Local - Access your data locally¶

class manta_light.local.Local(host: str | None = None, port: int | None = None, swarm_id: PairId | None = None, task_id: PairId | None = None)¶

Local service for accessing local data securely

The Local service allows the user to get local data.

Parameters should not be fulfill manually. Instead, it is managed during the deployment of the task by the Node.

Parameters:

host (Optional[str]) – Manager host
port (Optional[int]) – Manager port
swarm_id (Optional[PairId]) – Swarm ID
task_id (Optional[PairId]) – Task ID

get_numpy_data(dataset_name: str) → bytes¶

Get the numpy data

Parameters:: dataset_name (str) – The name of the dataset to get
Returns:: The numpy data
Return type:: bytes

Examples

Inside a Task class, you can access to local data by using the attribute self.local automatically created by Task <manta_light.task.Task:

>>> mnist_data = self.local.get_numpy_data("mnist.npz")
>>> data = np.load(mnist_data)

get_binary_data(dataset_name: str) → BytesIO¶

Get the binary data

Parameters:: dataset_name (str) – The name of the dataset to get
Returns:: The binary data
Return type:: io.BytesIO

Examples

Inside a Task class, you can access to local data by using the attribute self.local automatically created by Task <manta_light.task.Task:

>>> mnist_data = self.local.get_binary_data("mnist.npz")
>>> data = np.load(mnist_data)

list_directories(directory: str) → List[str]¶

List directories

Parameters:: directory (str) – Directory name
Returns:: List of directories
Return type:: List[str]

read_file_lines(file_name: str) → List[str]¶

Read file lines

Parameters:: file_name (str) – File name
Returns:: List of lines
Return type:: List[str]

async async_get_numpy_data(dataset_name: str) → BytesIO¶

Get Numpy Data

Parameters:: dataset_name (str) – Dataset name
Returns:: Binary data
Return type:: io.BytesIO

Examples

Same as select but asynchronous:

>>> mnist_data = await self.local.async_get_numpy_data("mnist.npz")
>>> data = np.load(mnist_data)

async async_get_binary_data(dataset_name: str) → BytesIO¶

Asynchronously get binary data

Parameters:: dataset_name (str) – Dataset name
Returns:: Binary data
Return type:: io.BytesIO

Examples

Same as select but asynchronous:

>>> mnist_data = await self.local.async_get_binary_data("mnist.npz")
>>> data = np.load(mnist_data)

async async_list_directories(directory: str) → List[str]¶

List directories

Parameters:: directory (str) – Directory name
Returns:: List of directories
Return type:: List[str]

async async_read_file_lines(file_name: str) → List[str]¶

Read file lines

Parameters:: file_name (str) – File name
Returns:: List of lines
Return type:: List[str]