Module agentforge4j.llm.fake

Package com.agentforge4j.llm.fake

Class FakeConfiguration

java.lang.Object

com.agentforge4j.llm.fake.FakeConfiguration

All Implemented Interfaces:

LlmClientConfiguration

public final class FakeConfiguration
extends Object
implements LlmClientConfiguration

Configuration for the fake provider. Unlike HTTP providers, the load-bearing field is the FakeResponseSource — the shared per-run store that the registration API (test runner / demo flow) also holds — so the FakeLlmClientFactory-built client and the registration API resolve against one store.

Wire this through the embedding application or programmatic withLlmProvider(...); the env/system-property bootstrap discovery path supplies a generic configuration and cannot carry a source, so it does not configure the fake provider (the same way Ollama requires its own configuration subtype).

Constructor Summary

Constructors

Constructor	Description
FakeConfiguration(FakeResponseSource responseSource)	Creates a configuration with the default model id and a zero connect timeout.
FakeConfiguration(FakeResponseSource responseSource, String defaultModel, Duration connectTimeout)	Creates a configuration with an explicit default model and connect timeout.

Method Summary

Modifier and Type	Method	Description
Duration	getConnectTimeout()	HTTP connect timeout for outbound requests to this provider.
String	getDefaultModel()	Default model when a request does not set LlmExecutionRequest.model().
String	getProviderName()	Provider id for this configuration (must align with the matching LlmClientFactory).
FakeResponseSource	responseSource()	Returns the shared scripted response source the client resolves against.

Methods inherited from class java.lang.Object

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

Methods inherited from interface com.agentforge4j.llm.LlmClientConfiguration

getApiKeyReference, getBaseUrl, getOptions, getRetryPolicy, requireBaseUrl

Constructor Details

FakeConfiguration

public FakeConfiguration(FakeResponseSource responseSource)

Creates a configuration with the default model id and a zero connect timeout.

Parameters:

responseSource - the scripted response source; must not be null

FakeConfiguration

public FakeConfiguration(FakeResponseSource responseSource,
 String defaultModel,
 Duration connectTimeout)

Creates a configuration with an explicit default model and connect timeout.

Parameters:

responseSource - the scripted response source; must not be null

defaultModel - default model id reported to callers; must not be blank

connectTimeout - nominal connect timeout (unused by the fake transport); must not be null

Method Details

getProviderName

public String getProviderName()

Description copied from interface: LlmClientConfiguration

Provider id for this configuration (must align with the matching LlmClientFactory).

Specified by:

getProviderName in interface LlmClientConfiguration

Returns:

non-blank provider id such as "openai" or "ollama"

getDefaultModel

public String getDefaultModel()

Description copied from interface: LlmClientConfiguration

Default model when a request does not set LlmExecutionRequest.model().

Specified by:

getDefaultModel in interface LlmClientConfiguration

Returns:

non-blank default model id for this provider

getConnectTimeout

public Duration getConnectTimeout()

Description copied from interface: LlmClientConfiguration

HTTP connect timeout for outbound requests to this provider.

Specified by:

getConnectTimeout in interface LlmClientConfiguration

Returns:

connect timeout duration

responseSource

public FakeResponseSource responseSource()

Returns the shared scripted response source the client resolves against.

Returns:

the response source; never null