Skip to content

Observation for Vertex gemini chat model streaming operation #1492

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
wants to merge 1 commit into from
Closed

Observation for Vertex gemini chat model streaming operation #1492

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
88 changes: 60 additions & 28 deletions ...-gemini/src/main/java/org/springframework/ai/vertexai/gemini/VertexAiGeminiChatModel.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,6 +37,7 @@
import org.springframework.ai.chat.model.ChatModel;
import org.springframework.ai.chat.model.ChatResponse;
import org.springframework.ai.chat.model.Generation;
import org.springframework.ai.chat.model.MessageAggregator;
import org.springframework.ai.chat.observation.ChatModelObservationContext;
import org.springframework.ai.chat.observation.ChatModelObservationConvention;
import org.springframework.ai.chat.observation.ChatModelObservationDocumentation;
Expand All @@ -59,7 +60,9 @@
import org.springframework.util.CollectionUtils;
import org.springframework.util.StringUtils;

import io.micrometer.observation.Observation;
import io.micrometer.observation.ObservationRegistry;
import io.micrometer.observation.contextpropagation.ObservationThreadLocalAccessor;
import reactor.core.publisher.Flux;

import java.util.ArrayList;
Expand Down Expand Up @@ -101,7 +104,7 @@ public class VertexAiGeminiChatModel extends AbstractToolCallSupport implements
/**
* Conventions to use for generating observations.
*/
private final ChatModelObservationConvention observationConvention = DEFAULT_OBSERVATION_CONVENTION;
private ChatModelObservationConvention observationConvention = DEFAULT_OBSERVATION_CONVENTION;

public enum GeminiMessageType {

Expand Down Expand Up @@ -242,38 +245,58 @@ public ChatResponse call(Prompt prompt) {

@Override
public Flux<ChatResponse> stream(Prompt prompt) {
try {

return Flux.deferContextual(contextView -> {
VertexAiGeminiChatOptions vertexAiGeminiChatOptions = vertexAiGeminiChatOptions(prompt);
var request = createGeminiRequest(prompt, vertexAiGeminiChatOptions);

ResponseStream<GenerateContentResponse> responseStream = request.model
.generateContentStream(request.contents);

return Flux.fromStream(responseStream.stream()).switchMap(response -> {

List<Generation> generations = response.getCandidatesList()
.stream()
.map(this::responseCandiateToGeneration)
.flatMap(List::stream)
.toList();
ChatModelObservationContext observationContext = ChatModelObservationContext.builder()
.prompt(prompt)
.provider(VertexAiGeminiConstants.PROVIDER_NAME)
.requestOptions(vertexAiGeminiChatOptions)
.build();

ChatResponse chatResponse = new ChatResponse(generations, toChatResponseMetadata(response));
Observation observation = ChatModelObservationDocumentation.CHAT_MODEL_OPERATION.observation(
this.observationConvention, DEFAULT_OBSERVATION_CONVENTION, () -> observationContext,
this.observationRegistry);

if (!isProxyToolCalls(prompt, this.defaultOptions) && isToolCall(chatResponse,
Set.of(FinishReason.STOP.name(), FinishReason.FINISH_REASON_UNSPECIFIED.name()))) {
var toolCallConversation = handleToolCalls(prompt, chatResponse);
// Recursively call the stream method with the tool call message
// conversation that contains the call responses.
return this.stream(new Prompt(toolCallConversation, prompt.getOptions()));
}
observation.parentObservation(contextView.getOrDefault(ObservationThreadLocalAccessor.KEY, null)).start();
var request = createGeminiRequest(prompt, vertexAiGeminiChatOptions);

return Flux.just(chatResponse);
});
}
catch (Exception e) {
throw new RuntimeException("Failed to generate content", e);
}
try {
ResponseStream<GenerateContentResponse> responseStream = request.model
.generateContentStream(request.contents);

return Flux.fromStream(responseStream.stream()).switchMap(response -> {

List<Generation> generations = response.getCandidatesList()
.stream()
.map(this::responseCandiateToGeneration)
.flatMap(List::stream)
.toList();

ChatResponse chatResponse = new ChatResponse(generations, toChatResponseMetadata(response));

if (!isProxyToolCalls(prompt, this.defaultOptions) && isToolCall(chatResponse,
Set.of(FinishReason.STOP.name(), FinishReason.FINISH_REASON_UNSPECIFIED.name()))) {
var toolCallConversation = handleToolCalls(prompt, chatResponse);
// Recursively call the stream method with the tool call message
// conversation that contains the call responses.
return this.stream(new Prompt(toolCallConversation, prompt.getOptions()));
}

Flux<ChatResponse> chatResponseFlux = Flux.just(chatResponse)
.doOnError(observation::error)
.doFinally(s -> {
observation.stop();
})
.contextWrite(ctx -> ctx.put(ObservationThreadLocalAccessor.KEY, observation));

return new MessageAggregator().aggregate(chatResponseFlux, observationContext::setResponse);
});
}
catch (Exception e) {
throw new RuntimeException("Failed to generate content", e);
}
});
}

protected List<Generation> responseCandiateToGeneration(Candidate candidate) {
Expand Down Expand Up @@ -619,4 +642,13 @@ public void destroy() throws Exception {
}
}

