shannon-entropy-rs/src/lib.rs

pub const BASE64_ALPHABET : &str = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=-_";
pub const HEX_ALPHABET : &str = "ABCDEFabcdef0123456789";

// Calculate the shannon entropy of a given byte slice
pub fn shannon_entropy(bytes: &[u8]) -> f32 {
    let mut entropy = 0.0;
    let mut counts = [0; 256];

    for &b in bytes {
        counts[b as usize] += 1;
    }

    for &count in counts.iter() {
        if count == 0 { continue }

        let p: f32 = (count as f32) / (bytes.len() as f32);
        entropy -= p * p.log(2.0);
    }

    entropy
}

// Calculate the shannon entropy of a given byte slice for a specific charset
pub fn shannon_entropy_charset(bytes: &[u8], charset: &str) -> f32 {
    let mut entropy = 0.0;

    for single_char in charset.chars() {
        let count = bytes.iter().filter(|&&n| n as char == single_char).count();
        if count == 0 { continue }
        let p: f32 = (count as f32) / (bytes.len() as f32);
        entropy -= p * p.log(2.0);
    }

    entropy
}

// Determine if a string is made up only of hexdigits
pub fn is_hex_str(s: &str) -> bool {
    s.chars().filter(|&n| !(n.is_digit(16))).count() == 0
}

// Determine if a string is made up only of base64/base64url digits
pub fn is_base64_str(s: &str) -> bool {
    let alphabet: Vec<_> = BASE64_ALPHABET.chars().collect();
    s.chars().filter(|n| !(alphabet.contains(n))).count() == 0
}

// Calculate the shannon entropy of a given string
pub fn str_entropy(s: &str) -> f32 {
    if is_hex_str(s) {
        return shannon_entropy_charset(s.as_bytes(), HEX_ALPHABET);
    }

    if is_base64_str(s) {
        return shannon_entropy_charset(s.as_bytes(), BASE64_ALPHABET);
    }

    shannon_entropy(s.as_bytes())
}

// Tests
#[cfg(test)]
mod tests {
    use super::{shannon_entropy, str_entropy, is_hex_str, is_base64_str};

    #[test]
    fn test_entropy() {
        let test_strings = vec![
            ("hello world", 2.8453512),
            ("hello worldd", 2.8553884),
            ("a", 0.0),
            ("aaaaa", 0.0),
            ("ab", 1.0),
            ("aab", 0.9182958),
            ("", 0.0),
        ];

        for (test, answer) in test_strings {
            let entropy: f32 = str_entropy(test);
            assert_eq!(entropy, answer);
        }  
    }

    #[test]
    fn test_entropy_equal_distribution1() {
        let mut bytes = [0u8; 256];
        for i in 0..256 {
            bytes[i] = i as u8;
        }

        let h = shannon_entropy(&bytes);
        assert_eq!(h, 8.0);
    }

    #[test]
    fn test_entropy_equal_distribution2() {
        let mut bytes = [0u8; 256*2];
        for i in 0..bytes.len() {
            bytes[i] = (i % 256) as u8;
        }

        let h = shannon_entropy(&bytes);
        assert_eq!(h, 8.0);
    }

    #[test]
    fn test_entropy_helloworld() {
        let h = str_entropy("hello, world");
        assert_eq!(h, 3.0220551);
        let h = str_entropy("hello world");
        assert_eq!(h, 2.8453512);
    }

    #[test]
    fn test_hex_str_reconizer() {
        let s = is_hex_str("0123456789abcdef0123456789abcdef");
        assert_eq!(s, true);
        let s = is_hex_str("68656c6c6f20776f726c64");
        assert_eq!(s, true);
        let s = is_hex_str("g");
        assert_eq!(s, false);
    }

    #[test]
    fn test_base64_str_reconizer() {
        let s = is_base64_str("aGVsbG8gd29ybGQ=");
        assert_eq!(s, true);
        let s = is_base64_str("#@$");
        assert_eq!(s, false);
    }
}