Description
[ ] I checked the documentation and related resources and couldn't find an answer to my question.
My Question
https://docs.ragas.io/en/stable/concepts/metrics/available_metrics/context_precision/#llm-based-context-precision
In the docs of Context Precision without reference - it clearly states:
""LLMContextPrecisionWithoutReference metric can be used when you have both retrieved contexts and also reference contexts associated with a user_input.""
However, in the code snippet, only the retrieved_context is used. Instead of the reference_contexts, as suggested in the text, the response is provided.
I'm trying to understand why reference_contexts aren't used here. From my understanding, the correct approach would be to use the reference contexts, as the documentation suggests, instead of the response — especially since we are calculating context recall. To calculate recall, we need a reference to determine which retrieved contexts are correct, so we can say, for example, that x% of the retrieved contexts are relevant.
Why isn't that the case here?