ch5mpy.H5Array

class ch5mpy.H5Array(dset)[source]

Wrapper around Dataset objects to interface with numpy’s API.

Attributes

`H5Array.MAX_MEM_USAGE`
`H5Array.chunk_size`	Get the size of a chunk (i.e.
`H5Array.dset`
`H5Array.dtype`
`H5Array.is_chunked`
`H5Array.ndim`
`H5Array.shape`
`H5Array.size`

Methods

`H5Array.__init__`
`H5Array.astype`	Cast an H5Array to a specified dtype.
`H5Array.contract`	Resize an H5Array by removing amount elements along the selected axis.
`H5Array.copy`	rtype: `ndarray`[`Any`, `dtype`[`TypeVar`(`_T`, bound= `generic`, covariant=True)]]
`H5Array.expand`	Resize an H5Array by adding amount elements along the selected axis.
`H5Array.iter_chunks`	rtype: `ChunkIterator`
`H5Array.iter_chunks_with`	rtype: `PairedChunkIterator`
`H5Array.maptype`	Cast an H5Array to any object type.
`H5Array.max`	rtype: `Union`[`TypeVar`(`_T`, bound= `generic`, covariant=True), `ndarray`[`Any`, `dtype`[`TypeVar`(`_T`, bound= `generic`, covariant=True)]]]
`H5Array.mean`	rtype: `Union`[`Any`, `ndarray`[`Any`, `dtype`[`Any`]]]
`H5Array.min`	rtype: `Union`[`TypeVar`(`_T`, bound= `generic`, covariant=True), `ndarray`[`Any`, `dtype`[`TypeVar`(`_T`, bound= `generic`, covariant=True)]]]
`H5Array.read_direct`	rtype: `None`

Parameters:: dset (Dataset[_T] | DatasetWrapper[_T]) –

Attributes

H5Array.MAX_MEM_USAGE: int | str = '250M'

H5Array.chunk_size

Get the size of a chunk (i.e. the nb of elements that can be read/written at a time).

H5Array.dset

H5Array.dtype

H5Array.is_chunked

H5Array.ndim

H5Array.shape

H5Array.size

Methods

H5Array.__init__(dset)[source]

Parameters:

dset (Dataset[_T] | DatasetWrapper[_T]) –

H5Array.astype(dtype, inplace=False)[source]

Cast an H5Array to a specified dtype. This does not perform a copy, it returns a wrapper around the underlying H5 dataset.

Return type:

H5Array[Any]

Parameters:

dtype (dtype[Any] | None | type[Any] | _SupportsDType[dtype[Any]] | str | tuple[Any, int] | tuple[Any, Union[SupportsIndex, collections.abc.Sequence[SupportsIndex]]] | list[Any] | _DTypeDict | tuple[Any, Any]) –

inplace (bool) –

H5Array.contract(amount, axis=None)[source]

Resize an H5Array by removing amount elements along the selected axis.

Raises:

TypeError – if the H5Array does not wrap a chunked Dataset.

Return type:

None

Parameters:

amount (int) –

axis (int | tuple[int, ...] | None) –

H5Array.copy()[source]

Return type:

ndarray[Any, dtype[TypeVar(_T, bound= generic, covariant=True)]]

H5Array.expand(amount, axis=None)[source]

Resize an H5Array by adding amount elements along the selected axis.

Raises:

TypeError – if the H5Array does not wrap a chunked Dataset.

Return type:

None

Parameters:

amount (int) –

axis (int | tuple[int, ...] | None) –

H5Array.iter_chunks(keepdims=False)[source]

Return type:

ChunkIterator

Parameters:

keepdims (bool) –

H5Array.iter_chunks_with(other, keepdims=False)[source]

Return type:

PairedChunkIterator

Parameters:

other (ndarray[Any, dtype[Any]] | H5Array[Any]) –

keepdims (bool) –

H5Array.maptype(otype)[source]

Cast an H5Array to any object type. This extends H5Array.astype() to any type <T>, where it is required that an object <T> can be constructed as T(v) for any value <v> in the dataset.

Return type:

H5Array[Any]

Parameters:

otype (type[Any]) –

H5Array.max(axis=None)[source]

Return type:

Union[TypeVar(_T, bound= generic, covariant=True), ndarray[Any, dtype[TypeVar(_T, bound= generic, covariant=True)]]]

Parameters:

axis (int | tuple[int, ...] | None) –

H5Array.mean(axis=None)[source]

Return type:

Union[Any, ndarray[Any, dtype[Any]]]

Parameters:

axis (int | tuple[int, ...] | None) –

H5Array.min(axis=None)[source]

Return type:

Union[TypeVar(_T, bound= generic, covariant=True), ndarray[Any, dtype[TypeVar(_T, bound= generic, covariant=True)]]]

Parameters:

axis (int | tuple[int, ...] | None) –

H5Array.read_direct(dest, source_sel, dest_sel)[source]

Return type:

None

Parameters:

dest (ndarray[Any, dtype[_T]]) –

source_sel (tuple[slice, ...]) –

dest_sel (tuple[slice, ...]) –