.containsCJK( s )
• Text.removeDelimited( s )
• Text.getPlain( s )
• Text.isLatinRange( s )
• Text.isQuote( c )
• Text.listToText( table, format )
• Text.quote( s, lang, mode )
• Text.quoteUnquoted( s, lang, mode )
• Text.removeDiacritics( s )
• Text.sentenceTerminated( s )
• Text.split( text, pattern, plain ) – non Unicode version of mw.text.split
• Text.gsplit( text, pattern, plain ) – non Unicode version of mw.text.gsplit
• Text.ucfirstAll( s )
• Text.uprightNonlatin( s )
• Text.zip(…)
• Text.test( s )

Usage

This is a general library; use it anywhere.

Dependencies

• Module:Yesno
• Module:Text/data --- Lua patterns and information about quotes

See also

---

local yesNo = require("Module:Yesno")
local Text = { serial = "2022-07-21",
               suite  = "Text" }
--[=[
Text utilities
]=]

local function fiatQuote( apply, alien, advance )
    -- Quote text
    -- Parameter:
    --     apply    -- string, with text
    --     alien    -- string, with language code
    --     advance  -- number, with level 1 or 2
    local r = apply and tostring(apply) or ""
    alien = alien or "en"
    advance = tonumber(advance) or 0
    local suite
    local data = mw.loadData('Module:Text/data')
    local QuoteLang = data.QuoteLang
    local QuoteType = data.QuoteType
    local slang = alien:match( "^(%l+)-" )
    suite = QuoteLang[alien] or slang and QuoteLang[slang] or QuoteLang["en"]
    if suite then
        local quotes = QuoteType[ suite ]
        if quotes then
            local space
            if quotes[ 3 ] then
                space = " "
            else
                space = ""
            end
            quotes = quotes[ advance ]
            if quotes then
                r = mw.ustring.format( "%s%s%s%s%s",
                                       mw.ustring.char( quotes[ 1 ] ),
                                       space,
                                       apply,
                                       space,
                                       mw.ustring.char( quotes[ 2 ] ) )
            end
        else
            mw.log( "fiatQuote() " .. suite )
        end
    end
    return r
end -- fiatQuote()



Text.char = function ( apply, again, accept )
    -- Create string from codepoints
    -- Parameter:
    --     apply   -- table (sequence) with numerical codepoints, or nil
    --     again   -- number of repetitions, or nil
    --     accept  -- true, if no error messages to be appended
    -- Returns: string
    local r = ""
    apply = type(apply) == "table" and apply or {}
    again = math.floor(tonumber(again) or 1)
    if again < 1 then
    	return ""
    end
    local bad   = { }
    local codes = { }
    for _, v in ipairs( apply ) do
    	local n = tonumber(v)
    	if not n or (n < 32 and n ~= 9 and n ~= 10) then
    		table.insert(bad, tostring(v))
    	else
    		table.insert(codes, math.floor(n))
		end
    end 
    if #bad > 0 then
    	if not accept then
    		r = tostring(  mw.html.create( "span" )
                    		:addClass( "error" )
                    		:wikitext( "bad codepoints: " .. table.concat( bad, " " )) )
    	end
    	return r
    end
    if #codes > 0 then
    	r = mw.ustring.char( unpack( codes ) )
    	if again > 1 then
    		r = r:rep(again)
    	end
	end
    return r
end -- Text.char()

local function trimAndFormat(args, fmt)
	local result = {}
	if type(args) ~= 'table' then
		args = {args}
	end
	for _, v in ipairs(args) do
		v = mw.text.trim(tostring(v))
		if v ~= "" then
			table.insert(result,fmt and mw.ustring.format(fmt, v) or v)
		end
	end
	return result
end

Text.concatParams = function ( args, apply, adapt )
    -- Concat list items into one string
    -- Parameter:
    --     args   -- table (sequence) with numKey=string
    --     apply  -- string (optional); separator (default: "|")
    --     adapt  -- string (optional); format including "%s"
    -- Returns: string
    local collect = { }
    return table.concat(trimAndFormat(args,adapt), apply or "|")
end -- Text.concatParams()



