Add ability to disable individual Neuropixels 2.0e probes

[bparks13]

Instead of assuming that because a Neuropixels 2.0 probe is connected to the Neuropixels 2.0e headstage that the user wants to record from it, we add an Enable button to each of the probe settings tabs that controls whether or not the probe is recorded from. This is separate from the Enable button that controls whether or not the whole headstage is recorded from, since we can reset both probes as normal through that route.

We will need to write to an OP_MODE register to tell individual probes not to turn on. This should also be tested to ensure that it doesn't cause any deadlocks or memory buffer issues due to the data not being propagated through the pipeline.

As part of this update, we can also consolidate the tabs. Right now, there is one Neuropixels 2.0 tab, with two probe tabs (A/B) underneath it. Instead, when we can turn individual probes on or off, we can abstract this so that there are two Neuropixels 2.0 probe tabs at the top level. See the Neuropixels 1.0f headstage for reference.

[bparks13]

From Jon:

We can send 1 << 8 to the OP_MODE register for both 1.0 and 2.0 probes and it will shut the whole probe down. However, it is undocumented if doing this will require power cycling the probe to get it out of this power down state, or if it leaves the digital interface alive and is just shutting down the ADCs, etc.

Another option: Send 1 << 6 to REC_MODE to hold the probe's digital interface in reset so that it never talks to the host PC.

[bparks13]

@jonnew before I start working on this one, I wanted to verify which option we wanted to pursue, and if there was any more information available to help us decide. Based on my last comment, it seems like sending 1 << 6 to REC_MODE might be more applicable here, but I wanted to confirm before spending time testing this.

[jonnew]

Let start by trying to use 1 << 7 to OP_MODE since this will result in the most power savings and will cut data off at production rather than some place down stream. I tested this in bonsai and it works. There are secondary problems that have to do with the way that library is implemented. However the answer to the question:

"can we have a different enable for each probe that is tied to 1 << 7 to OP_MODE?"

is yes. We will see if writing 0 << 7 to OP_MODE?" to turn the probe back on causes issues.

[jonnew]

I've made an issue over in the bonsai-onix1 detailing logical issues with the current enable property for neuropixels streams

open-ephys/bonsai-onix1#468

It might be useful