Class GoogleAiGeminiChatResponseMetadata

java.lang.Object

dev.langchain4j.model.chat.response.ChatResponseMetadata

dev.langchain4j.model.googleai.GoogleAiGeminiChatResponseMetadata

public class GoogleAiGeminiChatResponseMetadata
extends ChatResponseMetadata

Gemini-specific metadata for ChatResponse.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	GoogleAiGeminiChatResponseMetadata.Builder

Method Summary

Modifier and Type	Method	Description
String	blockReason()	The reason Gemini refused to process the prompt, for example "SAFETY", "PROHIBITED_CONTENT", "BLOCKLIST" or "IMAGE_SAFETY".
static GoogleAiGeminiChatResponseMetadata.Builder	builder()
boolean	equals(Object o)
GroundingMetadata	groundingMetadata()
int	hashCode()
List<GeminiSafetyRating>	promptSafetyRatings()	The safety assessments of the prompt, as reported by Gemini's promptFeedback.
List<GeminiSafetyRating>	safetyRatings()	The safety assessments of the generated content, one per harm category that Gemini evaluated.
GoogleAiGeminiChatResponseMetadata.Builder	toBuilder()
String	toString()
UrlContextMetadata	urlContextMetadata()

Methods inherited from class ChatResponseMetadata

finishReason, id, modelName, toBuilder, tokenUsage

Methods inherited from class Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Method Details

groundingMetadata

public GroundingMetadata groundingMetadata()

urlContextMetadata

public UrlContextMetadata urlContextMetadata()

safetyRatings

public List<GeminiSafetyRating> safetyRatings()

The safety assessments of the generated content, one per harm category that Gemini evaluated.

Use these to find out why generation stopped when ChatResponseMetadata.finishReason() is FinishReason.CONTENT_FILTER. To inspect the safety of the prompt instead, use promptSafetyRatings().

Returns:

an unmodifiable list of safety ratings, empty when the response carries none

promptSafetyRatings

public List<GeminiSafetyRating> promptSafetyRatings()

The safety assessments of the prompt, as reported by Gemini's promptFeedback.

These describe the request rather than the generated content, and are typically populated when Gemini refuses to answer at all. In that case blockReason() is set and safetyRatings() is empty.

Returns:

an unmodifiable list of safety ratings for the prompt, empty when the response carries none

blockReason

public String blockReason()

The reason Gemini refused to process the prompt, for example "SAFETY", "PROHIBITED_CONTENT", "BLOCKLIST" or "IMAGE_SAFETY".

When this is set, Gemini returned no content at all: the AiMessage carries no text, ChatResponseMetadata.finishReason() is FinishReason.CONTENT_FILTER, and promptSafetyRatings() explains which harm categories were involved.

Returns:

the block reason, or null when the prompt was not blocked

toBuilder

public GoogleAiGeminiChatResponseMetadata.Builder toBuilder()

Overrides:

toBuilder in class ChatResponseMetadata

equals

public boolean equals(Object o)

Overrides:

equals in class ChatResponseMetadata

hashCode

public int hashCode()

Overrides:

hashCode in class ChatResponseMetadata

toString

public String toString()

Overrides:

toString in class ChatResponseMetadata

builder

public static GoogleAiGeminiChatResponseMetadata.Builder builder()