orkgnlp.common.service.base.PipelineExecutor

class PipelineExecutor(encoder, runner, decoder)[source]

Bases: object

The PipelineExecutor executes a full service workflow given its encoder, runner and decoder. See the run function description for further information.

Parameters:

encoder (ORKGNLPBaseEncoder) – Service’s encoder.
runner (ORKGNLPBaseRunner) – Service’s runner.
decoder (ORKGNLPBaseDecoder) – Service’s decoder.

Methods

`release_memory`	Releases the memory of all available pipeline components.
`run`	Executes a full pipline of the common service workflow:

release_memory()[source]: Releases the memory of all available pipeline components.

run(raw_input, **kwargs)[source]

Executes a full pipline of the common service workflow:

Runs the service encoder with the user’s input.
The encoded input is passed to the model runner, which in turn is executed.
The model’s output is decoded to a user-friendly format using the service’s decoder.

Note that the kwargs can be updated and passed through the pipeline components.

Parameters:

raw_input (Any) – User’s input to be encoded.
kwargs (Any) – Named parameters for further processing config. Please check your used component documentation for specific parameter description.

Return type:

Any

Returns:

The decoded user-friendly output.

Raises:

orkgnlp.common.util.exceptions.ORKGNLPIllegalStateException – If either [Encoder, Runner, Decoder] is not initialized.