Package org.springframework.ai.vertexai.embedding.multimodal

Class VertexAiMultimodalEmbeddingOptions

java.lang.Object

org.springframework.ai.vertexai.embedding.multimodal.VertexAiMultimodalEmbeddingOptions

All Implemented Interfaces:

org.springframework.ai.embedding.EmbeddingOptions, org.springframework.ai.model.ModelOptions

public class VertexAiMultimodalEmbeddingOptions
extends Object
implements org.springframework.ai.embedding.EmbeddingOptions

Class representing the options for Vertex AI Multimodal Embedding.

The options include the embedding model name, the number of dimensions of the resulting output, the start and end offset of the video segment, and the interval of the video for embedding generation.

The supported embedding models are text-embedding-004, text-multilingual-embedding-002, and multimodalembedding@001.

The number of dimensions is used to specify the size of the resulting output embeddings. This can be useful for storage optimization purposes. Supported for model version 004 and later.

The video start offset and end offset specify the segment of the video to be used for embedding generation. If not specified, the default values are calculated based on the video length and are adjusted to ensure a minimum segment of 120 seconds.

The video interval specifies the period of the video over which embeddings will be generated. The minimum value is 4, and if it is lower, an InvalidArgumentError is returned. There is no maximum limit for the interval value, but if it exceeds the video length or 120 seconds, it may impact the quality of the generated embeddings. The default value is 16.

Since:

1.0.0

Author:

Christian Tzolov, Ilayaperumal Gopinathan, Sebastien Deleuze

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static final class	VertexAiMultimodalEmbeddingOptions.Builder

Field Summary

Fields

Modifier and Type	Field	Description
static final String	DEFAULT_MODEL_NAME

Constructor Summary

Constructors

Modifier	Constructor	Description
protected	VertexAiMultimodalEmbeddingOptions(@Nullable String model, @Nullable Integer dimensions, @Nullable Integer videoStartOffsetSec, @Nullable Integer videoEndOffsetSec, @Nullable Integer videoIntervalSec)

Method Summary

Modifier and Type	Method	Description
static VertexAiMultimodalEmbeddingOptions.Builder	builder()
@Nullable Integer	getDimensions()
String	getModel()
@Nullable Integer	getVideoEndOffsetSec()
@Nullable Integer	getVideoIntervalSec()
@Nullable Integer	getVideoStartOffsetSec()

Methods inherited from class java.lang.Object

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

Field Details

DEFAULT_MODEL_NAME

public static final String DEFAULT_MODEL_NAME

Constructor Details

VertexAiMultimodalEmbeddingOptions

protected VertexAiMultimodalEmbeddingOptions(@Nullable String model,
 @Nullable Integer dimensions,
 @Nullable Integer videoStartOffsetSec,
 @Nullable Integer videoEndOffsetSec,
 @Nullable Integer videoIntervalSec)

Method Details

builder

public static VertexAiMultimodalEmbeddingOptions.Builder builder()

getModel

public String getModel()

Specified by:

getModel in interface org.springframework.ai.embedding.EmbeddingOptions

getDimensions

public @Nullable Integer getDimensions()

Specified by:

getDimensions in interface org.springframework.ai.embedding.EmbeddingOptions

getVideoStartOffsetSec

public @Nullable Integer getVideoStartOffsetSec()

getVideoEndOffsetSec

public @Nullable Integer getVideoEndOffsetSec()

getVideoIntervalSec

public @Nullable Integer getVideoIntervalSec()