Adapt to new prescient plugin system

[xiangao1]

Changes in coordinator.py:

1. Register the customized command-line options using Prescient's new plugin system.
2. Register the callback functions using Prescient's new plugin system.

Changes in thermal_generator_prescient_plugin.py:

1. Demonstrate an example with the new plugin system.

[bknueven]

plugin_config.bidding_generator -- I would refactor this class slightly and store the plugin_config passed into register_plugins somewhere. You probably also don't need the _register_*_callbacks methods anymore.

The declare_as_argument method just has to do with how the command line option is created. Using the key in the name is optional.

[xiangao1]

Thanks Ben! I just pushed a small change to store the plugin_config as a property. I think I should get rid of the bidding_generator option, because the coordinator should get this information from both bidder and tracker.

The _register_*_callbacks methods make it a little easier for me to manage the plugin functions, in case we have more than one callbacks at some plugin points.

[bknueven]

Note that large Prescient PR also enabled one to run Prescient directly from Python, see: https://github.com/grid-parity-exchange/Prescient/blob/06c8b1d9416baf1ef85cfa880935674db25487cd/tests/simulator_tests/test_sim_rts_mod.py#L138-L198.

So it's possible to create an overall "driver" which runs Prescient along with the configured classes.

[xiangao1]

Note that large Prescient PR also enabled one to run Prescient directly from Python, see: https://github.com/grid-parity-exchange/Prescient/blob/06c8b1d9416baf1ef85cfa880935674db25487cd/tests/simulator_tests/test_sim_rts_mod.py#L138-L198.

So it's possible to create an overall "driver" which runs Prescient along with the configured classes.

Now my example uses the Python API.

[bknueven]

Now my example uses the Python API.

Excellent! Related, though I'm not sure when/where you want to include it, is the new shortcut simulator: grid-parity-exchange/Prescient#113, which just needs a single generator and two LMP time series (one for DA, and another for RT).

[xiangao1]

Now my example uses the Python API.

Excellent! Related, though I'm not sure when/where you want to include it, is the new shortcut simulator: grid-parity-exchange/Prescient#113, which just needs a single generator and two LMP time series (one for DA, and another for RT).

Perfect! I'll have to talk to @adowling2 and @nicolepcortes about the shortcut simulator.

[adowling2]

@bknueven, thanks for the update. @nicolepcortes will reach out to you about the shortcut simulator soon.