Permutations of strings takes too long to solve - ruby

Question

I'm creating an array of permutated and unique letters in a string, only to sort them alphabetically and find the middle element in the set.
def middle_permutation(string)
length = string.length
permutation_set = string.split("").permutation(length).to_a.map{|item| item.join}.sort
permutation_set.length.even? ? permutation_set[(permutation_set.length)/2-1] : permutation_set[(permutation_set.length/2)+1]
end
For example:
middle_permutation("zxcvbnmasd") should equal "mzxvsndcba"
Even for small strings (N >=10), the calculations take pretty long to finish, and I can forget about anything double that; is there a quicker way?

Answer 1

I'm assuming the letters are unique, as in the OP's question.
Sort
Pluck the middle letter of the sorted string (rounded down). This is the first letter of the middle permutation.
If the original list had an even number of letters, the rest of the permutation is the reverse sort of the remaining letters.
If not, take the middle letter again. Now the rest of the result is the reverse sort of the remaining letters.

Answer 2

The method below returns the desired permutation directly, without iterating through permutations.
The asker has stated that the string contains no duplicated letters, which is a requirement for this method. I assume the characters of the string are sorted. If they are not, the creation of a sorted string would be the first step:
str = "ebadc".chars.sort.join
#=> "abcde"
Code
def mid_perm(str)
return mid_perm_even_length_strings(str) if str.size.even?
first_char_index = str.size/2
str[first_char_index] << mid_perm_even_length_strings(str[0,first_char_index] +
str[first_char_index+1..-1])
end
def mid_perm_even_length_strings(str)
first_char_index = str.size/2-1
str[first_char_index] + (str[0,first_char_index] + str[first_char_index+1..-1]).reverse
end
Examples
mid_perm 'abcd'
#=> "bdca"
mid_perm 'abcde'
#=> "cbeda"
mid_perm 'abcdefghijklmnopqrstuvwxyz'
#=> "mzyxwvutsrqponlkjihgfedcba"
Explanation
Let's start by defining a method to produce permutations of the letters of a string.
def perms(str)
str.chars.permutation(str.size).map(&:join)
end
Strings containing an even number of characters
Consider
a = perms "abcd"
#=> ["abcd", "abdc", "acbd", "acdb", "adbc", "adcb",
# "bacd", "badc", "bcad", "bcda", "bdac", "bdca",
# "cabd", "cadb", "cbad", "cbda", "cdab", "cdba",
# "dabc", "dacb", "dbac", "dbca", "dcab", "dcba"]
a contains 4! #=> 4*3*2 => 24 elements, 4 being the length of the string.
Notice that since the characters in perms' argument are sorted, the array returned is also sorted1.
a == a.sort #=>true
As a.size #=> 24, the "middle" element is either a[11] #=> "bdca" or a[12] #=> "cabd" (where 11 = (24-1)/2 and 12 = 24/2), depending on how we want to round. The question stipulates that, for even-length strings, we are to round down, so that would be "bdca".
Now let's slice a into str.size equal arrays, each containing a.size/str.size #=> 24/4 => 6 elements:
b = a.each_slice(a.size/str.size).to_a
#=> [["abcd", "abdc", "acbd", "acdb", "adbc", "adcb"],
# ["bacd", "badc", "bcad", "bcda", "bdac", "bdca"],
# ["cabd", "cadb", "cbad", "cbda", "cdab", "cdba"],
# ["dabc", "dacb", "dbac", "dbca", "dcab", "dcba"]]
The desired element is therefore
b[(a.size/str.size-1)/2-1][-1]
#=> "bdca"
This value can be computed more directly as follows.
first_char_index = str.size/2-1
#=> 1
first_char = str[first_char_index]
#=> "b"
remaining_chars = (str[0,first_char_index] + str[first_char_index+1..-1]).reverse
#=> "dca"
first_char + remaining_chars
#=> "bdca"
The same logic applies to all strings having an even number of characters. We therefore can write the method mid_perm_even_length_strings shown in the Code section above.
For example (for a 12-character string)
mid_perm_even_length_strings 'abcdefghijkl'
#=> "flkjihgedcba"
Strings containing an odd number of characters
Now consider
str = "abcde"
a = perms str
#=> ["abcde", "abced", "abdce", "abdec", "abecd", "abedc",
# "acbde", "acbed", "acdbe", "acdeb", "acebd", "acedb",
# "adbce", "adbec", "adcbe", "adceb", "adebc", "adecb",
# "aebcd", "aebdc", "aecbd", "aecdb", "aedbc", "aedcb",
# "bacde", "baced", "badce", "badec", "baecd",..., "bedca",
# "cabde", "cabed", "cadbe", "cadeb", "caebd", "caedb",
# "cbade", "cbaed", "cbdae", "cbdea", "cbead", "cbeda",
# "cdabe", "cdaeb", "cdbae", "cdbea", "cdeab", "cdeba",
# "ceabd", "ceadb", "cebad", "cebda", "cedab", "cedba",
# "dabce", "dabec", "dacbe", "daceb", "daebc",..., "decba",
# "eabcd", "eabdc", "eacbd", "eacdb", "eadbc",..., "edcba"]
Here the permutation contains 5! #=> 100 elements, in 5 blocks of 20. (Again, a.each_cons(2).all? { |s1,s2| s1 < s2 } #=> true.)
The middle element of a is clearly the middle element of the block of elements that begin with
str[str.size/2] #=> "c"
That block would be the array
b = a.each_slice(a.size/str.size).to_a[str.size/2]
#=> ["cabde", "cabed", "cadbe", "cadeb", "caebd", "caedb",
# "cbade", "cbaed", "cbdae", "cbdea", "cbead", "cbeda",
# "cdabe", "cdaeb", "cdbae", "cdbea", "cdeab", "cdeba",
# "ceabd", "ceadb", "cebad", "cebda", "cedab", "cedba"]
which would be 'c' plus the middle element of the array
["abde", "abed", "adbe", "adeb", "aebd", "aedb",
"bade", "baed", "bdae", "bdea", "bead", "beda",
"dabe", "daeb", "dbae", "dbea", "deab", "deba",
"eabd", "eadb", "ebad", "ebda", "edab", "edba"]
That array is merely the permutations of the string "abde". Since that string contains an even number characters, its middle element is
mid_perm_even_length_strings 'abde'
#=> "beda"
It follows that the middle element of the permutations of the letters of "abcde" is therefore
'c' + 'abde'
#=> "cabde"
This clearly applies to all strings containing an odd number of characters.
1. The doc for Array#permutation states, "The implementation makes no guarantees about the order in which the permutations are yielded.". We therefore might need to tack .sort to the end of the operative line of perms, but with Ruby v2.4 (and I suspect, earlier versions) that is, in fact not necessary here.