Text.containsCJK = function ( s )
    -- Is any CJK code within?
    -- Parameter:
    --     s  -- string
    -- Returns: true, if CJK detected
    s = s and tostring(s) or ""
    local patternCJK = mw.loadData('Module:Text/data').PatternCJK
    return mw.ustring.find( s, patternCJK ) ~= nil
end -- Text.containsCJK()

Text.removeDelimited = function (s, prefix, suffix)
	-- Remove all text in s delimited by prefix and suffix (inclusive)
	-- Arguments:
	--    s = string to process
	--    prefix = initial delimiter
	--    suffix = ending delimiter
	-- Returns: stripped string
	s = s and tostring(s) or ""
	prefix = prefix and tostring(prefix) or ""
	suffix = suffix and tostring(suffix) or ""
	local prefixLen = mw.ustring.len(prefix)
	local suffixLen = mw.ustring.len(suffix)
	if prefixLen == 0 or suffixLen == 0 then
		return s
	end
	local i = s:find(prefix, 1, true)
	local r = s
	local j
	while i do
		j = r:find(suffix, i + prefixLen)
		if j then
			r = r:sub(1, i - 1)..r:sub(j+suffixLen)
		else
			r = r:sub(1, i - 1)
		end
		i = r:find(prefix, 1, true)
	end
	return r
end

Text.getPlain = function ( adjust )
    -- Remove wikisyntax from string, except templates
    -- Parameter:
    --     adjust  -- string
    -- Returns: string
    local r = Text.removeDelimited(adjust,"<!--","-->")
    r = r:gsub( "(</?%l[^>]*>)", "" )
         :gsub( "'''", "" )
         :gsub( "''", "" )
         :gsub( "&nbsp;", " " )
    return r
end -- Text.getPlain()

Text.isLatinRange = function (s)
    -- Are characters expected to be latin or symbols within latin texts?
    -- Arguments:
    --  s = string to analyze
    -- Returns: true, if valid for latin only
    s = s and tostring(s) or ""  --- ensure input is always string
    local PatternLatin = mw.loadData('Module:Text/data').PatternLatin
    return mw.ustring.match(s, PatternLatin) ~= nil
end -- Text.isLatinRange()



Text.isQuote = function ( s )
    -- Is this character any quotation mark?
    -- Parameter:
    --     s = single character to analyze
    -- Returns: true, if s is quotation mark
    s = s and tostring(s) or ""
    if s == "" then
    	return false
    end
    local SeekQuote = mw.loadData('Module:Text/data').SeekQuote
    return mw.ustring.find( SeekQuote, s, 1, true ) ~= nil
end -- Text.isQuote()



Text.listToText = function ( args, adapt )
    -- Format list items similar to mw.text.listToText()
    -- Parameter:
    --     args   -- table (sequence) with numKey=string
    --     adapt  -- string (optional); format including "%s"
    -- Returns: string
    return mw.text.listToText(trimAndFormat(args, adapt))
end -- Text.listToText()



Text.quote = function ( apply, alien, advance )
    -- Quote text
    -- Parameter:
    --     apply    -- string, with text
    --     alien    -- string, with language code, or nil
    --     advance  -- number, with level 1 or 2, or nil
    -- Returns: quoted string
    apply = apply and tostring(apply) or ""
    local mode, slang
    if type( alien ) == "string" then
        slang = mw.text.trim( alien ):lower()
    else
        slang = mw.title.getCurrentTitle().pageLanguage
        if not slang then
            -- TODO FIXME: Introduction expected 2017-04
            slang = mw.language.getContentLanguage():getCode()
        end
    end
    if advance == 2 then
        mode = 2
    else
        mode = 1
    end
    return fiatQuote( mw.text.trim( apply ), slang, mode )
end -- Text.quote()



