I've been attempting to refactor my original solution to a getPermutations() function using a point-free approach using Ramda.js. Is it possible to refactor it further, towards a point-free style. It looks like I just made an even bigger mess. Also, there is currently a bug in the refactored version when I run the tests: TypeError: reduce: list must be array or iterable.
Original Solution:
// getPermutations :: String -> [String]
function getPermutations(string) {
function permute(combination, permutations, s) {
if (!s.length) {
return permutations[combination] = true;
}
for (var i = 0; i < s.length; i++) {
permute( combination.concat(s[i])
, permutations
, (s.slice(0, i) + s.slice(i+1))
);
}
return Object.keys(permutations);
}
return permute('', {}, string);
}
My attempt to refactor with Ramda.js:
var _ = require('ramda');
// permute :: String -> {String: Boolean} -> String -> [String]
var permute = _.curry(function (combination, permutations, string) {
// callPermute :: String -> ({String: Bool} -> Char -> Int -> String) -> IO
var callPermute = function (combination) {
return function (acc, item, i, s) {
return permute( _.concat(combination, item)
, acc
, _.concat(_.slice(0, i, s), _.slice(i + Infinity, s))
);
};
};
var storeCombination = function () {
return permutations[combination] = true;
};
// should be an ifElse, incorporating compose below
_.when(_.not(string.length), storeCombination);
return _.compose( _.keys
, _.addIndex(_.reduce(callPermute(''), {}))
) (string.split(''));
});
// getPermutations :: String -> [String]
var getPermutations = permute('', {});
There seem to be several problems with your solution, and I'm afraid I don't have time to chase them down. (The first thing I see is that you are using addIndex
incorrectly.)
But if you want to see a working permutation
function in Ramda, I wrote this a while ago:
// permutations :: [a] -> [[a]]
const permutations = (tokens, subperms = [[]]) =>
R.isEmpty(tokens) ?
subperms :
R.addIndex(R.chain)((token, idx) => permutations(
R.remove(idx, 1, tokens),
R.map(R.append(token), subperms)
), tokens);
R.map(R.join(''), permutations(['A', 'B', 'C']));
//=> ["ABC", "ACB", "BAC", "BCA", "CAB", "CBA"]
(You can play with this on the Ramda REPL.)