Module:Switch by pattern

Module documentation[view] [edit] [history] [purge]

Unlike Switch parser function, this module matches any part of the input string or any Lua pattern in the input string. It returns the first match, or all matches separated by the chosen separator.

The module is used in {{Switch by pattern}} aka {{reSwitch}}.

Usage

{{#invoke:Switch by pattern|found_in|_input= |patterns1|patterns2|patterns3|keyA=patternsA |keyB=patternsB |_respectcase= |_returnall= |_sep= |_outputsep= |_default= }}

Module parameters:

• |_input= input string
• |key=patterns return key if any of the patterns is found in the input string; patterns can be one or multiple strings separated by # (or a separator given in |_sep=); if unnamed parameters are used, the keys returned are their respective positions (1, 2, 3…)
• |_respectcase= input string is lowercased by default, any value to avoid that
• |_returnall= return the keys of all patterns that matched, separated by # or _outputsep
• |_returncaptures= instead of keys, return captures/matches
• |_sep= which separator is used to separate patterns (defaults to #); no pattern should contain this character or characters
• |_outputsep= which separator to use to separate multiple keys returned when the _returnall option is on (defaults to #
• |_default= which string to return if no patterns matched (defaults to empty string)

The string in patterns can be a Lua pattern, or a word or part of a word (always interpreted as Lua patterns!), or several such patterns separated by #, as in this#that#whatever, which will be checked as aliases until the first match is found. (this is similar to regex matching this|that|whatever, which is not available in Lua) Note that all named parameters will have leading and trailing whitespace removed. This not only includes (|string key=pattern) but also |2=pattern and similar.

Examples

Matches parts of the input text

• {{#invoke:Switch by pattern|found_in|_input=[[Animalia|beasts]]|dark gray=viru|pink=animal#beast|brown=fung}} → pink

• {{#invoke:Switch by pattern|found_in|_input=[[Riboviria]] |dark gray=vir|pink=animal#beast |brown=fung}} → dark gray (keys may include whitespace)

With unnamed parameters, returning all matches

• {{#invoke:Switch by pattern|found_in|_input=navy and red |red|green;lime;olive|azure;blue;navy|_returnall=yes|_sep=;|_outputsep=#}} → 1#3

With Lua patterns and Ustring patterns

• {{#invoke:Switch by pattern|found_in|_input=name12|has two digits=%D%d%d$|has one digit=%D%d$}} → has two digits

Return captures

• {{#invoke:Switch by pattern|found_in|_input=''[[Equus zebra]]''|''%[*(%u%a+ ?%a+)%]*''#''%[*(%u%a+ %b() ?%a+)%]*''|_respectcase=y|_returncaptures=y}} → Equus zebra (matched first patterns, two words)
• {{#invoke:Switch by pattern|found_in|_input=''[[Equus (Hippotigris) hartmannae]]''|''%[*(%u%a+ ?%a+)%]*''#''%[*(%u%a+ %b() ?%a+)%]*''|_respectcase=y|_returncaptures=y}} → Equus (Hippotigris) hartmannae (matched second pattern: Word (word) word)
• {{#invoke:Switch by pattern|found_in|_input=''[[Anser cygnoides domesticus]]''|''%[*(%u%a+ ?%a+)%]*''#''%[*(%u%a+ %b() ?%a+)%]*''|_respectcase=y|_returncaptures=y}} → (no match, three words)

The above documentation is transcluded from Module:Switch by pattern/doc. (edit | history)
Editors can experiment in this module's sandbox (create | mirror) and testcases (create) pages.
Subpages of this module.

require('strict')
 
local p = {}

--will find either "human" or "homo" as part of the following _input, and return "orange"
--args = {_input="[[Homo sampiens|humans]]", ["pink"]="virus", ["green"]="plant", ["light gray"]="fung", ["orange"]="homo;human", _returnall="", _respectcase="", _sep=";"}
--aliases to search keys are separated by _sep, so _sep cannot be part of any key
--returns the first key found unless _returnall is nonempty, in which case it returns "all#keys#found" (for further processing?)
function p._found_in(args)
    local skip_args = "#_input#_respectcase#_returncaptures#_returnall#_sep#_outputsep#_default#" --each key must be surrounded by #'s'
    local res = {}

    local all = args["_returnall"] and args["_returnall"]~=""
    local returnkeys = not(args["_returncaptures"] and args["_returncaptures"]~="")
    local sep = args["_sep"]~="" and args["_sep"] or "#" --should not be blank
    local outputsep = args["_outputsep"] or "#" --can be blank
    local lowercase = not (args["_respectcase"] and args["_respectcase"]~="") --lowercase input (but not search keys, they can be regex patterns)
    local input = args["_input"] or ""
    if input == "" then return "" end
    
    if lowercase then input = mw.ustring.lower(input) end
    for k,v in pairs(args) do
    	if not mw.ustring.match(skip_args, "#"..k.."#" ) then
    		local aliases = mw.text.split(v, sep, true)
    		for _,a in ipairs(aliases) do
    			if a~="" then
    				local match = mw.ustring.match(input, a)
    				if match then
    					local key_or_capture = returnkeys and k  or  match
    					if not all then return key_or_capture end
    					table.insert(res, key_or_capture)
    					break --first found alias
    				end
    			end
    		end
    	end
    end
   
   if #res>0 then return table.concat(res, outputsep) end --returnall was not blank
   return args["_default"] or ""
end

function p.found_in(frame)
	local args = frame:getParent() and frame:getParent().args
	if not (args["_input"] and args["_input"]~="") then args = frame.args end
	if args["_input"] and args["_input"]~="" then return p._found_in(args) end
	return args["_default"] or ""
end

return p