Find the closest prime number of a character's ASCII value --- Python

Question

I get a string input from user, breaking it up into a character list, converting each character into its ASCII value and then searching for the nearest Prime Number. Convert the (prime number) ASCII value back into a list of characters, join it together and print it on the screen. Sounds pretty simple but I find it tricky trying to find the nearest prime number.

def is_prime(num):
    if num > 1:
        for n in range(2, num):
            if (num % n) == 0:
                return False
        else:
            return True
    else:
        return False


test_cases = int(input())


for _ in range(test_cases):
    string = str(input())
    array = list(string)
    result_array = []

    for i in array:
        if is_prime(ord(i)):
            result_array.append(i)
        else:
            pass

    print("".join(result_array))

On the last line the "pass" statement is just to have an error-free code till I find the solution. That's the part where the searching goes on.

classical mistake: don't loop till the number itself: `for n in range(2, num)` but to `for n in range(2, int(num**0.5+1))` you'll save iterations. — Jean-François Fabre, Jul 28 '17 at 20:13
what you need is to generate the list of the prime numbers from 32 to 127 and use `bisect` to see which is the closest one. the code you provided does nothing close to that. — Jean-François Fabre, Jul 28 '17 at 20:15
Thanks! Just on time... Was on google searching for a more efficient way to find if a number is prime or not. — Αστερης Παπαδακης, Jul 28 '17 at 20:16
@Jean-FrançoisFabre You don't need to generate the list, you can hard code it. Only 31 prime numbers are <= 128 — Gribouillis, Jul 28 '17 at 20:16
@Gribouillis let's settle on "generate a hardcoded list then" :) — Jean-François Fabre, Jul 28 '17 at 20:18
@Jean-FrançoisFabre The algorithm to generate this hard coded list is by googling. — Gribouillis, Jul 28 '17 at 20:19
@Jean-FrançoisFabre what you say can be implemented in the last "else" section.. I just remove the pass and type in the stuff.. My problem, is to search the nearest number... Could you explain "bisect" a bit more? Python documentation goes around the answer but never answers it — Αστερης Παπαδακης, Jul 28 '17 at 20:23
@Jean-FrançoisFabre Your code, finds the nearest ascii value above the current number... Input: AFREEN should have the output CGSCCO but instead gives CGSGGO — Αστερης Παπαδακης, Jul 28 '17 at 20:52
issue fixed, and very short answer variation added (thanks for your patience, tuning the indices was not that trivial) — Jean-François Fabre, Jul 28 '17 at 21:04
@Jean-FrançoisFabre how could it be done in order to include lower case letters also.. — Αστερης Παπαδακης, Jul 29 '17 at 13:04

Jean-François Fabre · Accepted Answer · 2017-07-28T21:34:17.630

Your algorithm cannot work because it never tackles with the "closest to a prime" issue (note that checking a prime like that is not efficient: the max boundary can be n**0.5+1)

That said, there's a flaw in this problem for lower-case letters:

prime numbers skip from 113 to 127. For a high lowercase letter, the closest prime is 127, which isn't printable.

So I'll stick to uppercase letters (unless we want to print the closest lowest prime, in which case lowercase is OK)

What I would do is:

generate/copy the list of primes from 0 to 97 (ASCII for Z is 96)
loop on the characters, and use bisect module to find the insertion position of the character ASCII code in the prime list (which must be sorted for bisect to work properly).
Then check if upper bound is closer than lower bound, and choose the index of the prime list.
Fiddle (a lot) with the insertion indices to avoid errors (like in my first version of the post)
Add to list. join in the end

code:

import bisect

# list sampled from https://primes.utm.edu/lists/small/1000.txt
# ASCII code for 'A' is 65, no need to go lower
primes = [int(x) for x in """61 67     71     73     79     83     89     97""".split()]

word = "ABCDEFGHIJKLMNOPRSTUVWXYZ"

primeword = []
for w in word:
    ow = ord(w)
    i = bisect.bisect_left(primes,ow)
    delta1 = abs(ow-primes[i])
    delta2 = abs(ow-primes[i-1])
    # select this index or next index (no risk for out of range here)
    primeword.append(chr(primes[i + int(delta2 > delta1) - 1]))

print("".join(primeword))

which gives me:

CCCCCGGGIIIIOOOOSSSSSYYYY

EDIT: since we generated the prime numbers, we could as well directly generate the lookup table for the characters and use str.translate:

primeword_dict = {65: 'C', 66: 'C', 67: 'C', 68: 'C', 69: 'C', 70: 'G', 71: 'G', 72: 'G', 73: 'I', 74: 'I', 75: 'I', 76: 'I', 77: 'O', 78: 'O', 79: 'O', 80: 'O', 82: 'S', 83: 'S', 84: 'S', 85: 'S', 86: 'S', 87: 'Y', 88: 'Y', 89: 'Y', 90: 'Y'}
print(word.translate(primeword_dict))

that would be even faster & shorter, and allows to pass strings like "HELLO WORLD. (with spaces & punctuation in it) and only have the letters changed & other symbols kept intact.

Find the closest prime number of a character's ASCII value --- Python

1 Answers1