Package org.codelibs.fess.llm

Interface LlmStreamCallback

All Known Implementing Classes:

PhaseAwareStreamCallback

Functional Interface:

This is a functional interface and can therefore be used as the assignment target for a lambda expression or method reference.

@FunctionalInterface
public interface LlmStreamCallback

Callback interface for receiving streaming chat responses from LLM.

Method Summary

Modifier and Type	Method	Description
void	onChunk(String chunk, boolean done)	Called for each chunk of the streaming response.
default void	onError(Throwable error)	Called when an error occurs during streaming.
default void	onRetry(String operation, int attempt, int maxAttempts, long sleepMs, Throwable cause)	Called when the LLM HTTP call is being retried.
default void	onWaiting(String reason, long elapsedMs, long timeoutMs)	Called when the request is waiting for a concurrency permit.
default void	onWarning(String code, String detail)	Called when an internal silent fallback occurs (e.g., reasoning model token exhaustion).

Method Details

onChunk

void onChunk(String chunk,
 boolean done)

Called for each chunk of the streaming response.

Parameters:

chunk - the text chunk from the LLM response

done - true if this is the final chunk

onError

default void onError(Throwable error)

Called when an error occurs during streaming.

Parameters:

error - the error that occurred

onRetry

default void onRetry(String operation,
 int attempt,
 int maxAttempts,
 long sleepMs,
 Throwable cause)

Called when the LLM HTTP call is being retried.

Parameters:

operation - log label for the operation (e.g., "chat", "streamChat")

attempt - 1-based current attempt index that just failed

maxAttempts - total attempts including the first

sleepMs - milliseconds the client will sleep before the next attempt

cause - the cause of the retry (RetryableHttpException or IOException)

onWaiting

default void onWaiting(String reason,
 long elapsedMs,
 long timeoutMs)

Called when the request is waiting for a concurrency permit.

Parameters:

reason - short reason code, currently always "concurrency_limit"

elapsedMs - milliseconds already spent waiting (0 on the first call)

timeoutMs - maximum milliseconds the client will wait before giving up

onWarning

default void onWarning(String code,
 String detail)

Called when an internal silent fallback occurs (e.g., reasoning model token exhaustion).

Parameters:

code - short machine-readable warning code (e.g., "reasoning_token_exhausted")

detail - human-readable detail (free-form)