Text.quoteUnquoted = function ( apply, alien, advance )
    -- Quote text, if not yet quoted and not empty
    -- Parameter:
    --     apply    -- string, with text
    --     alien    -- string, with language code, or nil
    --     advance  -- number, with level 1 or 2, or nil
    -- Returns: string; possibly quoted
    local r = mw.text.trim( apply and tostring(apply) or "" )
    local s = mw.ustring.sub( r, 1, 1 )
    if s ~= ""  and  not Text.isQuote( s, advance ) then
        s = mw.ustring.sub( r, -1, 1 )
        if not Text.isQuote( s ) then
            r = Text.quote( r, alien, advance )
        end
    end
    return r
end -- Text.quoteUnquoted()



Text.removeDiacritics = function ( adjust )
    -- Remove all diacritics
    -- Parameter:
    --     adjust  -- string
    -- Returns: string; all latin letters should be ASCII
    --                  or basic greek or cyrillic or symbols etc.
    local cleanup, decomposed
    local PatternCombined = mw.loadData('Module:Text/data').PatternCombined
    decomposed = mw.ustring.toNFD( adjust and tostring(adjust) or "" )
    cleanup    = mw.ustring.gsub( decomposed, PatternCombined, "" )
    return mw.ustring.toNFC( cleanup )
end -- Text.removeDiacritics()



Text.sentenceTerminated = function ( analyse )
    -- Is string terminated by dot, question or exclamation mark?
    --     Quotation, link termination and so on granted
    -- Parameter:
    --     analyse  -- string
    -- Returns: true, if sentence terminated
    local r
    local PatternTerminated = mw.loadData('Module:Text/data').PatternTerminated
    if mw.ustring.find( analyse, PatternTerminated ) then
        r = true
    else
        r = false
    end
    return r
end -- Text.sentenceTerminated()



Text.ucfirstAll = function ( adjust)
    -- Capitalize all words
    -- Arguments:
    --     adjust = string to adjust
    -- Returns: string with all first letters in upper case
    adjust = adjust and tostring(adjust) or ""
    local r = mw.text.decode(adjust,true)
    local i = 1
    local c, j, m
    m = (r ~= adjust)
    r = " "..r
    while i do
        i = mw.ustring.find( r, "%W%l", i )
        if i then
            j = i + 1
            c = mw.ustring.upper( mw.ustring.sub( r, j, j ) )
            r = string.format( "%s%s%s",
                               mw.ustring.sub( r, 1, i ),
                               c,
                               mw.ustring.sub( r, i + 2 ) )
            i = j
        end
    end -- while i
    r = r:sub( 2 )
    if m then
    	r = mw.text.encode(r)
    end
    return r
end -- Text.ucfirstAll()


