Crate re_hdf5 

Core HDF5-to-chunk loading logic for Rerun.

Reads an HDF5 file into a lazy stream of Rerun chunks: each HDF5 group maps to an entity, each leaf dataset to a component, with a single file-wide timeline (a designated 1-D index dataset, or a synthesized row_index sequence). HDF5 attributes are emitted as static components under a dedicated __hdf5_properties entity mirroring the source layout.

Scope is raw HDF5 → typed Arrow columns only — no semantic interpretation into archetypes (that is a downstream lens concern).

Structs

DatasetInfo

Structural metadata for a single HDF5 dataset.

Hdf5Config

Configuration for HDF5 loading.

IndexColumn

Specifies which HDF5 dataset provides the file-wide timeline.

Enums

AttrValue

Convenient attribute values for the write API.

DatasetDtype

Element type of an HDF5 dataset, as exposed by crate::DatasetInfo::dtype.

Hdf5Error

Errors that can occur during HDF5 loading.

IndexType

The type and scale of an index column.

TimeUnit

Scale of raw time values. Determines the multiplier to convert to nanoseconds.

Functions

list_datasets

Recursively list the datasets under group_path with their shape and dtype.

list_groups

Recursively list the full group paths under group_path ("/" = whole file).

load_hdf5

Load an HDF5 file and return a lazy iterator of chunks.

load_hdf5_from_bytes

Load HDF5 from in-memory bytes and return a lazy iterator of chunks.

read_attributes

Read the attributes of the object (group or dataset) at object_path ("/" = root), sorted by name.

validate_layout

Metadata-only structural validation of config against the file at path.