Would like Markov state machine generator

[tdunning]

Sometimes we would like to have transaction histories that represent plausible user actions. The simplest might be {login, do stuff, logout}* where you can only do stuff while logged in.

One simple way to simulate something like this is to have a state machine where transitions from state to state are selected stochastically based on specified transition probabilities.

The suggested syntax for the schema for such a generator would be something like this:

{
    "name": "history",
    "class": "markov-state-machine",
    "transitions": {
        "init": {"start":1}
        "start": {
            "abort": 0.1,
            "progress": 0.8,
            "finish": 0.1
        },
        "abort": {
            "start": 1
        },
        "progress": {
            "abort": 0.1,
            "progress": 0.8,
            "finish": 0.1
        },
        "finish": {
            "start": 1
        }
    }
}

Of course, the transition probabilities may not add up to 1 so they should be normalized.

The sample result would be a list of transactions much like the common-point-of-compromise generator produces.