add langfuse integration

litellm/integrations/langfuse.py

@@ -0,0 +1,58 @@
+#### What this does ####
+#    On success, logs events to Langfuse
+import dotenv, os
+import requests
+import requests
+from datetime import datetime
+
+dotenv.load_dotenv()  # Loading env variables using dotenv
+import traceback
+
+class LangFuseLogger:
+    # Class variables or attributes
+    def __init__(self):
+        from langfuse import Langfuse
+        # Instance variables
+        self.secret_key = os.getenv("LANGFUSE_SECRET_KEY")
+        self.public_key = os.getenv("LANGFUSE_PUBLIC_KEY")
+        self.Langfuse =  Langfuse(
+                        public_key=self.public_key,
+                        secret_key=self.secret_key,
+                        host="https://cloud.langfuse.com",
+                        # debug=True
+                    )
+
+    def log_event(self, kwargs, response_obj, start_time, end_time, print_verbose):
+        # Method definition
+        from langfuse.model import InitialGeneration, Usage
+        try:
+            print_verbose(
+                f"Langfuse Logging - Enters logging function for model {kwargs}"
+            )
+            # print(response_obj)
+            # print(response_obj['choices'][0]['message']['content'])
+            # print(response_obj['usage']['prompt_tokens'])
+            # print(response_obj['usage']['completion_tokens'])
+
+            generationStartTime = datetime.now()
+            self.Langfuse.generation(InitialGeneration(
+                name="litellm-completion",
+                startTime=start_time,
+                endTime=end_time,
+                model=kwargs['model'],
+                # modelParameters= kwargs,
+                prompt=[kwargs['messages']],
+                completion=response_obj['choices'][0]['message']['content'],
+                usage=Usage(
+                    prompt_tokens=response_obj['usage']['prompt_tokens'],
+                    completion_tokens=response_obj['usage']['completion_tokens']
+                ),
+            ))
+
+            print_verbose(
+                f"Langfuse Layer Logging - final response object: {response_obj}"
+            )
+        except:
+            # traceback.print_exc()
+            print_verbose(f"Langfuse Layer Error - {traceback.format_exc()}")
+            pass

litellm/tests/test_promptlayer_integration.py

@@ -28,4 +28,20 @@
 #     except Exception as e:
 #         print(e)
 
-# # test_chat_openai()
+
+
+# def test_chat_openai():
+#     litellm.success_callback = ["langfuse"]
+#     try:
+#         response = completion(model="gpt-3.5-turbo",
+#                               messages=[{
+#                                   "role": "user",
+#                                   "content": "Hi 👋 - i'm openai"
+#                               }])
+
+#         print(response)
+
+#     except Exception as e:
+#         print(e)
+
+# test_chat_openai()

litellm/utils.py

@@ -16,6 +16,7 @@ from .integrations.berrispend import BerriSpendLogger
 from .integrations.supabase import Supabase
 from .integrations.llmonitor import LLMonitorLogger
 from .integrations.prompt_layer import PromptLayerLogger
+from .integrations.langfuse import LangFuseLogger
 from .integrations.litedebugger import LiteDebugger
 from openai.error import OpenAIError as OriginalError
 from openai.openai_object import OpenAIObject
@@ -41,6 +42,7 @@ slack_app = None
 alerts_channel = None
 heliconeLogger = None
 promptLayerLogger = None
+langFuseLogger = None
 llmonitorLogger = None
 aispendLogger = None
 berrispendLogger = None
@@ -837,7 +839,7 @@ def load_test_model(
 
 
 def set_callbacks(callback_list):
-    global sentry_sdk_instance, capture_exception, add_breadcrumb, posthog, slack_app, alerts_channel, traceloopLogger, heliconeLogger, aispendLogger, berrispendLogger, supabaseClient, liteDebuggerClient, llmonitorLogger, promptLayerLogger
+    global sentry_sdk_instance, capture_exception, add_breadcrumb, posthog, slack_app, alerts_channel, traceloopLogger, heliconeLogger, aispendLogger, berrispendLogger, supabaseClient, liteDebuggerClient, llmonitorLogger, promptLayerLogger, langFuseLogger
     try:
         for callback in callback_list:
             print_verbose(f"callback: {callback}")
@@ -898,6 +900,8 @@ def set_callbacks(callback_list):
                 llmonitorLogger = LLMonitorLogger()
             elif callback == "promptlayer":
                 promptLayerLogger = PromptLayerLogger()
+            elif callback == "langfuse":
+                langFuseLogger = LangFuseLogger()
             elif callback == "aispend":
                 aispendLogger = AISpendLogger()
             elif callback == "berrispend":
@@ -1165,6 +1169,16 @@ def handle_success(args, kwargs, result, start_time, end_time):
                         end_time=end_time,
                         print_verbose=print_verbose,
                     )
+                elif callback == "langfuse":
+                    print_verbose("reaches langfuse for logging!")
+                    langFuseLogger.log_event(
+                        kwargs=kwargs,
+                        response_obj=result,
+                        start_time=start_time,
+                        end_time=end_time,
+                        print_verbose=print_verbose,
+                    )
+
                 elif callback == "aispend":
                     print_verbose("reaches aispend for logging!")
                     model = args[0] if len(args) > 0 else kwargs["model"]