Difference between revisions of "Module:TableTools"

From Meta
Jump to navigation Jump to search
(generate an error message when union and intersection functions are called with no arguments)
m (41 revisions imported)
 
(17 intermediate revisions by 2 users not shown)
Line 17: Line 17:
 
local infinity = math.huge
 
local infinity = math.huge
 
local checkType = libraryUtil.checkType
 
local checkType = libraryUtil.checkType
 
-- Define a unique value to represent NaN. This is because NaN cannot be used as a table key.
 
local nan = {}
 
  
 
--[[
 
--[[
Line 25: Line 22:
 
-- isPositiveInteger
 
-- isPositiveInteger
 
--
 
--
-- This function returns true if the given number is a positive integer, and false
+
-- This function returns true if the given value is a positive integer, and false
 
-- if not. Although it doesn't operate on tables, it is included here as it is
 
-- if not. Although it doesn't operate on tables, it is included here as it is
 
-- useful for determining whether a given table key is in the array part or the
 
-- useful for determining whether a given table key is in the array part or the
Line 31: Line 28:
 
------------------------------------------------------------------------------------
 
------------------------------------------------------------------------------------
 
--]]
 
--]]
function p.isPositiveInteger(num)
+
function p.isPositiveInteger(v)
if type(num) == 'number' and num >= 1 and floor(num) == num and num < infinity then
+
if type(v) == 'number' and v >= 1 and floor(v) == v and v < infinity then
 
return true
 
return true
 
else
 
else
Line 41: Line 38:
 
--[[
 
--[[
 
------------------------------------------------------------------------------------
 
------------------------------------------------------------------------------------
-- union
+
-- isNan
 
--
 
--
-- This returns the union of the key/value pairs of n tables. If any of the tables
+
-- This function returns true if the given number is a NaN value, and false
-- contain different values for the same table key, the table value is converted
+
-- if not. Although it doesn't operate on tables, it is included here as it is
-- to an array holding all of the different values.
+
-- useful for determining whether a value can be a valid table key. Lua will
 +
-- generate an error if a NaN is used as a table key.
 
------------------------------------------------------------------------------------
 
------------------------------------------------------------------------------------
 
--]]
 
--]]
function p.union(...)
+
function p.isNan(v)
local lim = select('#', ...)  
+
if type(v) == 'number' and tostring(v) == '-nan' then
if lim == 0 then
+
return true
error("no arguments passed to 'union'", 2)
+
else
 +
return false
 
end
 
end
local ret, trackArrays = {}, {}
+
end
for i = 1, lim do
 
local t = select(i, ...)
 
checkType('union', i, t, 'table')
 
for k, v in pairs(t) do
 
local retKey = ret[k]
 
if retKey == nil then
 
ret[k] = v
 
elseif retKey ~= v then
 
if trackArrays[k] then
 
local array = ret[k]
 
local valExists
 
for _, arrayVal in ipairs(array) do
 
if arrayVal == v then
 
valExists = true
 
break
 
end
 
end
 
if not valExists then
 
array[#array + 1] = v
 
ret[k] = array
 
end
 
else
 
ret[k] = {ret[k], v}
 
trackArrays[k] = true
 
end
 
end
 
end
 
end
 
return ret
 
end
 
  
 
--[[
 
--[[
 
------------------------------------------------------------------------------------
 
------------------------------------------------------------------------------------
-- valueUnion
+
-- shallowClone
 
--
 
--
-- This returns the union of the values of n tables, as an array. For example, for
+
-- This returns a clone of a table. The value returned is a new table, but all
-- the tables {1, 3, 4, 5, foo = 7} and {2, bar = 3, 5, 6}, union will return
+
-- subtables and functions are shared. Metamethods are respected, but the returned
-- {1, 2, 3, 4, 5, 6, 7}.
+
-- table will have no metatable of its own.
 
------------------------------------------------------------------------------------
 
------------------------------------------------------------------------------------
 
--]]
 
--]]
function p.valueUnion(...)
+
function p.shallowClone(t)
local lim = select('#', ...)
+
local ret = {}
if lim == 0 then
+
for k, v in pairs(t) do
error("no arguments passed to 'valueUnion'", 2)
+
ret[k] = v
end
 
local vals, ret = {}, {}
 
for i = 1, lim do
 
local t = select(i, ...)
 
checkType('valueUnion', i, t, 'table')
 
for k, v in pairs(t) do
 
if type(v) == 'number' and tostring(v) == '-nan' then
 
v = nan -- NaN cannot be a table key, so use a proxy variable.
 
end
 
vals[v] = true
 
end
 
end
 
for val in pairs(vals) do
 
if val == nan then
 
-- This ensures that we output a NaN when we had one as input, although
 
-- they may have been generated in a completely different way.
 
val = 0/0
 
end
 
ret[#ret + 1] = val
 
end
 
return ret
 
end
 
 
 
--[[
 
------------------------------------------------------------------------------------
 
-- intersection
 
--
 
-- This returns the intersection of the key/value pairs of n tables. Both the key
 
-- and the value must match to be included in the resulting table.
 
