Add SpikeEventSeries data type

[stephprince]

These changes add the SpikeEventSeries data type for recording threshold crossings during data acquisition. In addition to adding the new data type, these changes check whether the ElectrodesTable already exists on the NWBFile. (Note: what I've written so far assumes that once the electrodes table has been created, all electrodes have been added. I think the ElectrodesTable / DynamicTable classes would require further changes if we want to append to an existing table).

I think some additional tests are needed to check that the code :

• succeeds for different dimensions of the SpikeEventSeries ([nEvents x nSamples], [nEvents x nChannels x nSamples])
• succeeds if user makes file with both spike event series + electrical series

[oruebel]

Overall this looks good. The main question I have is how we can prevent duplication of 1) channels in the ElectrodeTable, 2) ElectrodeGroups , 3) Devices

[stephprince]

For duplication of channels in the ElectrodesTable, my first thought is to:

1. In ElectrodeTable->addElectrodes(), add a check using the global index/rowIds to only add channels not yet included
2. In ElectrodeTable->finalize(), if the columns already exist, use an addRow method to append additional channels, otherwise create the columns with all the data.

Thoughts on that approach?

[stephprince]

One thing I realized when adding tests for a file that has both ElectricalSeries and SpikeEventSeries was that we needed to have a way to specify that data came from the same source (e.g. the same device) but had different names within the acquisition group. I added this sourceName as a separate input to the Channel class as a way to accommodate this.

[oruebel]

Is sourceName the same as the name of the device?

[stephprince]

This is now updated so that groupName is used for the ElectrodeGroup and Device and is pulled from the channelVector input. The name of the dataset within the acquisition group is set by a recordingArrayNames input to createElectricalSeries and createSpikeEventSeries.

[stephprince]

Added tests and now also have it integrated with OpenEphys - see here for the latest version of that branch.

[oruebel]

For duplication of channels in the ElectrodesTable, my first thought is to:

1. In `ElectrodeTable->addElectrodes()`, add a check using the global index/rowIds to only add channels not yet included

2. In `ElectrodeTable->finalize()`, if the columns already exist, use an `addRow` method to append additional channels, otherwise create the columns with all the data.

Thoughts on that approach?

Seems reasonable. I think 2. assumes that there are no custom columns added to the table. We may want to at least check that we have data for all columns

[stephprince]

@oruebel this is ready for you to review again. I've created issues for what we want to address outside this PR

[stephprince]

Will fix #30

[oruebel]

For now this seems fine. I assume this is part of #93 to address the groupName?

[stephprince]

By address the groupName do you mean allowing different options for Device and ElectrodeGroup?

From what I had in my notes we also discussed having a single groupName for a channelVector. That would be part of #37

[oruebel]

Makes sense.