⚠️ Deprecation of Knowledge Graphs and Text2SparqlRetriever

[bilgeyucel]

Starting from version 1.15, BaseKnowledgeGraph, GraphDBKnowledgeGraph, InMemoryKnowledgeGraph, and Text2SparqlRetriever are being deprecated and will be removed from Haystack as of version 1.17.

How to use Knowledge Graphs?

If you're currently using older versions of Haystack (< v1.17), you don't need to make any changes and can continue to use them as before. Keep in mind that you will get a DeprecationWarning when you use these components starting from version 1.15.

If you want to use Haystack v1.17 and onwards, we recommend creating custom implementations to load and query knowledge graphs. See an example of knowledge graph functionality for a QA system with Haystack here.

Why are we doing this?

Haystack allows storing and querying knowledge graphs with the help of pre-trained Text2Sparql models that translate text queries to SPARQL queries. However, knowledge graphs have inherent limitations. The trained models can only generate SPARQL queries for the data included in the training set, and they cannot generalize beyond that. Additionally, Haystack does not provide tooling to train Text2Sparql models with custom data, making these components less practical to use and maintain with Haystack. For these reasons, we have decided to deprecate the functionality of knowledge graphs and Text2SparqlRetriever in Haystack.

Let us know 🙌

If you have any questions or comments regarding this deprecation, feel free to comment on this post or send a message to the #haystack channel on Discord.

[mariosconsta]

Hello, are there any plans on bringing Knowledge Graphs back to Haystack? Sorry for reviving this old discussion!

[bilgeyucel]

Hi @mariosconsta, thanks for bringing this up!
neo4j integration enables you to use graphs with Haystack. Are you looking for something extra or more specific?

[mariosconsta]

@bilgeyucel Oh neat, I didnt know that! Thank you for letting me know and sorry for the late response :D