Struct ZstdDictCompressor 

pub struct ZstdDictCompressor;

Stateless driver for zstd dictionary operations.

All methods take the dictionary by reference. No internal state is retained between calls; callers supply both the data and the dictionary each time.

The trained dictionary should be stored in crate::infrastructure::repositories::InMemoryDictionaryStore (or a custom crate::domain::ports::dictionary_store::DictionaryStore impl) and shared via Arc<ZstdDictionary>.

Examples

use pjson_rs::compression::zstd::{ZstdDictCompressor, N_TRAIN, MAX_DICT_SIZE};

let samples: Vec<Vec<u8>> = (0..N_TRAIN).map(|i| {
    format!("{{\"id\":{i},\"key\":\"value\",\"score\":{}}}", i * 3).into_bytes()
}).collect();

let dict = ZstdDictCompressor::train(&samples, MAX_DICT_SIZE).unwrap();

let data = b"{\"id\":99,\"key\":\"value\",\"score\":297}";
let compressed = ZstdDictCompressor::compress(data, &dict).unwrap();
let decompressed = ZstdDictCompressor::decompress(&compressed, &dict, data.len() * 2).unwrap();
assert_eq!(decompressed, data);

Implementations

impl ZstdDictCompressor

pub fn train( samples: &[Vec<u8>], max_dict_size: usize, ) -> Result<ZstdDictionary>

Train a zstd dictionary from a corpus of sample byte strings.

max_dict_size is clamped to MAX_DICT_SIZE before being passed to libzstd — even if the caller requests a larger dict, the type invariant of ZstdDictionary is always satisfied.

Libzstd requires at least 8 samples; the PJS convention is to call this after accumulating N_TRAIN (32) samples for better dictionary quality.

Errors

Returns Error::CompressionError if:

• samples.len() < 8 (libzstd hard minimum)
• libzstd training itself fails (e.g., samples too small or too uniform)

Examples

use pjson_rs::compression::zstd::{ZstdDictCompressor, N_TRAIN, MAX_DICT_SIZE};

let samples: Vec<Vec<u8>> = (0..N_TRAIN).map(|i| {
    format!("{{\"seq\":{i},\"payload\":\"aaabbbccc{i}\"}}").into_bytes()
}).collect();

let dict = ZstdDictCompressor::train(&samples, MAX_DICT_SIZE).unwrap();
assert!(dict.len() <= MAX_DICT_SIZE);

// Requesting a larger size is silently clamped.
let dict2 = ZstdDictCompressor::train(&samples, usize::MAX).unwrap();
assert!(dict2.len() <= MAX_DICT_SIZE);

// Insufficient samples are rejected before calling libzstd.
let few: Vec<Vec<u8>> = vec![b"data".to_vec(); 3];
assert!(ZstdDictCompressor::train(&few, MAX_DICT_SIZE).is_err());

pub fn compress(data: &[u8], dict: &ZstdDictionary) -> Result<Vec<u8>>

Compress data using the dictionary at the default level (DEFAULT_LEVEL).

Errors

Returns Error::CompressionError on libzstd failure.

Examples

use pjson_rs::compression::zstd::{ZstdDictCompressor, N_TRAIN, MAX_DICT_SIZE};

let samples: Vec<Vec<u8>> = (0..N_TRAIN)
    .map(|i| format!("{{\"n\":{i}}}").into_bytes())
    .collect();
let dict = ZstdDictCompressor::train(&samples, MAX_DICT_SIZE).unwrap();
let compressed = ZstdDictCompressor::compress(b"{\"n\":99}", &dict).unwrap();
assert!(!compressed.is_empty());

pub fn compress_with_level( data: &[u8], dict: &ZstdDictionary, level: i32, ) -> Result<Vec<u8>>

Compress data using the dictionary at an explicit compression level.

Level must be in [1, 22]; libzstd clamps out-of-range values silently.

Errors

Returns Error::CompressionError on libzstd failure.

Examples

use pjson_rs::compression::zstd::{ZstdDictCompressor, N_TRAIN, MAX_DICT_SIZE};

let samples: Vec<Vec<u8>> = (0..N_TRAIN)
    .map(|i| format!("{{\"n\":{i}}}").into_bytes())
    .collect();
let dict = ZstdDictCompressor::train(&samples, MAX_DICT_SIZE).unwrap();
let compressed = ZstdDictCompressor::compress_with_level(b"{\"n\":99}", &dict, 1).unwrap();
assert!(!compressed.is_empty());

pub fn decompress( data: &[u8], dict: &ZstdDictionary, max_output: usize, ) -> Result<Vec<u8>>

Decompress data using the dictionary, capping output at max_output bytes.

This is the standalone decompression path — for untrusted input routed through crate::compression::secure::SecureCompressor, use crate::compression::secure::ByteCodec::ZstdDict instead, which passes the output through crate::security::CompressionBombDetector.

Errors

Returns Error::CompressionError on libzstd failure.

Examples

use pjson_rs::compression::zstd::{ZstdDictCompressor, N_TRAIN, MAX_DICT_SIZE};

let samples: Vec<Vec<u8>> = (0..N_TRAIN)
    .map(|i| format!("{{\"n\":{i}}}").into_bytes())
    .collect();
let dict = ZstdDictCompressor::train(&samples, MAX_DICT_SIZE).unwrap();
let data = b"{\"n\":99}";
let compressed = ZstdDictCompressor::compress(data, &dict).unwrap();
let decompressed = ZstdDictCompressor::decompress(&compressed, &dict, 1024).unwrap();
assert_eq!(decompressed.as_slice(), data.as_slice());