[Feature request] support Dispatcher initialization from partial Schedule

[BrannonKing]

Feature request: allow me to give the dispatch solver a partial solution and have it finish it. Or is this already supported?

[Pabloo22]

Hello Brannon,
I'm unsure if you are referring to the Dispatcher class or the DispatchingRuleSolver.

The DispatchingRuleSolver already supports starting from partial solutions since both methods (step and solve) can receive an already initialized dispatcher as an argument. As a side note, this arg is optional in the solve method and was introduced in version 0.3.0.

The Dispatcher class doesn't support being initialized from a partially completed Schedule yet, I could add that feature in the new version if this is what you are asking for.

[BrannonKing]

The Dispatcher class doesn't support being initialized from a partially completed Schedule yet, I could add that feature in the new version if this is what you are asking for.

Yes, that's what I was wanting. It looks like it has an add method. Maybe the thing to do is just call the add method for my existing tasks in order.

[Pabloo22]

Yes, that's how I would implement it. You should just check that the operation is ready to be scheduled before. See Schedule.from_job_sequences as a reference. Internally it uses the Dispatcher class and does something similar to this using deques.

[dogukansert]

Can I see operation IDs or operation information on the gantt chart I created?

[Pabloo22]

@dogukansert No, currently there isn't an option for that. It might be possible to add it post hoc since the function returns both the matplotlib figure and axes. However, modifying the source code directly might be easier, and it would be a great feature if you are interested in contributing.

That said, this question is unrelated to the original topic. I opened a new issue to discuss this more (see #13).