Text.uprightNonlatin = function ( adjust )
    -- Ensure non-italics for non-latin text parts
    --     One single greek letter might be granted
    -- Precondition:
    --     adjust  -- string
    -- Returns: string with non-latin parts enclosed in <span>
    local r
    local data = mw.loadData('Module:Text/data')
    local PatternLatin = data.PatternLatin
    local RangesLatin = data.RangesLatin
    local NumLatinRanges = data.NumLatinRanges
    if mw.ustring.match( adjust, PatternLatin ) then
        -- latin only, horizontal dashes, quotes
        r = adjust
    else
        local c
        local j    = false
        local k    = 1
        local m    = false
        local n    = mw.ustring.len( adjust )
        local span = "%s%s<span dir='auto' style='font-style:normal'>%s</span>"
        local flat = function ( a )
                  -- isLatin
                  local range
                  -- NumLatinRanges has to be precomputed because # does not work from loadData
                  for i = 1, NumLatinRanges do
                      range = RangesLatin[ i ]
                      if a >= range[ 1 ]  and  a <= range[ 2 ] then
                          return true
                      end
                  end    -- for i
              end -- flat()
        local focus = function ( a )
                  -- char is not ambivalent
                  local r = ( a > 64 )
                  if r then
                      r = ( a < 8192  or  a > 8212 )
                  else
                      r = ( a == 38  or  a == 60 )    -- '&' '<'
                  end
                  return r
              end -- focus()
        local form = function ( a )
                return string.format( span,
                                      r,
                                      mw.ustring.sub( adjust, k, j - 1 ),
                                      mw.ustring.sub( adjust, j, a ) )
              end -- form()
        r = ""
        for i = 1, n do
            c = mw.ustring.codepoint( adjust, i, i )
            if focus( c ) then
                if flat( c ) then
                    if j then
                        if m then
                            if i == m then
                                -- single greek letter.
                                j = false
                            end
                            m = false
                        end
                        if j then
                            local nx = i - 1
                            local s  = ""
                            for ix = nx, 1, -1 do
                                c = mw.ustring.sub( adjust, ix, ix )
                                if c == " "  or  c == "(" then
                                    nx = nx - 1
                                    s  = c .. s
                                else
                                    break -- for ix
                                end
                            end -- for ix
                            r = form( nx ) .. s
                            j = false
                            k = i
                        end
                    end
                elseif not j then
                    j = i
                    if c >= 880  and  c <= 1023 then
                        -- single greek letter?
                        m = i + 1
                    else
                        m = false
                    end
                end
            elseif m then
                m = m + 1
            end
        end    -- for i
        if j  and  ( not m  or  m < n ) then
            r = form( n )
        else
            r = r .. mw.ustring.sub( adjust, k )
        end
    end
    return r
end -- Text.uprightNonlatin()


Text.test = function ( about )
    local r
    if about == "quote" then
        data = mw.loadData('Module:Text/data')
        r = { }
        r.QuoteLang = data.QuoteLang
        r.QuoteType = data.QuoteType
    end
    return r
end -- Text.test()

-- Non Unicode-aware version of mw.text.split and mw.text.gsplit
-- based on [[phab:diffusion/ELUA/browse/master/includes/Engines/LuaCommon/lualib/mw.text.lua]]
-- These run up to 60 times faster than the Unicode-aware versions
Text.split = function ( text, pattern, plain )
	local ret = {}
	for m in Text.gsplit( text, pattern, plain ) do
		ret[#ret+1] = m
	end
	return ret
end

Text.gsplit = function ( text, pattern, plain )
	local s, l = 1, string.len( text )
	return function ()
		if s then
			local e, n = string.find( text, pattern, s, plain )
			local ret
			if not e then
				ret = string.sub( text, s )
				s = nil
			elseif n < e then
				-- Empty separator!
				ret = string.sub( text, s, e )
				if e < l then
					s = e + 1
				else
					s = nil
				end
			else
				ret = e > s and string.sub( text, s, e - 1 ) or ''
				s = n + 1
			end
			return ret
		end
	end, nil, nil
end

-- Export
local p = { }

for _, func in ipairs({'containsCJK','isLatinRange','isQuote','sentenceTerminated'}) do
	p[func] = function (frame) 
		return Text[func]( frame.args[ 1 ] or "" ) and "1" or ""
	end
end

for _, func in ipairs({'getPlain','removeDiacritics','ucfirstAll','uprightNonlatin'}) do
	p[func] = function (frame) 
		return Text[func]( frame.args[ 1 ] or "" )
	end
end

function p.char( frame )
    local params = frame:getParent().args
    local story = params[ 1 ]
    local codes, lenient, multiple
    if not story then
        params = frame.args
        story  = params[ 1 ]
    end
    if story then
        local items = mw.text.split( mw.text.trim(story), "%s+" )
        if #items > 0 then
            local j
            lenient  = (yesNo(params.errors) == false)
            codes    = { }
            multiple = tonumber( params[ "*" ] )
            for _, v in ipairs( items ) do
            	j = tonumber((v:sub( 1, 1 ) == "x" and "0" or "") .. v)
                table.insert( codes,  j or v )
            end 
        end
    end
    return Text.char( codes, multiple, lenient )
end

function p.concatParams( frame )
    local args
    local template = frame.args.template
    if type( template ) == "string" then
        template = mw.text.trim( template )
        template = ( template == "1" )
    end
    if template then
        args = frame:getParent().args
    else
        args = frame.args
    end
    return Text.concatParams( args,
                              frame.args.separator,
                              frame.args.format )
end


function p.listToFormat(frame)
    local lists = {}
    local pformat = frame.args["format"]
    local sep = frame.args["sep"] or ";"

