I’m sorry I don’t have some example to share, but I view it as some basics of functional programming. Here’s some pseudo code (caution: I have not executed it) to think about
Enum.reduce(data_list, (:initial_state, nil), &transition/2)
def transition(data_element, {:initial_state, accumulated_data}) do
#do something and return {new_state, new_accumulated_data}
end
def transition(data_element, {:state2, accumulated_data}) do
#do something else and return {new_state, new_accumulated_data}
end
# ... plus as many more transition/2 function clauses you need
Now the result of the Enum.reduce is {final_state, final_accumulated_data} and if you use the async_stream approach I suggested, it will be sent to the calling process automatically. There is no need to store this (in process state or ETS) and later retrieve with some :get_job_result message. That is the whole advantage of the async_stream approach.
