Variable vectorIndexConst

Type declaration

• dynamic: ((opts?) => ModuleConfig<"dynamic", undefined | VectorIndexConfigDynamicCreate>)

  Create a ModuleConfig<'dynamic', VectorIndexConfigDynamicCreate | undefined> object when defining the configuration of the dynamic vector index.

  Use this method when defining the options.vectorIndexConfig argument of the configure.vectors method.

  Returns

  The configuration object.

  • • (opts?): ModuleConfig<"dynamic", undefined | VectorIndexConfigDynamicCreate>

    • Parameters

      • Optional opts: VectorIndexConfigDynamicCreateOptions

        The options available for configuring the dynamic vector index.

      Returns ModuleConfig<"dynamic", undefined | VectorIndexConfigDynamicCreate>

• flat: ((opts?) => ModuleConfig<"flat", undefined | VectorIndexConfigFlatCreate>)

  Create a ModuleConfig<'flat', VectorIndexConfigFlatCreate | undefined> object when defining the configuration of the FLAT vector index.

  Use this method when defining the options.vectorIndexConfig argument of the configure.vectors method.

  Returns

  The configuration object.

  • • (opts?): ModuleConfig<"flat", undefined | VectorIndexConfigFlatCreate>

    • Parameters

      • Optional opts: VectorIndexConfigFlatCreateOptions

        The options available for configuring the flat vector index.

      Returns ModuleConfig<"flat", undefined | VectorIndexConfigFlatCreate>

• hnsw: ((opts?) => ModuleConfig<"hnsw", undefined | VectorIndexConfigHNSWCreate>)

  Create a ModuleConfig<'hnsw', VectorIndexConfigHNSWCreate | undefined> object when defining the configuration of the HNSW vector index.

  Use this method when defining the options.vectorIndexConfig argument of the configure.vectors method.

  Returns

  The configuration object.

  • • (opts?): ModuleConfig<"hnsw", undefined | VectorIndexConfigHNSWCreate>

    • Parameters

      • Optional opts: VectorIndexConfigHNSWCreateOptions

        The options available for configuring the HNSW vector index.

      Returns ModuleConfig<"hnsw", undefined | VectorIndexConfigHNSWCreate>

• multiVector: {
      encoding: {
          muvera: ((options?) => {
              dprojections?: number;
              ksim?: number;
              repetitions?: number;
              type?: "muvera";
          });
      };
      multiVector: ((options?) => MultiVectorConfigCreate);
  }

  Define the configuration for a multi-vector index.

  • encoding: {
        muvera: ((options?) => {
            dprojections?: number;
            ksim?: number;
            repetitions?: number;
            type?: "muvera";
        });
    }

    Specify the encoding configuration for a multi-vector index.

    • muvera: ((options?) => {
          dprojections?: number;
          ksim?: number;
          repetitions?: number;
          type?: "muvera";
      })

      Create an object of type MuveraEncodingConfigCreate to be used when defining the encoding configuration of a multi-vector index using MUVERA.

      Returns

      The object of type MuveraEncodingConfigCreate.

      • • (options?): {
              dprojections?: number;
              ksim?: number;
              repetitions?: number;
              type?: "muvera";
          }

        • Parameters

          • Optional options: {
                dprojections?: number;
                ksim?: number;
                repetitions?: number;
            }

            • Optional dprojections?: number

              The number of projections to use. Default is undefined.

            • Optional ksim?: number

              The number of nearest neighbors to consider for similarity. Default is undefined.

            • Optional repetitions?: number

              The number of repetitions to use. Default is undefined.

          Returns {
              dprojections?: number;
              ksim?: number;
              repetitions?: number;
              type?: "muvera";
          }

          • Optional dprojections?: number

          • Optional ksim?: number

          • Optional repetitions?: number

          • Optional type?: "muvera"

  • multiVector: ((options?) => MultiVectorConfigCreate)

    Create an object of type MultiVectorConfigCreate to be used when defining the configuration of a multi-vector index.

    Returns

    The object of type MultiVectorConfigCreate.

