Add RDF formats to load already formatted data #784
Comments
|
Hey @malte-b! Thanks for the comment. It's nice to see you here. Yes, you are right. At the moment, we don't ingest data purely from Owl. We combine it with textual data to have a grounded knowledge of the text. We only use the owl subgraphs matching the LLM extracted knowledge graph entities or entity types. You need an owl to a knowledge graph parser to do that, or if you want to use LLMs, it's also possible to change the pydantic and extract triplets and the needed information from your owl data. Our low-level pipeline example is similar but with different Pydantic objects. |
|
Thanks for the reply @hajdul88! I see. Is there any reason for why it wouldn't be possible using only OWL? AFAIK in the ontology there can also be descriptions and comments for grounded knowledge. Or do you disregard those fields? |
|
There is no reason not to do it. We just aimed at grounding LLM generated graphs with OWL, and this definitely can be extended and an adapter between OWL and our Pydantic DataPoint model can be created. OWL + LLM helped us solve some customer issues on our end, and we will definitely continue working on it. Realistically, is not very high on our list of priorities at the moment, as we are working on association layers for Agentic memory right now. But, we are open to contributions, and if this is something you'd like to see happen, we are happy support you with the implementation |
My data is already in an RDF format which is knowledge graph ready (Turtle, N-Quads, JSON-LD). I did convert my data already to OWL but as far as I can see, there is no way to just use OWL as a data source, it always needs raw text additionally (see here).
The text was updated successfully, but these errors were encountered: