[Feature Request] Promptflow can't save some node output when exception happens

[vhan2kpmg]

Is your feature request related to a problem? Please describe. At the moment when exception happens from run as a function, for example, content management policy exception triggered all nodes output are lost:

f = load_flow(source="../../examples/flows/chat/chat-basic/")
f.context.streaming = True
try:
  result = f(
      chat_history=[
          {
              "inputs": {"chat_input": "Hi"},
              "outputs": {"chat_output": "Hello! How can I assist you today?"},
          }
      ],
      question="How are you?",
  )
except WrappedOpenAIError as exc:
  # we lost succeeded node output as well
  pass

answer = ""
# the result will be a generator, iterate it to get the result
for r in result["answer"]:
    answer += r

Describe the solution you'd like when exception happens can we have some nodes output saved somewhere? for example: node1->node2->node3 when node2 triggers exception we can still have output from node1?

Describe alternatives you've considered Not sure

Additional context Not sure

[D-W-]

Hi @vhan2kpmg , we currently model flow-as-function feature like a function call. Which means when calling the flow function, no intermedia data will be persisted. We won't change its behavior in near future. I've added a long term tag to track this.