com.google.cloud.vertexai.generativeai

Class GenerativeModel

java.lang.Object

com.google.cloud.vertexai.generativeai.GenerativeModel

public final class GenerativeModel
extends Object

This class holds a generative model that can complete what you provided. This class is thread-safe.

Note: The instances of ChatSession returned by startChat() are NOT thread-safe.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	GenerativeModel.Builder Builder class for GenerativeModel.

Constructor Summary

Constructors

Constructor and Description
GenerativeModel(String modelName, VertexAI vertexAi) Constructs a GenerativeModel instance.

Method Summary

Modifier and Type	Method and Description
CountTokensResponse	countTokens(Content content) Counts tokens in a single content.
CountTokensResponse	countTokens(List<Content> contents) Counts tokens in a list of contents.
CountTokensResponse	countTokens(String text) Counts tokens in a text message.
GenerateContentResponse	generateContent(Content content) Generates content from this model given a single content.
GenerateContentResponse	generateContent(List<Content> contents) Generates content from this model given a list of contents.
GenerateContentResponse	generateContent(String text) Generates content from generative model given a text.
com.google.api.core.ApiFuture<GenerateContentResponse>	generateContentAsync(Content content) Asynchronously generates content from generative model given a single Content.
com.google.api.core.ApiFuture<GenerateContentResponse>	generateContentAsync(List<Content> contents) Asynchronously generates content from generative model given a list of contents.
com.google.api.core.ApiFuture<GenerateContentResponse>	generateContentAsync(String text) Asynchronously generates content from generative model given a text.
ResponseStream<GenerateContentResponse>	generateContentStream(Content content) Generates content with streaming support from generative model given a single Content.
ResponseStream<GenerateContentResponse>	generateContentStream(List<Content> contents) Generates content with streaming support from generative model given a list of contents.
ResponseStream<GenerateContentResponse>	generateContentStream(String text) Generates content with streaming support from generative model given a text.
GenerationConfig	getGenerationConfig() Returns the GenerationConfig of this generative model.
String	getModelName() Returns the model name of this generative model.
com.google.common.collect.ImmutableList<SafetySetting>	getSafetySettings() Returns a list of SafetySetting of this generative model.
Optional<Content>	getSystemInstruction() Returns the optional system instruction of this generative model.
Optional<ToolConfig>	getToolConfig() Returns the optional ToolConfig of this generative model.
com.google.common.collect.ImmutableList<Tool>	getTools() Returns a list of Tool of this generative model.
ChatSession	startChat() Returns a new ChatSession instance that can be used to start a chat with this model.
GenerativeModel	withGenerationConfig(GenerationConfig generationConfig) Creates a copy of the current model with updated GenerationConfig.
GenerativeModel	withSafetySettings(List<SafetySetting> safetySettings) Creates a copy of the current model with updated safetySettings.
GenerativeModel	withSystemInstruction(Content systemInstruction) Creates a copy of the current model with updated system instructions.
GenerativeModel	withToolConfig(ToolConfig toolConfig) Creates a copy of the current model with updated tool config.
GenerativeModel	withTools(List<Tool> tools) Creates a copy of the current model with updated tools.

Methods inherited from class java.lang.Object

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

Constructor Detail

GenerativeModel

public GenerativeModel(String modelName,
                       VertexAI vertexAi)

Constructs a GenerativeModel instance.

Parameters:

modelName - the name of the generative model. Supported format: "gemini-pro", "models/gemini-pro", "publishers/google/models/gemini-pro", where "gemini-pro" is the model name. Valid model names can be found at https://cloud.google.com/vertex-ai/docs/generative-ai/learn/models#gemini-models

vertexAi - a VertexAI that contains the default configs for the generative model

Method Detail

withGenerationConfig

public GenerativeModel withGenerationConfig(GenerationConfig generationConfig)

Creates a copy of the current model with updated GenerationConfig.

Parameters:

generationConfig - a GenerationConfig that will be used in the new model.

Returns:

a new GenerativeModel instance with the specified GenerationConfig.

withSafetySettings

public GenerativeModel withSafetySettings(List<SafetySetting> safetySettings)

Creates a copy of the current model with updated safetySettings.

Parameters:

safetySettings - a list of SafetySetting that will be used in the new model.

Returns:

a new GenerativeModel instance with the specified safetySettings.

withTools

public GenerativeModel withTools(List<Tool> tools)

Creates a copy of the current model with updated tools.

Parameters:

tools - a list of Tool that will be used in the new model.

Returns:

a new GenerativeModel instance with the specified tools.

withToolConfig

public GenerativeModel withToolConfig(ToolConfig toolConfig)

Creates a copy of the current model with updated tool config.

Parameters:

toolConfig - a ToolConfig that will be used in the new model.

Returns:

a new GenerativeModel instance with the specified tool config.

withSystemInstruction

public GenerativeModel withSystemInstruction(Content systemInstruction)

Creates a copy of the current model with updated system instructions.

Parameters:

systemInstruction - a Content containing system instructions.

