Crate rustyhdf5_format 

Pure-Rust HDF5 binary format parsing and writing.

rustyhdf5-format is a zero-dependency (no C libhdf5) crate for reading and writing HDF5 files. It supports no_std environments with the alloc crate.

Writing files

Use file_writer::FileWriter to create HDF5 files:

use rustyhdf5_format::file_writer::{FileWriter, AttrValue};

let mut fw = FileWriter::new();
fw.create_dataset("data")
    .with_f64_data(&[1.0, 2.0, 3.0])
    .with_shape(&[3])
    .set_attr("unit", AttrValue::String("m/s".into()));
let bytes = fw.finish().unwrap();

Reading files

Parsing follows the HDF5 object model: superblock → object header → messages.

use rustyhdf5_format::{signature, superblock, object_header, group_v2,
    datatype, dataspace, data_layout, data_read, message_type::MessageType};

let file_data = std::fs::read("output.h5").unwrap();
let sig = signature::find_signature(&file_data).unwrap();
let sb  = superblock::Superblock::parse(&file_data, sig).unwrap();
let addr = group_v2::resolve_path_any(&file_data, &sb, "data").unwrap();
let hdr = object_header::ObjectHeader::parse(
    &file_data, addr as usize, sb.offset_size, sb.length_size).unwrap();

Features

Feature	Default	Description
std	yes	Standard library support
checksum	yes	Jenkins lookup3 checksum validation
deflate	yes	Deflate (gzip) compression via flate2
provenance	yes	SHINES provenance — SHA-256 hashing & verification

Modules

attribute

HDF5 Attribute message parsing (message type 0x000C).

attribute_info

HDF5 Attribute Info message parsing (message type 0x0015).

btree_v1

HDF5 B-tree v1 parsing (type 0 for groups).

btree_v2

HDF5 B-tree v2 parsing.

checksum

HDF5 metadata checksum: Jenkins lookup3 hashlittle and CRC32.

chunk_cache

Chunk cache with hash-based index and LRU eviction.

chunked_read

Chunked dataset reading: B-tree v1 type 1 traversal and chunk assembly.

chunked_write

Chunked dataset writing: chunk splitting, compression, index building.

data_layout

HDF5 Data Layout message parsing (message type 0x0008).

data_read

Raw data reading and typed conversion for HDF5 datasets.

dataspace

HDF5 Dataspace message parsing (message type 0x0001).

datatype

HDF5 Datatype message parsing (message type 0x0003).

error

Error types for HDF5 format parsing.

extensible_array

HDF5 Extensible Array index parsing for chunked datasets (v4 index type 4).

file_writer

HDF5 file creation (write pipeline).

filter_pipeline

HDF5 Filter Pipeline message parsing (message type 0x000B).

filters

HDF5 filter implementations: deflate, shuffle, fletcher32.

fixed_array

HDF5 Fixed Array index parsing for chunked datasets (v4 index type 3).

fractal_heap

HDF5 Fractal Heap parsing for v2 group link storage.

global_heap

HDF5 Global Heap collection parsing.

group_info

HDF5 Group Info message parsing (message type 0x000A).

group_v1

V1 group traversal: resolve group children and navigate paths.

group_v2

V2 group traversal: resolve group children and navigate paths.

link_info

HDF5 Link Info message parsing (message type 0x0002).

link_message

HDF5 Link message parsing (message type 0x0006).

local_heap

HDF5 Local Heap parsing.

message_type

HDF5 object header message type identifiers.

metadata_index

Index-table + metadata-block architecture for independent parallel dataset creation.

object_header

HDF5 Object Header parsing (v1 and v2).

object_header_writer

Object header writer for v2 format.

provenance

SHINES provenance: SHA-256 content hashing, provenance attributes, and data-integrity verification.

shared_message

HDF5 Shared Object Header Message resolution.

signature

HDF5 file signature (magic bytes) detection.

superblock

HDF5 Superblock parsing for versions 0, 1, 2, and 3.

symbol_table

HDF5 Symbol Table Message and Symbol Table Node (SNOD) parsing.

type_builders

Builder types for HDF5 datatypes, attributes, datasets, and groups.

vl_data

Variable-length data reading (VL strings & VL sequences).