Fix AI service model selection

[graemerocher]

Summary

• configure only the chat model for non-streaming AI services and only the streaming chat model for streaming AI services
• retain support for mixed service interfaces by configuring both models when a bean exposes both return styles
• add regression coverage for chat-only and streaming-only AI service interfaces

Verification

• ./gradlew :micronaut-langchain4j-core:test --tests io.micronaut.langchain4j.aiservices.AiServiceFactoryModelSelectionTest

Resolves #233

[Copilot]

Pull request overview

Updates AiServiceFactory to avoid configuring both ChatModel and StreamingChatModel for AI services that only need one model type, aligning Micronaut’s wiring with LangChain4j’s expectations and preventing builder errors.

Changes:

• Add model-selection logic that inspects an @AiService bean’s executable methods to decide whether to configure ChatModel, StreamingChatModel, or both.
• Use LangChain4j TokenStreamAdapter SPI to classify streaming return types beyond TokenStream.
• Add regression tests for chat-only and streaming-only AI service interfaces.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 3 comments.

File	Description
micronaut-langchain4j-core/src/main/java/io/micronaut/langchain4j/aiservices/AiServiceFactory.java	Introduces method-based model selection and conditional wiring of chat vs streaming chat models.
micronaut-langchain4j-core/src/test/java/io/micronaut/langchain4j/aiservices/AiServiceFactoryModelSelectionTest.java	Adds tests ensuring only the needed model is configured for chat-only vs streaming-only services.

[graemerocher]

@copilot apply changes based on the comments in this thread

[Copilot]

@copilot apply changes based on the comments in this thread

Applied all three review changes in f4567ca:

1. isAssignableFrom fix – isStreamingReturnType now uses TokenStream.class.isAssignableFrom(returnType.getType()) to correctly handle subtype returns.
2. Deterministic error handling – the streaming test now captures errors in an AtomicReference<Throwable> and counts down the latch in both success and error paths, then re-throws after await().
3. BOTH test – added configuresBothModelsForMixedServices covering a MixedAssistant interface with both String chat(String) and TokenStream streamChat(String) methods, asserting ModelSelection.BOTH and validating both call styles.

All 3 tests pass.

Warning

Firewall rules blocked me from connecting to one or more addresses (expand for details)

I tried to connect to the following addresses, but was blocked by firewall rules:

• ge.micronaut.io
  • Triggering command: /usr/lib/jvm/temurin-25-jdk-amd64/bin/java /usr/lib/jvm/temurin-25-jdk-amd64/bin/java --add-opens=java.base/java.lang=ALL-UNNAMED --add-opens=java.base/java.lang.invoke=ALL-UNNAMED --add-opens=java.base/java.util=ALL-UNNAMED --add-opens=java.prefs/java.util.prefs=ALL-UNNAMED --add-exports=jdk.compiler/com.sun.tools.javac.api=ALL-UNNAMED --add-exports=jdk.compiler/com.sun.tools.javac.util=ALL-UNNAMED --add-opens=java.base/java.nio.charset=ALL-UNNAMED --add-opens=java.base/java.net=ALL-UNNAMED --add-opens=java.base/java.util.concurrent=ALL-UNNAMED --add-opens=java.base/java.util.concurrent.atomic=ALL-UNNAMED --add-opens=java.xml/javax.xml.namespace=ALL-UNNAMED --add-opens=java.base/java.time=ALL-UNNAMED --enable-native-access=ALL-UNNAMED -Xmx1g -Dfile.encoding=UTF-8 -Duser.country -Duser.language=en -Duser.variant -cp (dns block)

If you need me to access, download, or install something from one of these locations, you can either:

• Configure Actions setup steps to set up my environment, which run before the firewall is enabled
• Add the appropriate URLs or hosts to the custom allowlist in this repository's Copilot coding agent settings (admins only)

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
85.7% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud