Class VoyageAiScoringModel.Builder

java.lang.Object
dev.langchain4j.model.voyageai.VoyageAiScoringModel.Builder
Enclosing class:
VoyageAiScoringModel
public static class VoyageAiScoringModel.Builder extends Object

  • Constructor Details

    • Builder

      public Builder()

  • Method Details

    • httpClientBuilder

      public VoyageAiScoringModel.Builder httpClientBuilder(HttpClientBuilder httpClientBuilder)
      Sets a custom HTTP client builder, allowing fine-grained control over the HTTP client configuration such as timeouts and proxy settings.
      Parameters:
      httpClientBuilder - the HTTP client builder
      Returns:
      this

    • customHeaders

      public VoyageAiScoringModel.Builder customHeaders(Map<String,String> customHeaders)
      Sets custom HTTP headers.

    • customHeaders

      public VoyageAiScoringModel.Builder customHeaders(Supplier<Map<String,String>> customHeadersSupplier)
      Sets a supplier for custom HTTP headers. The supplier is called before each request, allowing dynamic header values. For example, this is useful for OAuth2 tokens that expire and need refreshing.

    • baseUrl

      public VoyageAiScoringModel.Builder baseUrl(String baseUrl)
      Sets the base URL of the Voyage AI API. Defaults to "https://api.voyageai.com/v1/".
      Parameters:
      baseUrl - the base URL
      Returns:
      this

    • timeout

      public VoyageAiScoringModel.Builder timeout(Duration timeout)
      Sets the HTTP request timeout. Defaults to 60 seconds.
      Parameters:
      timeout - the request timeout
      Returns:
      this

    • maxRetries

      public VoyageAiScoringModel.Builder maxRetries(Integer maxRetries)
      Sets the maximum number of retries on transient errors. Defaults to 3.
      Parameters:
      maxRetries - the maximum number of retries
      Returns:
      this

    • apiKey

      public VoyageAiScoringModel.Builder apiKey(String apiKey)
      Sets the Voyage AI API key used to authenticate requests. See Voyage AI dashboard to obtain a key.
      Parameters:
      apiKey - the Voyage AI API key
      Returns:
      this

    • modelName

      public VoyageAiScoringModel.Builder modelName(VoyageAiScoringModelName modelName)
      Name of the model.
      Parameters:
      modelName - Name of the model.
      See Also:

    • modelName

      public VoyageAiScoringModel.Builder modelName(String modelName)
      Name of the model.
      Parameters:
      modelName - Name of the model.
      See Also:

    • topK

      @Deprecated(forRemoval=true, since="1.20.0") public VoyageAiScoringModel.Builder topK(Integer topK)
      Deprecated, for removal: This API element is subject to removal in a future version.
      topK cannot be used together with ScoringModel, which requires exactly one score per input segment (see VoyageAiScoringModel.scoreAll(List, String)). Asking Voyage AI for fewer documents than were sent makes it impossible to tell which segment each returned score belongs to, so scoreAll now throws an IllegalArgumentException when topK is smaller than the number of segments. Any other value has no effect. To limit how many results a RAG pipeline keeps after re-ranking, use ReRankingContentAggregator.ReRankingContentAggregatorBuilder.maxResults(Integer) instead: 
      ReRankingContentAggregator.builder()
        .scoringModel(scoringModel)
        .maxResults(3)
        .build();
      The number of most relevant documents Voyage AI should return.
      Parameters:
      topK - the number of most relevant documents to return.

    • truncation

      public VoyageAiScoringModel.Builder truncation(Boolean truncation)
      Whether to truncate the input to satisfy the "context length limit" on the query and the documents. Defaults to true.
      • If true, the query and documents will be truncated to fit within the context length limit, before processed by the reranker model.
      • If false, an error will be raised when the query exceeds 1000 tokens for rerank-lite-1 and 2000 tokens for rerank-1, or the sum of the number of tokens in the query and the number of tokens in any single document exceeds 4000 for rerank-lite-1 and 8000 for rerank-1.
      Parameters:
      truncation - Whether to truncate the input to satisfy the "context length limit" on the query and the documents.

    • logRequests

      public VoyageAiScoringModel.Builder logRequests(Boolean logRequests)
      Enables debug logging of request bodies sent to the Voyage AI API.
      Parameters:
      logRequests - true to enable request logging
      Returns:
      this

    • logResponses

      public VoyageAiScoringModel.Builder logResponses(Boolean logResponses)
      Enables debug logging of response bodies received from the Voyage AI API.
      Parameters:
      logResponses - true to enable response logging
      Returns:
      this

    • logger

      public VoyageAiScoringModel.Builder logger(org.slf4j.Logger logger)
      Parameters:
      logger - an alternate Logger to be used instead of the default one provided by Langchain4J for logging requests and responses.
      Returns:
      this.

    • build

      public VoyageAiScoringModel build()