jiriOndrusek-agent commented on PR #27289: URL: https://github.com/apache/camel/pull/27289#issuecomment-5990008796
Thanks for the review @davsclaus. Addressed in 8740a78dd819, kept as a separate commit. **Fixed** - `documentFilter` description now says the predicate sees the body as the consumer delivered it, not yet read as text or bytes. - `maxDocumentSize` description now reads "characters of the text about to be split, or bytes of a media body with modality=media"; both setter javadocs updated. JSON, catalog and DSL regenerated. **Suggestions taken** - Size check before the read: with `modality=media` the length a file consumer announces in `CamelFileLength` is checked against `maxDocumentSize` before the file is read; any other body is checked once in memory. Covered by a test with an unreadable body. - The placeholder segment now carries the MIME type under `camel_ingest_content_type` (media segments only), so a retriever on a shared store can tell a media hit from text. - Listener class-name check: follow-up [CAMEL-25332](https://issues.apache.org/jira/browse/CAMEL-25332), referenced from the TODO next to the check. **Questions** - Automatic listeners: neither quarkus-langchain4j nor langchain4j-spring references `EmbeddingModelListener` today; their listener wiring is for chat models. Added a sentence to the doc; the opt-out is a model configured without listeners. - `image/svg+xml`: intended. The MIME family decides the medium and the provider decides which formats within it, so Camel keeps no per-provider format list. Documented in the media section. _Written with Claude Code on behalf of Jiří Ondrušek._ -- 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]
