Module:User:Saph/module parser

local export = {}

--[[ This module provides two main utilities: tools for parsing other modules,
	 and, to be added, a linter using those tools. The linter is meant to be 
	 used on documentation pages and the tools are invocable from modules. ]]

local find = string.find
local gmatch = string.gmatch
local gsub = string.gsub
local match = string.match
local setmetatable = setmetatable
local sub = string.sub

----------------------
--		Keys		--
----------------------

-- Prototype for keys.
local Key = {
	declaration = "",	-- This contains the raw declaration, like "local foo = 'bar'" for example.
	global = false,		-- Whether the variable is global or local.
	declared_at = 0,	-- The line # the key is declared at.
	name = "",			-- Name of the key.
	--scope = {},		-- A `scope` object in which the variable applies. TODO!
	--type = "",		-- The type of the variable at the point it's declared. TODO!
	uses = 0,			-- The number of times it is used throughout the module.
	value = "",			-- Its value.
}

-- For internal use only.
function Key:new(data)
	return setmetatable(data, Key)
end

--[[ Keys are anything defined within the module: local/global variables, 
	 functions, etc. This takes the full text of the module. ]]
function export.grab_keys(text)
	local keys = {}
	local n = 1
	
	-- Remove comments.
	text = gsub(text, "%-%-+", "")
	text = gsub(text, "%--%%] ?%--", "")
	
	for declaration in gmatch(text, "%s-(- ?= ?)") do
		repeat -- So that keys which can't be assigned values can be skipped.
			-- Filter out rubbish: conditionals, returns, etc.
			if find(text, sub(declaration, 1, -2) .. "=", declaration_index, true)
			or find(declaration, "return")
			or find(declaration, "~ -=") then
				break
			end
			
			local data = {}
			
			local declaration_index = find(text, declaration, nil, true)
			if not find(sub(text, declaration_index - 7, declaration_index + #declaration), "local " .. declaration, nil, true) then
				data.global = true
			end
			
			--[[ Expand the declaration until the full value is found. This can
				 be somewhat difficult to find. This looks for:
					- Matched brackets, as in a table declaration
					- End of function
					- Comma-separated declarations of multiple variables ]]
				
			-- Simplest case first - regular, single-line declarations, not in tables.
			data.value = match(text, declaration .. "(+)", declaration_index)
			
			-- Table declarations.
			if not data.value then
				data.value = match(text, declaration .. "(%b{}.-)", declaration_index)
				
				-- End of function.
				if not data.value then
					data.value = match(text, "(" .. declaration .. ".-) end")
					
					-- Comma-separated. TODO!
					--if not data.value then
					--	
					--end
					
					if not data.value then
						break -- skip
					end
				end
			end
			
			data.name = match(declaration, "local (-) ?= ?") or match(declaration, "(-) ?= ?")
			data.declaration = ((not data.global) and "local " or "") .. data.name .. " = " .. data.value
			_, data.uses = gsub(text, "" .. data.name .. "", "") -- dummy gsub to get count
			
			-- Find the number of times it's declared and subtract that from uses.
			local _, declarations = gsub(text, data.name .. " ?= ?(+)", "")
			data.uses = data.uses - declarations
				
			_, data.declared_at = gsub("\n" .. sub(text, 1, declaration_index), "\n", "")
			
			keys = Key:new(data)
			n = n + 1
		until true
	end
	
	keys.amount = #keys
	
	return keys
end

function export.test(frame)
	return frame:preprocess('<syntaxhighlight lang="Lua">\n' .. mw.dumpObject(export.grab_keys(frame.args)) .. '\n</syntaxhighlight>')
end

return export