gilinachum / bedrock-latency Goto Github PK

def consume_openai_stream(response):
    first_byte = None
    stop_reason = None
    for chunk in response:
        if chunk.choices[0].finish_reason is not None:
            if not first_byte: 
                first_byte = time.time() # update the time to first byte
            stop_reason = chunk.choices[0].finish_reason
    return first_byte, stop_reason

def consume_openai_stream(response):
    first_byte = None
    stop_reason = None
    for chunk in response:
        #print(chunk)
        if not first_byte: 
                first_byte = time.time() # update the time to first byte
        if chunk.choices[0].finish_reason is not None:
            stop_reason = chunk.choices[0].finish_reason
    return first_byte, stop_reason