Crate rust_hdf5 

Pure Rust HDF5 library with full read/write support.

This crate provides a high-level API for creating and reading HDF5 files without any C dependencies. It supports contiguous and chunked datasets, deflate compression (with optional shuffle), SWMR streaming, and hyperslab (slice) I/O.

Quick start

Writing a dataset

use rust_hdf5::H5File;

let file = H5File::create("output.h5").unwrap();
let ds = file.new_dataset::<f64>()
    .shape(&[100, 200])
    .create("matrix")
    .unwrap();
let data: Vec<f64> = (0..20_000).map(|i| i as f64).collect();
ds.write_raw(&data).unwrap();
file.close().unwrap();

Reading a dataset

use rust_hdf5::H5File;

let file = H5File::open("output.h5").unwrap();
let ds = file.dataset("matrix").unwrap();
let data = ds.read_raw::<f64>().unwrap();
assert_eq!(data.len(), 20_000);

Chunked + compressed streaming

use rust_hdf5::H5File;

let file = H5File::create("stream.h5").unwrap();
let ds = file.new_dataset::<f32>()
    .shape(&[0usize, 64])
    .chunk(&[1, 64])
    .max_shape(&[None, Some(64)])
    .deflate(6)
    .create("sensor")
    .unwrap();

for frame in 0..1000u64 {
    let vals: Vec<f32> = (0..64).map(|i| (frame * 64 + i) as f32).collect();
    let raw: Vec<u8> = vals.iter().flat_map(|v| v.to_le_bytes()).collect();
    ds.write_chunk(frame as usize, &raw).unwrap();
}
ds.extend(&[1000, 64]).unwrap();
file.close().unwrap();

Hyperslab (slice) I/O

use rust_hdf5::H5File;

// Write a sub-region
let file = H5File::create("slice.h5").unwrap();
let ds = file.new_dataset::<i32>()
    .shape(&[10usize, 10])
    .create("grid")
    .unwrap();
ds.write_raw(&[0i32; 100]).unwrap();
ds.write_slice(&[2, 3], &[3, 4], &[1i32, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]).unwrap();
file.close().unwrap();

// Read a sub-region
let file = H5File::open("slice.h5").unwrap();
let ds = file.dataset("grid").unwrap();
let region = ds.read_slice::<i32>(&[2, 3], &[3, 4]).unwrap();
assert_eq!(region.len(), 12);

Re-exports

pub use attribute::H5Attribute;

pub use dataset::H5Dataset;

pub use error::Hdf5Error;

pub use error::Result;

pub use file::H5File;

pub use group::H5Group;

pub use types::Complex32;

pub use types::Complex64;

pub use types::CompoundType;

pub use types::H5Type;

pub use types::HBool;

pub use types::VarLenUnicode;

Modules

attribute

Attribute support.

dataset

Dataset creation and I/O.

error

Error types for the hdf5 public API crate.

file

HDF5 file handle — the main entry point for the public API.

format

Pure Rust HDF5 on-disk format codec.

group

Group support.

swmr

Single Writer / Multiple Reader (SWMR) API.

types

Mapping from Rust types to HDF5 datatype messages.