More actions
add unstripNoWiki, update parseTemplateFormat to use gsplit options, update documentation |
m 1 revision imported |
||
(No difference)
| |||
Latest revision as of 07:08, 7 June 2024
--- Miscellaneous useful functions.
local lib = {}
local util = require('libraryUtil')
local checkType = util.checkType
local checkTypeMulti = util.checkTypeMulti
local NIL_OK = true
--- Choose one of two values to return.
-- @param {boolean} cond Determines which value to return.
-- @param T The value to return if `cond` is true (or truthy).
-- @param F The value to return if `cond` is false (or falsey).
function lib.ternary(cond, T, F)
if cond then
return T
end
return F
end
--- Some functions from `mw.text` are slow or may not always work as intended.
-- This group of functions provides better alternatives for them.
-- @section `mw.text` replacements
---
-- Removes ASCII whitespace from the start and end of `text`.
-- Slightly more efficient than the `mw.text` implementation.
-- @see https://help.fandom.com/wiki/Extension:Scribunto#mw.text.trim_is_slow
-- @param {string} text The text to trim.
-- @return {string} The trimmed text.
function lib.trim(text)
return (text:gsub( '^[\t\r\n\f ]+', '' ):gsub( '[\t\r\n\f ]+$', '' ))
-- the "extra" parentheses are important for removing the second value returned from `gsub`
end
--- Returns an iterator over substrings that would be returned by @{lib.split}.
-- @see @{lib.split} for argument documentation.
-- @return {function} Iterator over substrings of `text` separated by `delim`.
function lib.gsplit(text, delim, opt)
checkType('Feature.gsplit', 1, text, 'string')
checkType('Feature.gsplit', 2, delim, 'string', NIL_OK)
checkType('Feature.gsplit', 3, opt, 'table', NIL_OK)
if delim == nil then delim = " " end
if opt == nil then opt = {} end
-- the mediawiki implementation uses ustring, which is slower than string
-- and also not necessary if delim isn't a pattern.
-- https://help.fandom.com/wiki/Extension:Scribunto#mw.text.split_is_very_slow
-- local g = mw.text.gsplit(text, delim, true)
-- local function f()
-- local value = g()
-- if value and not opt.noTrim then -- value is nil when the generator ends
-- value = mw.text.trim(value)
-- end
-- if value == "" and opt.removeEmpty then
-- return f()
-- end
-- return value
-- end
-- return f, nil, nil
-- based on https://github.com/wikimedia/mediawiki-extensions-Scribunto/blob/1eecdac6def6418fb36829cc2f20b464c30e4b37/includes/Engines/LuaCommon/lualib/mw.text.lua#L222
local s, l = 1, #text
local function f()
if s then
local e, n = string.find( text, delim, s, true )
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
if not opt.noTrim then
ret = lib.trim(ret)
end
if ret == '' and opt.removeEmpty then
return f()
end
return ret
end
end
return f, nil, nil
end
--- Returns a table containing the substrings of `text` that are separated by `delim`.
-- (Compared to @{mw.text.split}, this function always treats `delim` as a literal
-- string rather than a pattern, and it trims output substrings using @{lib.trim} by default.)
-- @param {string} text The string to split.
-- @param[opt] {string} delim The delimiter string to use when splitting `text`.
-- (Using an empty string will split `text` into individual characters.)
-- @param[opt] {table} opt Extra options:
-- @param[opt] {boolean} opt.noTrim Set true to disable trimming of generated substrings
-- using @{lib.trim}.
-- @param[opt] {boolean} opt.removeEmpty Set true to omit empty substrings
-- (i.e., when multiple `delim` appear consecutively, or when
-- `delim` appears at the start or end of `text`).
-- @return {table} Substrings of `text separated by `delim`.
function lib.split(text, delim, opt)
checkType('Feature.split', 1, text, 'string')
checkType('Feature.split', 2, delim, 'string', NIL_OK)
checkType('Feature.split', 3, opt, 'table', NIL_OK)
local output = {}
for item in lib.gsplit(text, delim, opt) do
table.insert(output, item)
end
return output
end
--- A wrapper around @{mw.text.unstripNoWiki} that un-escapes
-- characters/sequences that <nowiki> escapes.
-- @see https://github.com/wikimedia/mediawiki/blob/c22d01f23b7fe754ef106e97bae32c3966f8db3e/includes/parser/CoreTagHooks.php#L146
-- for MediaWiki source code for <nowiki>
function lib.unstripNoWiki(str)
return (mw.text.unstripNoWiki(str)
:gsub('<', '<'):gsub('>', '>')
:gsub('-{', '-{'):gsub('}-', '}-'))
-- the "extra" parentheses are important for removing the second value returned from `gsub`
end
--- @section end
--- Returns an iterator over `tbl` that outputs items in the order defined by `order`.
-- @param {table} tbl The table to iterate over.
-- @param[opt] {table|function} order The iteration order.
--
-- Can be specified either as an ordered list (table) of keys from `tbl`
-- or as an ordering function that accepts `tbl`, `keyA`, and `keyB`
-- and returns true when the entry in `tbl` associated wity `keyA`
-- should come before the one for `keyB`.
--
-- If not specified, the keys' natural ordering is used
-- (i.e., `function(tbl, a, b) return a < b end`).
