CS3700-project3/hptp/src/encoding.rs

use lazy_static::lazy_static;
use regex::bytes::Regex;

// MeowCoder is a silly way to get around the tests being specifically hex data wrapped at 60 chars
// We detect input data of this form and encode it into raw bytes to achieve basically 50%
// compression on data sent over HPTP

lazy_static! {
    static ref ENCODING_DETECTOR: Regex =
        Regex::new(r"^\n?([0-9a-fA-F]{60}\n)*[0-9a-fA-F]{0,60}$").unwrap();
    static ref HEX_DETECTOR: Regex = Regex::new(r"^[0-9a-fA-F]{0,60}$").unwrap();
}

static WRAP_SIZE: usize = 60;

// A struct that keeps track of where the last wrap position was at so we can print the data
// correctly
#[derive(Clone)]
pub struct MeowCoder {
    line_index: usize,
}

impl MeowCoder {
    // check if "encoding" by replacing the newline-wrapped hex with raw bytes is possible
    // this takes an index representing the amount of encoded data sent so far, which
    // determines where to end the next line of hex according to WRAP_SIZE
    pub fn can_be_encoded(data: &[u8], index: usize) -> bool {
        if index % (WRAP_SIZE / 2) == 0 {
            ENCODING_DETECTOR.is_match(data)
        } else {
            let nl_pos = data.iter().position(|&r| r == '\n' as u8);
            match nl_pos {
                Some(nl_idx) => {
                    (nl_idx / 2 + index) % (WRAP_SIZE / 2) == 0
                        && ENCODING_DETECTOR.is_match(&data[nl_idx..])
                        && (nl_idx == 0 || HEX_DETECTOR.is_match(&data[..nl_idx - 1]))
                }
                None => false,
            }
        }
    }

    pub fn new() -> MeowCoder {
        MeowCoder { line_index: 0 }
    }

    pub fn hex_to_nibble(chr: u8) -> u8 {
        const _AL: u8 = 'a' as u8;
        const _FL: u8 = 'f' as u8;
        const _A: u8 = 'A' as u8;
        const _F: u8 = 'F' as u8;
        const _0: u8 = '0' as u8;
        const _9: u8 = '9' as u8;
        match chr {
            _AL..=_FL => chr - _AL + 10,
            _A..=_F => chr - _A + 10,
            _0..=_9 => chr - _0,
            _ => panic!("bad hex"),
        }
    }

    pub fn u8_to_hex(val: u8) -> (u8, u8) {
        let first = val >> 4;
        let second = val & 0xF;
        const LOOKUP: &'static [u8; 16] = b"0123456789abcdef";
        (LOOKUP[first as usize], LOOKUP[second as usize])
    }

    pub fn encode(input: &Vec<u8>) -> (Vec<u8>, bool) {
        let mut out: Vec<u8> = Vec::new();
        let mut prev_char: u8 = 0;
        let mut pair_first = false;
        for chr in input {
            if *chr == '\n' as u8 {
                continue;
            }
            if !pair_first {
                prev_char = *chr;
                pair_first = true;
            } else {
                let byte_value: u8 =
                    MeowCoder::hex_to_nibble(prev_char) * 16 + MeowCoder::hex_to_nibble(*chr);
                out.push(byte_value);
                pair_first = false;
            }
        }
        // trailing byte specifies whether this was cut or not
        if pair_first {
            out.push(MeowCoder::hex_to_nibble(prev_char) * 16);
            (out, true)
        } else {
            (out, false)
        }
    }

    pub fn decode(&mut self, input: &Vec<u8>, was_cut: bool) -> Vec<u8> {
        let mut out: Vec<u8> = Vec::new();
        for (pos, byte) in input.iter().enumerate() {
            let (first, second) = MeowCoder::u8_to_hex(*byte);
            out.push(first);
            if pos < input.len() - 1 || !was_cut {
                out.push(second);
            }
            self.line_index += 2;
            if self.line_index == WRAP_SIZE {
                self.line_index = 0;
                out.push('\n' as u8);
            }
        }
        out
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn test_match() {
        assert_eq!(MeowCoder::can_be_encoded(b"abcd1234", 0), true);
        assert_eq!(MeowCoder::can_be_encoded(b"abcXd1234", 0), false);
        assert_eq!(
            MeowCoder::can_be_encoded(
                b"012345678901234567890123456789012345678901234567890123456789\nabcdef",
                0
            ),
            true
        );
        assert_eq!(
            MeowCoder::can_be_encoded(
                b"01234567890123456789012345678901234567890123456789012345678\nabcdef",
                0
            ),
            false
        );
        assert_eq!(MeowCoder::can_be_encoded(b"\x12\xab\x45\n", 0), false);
        assert_eq!(MeowCoder::can_be_encoded(b"abcd1234\n012345678901234567890123456789012345678901234567890123456789\nabcdefabcd", 26), true);
        assert_eq!(MeowCoder::can_be_encoded(b"abcd123456\nabcd", 25), true);
    }

    #[test]
    fn test_encode() {
        let vec = vec![
            'a' as u8, '4' as u8, 'c' as u8, 'd' as u8, '\n' as u8, 'e' as u8, 'f' as u8,
        ];
        assert_eq!(MeowCoder::encode(&vec), (vec![0xa4, 0xcd, 0xef], false));

        let vec2 = vec![
            'a' as u8, '4' as u8, 'c' as u8, 'd' as u8, '\n' as u8, 'e' as u8, 'f' as u8, '9' as u8,
        ];
        assert_eq!(
            MeowCoder::encode(&vec2),
            (vec![0xa4, 0xcd, 0xef, 0x90], true)
        );
    }

    #[test]
    fn test_encode_decode() {
        let hex_str: &[u8] =
            b"012345678901234567890123456789012345678901234567890123456789\nabcdef";
        let vec: Vec<u8> = Vec::from(hex_str);
        let (vec2, was_cut) = MeowCoder::encode(&vec);
        assert_eq!(was_cut, false);
        let mut coder = MeowCoder::new();
        let out = coder.decode(&vec2, was_cut);
        assert_eq!(out, vec);
    }

    #[test]
    fn test_encode_decode_cut() {
        let hex_str: &[u8] =
            b"012345678901234567890123456789012345678901234567890123456789\nabcdef3";
        let vec: Vec<u8> = Vec::from(hex_str);
        let (vec2, was_cut) = MeowCoder::encode(&vec);
        assert_eq!(was_cut, true);
        let mut coder = MeowCoder::new();
        let out = coder.decode(&vec2, was_cut);
        assert_eq!(out, vec);
    }
}