atproto-plc / src / encoding.rs

//! Encoding utilities for base32, base64url, and DAG-CBOR

use crate::error::{PlcError, Result};
use base64::engine::general_purpose::URL_SAFE_NO_PAD;
use base64::Engine;
use cid::Cid;
use data_encoding::BASE32_NOPAD;
use multihash::Multihash;
use serde::{Deserialize, Serialize};
use sha2::{Digest, Sha256};

/// Base32 alphabet used for did:plc identifiers
/// Lowercase, excludes 0,1,8,9
const BASE32_ALPHABET: &str = "abcdefghijklmnopqrstuvwxyz234567";

/// Maximum size for an operation in bytes
pub const MAX_OPERATION_SIZE: usize = 7500;

/// Encode bytes to base32 using the lowercase alphabet
///
/// # Examples
///
/// ```
/// use atproto_plc::encoding::base32_encode;
///
/// let data = b"hello world";
/// let encoded = base32_encode(data);
/// assert!(!encoded.is_empty());
/// ```
pub fn base32_encode(data: &[u8]) -> String {
    BASE32_NOPAD.encode(data).to_lowercase()
}

/// Decode base32 string to bytes
///
/// # Errors
///
/// Returns `PlcError::InvalidBase32` if the input contains invalid characters
pub fn base32_decode(s: &str) -> Result<Vec<u8>> {
    // Validate that all characters are in the allowed alphabet
    if !s.chars().all(|c| BASE32_ALPHABET.contains(c)) {
        return Err(PlcError::InvalidBase32(format!(
            "String contains invalid characters. Allowed: {}",
            BASE32_ALPHABET
        )));
    }

    BASE32_NOPAD
        .decode(s.to_uppercase().as_bytes())
        .map_err(|e| PlcError::InvalidBase32(e.to_string()))
}

/// Encode bytes to base64url without padding
///
/// # Examples
///
/// ```
/// use atproto_plc::encoding::base64url_encode;
///
/// let data = b"hello world";
/// let encoded = base64url_encode(data);
/// assert!(!encoded.contains('='));
/// ```
pub fn base64url_encode(data: &[u8]) -> String {
    URL_SAFE_NO_PAD.encode(data)
}

/// Decode base64url string to bytes
///
/// # Errors
///
/// Returns `PlcError::InvalidBase64Url` if the input is not valid base64url
pub fn base64url_decode(s: &str) -> Result<Vec<u8>> {
    URL_SAFE_NO_PAD
        .decode(s.as_bytes())
        .map_err(|e| PlcError::InvalidBase64Url(e.to_string()))
}

/// Encode a value to DAG-CBOR format
///
/// # Errors
///
/// Returns `PlcError::DagCborError` if serialization fails or the result exceeds MAX_OPERATION_SIZE
pub fn dag_cbor_encode<T: Serialize>(value: &T) -> Result<Vec<u8>> {
    let bytes = serde_ipld_dagcbor::to_vec(value)
        .map_err(|e| PlcError::DagCborError(e.to_string()))?;

    if bytes.len() > MAX_OPERATION_SIZE {
        return Err(PlcError::OperationTooLarge(bytes.len()));
    }

    Ok(bytes)
}

/// Decode a value from DAG-CBOR format
///
/// # Errors
///
/// Returns `PlcError::DagCborDecodeError` if deserialization fails
pub fn dag_cbor_decode<T: for<'de> Deserialize<'de>>(data: &[u8]) -> Result<T> {
    serde_ipld_dagcbor::from_slice(data)
        .map_err(|e| PlcError::DagCborDecodeError(e.to_string()))
}

/// Compute the CID (Content Identifier) of data using SHA-256 and dag-cbor codec
///
/// The CID is computed as:
/// 1. Hash the data with SHA-256
/// 2. Create a multihash with the hash
/// 3. Create a CIDv1 with dag-cbor codec
/// 4. Encode as base32
///
/// # Examples
///
/// ```
/// use atproto_plc::encoding::compute_cid;
///
/// let data = b"hello world";
/// let cid = compute_cid(data).unwrap();
/// assert!(cid.starts_with("bafy"));
/// ```
pub fn compute_cid(data: &[u8]) -> Result<String> {
    // Hash the data with SHA-256
    let hash_bytes = sha256(data);

    // Create multihash (0x12 = SHA-256, followed by length and hash)
    let mut multihash_bytes = Vec::with_capacity(34); // 2 bytes header + 32 bytes hash
    multihash_bytes.push(0x12); // SHA-256 code
    multihash_bytes.push(32); // Hash length
    multihash_bytes.extend_from_slice(&hash_bytes);

    // Create multihash
    let multihash = Multihash::from_bytes(&multihash_bytes)
        .map_err(|e| PlcError::InvalidCid(format!("Failed to create multihash: {:?}", e)))?;

    // Create CIDv1 with dag-cbor codec (0x71)
    let cid = Cid::new_v1(0x71, multihash);

    Ok(cid.to_string())
}

/// Hash data with SHA-256 and return the digest
pub fn sha256(data: &[u8]) -> [u8; 32] {
    let mut hasher = Sha256::new();
    hasher.update(data);
    hasher.finalize().into()
}

/// Validate that a string is a valid base32 encoding
///
/// Returns `true` if all characters are in the allowed alphabet
pub fn is_valid_base32(s: &str) -> bool {
    !s.is_empty() && s.chars().all(|c| BASE32_ALPHABET.contains(c))
}

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

    #[test]
    fn test_base32_roundtrip() {
        let data = b"hello world";
        let encoded = base32_encode(data);
        let decoded = base32_decode(&encoded).unwrap();
        assert_eq!(data, decoded.as_slice());
    }

    #[test]
    fn test_base32_invalid_chars() {
        assert!(base32_decode("0189").is_err()); // Invalid chars: 0, 1, 8, 9
        assert!(base32_decode("ABCD").is_err()); // Uppercase not allowed
    }

    #[test]
    fn test_base64url_roundtrip() {
        let data = b"hello world";
        let encoded = base64url_encode(data);
        let decoded = base64url_decode(&encoded).unwrap();
        assert_eq!(data, decoded.as_slice());
        assert!(!encoded.contains('='));
    }

    #[test]
    fn test_is_valid_base32() {
        assert!(is_valid_base32("abcdefghijklmnopqrstuvwxyz234567"));
        assert!(!is_valid_base32("0189"));
        assert!(!is_valid_base32("ABCD"));
        assert!(!is_valid_base32(""));
    }

    #[test]
    fn test_sha256() {
        let data = b"hello world";
        let hash = sha256(data);
        assert_eq!(hash.len(), 32);
    }

    #[test]
    fn test_compute_cid() {
        let data = b"hello world";
        let cid = compute_cid(data).unwrap();
        assert!(cid.starts_with("b")); // CIDv1 starts with 'b' in base32
    }
}