Python SDK
Observability
Manage and observe your R2R system.
Analytics and Observability
R2R provides various tools for analytics and observability to help you monitor and improve the performance of your RAG system.
Completion Scores
Scores act as an observability metric and are attached to RAG completions in R2R.
Scores are bound to values between -1 (a strong negative association) and 1 (a strong positive association), where:
- A score close to 1 may indicate a highly relevant and accurate completion
- A score close to 0 may indicate a neutral or uncertain quality
- A score close to -1 may indicate a potentially irrelevant or incorrect completion
response = client.score_completion("message_id_here", 0.75)
message_id
uuid.UUID
default: "None"requiredThe unique identifier of the completion message to be scored. Found in logs for a message.
score
float
default: "None"requiredThe score to assign to the completion, ranging from -1 to 1.
Was this page helpful?