Chunks and Cache — Weave Docs

ChunkCache caches packed chunks by Strand sequence and tracks bytes, entries, hits and misses. BlobConfig.compression is an Option<CompressionType>; availability of compression algorithms follows Cargo features. A deduplication hash and a cache hit do not certify a peer’s identity or remote availability.

The example requires the zstd Cargo feature on strand-blobs. It is checked against source signatures; it was not compiled or run during this documentation pass.

use bytes::Bytes;
use strand_blobs::{BlobConfig, ChunkCache, ChunkIndex, CompressionType};

fn main() {
    // The blob store caches recently read chunks in an LRU keyed by chunk index.
    // Example capacity: 4 MiB of cached chunk bytes.
    let cache = ChunkCache::new(4 * 1024 * 1024);
    cache.insert(0, Bytes::from_static(b"chunk-0"));
    cache.insert(1, Bytes::from_static(b"chunk-1"));

    // Lookups update hit/miss counters.
    let hit = cache.get(&0);
    let miss = cache.get(&99);
    assert!(hit.is_some());
    assert!(miss.is_none());

    // BlobConfig controls the chunk size and compression applied at put time.
    let cfg = BlobConfig {
        chunk_size: 1024 * 1024,
        compression: Some(CompressionType::Zstd),
        ..BlobConfig::default()
    };
    let _ = ChunkIndex::default();

    println!(
        "cache_entries={} cache_hits={} cache_misses={} chunk_size={}",
        cache.len(),
        cache.hits(),
        cache.misses(),
        cfg.chunk_size,
    );
}

Source declarations