Move arming the eiger out of stage

[DominicOram]

As documented on the bluesky wiki stage is not expected to take a meaningful amount of time. Bluesky therefore waits on any stage being complete before continuing. Arming the detector takes ~6s and we currently do it in the stage, we should move it out of the stage so that we can do other things concurrently.

Acceptance Criteria

• Staging the eiger takes little time
• Starting to arm the detector can be done in parrallel with other tasks
• The staging/arming are done as soon as possible in the plan

[bentom08]

Just to note here - if we move the detector arm outside the stage method, the rest of the Eiger stage will still need to be completed before we start arming. If we want to save time, at best we can start staging the Eiger immediately so we can begin arming as soon as possible as well.

[DominicOram]

Yes, I think we should do the following:

• Split the stage in the eiger into two:
  1. Setting up for the arm, e.g. check stale params etc. (everything that needs to be done before the arm)
  2. Doing the arm (can probably stay in the stage)
• We should then do 1. concurrently as that takes more time compared to 2.
• We need to make sure we're waiting on 1. to finish before we actually do the stage though

[evalott100]

I wasn't sure if the seperation in the above commit is what we're looking for. The arm is the only thing in the stage function, everything before is contained in the new pre_arm function, which is ran in a new thread in the fast grid scan plan

[DominicOram]

I'm going to move this back into To Do as I think it a chat as a group and some further thought

[DominicOram]

Suggestion from @callumforrester stage/unstage should arm/disarm if not already in that state.

[DominicOram]

Odin currently assumes that frames are coming in very soon after arming. This could be a bad assumption.