Returns:

a new GenerativeModel instance with the specified tools.

countTokens

@BetaApi
public CountTokensResponse countTokens(String text)
                                         throws IOException

Counts tokens in a text message.

Parameters:

text - a text message to count tokens

Returns:

a CountTokensResponse instance that contains the total tokens and total billable characters of the given list of contents

Throws:

IOException - if an I/O error occurs while making the API call

countTokens

@BetaApi
public CountTokensResponse countTokens(Content content)
                                         throws IOException

Counts tokens in a single content.

Parameters:

content - a Content to count tokens

Returns:

a CountTokensResponse instance that contains the total tokens and total billable characters of the given list of contents

Throws:

IOException - if an I/O error occurs while making the API call

countTokens

@BetaApi
public CountTokensResponse countTokens(List<Content> contents)
                                         throws IOException

Counts tokens in a list of contents.

Parameters:

contents - a list of Content to count tokens

Returns:

a CountTokensResponse instance that contains the total tokens and total billable characters of the given list of contents

Throws:

IOException - if an I/O error occurs while making the API call

generateContent

public GenerateContentResponse generateContent(String text)
                                        throws IOException

Generates content from generative model given a text.

Parameters:

text - a text message to send to the generative model

Returns:

a GenerateContentResponse instance that contains response contents and other metadata

Throws:

IOException - if an I/O error occurs while making the API call

generateContent

public GenerateContentResponse generateContent(Content content)
                                        throws IOException

Generates content from this model given a single content.

Parameters:

content - a Content to send to the generative model

Returns:

a GenerateContentResponse instance that contains response contents and other metadata

Throws:

IOException - if an I/O error occurs while making the API call

generateContent

public GenerateContentResponse generateContent(List<Content> contents)
                                        throws IOException

Generates content from this model given a list of contents.

Parameters:

contents - a list of Content to send to the generative model

Returns:

a GenerateContentResponse instance that contains response contents and other metadata

Throws:

IOException - if an I/O error occurs while making the API call

generateContentStream

public ResponseStream<GenerateContentResponse> generateContentStream(String text)
                                                              throws IOException

Generates content with streaming support from generative model given a text.

Parameters:

text - a text message to send to the generative model

Returns:

a ResponseStream that contains a streaming of GenerateContentResponse

Throws:

IOException - if an I/O error occurs while making the API call

generateContentStream

public ResponseStream<GenerateContentResponse> generateContentStream(Content content)
                                                              throws IOException

Generates content with streaming support from generative model given a single Content.

Parameters:

content - a Content to send to the generative model. The role of the content is "user".

Returns:

a ResponseStream that contains a streaming of GenerateContentResponse

Throws:

IOException - if an I/O error occurs while making the API call

generateContentStream

public ResponseStream<GenerateContentResponse> generateContentStream(List<Content> contents)
                                                              throws IOException

Generates content with streaming support from generative model given a list of contents.

Parameters:

contents - a list of Content to send to the generative model

Returns:

a ResponseStream that contains a streaming of GenerateContentResponse

Throws:

IOException - if an I/O error occurs while making the API call

generateContentAsync

public com.google.api.core.ApiFuture<GenerateContentResponse> generateContentAsync(String text)
                                                                            throws IOException

Asynchronously generates content from generative model given a text.

Parameters:

text - a text message to send to the generative model

Returns:

a ApiFuture represents the response of an asynchronous generateContent request

Throws:

IOException - if an I/O error occurs while making the API call

generateContentAsync

public com.google.api.core.ApiFuture<GenerateContentResponse> generateContentAsync(Content content)
                                                                            throws IOException

Asynchronously generates content from generative model given a single Content.

Parameters:

content - a Content to send to the generative model. The role of the content is "user".

Returns:

a ApiFuture represents the response of an asynchronous generateContent request

Throws:

IOException - if an I/O error occurs while making the API call

generateContentAsync

public com.google.api.core.ApiFuture<GenerateContentResponse> generateContentAsync(List<Content> contents)
                                                                            throws IOException

Asynchronously generates content from generative model given a list of contents.

Parameters:

contents - a list of Content to send to the generative model

Returns:

a ApiFuture represents the response of an asynchronous generateContent request

Throws:

IOException - if an I/O error occurs while making the API call

getModelName

public String getModelName()

Returns the model name of this generative model.

getGenerationConfig

public GenerationConfig getGenerationConfig()

Returns the GenerationConfig of this generative model.

getSafetySettings

public com.google.common.collect.ImmutableList<SafetySetting> getSafetySettings()

Returns a list of SafetySetting of this generative model.

getTools

public com.google.common.collect.ImmutableList<Tool> getTools()

Returns a list of Tool of this generative model.

getToolConfig

public Optional<ToolConfig> getToolConfig()

Returns the optional ToolConfig of this generative model.

getSystemInstruction

public Optional<Content> getSystemInstruction()

Returns the optional system instruction of this generative model.

startChat

public ChatSession startChat()

Returns a new ChatSession instance that can be used to start a chat with this model.

Note: the returned ChatSession instance is NOT thread-safe.