-- @return {function} The iterator.
function lib.spairs(tbl, order)
checkType('Feature.spairs', 1, tbl, 'table')
checkTypeMulti('Feature.spairs', 2, order, {'table', 'function', 'nil'})
local keys
if type(order) == "table" then
keys = order
else
-- collect the keys
keys = {}
for k in pairs(tbl) do table.insert(keys, k) end
-- sort the keys (using order function if given)
if order then
table.sort(keys, function(a, b) return order(tbl, a, b) end)
else
table.sort(keys)
end
end
-- return the iterator function
local i = 0
return function()
i = i + 1
local key = keys[i]
return key, tbl[key]
end
end
--[[
Parses Phantom Template Format strings into a list of maps.
@param {string} input A string formed by concatenating the output of Phantom Templates.
Usually, this string is generated by DPL.
@param[opt] {string} key_separator Separator between the entries (key-value pairs) of items in `input`. Defaults to ';'.
@param[opt] {string} end_separator Separator between items in `input`. Defaults to '$'.
@param[opt] {string} equal_separator Separator between the key and value of each entry in `input`. Defaults to '='.
@return {table} A list of items from `input`; each value is a map of the item's entries.
--]]
function lib.parseTemplateFormat (inputStr, key_separator, end_separator, equal_separator)
if key_separator == nil then key_separator = ";" end
if end_separator == nil then end_separator = "$" end
if equal_separator == nil then equal_separator = "=" end
local arg_format = "^%s*(.-)%s*" .. equal_separator .. "%s*(.-)%s*$"
local resultTable = {}
for str in lib.gsplit(inputStr, end_separator, {noTrim=true, removeEmpty=true}) do
local result = {}
for param in lib.gsplit(str, key_separator) do
local arg, val = param:match(arg_format)
if arg then
result[arg] = val
else
-- skip, i guess
-- mw.log("Warning: Lua module found extra " .. key_separator .. " or " .. end_separator .. " separators in DPL output.")
end
end
table.insert(resultTable, result)
end
return resultTable
end
--[=[
Parses Phantom Template Format strings into a list of ordered maps.
@param {string} input A string formed by concatenating the output of Phantom Templates.
Usually, this string is generated by DPL.
@param[opt] {string} key_separator Separator between the entries (key-value pairs) of items in `input`. Defaults to ';'.
@param[opt] {string} end_separator Separator between items in `input`. Defaults to '$'.
@param[opt] {string} equal_separator Separator between the key and value of each entry in `input`. Defaults to '='.
@return[name=output] {table} A list of items from `input`; each value is a list of the item's entries.
@return[name=output[i]] {table} The i-th item of `input`.
@return[name=output[i].page] {string} The value of the `page` key for this item.
@return[name=output[i][j]] {table} The j-th key-value pair of this item.
@return[name=output[i][j].key] {string} The j-th key of this item.
@return[name=output[i][j].value] The j-th value of this item.
--]=]
function lib.parseTemplateFormatOrdered (inputStr, key_separator, end_separator, equal_separator)
if key_separator == nil then key_separator = ";" end
if end_separator == nil then end_separator = "$" end
if equal_separator == nil then equal_separator = "=" end
local arg_format = "^%s*(.-)%s*" .. equal_separator .. "%s*(.-)%s*$"
local resultTable = {}
for str in lib.gsplit(inputStr, end_separator, {noTrim=true, removeEmpty=true}) do
local result = {}
for param in lib.gsplit(str, key_separator) do
local arg, val = param:match(arg_format)
if arg == 'page' then
result['page'] = val
else
table.insert(result,{
key = arg,
value = val
})
end
end
table.insert(resultTable, result)
end
return resultTable
end
-- searches ordered table and returns value
function lib.orderedTableSearch(tbl, search)
for i, obj in ipairs(tbl) do
if obj.key == search then
return obj.value
end
end
return false
end
--- Add thousands separator to number `n`.
-- @param {number|frame} n If a frame is given, then its first argument (`frame.args[1]`) will be used as input instead.
-- @return {string} The number formatted with commas for thousands separators.
-- @see https://stackoverflow.com/questions/10989788/format-integer-in-lua/10992898#10992898
function lib.thousandsSeparator(n)
if (n == mw.getCurrentFrame()) then
n = n.args[1]
elseif (type(n) == "table") then
n = n[1]
end
local i, j, minus, int, fraction = tostring(n):find('([-]?)(%d+)([.]?%d*)')
-- reverse the int-string and append a comma to all blocks of 3 digits
int = int:reverse():gsub("(%d%d%d)", "%1,")
-- reverse the int-string back remove an optional comma and put the optional minus and fractional part back
return minus .. int:reverse():gsub("^,", "") .. fraction
end
--- @return {boolean} true iff string or table is empty
-- @note May not be correct for tables with metatables.
function lib.isEmpty(item)
if item == nil or item == "" then
return true
end
if type(item) == "table" then
return next(item) == nil
end
return false
end
--- @return {boolean} true iff string or table is not empty
-- @note May not be correct for tables with metatables.
function lib.isNotEmpty(item)
return not lib.isEmpty(item)
end
--- @return nil if string or table is empty, otherwise return the value.
function lib.nilIfEmpty(item)
if lib.isEmpty(item) then
return nil
else
return item
end
end
---
-- @param {table} t A table of items
-- @param elm The item to search for
-- @returns true if `elm` is a value in `t`; false otherwise. (Does not check keys of `t`.)
-- @see http://stackoverflow.com/q/2282444
-- @see another implementation: Dev:TableTools.includes()
function lib.inArray(t, elm)
for _, v in pairs(t) do
if v == elm then
return true
end
end
return false
end
return lib