------------------------------------------------------------------------------------
 
--]]
 
function p.intersection(...)
 
local lim = select('#', ...)
 
if lim == 0 then
 
error("no arguments passed to 'intersection'", 2)
 
end
 
local ret, track, pairCounts = {}, {}, {}
 
for i = 1, lim do
 
local t = select(i, ...)
 
checkType('intersection', i, t, 'table')
 
for k, v in pairs(t) do
 
local trackVal = track[k]
 
if trackVal == nil then
 
track[k] = v
 
pairCounts[k] = 1
 
elseif trackVal == v then
 
pairCounts[k] = pairCounts[k] + 1
 
end
 
end
 
end
 
for k, v in pairs(track) do
 
if pairCounts[k] == lim then
 
ret[k] = v
 
end
 
 
end
 
end
 
return ret
 
return ret
Line 158: Line 73:
 
--[[
 
--[[
 
------------------------------------------------------------------------------------
 
------------------------------------------------------------------------------------
-- valueIntersection
+
-- removeDuplicates
 
--
 
--
-- This returns the intersection of the values of n tables, as an array. For
+
-- This removes duplicate values from an array. Non-positive-integer keys are
-- example, for the tables {1, 3, 4, 5, foo = 7} and {2, bar = 3, 5, 6},
+
-- ignored. The earliest value is kept, and all subsequent duplicate values are
-- intersection will return {3, 5}.
+
-- removed, but otherwise the array order is unchanged.
 
------------------------------------------------------------------------------------
 
------------------------------------------------------------------------------------
 
--]]
 
--]]
function p.valueIntersection(...)
+
function p.removeDuplicates(t)
local lim = select('#', ...)
+
checkType('removeDuplicates', 1, t, 'table')
if lim == 0 then
+
local isNan = p.isNan
error("no arguments passed to 'valueIntersection'", 2)
+
local ret, exists = {}, {}
end
+
for i, v in ipairs(t) do
local vals, ret = {}, {}
+
if isNan(v) then
for i = 1, lim do
+
-- NaNs can't be table keys, and they are also unique, so we don't need to check existence.
local t = select(i, ...)
+
ret[#ret + 1] = v
checkType('valueIntersection', i, t, 'table')
+
else
for k, v in pairs(t) do
+
if not exists[v] then
if type(v) == 'number' and tostring(v) == '-nan' then
+
ret[#ret + 1] = v
v = nan -- NaN cannot be a table key, so use a proxy variable.
+
exists[v] = true
 
end
 
end
local valCount = vals[v] or 0
+
end
vals[v] = valCount + 1
 
end
 
end
 
for val, count in pairs(vals) do
 
if count == lim then
 
if val == nan then
 
-- This ensures that we output a NaN when we had one as input, although
 
-- they may have been generated in a completely different way.
 
val = 0/0
 
end
 
ret[#ret + 1] = val
 
end
 
 
end
 
end
 
return ret
 
return ret
end
+
end
  
 
--[[
 
--[[
Line 230: Line 133:
 
checkType('affixNums', 2, prefix, 'string', true)
 
checkType('affixNums', 2, prefix, 'string', true)
 
checkType('affixNums', 3, suffix, 'string', true)
 
checkType('affixNums', 3, suffix, 'string', true)
 +
 +
local function cleanPattern(s)
 +
-- Cleans a pattern so that the magic characters ()%.[]*+-?^$ are interpreted literally.
 +
s = s:gsub('([%(%)%%%.%[%]%*%+%-%?%^%$])', '%%%1')
 +
return s
 +
end
 +
 
prefix = prefix or ''
 
prefix = prefix or ''
 
suffix = suffix or ''
 
suffix = suffix or ''
 +
prefix = cleanPattern(prefix)
 +
suffix = cleanPattern(suffix)
 
local pattern = '^' .. prefix .. '([1-9]%d*)' .. suffix .. '$'
 
local pattern = '^' .. prefix .. '([1-9]%d*)' .. suffix .. '$'
 +
 
local nums = {}
 
local nums = {}
 
for k, v in pairs(t) do
 
for k, v in pairs(t) do
Line 244: Line 157:
 
