Add AMSMonitor interface and unify both RMQ API (#32)

[lpottier]

AMSMonitor helps us monitor each stage of the python pipeline with simple decorator approach. Each call to AMSMonitor decorator or context manager populate a shared dictionary that track each call.

In addition of recording (by default, can be restricted with @AMSMonitor(record=["attr1", "attr2"])) all attributes of a class, AMSMonitor will also inject its own metrics, for example the duration of each block amsmonitor_duration. For example, the following code:

class ExampleTask1(Task):
  def __init__(self):
    self.total_bytes = 0
  @AMSMonitor()
  def __call__(self):
    i = 0
    with AMSMonitor(obj=self, tag="while_loop"):
      while (i<3):
        self.total_bytes += 10
        time.sleep(1)
        i += 1
    self.total_bytes += 5

will produce the following dictionary (dumped as JSON here):

{
  "ExampleTask1": {
    "while_loop": {
      "02/29/2024-19:27:53": {
        "total_bytes": 30,
        "amsmonitor_duration": 3.004607439041138
      }
    },
    "__call__": {
      "02/29/2024-19:29:24": {
        "total_bytes": 35,
        "amsmonitor_duration": 3.10461138
      }
    }
  }
}

This PR also addresses #32 and unify both RMQ API.

[koparasy]

LGTM. I have some minor comments.