meerkat.block package

Submodules

meerkat.block.abstract module

class AbstractBlock(*args, **kwargs)

Bases: object

classmethod consolidate(block_refs: Sequence[BlockRef]) → Tuple[AbstractBlock, Mapping[str, BlockIndex]]

classmethod from_block_data(data: object) → Tuple[AbstractBlock, BlockView]

classmethod from_column_data(data: object) → Tuple[AbstractBlock, BlockView]

classmethod read(path: str, *args, **kwargs)

write(path: str, *args, **kwargs)

property is_mmap

property signature: Hashable

class BlockView(block_index: 'BlockIndex', block: 'AbstractBlock')

Bases: object

block: AbstractBlock

block_index: int | slice | str

property data

meerkat.block.arrow_block module

class ArrowBlock(data: Table, *args, **kwargs)

Bases: AbstractBlock

class Signature(nrows: 'int', klass: 'type')

Bases: object

klass: type

nrows: int

classmethod from_block_data(data: Table) → List[BlockView]

classmethod from_column_data(data: Array) → BlockView

property signature: Hashable

meerkat.block.manager module

class BlockManager

Bases: MutableMapping

Manages all blocks in a DataPanel.

add_column(col: AbstractColumn, name: str)

Convert data to a meerkat column using the appropriate Column type.

apply(method_name: str = '_get', *args, **kwargs) → BlockManager

[summary]

Parameters:

fn (str) – a function that is applied to a block and column_spec and returns a new block and column_spec.

Returns:

[description]

Return type:

[type]

consolidate()

copy()

classmethod from_dict(data: Mapping[str, object])

get_block_ref(name: str)

classmethod read(path: str, columns: Sequence[str] | None = None, *args, **kwargs) → BlockManager

Load a DataPanel stored on disk.

remove(name)

reorder(order: Sequence[str])

update(block_ref: BlockRef)

data (): a single blockable object, potentially contains multiple columns.

view()

write(path: str)

property ncols

property nrows

meerkat.block.numpy_block module

class NumpyBlock(data, *args, **kwargs)

Bases: AbstractBlock

class Signature(dtype: 'np.dtype', nrows: 'int', shape: 'Tuple[int]', klass: 'type', mmap: 'Union[bool, int]')

Bases: object

dtype: dtype

klass: type

mmap: bool | int

nrows: int

shape: Tuple[int]

classmethod from_column_data(data: ndarray) → Tuple[NumpyBlock, BlockView]

[summary]

Parameters:

• data (np.ndarray) – [description]

• names (Sequence[str]) – [description]

Raises:

ValueError – [description]

Returns:

[description]

Return type:

Tuple[NumpyBlock, Mapping[str, BlockIndex]]

property is_mmap

property signature: Hashable

meerkat.block.pandas_block module

class PandasBlock(data: DataFrame, *args, **kwargs)

Bases: AbstractBlock

class Signature(nrows: 'int', klass: 'type')

Bases: object

klass: type

nrows: int

classmethod from_column_data(data: Series) → Tuple[PandasBlock, BlockView]

[summary]

Parameters:

• data (np.ndarray) – [description]

• names (Sequence[str]) – [description]

Raises:

ValueError – [description]

Returns:

[description]

Return type:

Tuple[PandasBlock, Mapping[str, BlockIndex]]

property signature: Hashable

meerkat.block.ref module

class BlockRef(columns: Mapping[str, AbstractColumn], block: AbstractBlock)

Bases: Mapping

apply(method_name: str = '_get', *args, **kwargs)

update(block_ref: BlockRef)

property block_indices

meerkat.block.tensor_block module

class TensorBlock(data, *args, **kwargs)

Bases: AbstractBlock

class Signature(device: 'torch.device', dtype: 'torch.dtype', nrows: 'int', shape: 'Tuple[int]', klass: 'type')

Bases: object

device: device

dtype: dtype

klass: type

nrows: int

shape: Tuple[int]

classmethod from_column_data(data: Tensor) → Tuple[TensorBlock, BlockView]

[summary]

Parameters:

• data (np.ndarray) – [description]

• names (Sequence[str]) – [description]

Raises:

ValueError – [description]

Returns:

[description]

Return type:

Tuple[NumpyBlock, Mapping[str, BlockIndex]]

property signature: Hashable