table.sort(nums)
 
table.sort(nums)
 
return nums
 
return nums
 +
end
 +
 +
--[[
 +
------------------------------------------------------------------------------------
 +
-- numData
 +
--
 +
-- Given a table with keys like ("foo1", "bar1", "foo2", "baz2"), returns a table
 +
-- of subtables in the format
 +
-- { [1] = {foo = 'text', bar = 'text'}, [2] = {foo = 'text', baz = 'text'} }
 +
-- Keys that don't end with an integer are stored in a subtable named "other".
 +
-- The compress option compresses the table so that it can be iterated over with
 +
-- ipairs.
 +
------------------------------------------------------------------------------------
 +
--]]
 +
function p.numData(t, compress)
 +
checkType('numData', 1, t, 'table')
 +
checkType('numData', 2, compress, 'boolean', true)
 +
local ret = {}
 +
for k, v in pairs(t) do
 +
local prefix, num = mw.ustring.match(tostring(k), '^([^0-9]*)([1-9][0-9]*)$')
 +
if num then
 +
num = tonumber(num)
 +
local subtable = ret[num] or {}
 +
if prefix == '' then
 +
-- Positional parameters match the blank string; put them at the start of the subtable instead.
 +
prefix = 1
 +
end
 +
subtable[prefix] = v
 +
ret[num] = subtable
 +
else
 +
local subtable = ret.other or {}
 +
subtable[k] = v
 +
ret.other = subtable
 +
end
 +
end
 +
if compress then
 +
local other = ret.other
 +
ret = p.compressSparseArray(ret)
 +
ret.other = other
 +
end
 +
return ret
 
end
 
end
  
Line 283: Line 237:
 
local key = nums[i]
 
local key = nums[i]
 
return key, t[key]
 
return key, t[key]
 +
else
 +
return nil, nil
 
end
 
end
 
end
 
end

Latest revision as of 10:21, 5 March 2018

Documentation for this module may be created at Module:TableTools/doc

--[[
------------------------------------------------------------------------------------
--                               TableTools                                       --
--                                                                                --
-- This module includes a number of functions for dealing with Lua tables.        --
-- It is a meta-module, meant to be called from other Lua modules, and should     --
-- not be called directly from #invoke.                                           --
------------------------------------------------------------------------------------
--]]

local libraryUtil = require('libraryUtil')

local p = {}

-- Define often-used variables and functions.
local floor = math.floor
local infinity = math.huge
local checkType = libraryUtil.checkType

--[[
------------------------------------------------------------------------------------
-- isPositiveInteger
--
-- This function returns true if the given value is a positive integer, and false
-- if not. Although it doesn't operate on tables, it is included here as it is
-- useful for determining whether a given table key is in the array part or the
-- hash part of a table.
------------------------------------------------------------------------------------
--]]
function p.isPositiveInteger(v)
	if type(v) == 'number' and v >= 1 and floor(v) == v and v < infinity then
		return true
	else
		return false
	end
end

--[[
------------------------------------------------------------------------------------
-- isNan
--
-- This function returns true if the given number is a NaN value, and false
-- if not. Although it doesn't operate on tables, it is included here as it is
-- useful for determining whether a value can be a valid table key. Lua will
-- generate an error if a NaN is used as a table key.
------------------------------------------------------------------------------------
--]]
function p.isNan(v)
	if type(v) == 'number' and tostring(v) == '-nan' then
		return true
	else
		return false
	end
end

--[[
------------------------------------------------------------------------------------
-- shallowClone
--
-- This returns a clone of a table. The value returned is a new table, but all
-- subtables and functions are shared. Metamethods are respected, but the returned
-- table will have no metatable of its own.
------------------------------------------------------------------------------------
--]]
function p.shallowClone(t)
	local ret = {}
	for k, v in pairs(t) do
		ret[k] = v
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- removeDuplicates
--
-- This removes duplicate values from an array. Non-positive-integer keys are
-- ignored. The earliest value is kept, and all subsequent duplicate values are
-- removed, but otherwise the array order is unchanged.
------------------------------------------------------------------------------------
--]]
function p.removeDuplicates(t)
	checkType('removeDuplicates', 1, t, 'table')
	local isNan = p.isNan
	local ret, exists = {}, {}
	for i, v in ipairs(t) do
		if isNan(v) then
			-- NaNs can't be table keys, and they are also unique, so we don't need to check existence.
			ret[#ret + 1] = v
		else
			if not exists[v] then
				ret[#ret + 1] = v
				exists[v] = true
			end
		end	
	end
	return ret
end			

