Hi all, I'm using the latest version 4.0.0 with db support. I'm using sqlite but I observe something which might be a bug:
When I run some chains and immediately run evaluation, get_qa_with_reference could not fetch the last trace while get_retrieved_documents could.
start_time = datetime.now()
LangChainInstrumentor().instrument()
OpenAIInstrumentor().instrument()
print("Running experiments to generate traces")
experiment_runner.run_experiment(test_data) # pseudocode.
print("Running LLM evaluation...")
LangChainInstrumentor().uninstrument()
OpenAIInstrumentor().uninstrument()
phoenix_client = px.Client()
qa_spans_df = get_qa_with_reference(phoenix_client, start_time=start_time - timedelta(seconds=30))
retriever_spans_df = get_retrieved_documents(phoenix_client, start_time=start_time - timedelta(seconds=30))
...
I tried to print qa_spans_df and indeed, it does not contain the latest trace.
nothing special. Maybe just starting from a freeform text field for manual annotation would be awesome!
