Introduction
General
Events are the building blocks of the analytics in Laminar. Events are a powerful way to mine metrics from your trace data. There are two types of events:
- Regular events
- Semantic events
Laminar Events are OpenTelemetry-compatible.
Regular events
Regular events are programmatically triggered events.
You can trigger them from your code manually and assign a value to it.
These events are useful when you want to track specific points in your code execution or associate metric values with specific points in your code.
Because of the OpenTelemetry compatibility, the value of the event can be Boolean, String, or Number.
In Laminar’s SDKs, value is optional. If value is not provided, it will be set to true.
This is useful for associating simple tags with spans.
Example of a regular event:
from lmnr import Laminar as L
L.initialize()
L.event("user_reaction", "left_conversation")
Semantic events
The values for semantic events are assigned by LLM. They can be either Boolean, String, or Number Learn more
First, you need to create a pipeline in Laminar that will assess the event. Typically, this will be an LLM or chain of LLMs that assesses whether the event has happened or assigns a value to it.
When you send an evaluate event from the code, the pipeline will be excuted on the Laminar backend, so there is no overhead for your code. An example of a semantic event would be user sentiment detection.