All Classes and Interfaces
Class
Description
Provides the basic building blocks that the generated Interface methods call into
Abstract class to be implemented in order to keep track of whatever information is useful for the application auditing.
Information about the AiService that is being audited
Allow applications to audit parts of the interactions with the LLM that interest them
Interface implemented by each AiService that allows the removal of chat memories from an AiService
Allows the application to manually control when a
ChatMemory should be removed from the underlying
ChatMemoryStore.Provides a way for an AiService to get its chat memory seeded.
Allows for user code to provide a custom strategy for estimating the cost of API calls
Meant to be injected where
ChatModelListener
is used in order to determine the cost of the API requestUsed to make Quarkus aware of classes being used in
AiServices.create(java.lang.Class<T>, dev.langchain4j.model.chat.ChatLanguageModel)A guardrail is a rule that is applied when interacting with an LLM either to the input (the user message) or to the output of
the model to ensure that they are safe and meet the expectations of the model.
Exception thrown when a input or output guardrail validation fails.
Represents the parameter passed to
Guardrail.validate(GuardrailParams)} in order to validate an interaction between a
user and the LLM.The result of the validation of an interaction between a user and the LLM.
The message and the cause of the failure of a single validation.
The possible results of a guardrails validation.
This annotation is useful when an AiService is meant to describe an image as the value of the method parameter annotated
with @ImageUrl
will be used as an
ImageContent.Creates the default
InMemoryChatMemoryStore store to be used by classes annotated with RegisterAiServiceAn input guardrail is a rule that is applied to the input of the model to ensure that the input (the user message) is
safe and meets the expectations of the model.
Represents the parameter passed to
InputGuardrail.validate(InputGuardrailParams).The result of the validation of an
InputGuardrailAn annotation to apply guardrails to the input of the model.
Creates metrics that follow the
Semantic Conventions
for GenAI Metrics
Marker annotation to select a named model
Configure the
name parameter to select the model instance.An implementation of
ChatMemory that does nothing.An output guardrail is a rule that is applied to the output of the model to ensure that the output is safe and meets the
expectations.
Represents the parameter passed to
OutputGuardrail.validate(OutputGuardrailParams).The result of the validation of an
OutputGuardrailAn annotation to apply guardrails to the output of the model.
Used to create LangChain4j's
AiServices in a declarative manner that the application can then use simply by
using the class as a CDI bean.Marker that is used to tell Quarkus to use the
ChatLanguageModel that has been configured as a CDI bean by
any of the extensions providing such capability (such as quarkus-langchain4j-openai and
quarkus-langchain4j-hugging-face).Marker that is used to tell Quarkus to use the retriever that the user has configured as a CDI bean.
Marker that is used to tell Quarkus to use the
AuditService that the user has configured as a CDI bean.Marker that is used to tell Quarkus to use the
ImageModel that the user has configured as a CDI bean.Marker that is used to tell Quarkus to use the
ModerationModel that the user has configured as a CDI bean.Marker that is used to tell Quarkus to use the
RetrievalAugmentor that the user has configured as a CDI bean.Marker that is used to tell Quarkus to use the
StreamingChatLanguageModel that has been configured as a CDI bean
by * any of the extensions providing such capability (such as quarkus-langchain4j-openai and
quarkus-langchain4j-hugging-face).Marker that is used when the user does not want any memory configured for the AiService
Marker that is used to tell Quarkus to not use any retrieval augmentor even if a CDI bean implementing
the `RetrievalAugmentor` interface exists.
Marker class to indicate that no retriever should be used
This implementation uses the state of the request scope as the default value
Provides a way for an AiService to get its chat memory seeded with examples interactions.
Creates a span that follows the
Semantic Conventions
for GenAI operations
When used on a method of an AiService annotated with
RegisterAiService, the method will the tool classes provided by
value
instead of the ones configured for the entire AiService (via RegisterAiService.tools())