search

phidatahq / phidata

Build AI Assistants with memory, knowledge and tools.
https://docs.phidata.com
Mozilla Public License 2.0
10.44k stars 1.51k forks source link

knowledge_base.load() function issue with WebsiteKnowledgeBase #1011

Open DanielChico opened 2 weeks ago

DanielChico commented 2 weeks ago

Hello. I noticed an inconsistency in the code implementation for the WebsiteKnowledgeBase class. This image shows de base load function image And this is the implementation for WebsiteKnowledgeBase image It seems to be using the recreate variable for both recreating and skipping documents

ysolanky commented 2 weeks ago

@DanielChico You're right, the implementation of the load function of WebsiteKnowledgeBase is different from the Base class. But the logic remains the same

DanielChico commented 2 weeks ago

@DanielChico You're right, the implementation of the load function of WebsiteKnowledgeBase is different from the Base class. But the logic remains the same

But is not using the "skip_existing" parameter to skip existing urls instead is using "reacreate" parameter i think this behavior is wrong. For instance, if i don't want to recreate the database table but i want to skip existing urls is not possible.