Merge pull request #4939 from BerriAI/litellm_log_transcription_resp_langfuse

[Feat-Proxy] - Langfuse log /audio/transcription on langfuse

litellm/integrations/langfuse.py

@@ -144,6 +144,10 @@ class LangFuseLogger:
                 f"Langfuse Logging - Enters logging function for model {kwargs}"
             )
 
+            # set default values for input/output for langfuse logging
+            input = None
+            output = None
+
             litellm_params = kwargs.get("litellm_params", {})
             litellm_call_id = kwargs.get("litellm_call_id", None)
             metadata = (
@@ -198,6 +202,11 @@ class LangFuseLogger:
             ):
                 input = prompt
                 output = response_obj["data"]
+            elif response_obj is not None and isinstance(
+                response_obj, litellm.TranscriptionResponse
+            ):
+                input = prompt
+                output = response_obj["text"]
             print_verbose(f"OUTPUT IN LANGFUSE: {output}; original: {response_obj}")
             trace_id = None
             generation_id = None

litellm/proxy/common_utils/http_parsing_utils.py

@@ -56,6 +56,9 @@ def check_file_size_under_limit(
 
     file_contents_size = file.size or 0
     file_content_size_in_mb = file_contents_size / (1024 * 1024)
+    if "metadata" not in request_data:
+        request_data["metadata"] = {}
+    request_data["metadata"]["file_size_in_mb"] = file_content_size_in_mb
     max_file_size_mb = None
 
     if llm_router is not None and request_data["model"] in router_model_names:

litellm/proxy/proxy_config.yaml

@@ -34,4 +34,4 @@ general_settings:
   max_response_size_mb: 10
 
 litellm_settings:
-  callbacks: ["otel"]
+  success_callback: ["langfuse"]

litellm/tests/test_alangfuse.py

@@ -245,6 +245,49 @@ async def test_langfuse_logging_without_request_response(stream, langfuse_client
         pytest.fail(f"An exception occurred - {e}")
 
 
+# Get the current directory of the file being run
+pwd = os.path.dirname(os.path.realpath(__file__))
+print(pwd)
+
+file_path = os.path.join(pwd, "gettysburg.wav")
+
+audio_file = open(file_path, "rb")
+
+
+@pytest.mark.asyncio
+async def test_langfuse_logging_audio_transcriptions(langfuse_client):
+    """
+    Test that creates a trace with masked input and output
+    """
+    import uuid
+
+    _unique_trace_name = f"litellm-test-{str(uuid.uuid4())}"
+    litellm.set_verbose = True
+    litellm.success_callback = ["langfuse"]
+    await litellm.atranscription(
+        model="whisper-1",
+        file=audio_file,
+        metadata={
+            "trace_id": _unique_trace_name,
+        },
+    )
+
+    langfuse_client.flush()
+    await asyncio.sleep(2)
+
+    # get trace with _unique_trace_name
+    trace = langfuse_client.get_trace(id=_unique_trace_name)
+    generations = list(
+        reversed(langfuse_client.get_generations(trace_id=_unique_trace_name).data)
+    )
+
+    print("generations for given trace=", generations)
+
+    assert len(generations) == 1
+    assert generations[0].name == "litellm-atranscription"
+    assert generations[0].output is not None
+
+
 @pytest.mark.asyncio
 async def test_langfuse_masked_input_output(langfuse_client):
     """