    -- Parameter parsen: Listen
    for k, v in pairs(frame.args) do
        local knum = tonumber(k)
        if knum then lists[knum] = v end
    end

    -- Listen splitten
    local maxListLen = 0
    for i = 1, #lists do
        lists[i] = mw.text.split(lists[i], sep)
        if #lists[i] > maxListLen then maxListLen = #lists[i] end
    end

    -- Ergebnisstring generieren
    local result = ""
    local result_line = ""
    for i = 1, maxListLen do
        result_line = pformat
        for j = 1, #lists do
            result_line = mw.ustring.gsub(result_line, "%%s", lists[j][i], 1)
        end
        result = result .. result_line
    end

    return result
end



function p.listToText( frame )
    local args
    local template = frame.args.template
    if type( template ) == "string" then
        template = mw.text.trim( template )
        template = ( template == "1" )
    end
    if template then
        args = frame:getParent().args
    else
        args = frame.args
    end
    return Text.listToText( args, frame.args.format )
end



function p.quote( frame )
    local slang = frame.args[2]
    if type( slang ) == "string" then
        slang = mw.text.trim( slang )
        if slang == "" then
            slang = false
        end
    end
    return Text.quote( frame.args[ 1 ] or "",
                       slang,
                       tonumber( frame.args[3] ) )
end



function p.quoteUnquoted( frame )
    local slang = frame.args[2]
    if type( slang ) == "string" then
        slang = mw.text.trim( slang )
        if slang == "" then
            slang = false
        end
    end
    return Text.quoteUnquoted( frame.args[ 1 ] or "",
                               slang,
                               tonumber( frame.args[3] ) )
end


function p.zip(frame)
    local lists = {}
    local seps = {}
    local defaultsep = frame.args["sep"] or ""
    local innersep = frame.args["isep"] or ""
    local outersep = frame.args["osep"] or ""

    -- Parameter parsen
    for k, v in pairs(frame.args) do
        local knum = tonumber(k)
        if knum then lists[knum] = v else
            if string.sub(k, 1, 3) == "sep" then
                local sepnum = tonumber(string.sub(k, 4))
                if sepnum then seps[sepnum] = v end
            end
        end
    end
    -- sofern keine expliziten Separatoren angegeben sind, den Standardseparator verwenden
    for i = 1, math.max(#seps, #lists) do
        if not seps[i] then seps[i] = defaultsep end
    end

    -- Listen splitten
    local maxListLen = 0
    for i = 1, #lists do
        lists[i] = mw.text.split(lists[i], seps[i])
        if #lists[i] > maxListLen then maxListLen = #lists[i] end
    end

    local result = ""
    for i = 1, maxListLen do
        if i ~= 1 then result = result .. outersep end
        for j = 1, #lists do
            if j ~= 1 then result = result .. innersep end
            result = result .. (lists[j][i] or "")
        end
    end
    return result
end


function p.split(frame)
	local text = frame.args.text or frame.args[1] or ''
	local pattern = frame.args.pattern or frame.args[2] or ''
	local plain = yesNo(frame.args.plain or frame.args[3])
	local index = tonumber(frame.args.index) or tonumber(frame.args[4]) or 1
	local a = Text.split(text, pattern, plain)
	if index < 0 then index = #a + index + 1 end
	return a[index]
end


function p.failsafe()
    return Text.serial
end


p.Text = function ()
    return Text
end -- p.Text

return p