io.pinecone.clients

Class Inference

java.lang.Object

io.pinecone.clients.Inference

public class Inference
extends java.lang.Object

The Inference class provides methods to interact with Pinecone's inference API through the Java SDK. It allows users to send input data to generate embeddings or rank documents using a specified model.

This class utilizes the InferenceApi to make API calls to the Pinecone inference service.

Constructor Summary

Constructors

Constructor and Description
Inference(PineconeConfig config) Constructs an instance of Inference class using PineconeConfig object.

Method Summary

Modifier and Type	Method and Description
EmbeddingsList	embed(java.lang.String model, java.util.Map<java.lang.String,java.lang.Object> parameters, java.util.List<java.lang.String> inputs) Sends input data and parameters to the embedding model and returns a list of embeddings.
RerankResult	rerank(java.lang.String model, java.lang.String query, java.util.List<java.util.Map<java.lang.String,java.lang.String>> documents) Reranks a list of documents based on the relevance to a query using the specified model.
RerankResult	rerank(java.lang.String model, java.lang.String query, java.util.List<java.util.Map<java.lang.String,java.lang.String>> documents, java.util.List<java.lang.String> rankFields, int topN, boolean returnDocuments, java.util.Map<java.lang.String,java.lang.String> parameters) Reranks a list of documents based on the relevance to a query using the specified model with additional options.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Inference

public Inference(PineconeConfig config)

Constructs an instance of Inference class using PineconeConfig object.

Parameters:

config - The Pinecone configuration object for interacting with inference api.

Method Detail

embed

public EmbeddingsList embed(java.lang.String model,
                            java.util.Map<java.lang.String,java.lang.Object> parameters,
                            java.util.List<java.lang.String> inputs)
                     throws ApiException

Sends input data and parameters to the embedding model and returns a list of embeddings.

Parameters:

model - The embedding model to use.

parameters - A map containing model-specific parameters.

inputs - A list of input strings to generate embeddings for.

Returns:

EmbeddingsList containing the embeddings for the provided inputs.

Throws:

ApiException - If the API call fails, an ApiException is thrown.

rerank

public RerankResult rerank(java.lang.String model,
                           java.lang.String query,
                           java.util.List<java.util.Map<java.lang.String,java.lang.String>> documents)
                    throws ApiException

Reranks a list of documents based on the relevance to a query using the specified model. Since rest of the parameters are optional, they are set to their default values.

Parameters:

model - The model to be used for reranking the documents.

query - The query string to rank the documents against.

documents - A list of maps representing the documents to be ranked. Each map should contain document attributes, such as "text".

Returns:

RerankResult containing the ranked documents and their scores.

Throws:

ApiException - If the API call fails, an ApiException is thrown.

rerank

public RerankResult rerank(java.lang.String model,
                           java.lang.String query,
                           java.util.List<java.util.Map<java.lang.String,java.lang.String>> documents,
                           java.util.List<java.lang.String> rankFields,
                           int topN,
                           boolean returnDocuments,
                           java.util.Map<java.lang.String,java.lang.String> parameters)
                    throws ApiException

Reranks a list of documents based on the relevance to a query using the specified model with additional options.

Parameters:

model - The model to be used for reranking the documents.

query - The query string to rank the documents against.

documents - A list of maps representing the documents to be ranked. Each map should contain document attributes, such as "text".

rankFields - A list of fields in the documents to be used for ranking, typically "text".

topN - The number of top-ranked documents to return.

returnDocuments - Whether to return the documents along with the ranking scores.

parameters - A map containing additional model-specific parameters for reranking.

Returns:

RerankResult containing the ranked documents and their scores.

Throws:

ApiException - If the API call fails, an ApiException is thrown.