Daily Challenge #300 - Username / Password Comparison

You are writing a password validator for a website. You want to discourage users from using their username as part of their password, or vice-versa, because it is insecure. Since its unreasonable to not allow them to have any letters in common, you come up with this rule:

For any password and username pair, the length of the longest common substring should be less than half the length of the shortest of the two.

In other words, you won't allow users to have half their password repeated in their username, or half their username repeated in their password.

Write a function validate(username, password) which returns true if your rule is followed, false otherwise.

Assume both the username and the password may contain uppercase letters, lowercase letters, numbers, spaces, and the following special/punctation characters: !@#$%^&*()_+{}:"<>?[];',. The username and password will each be less than 100 characters.

Examples

validate("newyorkcity", "yankees101"), TRUE
validate("username", "usernamePW"), FALSE

Tests

validate("Paa2", "Paaaaaa222!!!")
validate("wordpass", "password")
validate("superman", "persuzod")

---

Want to propose a challenge idea for a future post? Email [e-mail] with your suggestions!

Comments

[peter279k]

Here is the simple solution with Python and difflib module:

import difflib


def validate(username, password):
    if len(username) == 0 and len(password) == 0:
        return False
    if len(username) == 0 or len(password) == 0:
        return False
    limit = min([len(password), len(username)]) / 2

    matches = difflib.SequenceMatcher(None, username, password).get_matching_blocks()
    size = 0
    for match in matches:
        if match.size >= limit:
            size += match.size

    if size == 0:
        return True
    return False

[Michael Baas]

Here's a solution in Dyalog APL:

∇ z←validate(a b);l;m;s;subs
 →(z←0∊l←≢¨a b)↑0                               ⍝ l=lengths, exit if there's 0 among them
 m←⌊/l                                          ⍝ lowest length
 s←(l⍳m)⊃a b                                    ⍝ get the shortest string
 m←⌊m÷2                                         ⍝ divide by 2
 subs←m{(↓⍺,[1.5]⍳(≢⍵)-⍺){⍺[1]↑⍺[2]↓⍵}¨⊂⍵}¨a b ⍝ look for matches of substrings of a in b (and vice versa)  ⍝ get substrings (of length m)
 z←~1∊∊subs{⍺⍷¨⊂⍵}¨b a                          ⍝ return 0 if we found any matches
∇

Running it:

      validate'newyorkcity' 'yankees101'
1
      validate'username' 'usernamePW'
0
      validate'Paa2' 'Paaaaaa222!!!'
0
      validate'wordpass' 'password'
0
      validate'superman' 'persuzod'
1
      validate'abc' 'defghijklmnopab'  ⍝ added this testcase
0

But don't take my word for it - Try It Online!

[Mihail Malo]

Rust

#![feature(once_cell)]

use aho_corasick::AhoCorasick;
use regex::Regex;
use std::collections::HashSet;
use std::lazy::SyncLazy;

static REGEX: SyncLazy<Regex> = SyncLazy::new(|| Regex::new(r"^[!-z&&[^\-/=\\]]{2,100}$").unwrap());

fn validate(u: &str, p: &str) -> bool {
    if [u, p].iter().any(|x| !REGEX.is_match(x)) {
        return false;
    };
    let (needle, haystack) = if u.len() < p.len() { (u, p) } else { (p, u) };

    let len = needle.len() / 2;
    let patterns = (0..needle.len() - len)
        .into_iter()
        .map(|i| &needle[i..i + len])
        .collect::<HashSet<_>>();
    let corasick = AhoCorasick::new(patterns);
    !corasick.is_match(haystack)
}

fn main() {
    assert_eq!(validate("newyorkcity", "yankees101"), true);
    assert_eq!(validate("username", "usernamePW"), false);
    dbg!(validate("Paa2", "Paaaaaa222!!!"));
    dbg!(validate("wordpass", "password"));
    dbg!(validate("superman", "persuzod"));
}

Look at it go!

A silly one.
Regex with range subtraction (all characters '!' to 'z' are accepted, except - / = and \), also checks for length.
So that regex is compiled only once, we construct it lazily.

Then we pick the shorter of the two strings, get all the overlapping windows into it, and colleсt them into a HashSet for uniqueness.

Then we use an Aho-Corasick automaton with all of these substrings to match on them simultaneously in one pass.

However, if you think about it, the same thing can be written without dependencies/nightly features
This trades CPU for memory, for instance:

fn ok_char(c: char) -> bool {
    match c {
        '-' | '/' | '=' | '\\' => false,
        '!'..='z' => true,
        _ => false,
    }
}

fn validate(u: &str, p: &str) -> bool {
    if [u, p].iter().any(|x| x.chars().any(|c| !ok_char(c))) {
        return false;
    }
    let (needle, haystack) = if u.len() < p.len() { (u, p) } else { (p, u) };

    let len = needle.len() / 2;
    !(0..needle.len() - len)
        .into_iter()
        .map(|i| &needle[i..i + len])
        .any(|w| haystack.contains(w))
}

fn main() {
    assert_eq!(validate("newyorkcity", "yankees101"), true);
    assert_eq!(validate("username", "usernamePW"), false);
    dbg!(validate("Paa2", "Paaaaaa222!!!"));
    dbg!(validate("wordpass", "password"));
    dbg!(validate("superman", "persuzod"));
}

Look at it go!