jiriOndrusek-agent opened a new issue, #9225:
URL: https://github.com/apache/camel-quarkus/issues/9225

   The docs promise a default `RetrievalAugmentor` from the `@Default` 
`EmbeddingStore` and `EmbeddingModel` beans. With a Quarkus LangChain4j store 
(e.g. `quarkus-langchain4j-pgvector`) none is produced: injecting 
`RetrievalAugmentor` fails with `Unsatisfied dependency for type 
dev.langchain4j.rag.RetrievalAugmentor`, and `@RegisterAiService` silently runs 
without RAG.
   
   Cause: `Langchain4jEmbeddingstoreRagProcessor.BeanCensus` counts beans from 
`BeanDiscoveryFinishedBuildItem`, which by its javadoc excludes synthetic 
beans. All Quarkus LangChain4j store and model beans are synthetic. The 
existing IT passes only because its store is an application-produced 
`InMemoryEmbeddingStore`.
   
   Proposed fix: also count Quarkus LangChain4j's `EmbeddingStoreBuildItem` 
(stores) and `EmbeddingModelProviderCandidateBuildItem` / 
`InProcessEmbeddingBuildItem` (models). `SynthesisFinishedBuildItem` is not an 
option, as the step itself produces a synthetic bean.
   
   Reproducer: the `langchain4j-ingest-rag` example (PR to follow), with 
`DocumentsAugmentor` removed. camel-quarkus main, quarkus-langchain4j 1.13.3.
   
   _Claude Code on behalf of @JiriOndrusek_
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to