Honzo applies compression at the chunk level. Each TOC entry selects its own compression algorithm independently of every other chunk.

Compression Codes

LZ4

LZ4 is a fast compression algorithm that prioritizes decode speed over compression ratio. Honzo uses it for text heavy chunks where decode speed matters more than squeezing every byte.

Typical compression ratios for chapter text:

Per Chunk Tradeoff

Each TOC entry selects its own compression. This lets you optimize per chunk.

Leave images uncompressed. JPEG and WebP are already compressed. Running LZ4 on them wastes CPU and may increase size slightly.

Compress chapters. Text benefits from LZ4. The decode speed is fast enough that page turns feel instant.

Do not compress fonts. WOFF2 is already compressed. TTF and OTF may see modest gains but the overhead rarely matters.

Size Tracking in the TOC

Each TOC entry tracks two sizes:

When compression equals 1, size is the compressed length. orig_size is the decompressed length and is non zero.

When compression equals 0, size equals orig_size. Alternatively orig_size is 0.

Implementation

use honzo_io::{HonzoBuilder, Compression, MarkupType};

HonzoBuilder::new()
    .add_chapter("Chapter 1", Compression::Lz4, MarkupType::Markdown)
    .add_chapter("Chapter 2", Compression::None, MarkupType::HTML)
    .finalize()

use honzo_core::HonzoParser;

let p = HonzoParser::new(&data, 1).unwrap();
for entry in p.toc_entries() {
    let compressed = entry.compression == 1;
    let chunk_data = p.read_chunk(&entry).unwrap();
}

Next Steps