lordrip commented on PR #19261: URL: https://github.com/apache/camel/pull/19261#issuecomment-3311778266
> [ibm-granite published docling on hugging face](https://huggingface.co/ibm-granite/granite-docling-258M), and they are going to create an ollama model too. Instead of creating a docling component that uses the python docling behind the hood, would it make sense to create a specialized Kamelet? In particular, I think that we can reuse the camel-langchain4j component to interact with the ollama model. The Inputs to the Kamelet should be the images and a text and the camel-langchain4j component should be configured accordingly. Thanks for the heads up @Croway, I think I agree with @oscerd in the sense that it's ok for the component to depend on docling directly (_I hope there's a configuration to specify the binary location somehow, just in case_) and then having something else like a Kamelet to handle it in a more generic way as you said, I like that idea. I wonder if we can create a Kamelet that interacts with `ollama` in a programmatic way, it could behave like Camel bindings for ollama, so as a user, instead of using ollama's REST API directly, we could use ollama through a Kamelet, so we give Camel access to ollama. Note: Maybe this is already possible with langchain4j components and I don't know about it :smile: -- 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]
