🔨 update langfuse.mdc rule to reflect v3.x API migration

Author: Miyamura80

.cursor/rules/langfuse.mdc

@@ -3,17 +3,28 @@ description:
 globs: *.py
 alwaysApply: false
 ---
-LangFuse is an LLM observability tool, to see how LLMs are behaving in our application. It uses callbacks to reord LLM inference. You should always use it when using LLM applications, to track if the LLM is behaving as we want it to.
+Langfuse is an LLM observability tool used to monitor LLM behavior in our application. It uses decorators and callbacks to record LLM inference. Always use it in LLM-related modules to ensure traceability.
 
-Below is the most typical usage.
+### Langfuse v3 Migration Guide
+As of Langfuse SDK v3.x, the `langfuse.decorators` module structure has changed.
+
+**Typical Usage:**
 
 ```python
-from langfuse.decorators import observe, langfuse_context
+from langfuse import observe, get_client
 
-@observe
+@observe()
 def function_name(...):
+    # To rename the span/observation or update metadata:
+    get_client().update_current_span(name=f"descriptive_name_{id}")
 
-    # If we don't want the name of the span to be "function_name", and want to give a more descriptive name like with email, we should use the following
-    langfuse_context.update_current_observation(name=f"{email}")
+    # To get trace or observation IDs:
+    trace_id = get_client().get_current_trace_id()
+    observation_id = get_client().get_current_observation_id()
 ```
 
+**Key Changes from v2:**
+1. Import `observe` and `get_client` directly from `langfuse`.
+2. Replace `langfuse_context` with `get_client()`.
+3. Use `update_current_span()` instead of `update_current_observation()`.
+