Answer 3

I was able to compact it like this:
def middle_permutation(string)
list = string.chars.permutation.map(&:join).sort
list[list.length / 2 - (list.length.even? ? 1 : 0)]
end
Which yields:
middle_permutation('zxcvbnmasd')
# => "mzxvsndcba"

Answer 4

You don't need to generate all permutations. Just find overall number of permutations as PN = N! where N is string (of different chars) length and calculate only needed PN/2-th permutation by its number - for example, using this approach
public static int[] perm(int n, int k)
{
int i, ind, m=k;
int[] permuted = new int[n];
int[] elems = new int[n];
for(i=0;i<n;i++) elems[i]=i;
for(i=0;i<n;i++)
{
ind=m%(n-i);
m=m/(n-i);
permuted[i]=elems[ind];
elems[ind]=elems[n-i-1];
}
return permuted;
}

Answer 5

So it turns out there are two tracks to this, odd strings and even strings.
For odd strings, you take out the middle character Element of the sorted array and the one before it, in that order. When you do that you have two remaining arrays, the one the right and left, both alphabetically sorted. You tack on elements of the right array, starting with the last element, then do the same for the one on the left.
For even strings, Do the same but only take one character in the first step: the (N/2) element.
Here's my solution:
def middle_permutation(string)
string_array = string.chars.sort
mid_string = []
length = string.length
if length.even?
mid_string << string_array[length/2-1]
string_array.delete_at(length/2-1)
(mid_string << string_array.reverse).flatten.join
else
mid_string << string_array[(length/2)-1..length/2].reverse
string_array.slice!((length/2)-1, 2)
(mid_string << string_array.reverse).flatten.join
end
end