Singles (SIN) messages from hsmonitor non-standard for slaves.

[tomkooij]

See also: HiSPARC/datastore#18

Issue: Singles data columns for two detector stations are still set to zero instead of -1. This is non-standard. The columns should be -1 for missing slave data.

HiSPARC/datastore#20 changed the behaviour of the datastore: When no data is reported for the slave, it is set to -1.
Unfortunately, this turns out to be a firmware/DAQ/monitor issue as well. The incoming data always has the slave columns set to some values. (zero if the station has no slave).

Analysis:
The singles data is generated from the one second messages.
These are interpreted by the hs monitor.

Output of the above code for a two detector station (8004):

{'datalist': [{'data': 243, 'data_uploadcode': 'MAS_CH1_LOW'},
    {'data': 115, 'data_uploadcode': 'MAS_CH1_HIGH'},
    {'data': 219, 'data_uploadcode': 'MAS_CH2_LOW'},
    {'data': 99, 'data_uploadcode': 'MAS_CH2_HIGH'},
    {'data': 0, 'data_uploadcode': 'SLV_CH1_LOW'},
    {'data': 0, 'data_uploadcode': 'SLV_CH1_HIGH'},
    {'data': 0, 'data_uploadcode': 'SLV_CH2_LOW'},
    {'data': 0, 'data_uploadcode': 'SLV_CH2_HIGH'}],
   'header': {'date': '2017-02-26',
    'datetime': datetime.datetime(2017, 2, 26, 7, 26, 59),
    'eventtype_uploadcode': 'SIN',
    'nanoseconds': 0,
    'time': '07:26:59'}},

So for all new data recieved from a station, the slave columns will always be zero for a missing slave.

The current status is:

• All data in the datastore before the feb 13st 2017 migration: slave columns set to -1 for missing slaves
• All data that arrived after the migration: slave columns set to 0 for missing slave.

[153957]

And this is how the electronics send data to the DAQ:
http://docs.hisparc.nl/firmware/messages.html#one-second-structure
http://docs.hisparc.nl/firmware/messages.html#threshold-counters-over-threshold-information

Here is the code that used to read the singles values in the HiSPARC Monitor from the DAQ Singles messages:
https://github.com/HiSPARC/station-software/blob/master/user/hsmonitor/HiSPARCSingles.py

Using the following format: >8H, H = unsinged short
That seems to be part of the problem. But perhaps that is because the DAQ sends the data that way.. ask Bob to check the DAQ implementation..