`rhino_health.lib.endpoints.code_run.code_run_endpoints`#

Module Contents#

Classes#

CodeRunEndpoints

Endpoints for interacting with CodeRuns

class rhino_health.lib.endpoints.code_run.code_run_endpoints.CodeRunEndpoints(session)#

Bases: LTSCodeRunEndpoints

Endpoints for interacting with CodeRuns

get_code_run(code_run_uid: str)#

Returns a CodeRun dataclass

Parameters:

code_run_uid: str: UID for the CodeRun

Returns:

code_run: CodeRun: CodeRun dataclass

Examples

>>> session.code_object.get_code_run(code_run_uid)
CodeRun()

get_model_params(code_run_uid: str, model_weights_files: List[str] | None = None) → io.BytesIO#

Returns the contents of one or more model params file(s) associated with a model result.

The return value is an open binary file-like object, which can be read or written to a file.

The contents are for a single file. This is either the model params file if there was only one available or selected, or a Zip file containing multiple model params files.

Parameters:

code_run_uid: str: UID for the CodeRun
model_weights_files: List(str): List of paths within S3 of model weight files to download. If multiple files are supplied, download as zip. If the argument is not specified, download all model weight files found for the given CodeRun.

Returns:

model_params: BytesIO: A Python BytesIO Buffer

Examples

>>> with open("my_output_file.out", "wb") as output_file:
>>>     model_params_buffer = session.code_run.get_model_params(code_run_uid, model_weights_files)
>>>     output_file.write(model_params_buffer.getbuffer())

halt_code_run(code_run_uid: str)#

Send a halting request to a Code Run.

This triggers the halting process but does not wait for halting to complete.

If triggering the halting process fails, a message specifying the error is returned.

Warning

This feature is under development and the interface may change

Parameters:

code_run_uid: str: UID of the CodeRun to halt.

Returns:

json response in the format of:

“status”: the request’s status code -

200: valid request, halting innitiated.
400: invalid request, the model can not be halted, or does not exist.
500: error while initiating halting.

“data”: message specifying if the halting was initiated or failed. In case the request failed,

the error message is also displayed.

Examples

>>> session.code_run.halt_code_run(code_run_uid)

run_inference(code_run_uid: str, validation_dataset_uids: List[str], validation_datasets_suffix: str, timeout_seconds: int)#

Start running inference on one or more datasets using a previously trained NVFlare Model.

Parameters:

code_run_uid: str: UID for the code run
validation_dataset_uids: List[str]: List of dataset UIDs to run inference on
validation_datasets_suffix: str: Suffix for the validation datasets
timeout_seconds: int: Timeout in seconds

Returns:

ModelInferenceAsyncResponse()

Examples

>>> s = session.code_run.run_inference(code_run_uid, validation_dataset_uids, validation_datasets_suffix, timeout_seconds)
>>> s.code_run() # Get the asynchronous result
>>> s.code_run_uid # Get the code run UID
See Also
--------
rhino_health.lib.endpoints.code_object.code_object_dataclass.ModelInferenceAsyncResponse : ModelInferenceAsyncResponse Dataclass

rhino_health.lib.endpoints.code_run.code_run_endpoints#

Module Contents#

Classes#

`rhino_health.lib.endpoints.code_run.code_run_endpoints`#