Add langfuse tracing additionally when invoking llm_with_tools and graph

app.py

@@ -9,6 +9,9 @@ from langgraph.checkpoint.memory import MemorySaver
 from langgraph.graph import MessagesState, StateGraph, END
 from langgraph.prebuilt import ToolNode, tools_condition
 from langgraph.prebuilt import ToolNode
+from langfuse import Langfuse, get_client
+from langfuse.langchain import CallbackHandler
+
 import structlog
 
 from qdrant_client import QdrantClient
@@ -21,6 +24,21 @@ from ingestion.main import ingest_document
 
 from config import app_settings
 
+ 
+# Initialize Langfuse client with constructor arguments
+Langfuse(
+    public_key=app_settings.langfuse_public_api_key,
+    secret_key=app_settings.langfuse_secret_api_key,
+    host=app_settings.langfuse_host
+)
+ 
+# Get the configured client instance
+langfuse = get_client()
+ 
+# Initialize the Langfuse handler
+langfuse_handler = CallbackHandler()
+
+
 # Create a logger instance
 logger = structlog.get_logger(__name__)
 
@@ -65,7 +83,7 @@ def retrieve(query: str):
 def query_or_respond(state: MessagesState):
     """Generate tool call for retrieval or respond."""
     llm_with_tools = llm.bind_tools([retrieve])
-    response = llm_with_tools.invoke(state["messages"])
+    response = llm_with_tools.invoke(state["messages"], config={"callbacks": [langfuse_handler]})
     # MessagesState appends messages to state instead of overwriting
     return {"messages": [response]}
 
@@ -116,7 +134,7 @@ def bot(message, history) -> list[Any]:
     
     # create text response
     # TODO: see how state can be set in chat interface
-    config = {"configurable": {"thread_id": "abc123"}}
+    config = {"configurable": {"thread_id": "abc123"}, "callbacks": [langfuse_handler]}
     response = graph.invoke(
         {"messages": [{"role": "user", "content": message.get("text")}]},
         config=config,