    • • (options?): MultiVectorConfigCreate

      • Parameters

        • Optional options: {
              aggregation?: string;
              encoding?: {
                  dprojections?: number;
                  ksim?: number;
                  repetitions?: number;
                  type?: "muvera";
              };
          }

          • Optional aggregation?: string

            The aggregation method to use. Default is 'maxSim'.

          • Optional encoding?: {
                dprojections?: number;
                ksim?: number;
                repetitions?: number;
                type?: "muvera";
            }

            The encoding configuration for the multi-vector index. Default is undefined.

            • Optional dprojections?: number

            • Optional ksim?: number

            • Optional repetitions?: number

            • Optional type?: "muvera"

        Returns MultiVectorConfigCreate

• quantizer: {
      bq: ((options?) => QuantizerRecursivePartial<BQConfig>);
      none: (() => UncompressedConfig);
      pq: ((options?) => QuantizerRecursivePartial<PQConfig>);
      rq: ((options?) => QuantizerRecursivePartial<RQConfig>);
      sq: ((options?) => QuantizerRecursivePartial<SQConfig>);
  }

  Define the quantizer configuration to use when creating a vector index.

  • bq: ((options?) => QuantizerRecursivePartial<BQConfig>)

    Create an object of type BQConfigCreate to be used when defining the quantizer configuration of a vector index.

    Returns

    The object of type BQConfigCreate.

    • • (options?): QuantizerRecursivePartial<BQConfig>

      • Parameters

        • Optional options: {
              cache?: boolean;
              rescoreLimit?: number;
          }

          • Optional cache?: boolean

            Whether to cache the quantizer. Default is false.

          • Optional rescoreLimit?: number

            The rescore limit. Default is 1000.

        Returns QuantizerRecursivePartial<BQConfig>

  • none: (() => UncompressedConfig)

    Create an object of type UncompressedConfig to be used when defining the quantizer configuration of a vector index.

    This is useful for disabling the default quantization present in Weaviate>=1.33.0.

    Returns

    The object of type UncompressedConfig.

    • • (): UncompressedConfig

      • Returns UncompressedConfig

  • pq: ((options?) => QuantizerRecursivePartial<PQConfig>)

    Create an object of type PQConfigCreate to be used when defining the quantizer configuration of a vector index.

    Returns

    The object of type PQConfigCreate.

    • • (options?): QuantizerRecursivePartial<PQConfig>

      • Parameters

        • Optional options: {
              bitCompression?: boolean;
              centroids?: number;
              encoder?: {
                  distribution?: PQEncoderDistribution;
                  type?: PQEncoderType;
              };
              segments?: number;
              trainingLimit?: number;
          }

          • Optional bitCompression?: boolean

            Whether to use bit compression.

          • Optional centroids?: number

            The number of centroids.

          • Optional encoder?: {
                distribution?: PQEncoderDistribution;
                type?: PQEncoderType;
            }

            • Optional distribution?: PQEncoderDistribution

            • Optional type?: PQEncoderType

          • Optional segments?: number

            The number of segments.

          • Optional trainingLimit?: number

            The training limit.

        Returns QuantizerRecursivePartial<PQConfig>

  • rq: ((options?) => QuantizerRecursivePartial<RQConfig>)

    Create an object of type RQConfigCreate to be used when defining the quantizer configuration of a vector index.

    Returns

    The object of type RQConfigCreate.

    • • (options?): QuantizerRecursivePartial<RQConfig>

      • Parameters

        • Optional options: {
              bits?: number;
              rescoreLimit?: number;
          }

          • Optional bits?: number

            Number of bits to user per vector element.

          • Optional rescoreLimit?: number

            The rescore limit. Default is 1000.

        Returns QuantizerRecursivePartial<RQConfig>

  • sq: ((options?) => QuantizerRecursivePartial<SQConfig>)

    Create an object of type SQConfigCreate to be used when defining the quantizer configuration of a vector index.

    Returns

    The object of type SQConfigCreate.

    • • (options?): QuantizerRecursivePartial<SQConfig>

      • Parameters

        • Optional options: {
              rescoreLimit?: number;
              trainingLimit?: number;
          }

          • Optional rescoreLimit?: number

            The rescore limit.

          • Optional trainingLimit?: number

            The training limit.

        Returns QuantizerRecursivePartial<SQConfig>