--[[
------------------------------------------------------------------------------------
-- numKeys
--
-- This takes a table and returns an array containing the numbers of any numerical
-- keys that have non-nil values, sorted in numerical order.
------------------------------------------------------------------------------------
--]]
function p.numKeys(t)
	checkType('numKeys', 1, t, 'table')
	local isPositiveInteger = p.isPositiveInteger
	local nums = {}
	for k, v in pairs(t) do
		if isPositiveInteger(k) then
			nums[#nums + 1] = k
		end
	end
	table.sort(nums)
	return nums
end

--[[
------------------------------------------------------------------------------------
-- affixNums
--
-- This takes a table and returns an array containing the numbers of keys with the
-- specified prefix and suffix. For example, for the table
-- {a1 = 'foo', a3 = 'bar', a6 = 'baz'} and the prefix "a", affixNums will
-- return {1, 3, 6}.
------------------------------------------------------------------------------------
--]]
function p.affixNums(t, prefix, suffix)
	checkType('affixNums', 1, t, 'table')
	checkType('affixNums', 2, prefix, 'string', true)
	checkType('affixNums', 3, suffix, 'string', true)

	local function cleanPattern(s)
		-- Cleans a pattern so that the magic characters ()%.[]*+-?^$ are interpreted literally.
		s = s:gsub('([%(%)%%%.%[%]%*%+%-%?%^%$])', '%%%1')
		return s
	end

	prefix = prefix or ''
	suffix = suffix or ''
	prefix = cleanPattern(prefix)
	suffix = cleanPattern(suffix)
	local pattern = '^' .. prefix .. '([1-9]%d*)' .. suffix .. '$'

	local nums = {}
	for k, v in pairs(t) do
		if type(k) == 'string' then			
			local num = mw.ustring.match(k, pattern)
			if num then
				nums[#nums + 1] = tonumber(num)
			end
		end
	end
	table.sort(nums)
	return nums
end

--[[
------------------------------------------------------------------------------------
-- numData
--
-- Given a table with keys like ("foo1", "bar1", "foo2", "baz2"), returns a table
-- of subtables in the format 
-- { [1] = {foo = 'text', bar = 'text'}, [2] = {foo = 'text', baz = 'text'} }
-- Keys that don't end with an integer are stored in a subtable named "other".
-- The compress option compresses the table so that it can be iterated over with
-- ipairs.
------------------------------------------------------------------------------------
--]]
function p.numData(t, compress)
	checkType('numData', 1, t, 'table')
	checkType('numData', 2, compress, 'boolean', true)
	local ret = {}
	for k, v in pairs(t) do
		local prefix, num = mw.ustring.match(tostring(k), '^([^0-9]*)([1-9][0-9]*)$')
		if num then
			num = tonumber(num)
			local subtable = ret[num] or {}
			if prefix == '' then
				-- Positional parameters match the blank string; put them at the start of the subtable instead.
				prefix = 1
			end
			subtable[prefix] = v
			ret[num] = subtable
		else
			local subtable = ret.other or {}
			subtable[k] = v
			ret.other = subtable
		end
	end
	if compress then
		local other = ret.other
		ret = p.compressSparseArray(ret)
		ret.other = other
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- compressSparseArray
--
-- This takes an array with one or more nil values, and removes the nil values
-- while preserving the order, so that the array can be safely traversed with
-- ipairs.
------------------------------------------------------------------------------------
--]]
function p.compressSparseArray(t)
	checkType('compressSparseArray', 1, t, 'table')
	local ret = {}
	local nums = p.numKeys(t)
	for _, num in ipairs(nums) do
		ret[#ret + 1] = t[num]
	end
	return ret
end

--[[
------------------------------------------------------------------------------------
-- sparseIpairs
--
-- This is an iterator for sparse arrays. It can be used like ipairs, but can
-- handle nil values.
------------------------------------------------------------------------------------
--]]
function p.sparseIpairs(t)
	checkType('sparseIpairs', 1, t, 'table')
	local nums = p.numKeys(t)
	local i = 0
	local lim = #nums
	return function ()
		i = i + 1
		if i <= lim then
			local key = nums[i]
			return key, t[key]
		else
			return nil, nil
		end
	end
end

--[[
------------------------------------------------------------------------------------
-- size
--
-- This returns the size of a key/value pair table. It will also work on arrays,
-- but for arrays it is more efficient to use the # operator.
------------------------------------------------------------------------------------
--]]
function p.size(t)
	checkType('size', 1, t, 'table')
	local i = 0
	for k in pairs(t) do
		i = i + 1
	end
	return i
end

return p