Add Levenshtein distance functions to string lib

[Beherith]

--- Calculates edit distance of two strings, O(n*m) time, O(n) memory
	---@param a string
	---@param b string
	---@return number 

	function string.Levenshtein(a,b)

	--- Finds string that is closest to a in a table
	---@param a string
	---@param t table, primarily values are strings, keys can be strings too
	---@return string, number bestresult, bestscore
	function string.FindClosest(a,t)

[salinecitrine]

What's the use case for this? Probably worth mentioning in the PR description.

[salinecitrine]

rename to match other functions, and to make it a little more clear what it does

Suggested change

	if not string.Levenshtein then
	if not string.levenshteinDistance then

(and changes on related lines for this and FindClosest)

[salinecitrine]

already defined above

Suggested change

local ssub = string.sub

[salinecitrine]

Suggested change

	-- print(string.Levenshtein(string.rep("asdfasdfasdfasdf", 1000), string.rep("asdfasdfasdfasda", 1000))) -- 5 seconds
	-- print(string.FindClosest("apple", {"pear", "popple","bear","ple"}))

[salinecitrine]

Suggested change

--print(table.unpack(Levenshtein0))

[salinecitrine]

The function gives incorrect output for at least these test cases:

{
	name = "returns the length of the non-empty string for one empty string",
	args = { "", "apple" },
	expected = 5,
},
{
	name = "returns the length of the non-empty string for one empty string (reverse case)",
	args = { "apple", "" },
	expected = 5,
},
{
	name = "returns 1 for strings differing by one character",
	args = { "apple", "appla" },
	expected = 1,
},
{
	name = "returns the length of the longer string for completely different strings",
	args = { "cat", "dogs" },
	expected = 4,
},

This change fixes those tests:

Suggested change

	return Levenshtein1[lenb]
	return Levenshtein0[lenb + 1]

[WatchTheFort]

@Beherith What is the status of this?