Memify Problems

[LStromann]

=I am having some problems with cognee.memify() method (cognee version 0.5.1).
I ran with different ollama models but in general the problems are the same.
I am running first the methods .add(), .cognify() and then memify().

3 validation errors for RuleSet id Input should be a valid UUID, invalid character: expected an optional prefix of urn:uuid: followed by [0-9a-fA-F-], found u at 1 [type=uuid_parsing,
input_value='uuid:abcd-efgh-ijkl-mno1', input_type=str]

Does anyone have any suggestions?

This discussion was automatically pulled from Discord.

[LStromann]

hey <@1069659243810144386> this looks like an issue with ollama models generating invalid UUIDs during the memify() step.

when memify() runs it uses the llm to generate ruleset objects via structured output. some ollama models occasionally produce UUIDs that don’t match the expected format, which then fails pydantic validation.

a couple things you can try:

• run a larger ollama model likellama3.1:8b or bigger.
• try switching to a cloud LLM provider, which usually handles structured output better.
• double-check your structured output config. make sure this is set in your .env: STRUCTURED_OUTPUT_FRAMEWORK=instructor

please refer to these docs -
https://docs.cognee.ai/setup-configuration/llm-providers#ollama-local
https://docs.cognee.ai/setup-configuration/structured-output-backends

Let us know if you still having and error.

[LStromann]

<@1373135538147692604> Thank you for your help!!! Problem solved!