ontolearn.scripts.litserve_neural_reasoner

Attributes

parser

Classes

NeuralReasonerAPI

NeuralReasonerAPI is a LitAPI implementation that handles requests to a neural reasoner

Module Contents

class ontolearn.scripts.litserve_neural_reasoner.NeuralReasonerAPI(path_neural_embedding, gamma=0.9)

Bases: litserve.LitAPI

NeuralReasonerAPI is a LitAPI implementation that handles requests to a neural reasoner using OWL expressions. It utilizes a neural embedding model for reasoning over ontology data.

path_neural_embedding

Path to the neural embedding.

Type:: str

gamma

Minimum confidence threshold for the reasoning model, defaults to 0.9.

Type:: float

path_neural_embedding

gamma

setup(device): Sets up the neural reasoner instance.

decode_request(request)

Decodes an incoming request to extract the DL expression and namespace.

Parameters:: request (dict) – A dictionary containing the request data, with ‘expression’ and ‘namespace’ keys.
Returns:: A tuple with the DL expression (str) and namespace (str).
Return type:: tuple

predict(data)

Predicts individuals of the given OWL expression using the neural reasoner.

Parameters:: data (tuple) – A tuple containing the DL expression (str) and namespace (str).
Returns:: A set of individuals satisfying the given OWL expression.
Return type:: set

encode_response(output)

Encodes the output from the reasoner back into a DL expression format for response.

Parameters:: output (set) – A set of OWL expressions representing the individuals.
Returns:: A dictionary with ‘retrieval_result’ key containing a list of DL expressions as strings.
Return type:: dict

ontolearn.scripts.litserve_neural_reasoner.parser