Recording already finished, but want to test beamlime.

[YooSunYoung]

As explained in the #239, there are different scenarios of run-start/run-stop status.

2, 3 will be simple to implement and it'll support most cases for users
but 1 should be implemented first since it'll enable the fast/quick tests with the existing records.

1. Listener should find the latest run-start message from the topic and retrieve the timestamp and the job-id.
2. Listener should find set the time-offset for finding the run-stop message and match it with the job-id.
3. Listener should start consuming the event/log data between those two time-stamps

Here is a script that quickly check last 10 messages in coda.

import os

def test():
    from rich import print
    from confluent_kafka import Consumer
    from streaming_data_types.run_stop_6s4t import deserialise_6s4t
    from streaming_data_types.run_start_pl72 import deserialise_pl72

    consumer = Consumer(
        {
            "bootstrap.servers": f"{os.environ['KAFKA_SERVER']}:{os.environ['KAFKA_PORT']}",
            "group.id": "beamlime",
            "auto.offset.reset": "earliest",
            "enable.auto.commit": False,  # Not to leave commit of this testing-group.
        }
    )
    try:
        consumer.subscribe(["ess_filewriter_pool"])
        for _ in range(10):
            msg = consumer.poll(timeout=2.0)
            print(msg)
            print(msg.value()[4:8].decode())
            pl72 = deserialise_pl72(msg.value())
            print(pl72.instrument_name)
    except Exception as e:
        consumer.close()
        raise e
    else:
        consumer.close()

if __name__ == "__main__":
    test()