/**
* Use the provided convention for reporting observation data
* @param observationConvention The provided convention
*/
public void setObservationConvention(ChatModelObservationConvention observationConvention) {
Assert.notNull(observationConvention, "observationConvention cannot be null");
this.observationConvention = observationConvention;
}

}
34 changes: 34 additions & 0 deletions .../src/test/java/org/springframework/ai/vertexai/gemini/VertexAiChatModelObservationIT.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,6 +19,7 @@
import static org.assertj.core.api.Assertions.assertThat;

import java.util.List;
import java.util.stream.Collectors;

import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;
Expand All @@ -41,6 +42,7 @@
import com.google.cloud.vertexai.VertexAI;
import io.micrometer.observation.tck.TestObservationRegistry;
import io.micrometer.observation.tck.TestObservationRegistryAssert;
import reactor.core.publisher.Flux;

/**
* @author Soby Chacko
Expand Down Expand Up @@ -83,6 +85,38 @@ void observationForChatOperation() {
validate(responseMetadata);
}

@Test
void observationForStreamingOperation() {

var options = VertexAiGeminiChatOptions.builder()
.withModel(VertexAiGeminiChatModel.ChatModel.GEMINI_1_5_PRO.getValue())
.withTemperature(0.7)
.withStopSequences(List.of("this-is-the-end"))
.withMaxOutputTokens(2048)
.withTopP(1.0)
.build();

Prompt prompt = new Prompt("Why does a raven look like a desk?", options);

Flux<ChatResponse> chatResponse = chatModel.stream(prompt);
List<ChatResponse> responses = chatResponse.collectList().block();
assertThat(responses).isNotEmpty();
assertThat(responses).hasSizeGreaterThan(1);

String aggregatedResponse = responses.subList(0, responses.size() - 1)
.stream()
.map(r -> r.getResult().getOutput().getContent())
.collect(Collectors.joining());
assertThat(aggregatedResponse).isNotEmpty();

ChatResponse lastChatResponse = responses.get(responses.size() - 1);

ChatResponseMetadata responseMetadata = lastChatResponse.getMetadata();
assertThat(responseMetadata).isNotNull();

validate(responseMetadata);
}

private void validate(ChatResponseMetadata responseMetadata) {
TestObservationRegistryAssert.assertThat(observationRegistry)
.doesNotHaveAnyRemainingCurrentObservation()
Expand Down
14 changes: 11 additions & 3 deletions ...org/springframework/ai/autoconfigure/vertexai/gemini/VertexAiGeminiAutoConfiguration.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,10 +19,12 @@
import java.util.List;

import org.springframework.ai.autoconfigure.retry.SpringAiRetryAutoConfiguration;
import org.springframework.ai.chat.observation.ChatModelObservationConvention;
import org.springframework.ai.model.function.FunctionCallback;
import org.springframework.ai.model.function.FunctionCallbackContext;
import org.springframework.ai.model.function.FunctionCallbackContext.SchemaType;
import org.springframework.ai.vertexai.gemini.VertexAiGeminiChatModel;
import org.springframework.beans.factory.ObjectProvider;
import org.springframework.boot.autoconfigure.AutoConfiguration;
import org.springframework.boot.autoconfigure.ImportAutoConfiguration;
import org.springframework.boot.autoconfigure.condition.ConditionalOnClass;
Expand All @@ -38,6 +40,7 @@

import com.google.auth.oauth2.GoogleCredentials;
import com.google.cloud.vertexai.VertexAI;
import io.micrometer.observation.ObservationRegistry;

/**
* Auto-configuration for Vertex AI Gemini Chat.
Expand Down Expand Up @@ -86,12 +89,17 @@ public VertexAI vertexAi(VertexAiGeminiConnectionProperties connectionProperties
@ConditionalOnProperty(prefix = VertexAiGeminiChatProperties.CONFIG_PREFIX, name = "enabled", havingValue = "true",
matchIfMissing = true)
public VertexAiGeminiChatModel vertexAiGeminiChat(VertexAI vertexAi, VertexAiGeminiChatProperties chatProperties,
List<FunctionCallback> toolFunctionCallbacks, ApplicationContext context, RetryTemplate retryTemplate) {
List<FunctionCallback> toolFunctionCallbacks, ApplicationContext context, RetryTemplate retryTemplate,
ObjectProvider<ObservationRegistry> observationRegistry,
ObjectProvider<ChatModelObservationConvention> observationConvention) {

FunctionCallbackContext functionCallbackContext = springAiFunctionManager(context);

return new VertexAiGeminiChatModel(vertexAi, chatProperties.getOptions(), functionCallbackContext,
toolFunctionCallbacks, retryTemplate);
VertexAiGeminiChatModel chatModel = new VertexAiGeminiChatModel(vertexAi, chatProperties.getOptions(),
functionCallbackContext, toolFunctionCallbacks, retryTemplate,
observationRegistry.getIfUnique(() -> ObservationRegistry.NOOP));
observationConvention.ifAvailable(chatModel::setObservationConvention);
return chatModel;
}

/**
Expand Down