spiketag output labelled spike data for both decoder and FPGA

[chongxi]

spiketag output the labelled spike data to two different places:

1. The FPGA: vector quantization sets with their spike_id labels
2. The decoder: spike times with their spike_id labels

To be finished in this issue:

• [x] fpga.n_units as a variable store at the FPGA for different parts of the software to get access to this information and it is set every time the new VQ model is downloaded into the FPGA in ctrl.done().
• [x] Each group of VQ starts from label 0 to its unique label space which accumulates as the group number grows in a non-overlapping manner.
• [x] The VQ labels in FPGA and the spike trains output, share the same label.
• [x] The direct output spike count vector (ctrl.spk_count_vector()) from spiketag to its decoder training use the label that the BMI and the binner use (the VQ labels)

[chongxi]

For FPGA, check

ctrl.vq['labels'][group_id]
ctrl.fpga.label[group_id]

[chongxi]

The group-unit relationship:

Adding together is 102 units, their label is from 0-101

[chongxi]

The group done:

[chongxi]

[chongxi]

... ... ...

[chongxi]

Make spike times numpy array

@property
def spk_times_all_in_one_array(self):
    return np.array(list(self.spk_times_all_in_one.values()))

[chongxi]