Updating plugins using nidaqmx and not PyDAQmx

[jerlfan]

Hi guys,
I am actually working on the update of the plugin to replace PyDAQmx by nidaqmx (from NI).
The current state of the work can be find on my fork : https://github.com/jerlfan/pymodaq_plugins_daqmx
I created daqmx-ni wrapper using nidaqmx (in hardware) to replace daqmx that uses PyDAQmx
If you can have a look on what I have done at this stage, I would be grateful. It is not fuly tested I have to implement the reading methods (see below)

To finish the work I am actually facing one problem:
reading procedure is much simpler with nidaqmx (no needs to use ctype, or byref...): i.e. task.read(Nsamples, Readout time) and it is independant from the channel type. The time is defined elsewhere (task.timing).
I will start by implementing the counter channels and I would your advices.
There is two ways to use it:
1- we feed a clock to the counter. Usually you use another counter channel for that (Aurore's way). It is really good for short time but you loose a counter channel :-( (at least on the NI board we have in the Lab, I have a USB6341).
2- You can simply use a "machine clock". This is where I may need help. I propose the following solution:
init=time.perf_counter()
task.start()
while time.perf_counter()-init <=Acquisition_time:
pass
counts=task.read()
print(counts)
task.stop()

What do you think? I think it is not really elegant. So sugestions are welcomed!

I will then create a daq_viewer_0D counter-fast and counter-slow.

Cheers

[seb5g]

concerning the timing, you can use a timer:
https://docs.python.org/fr/3/library/threading.html?highlight=timer#threading.Timer
this will call a callback (function) when the timer has finished its count down.
Otherwise I would suggest you put you functionnality into a separated thread but then it is probably the same as the threading.Timer class

[jerlfan]

Thanks,

threading.Timer seems indeed to be exactly what I want. However, I don't understand where I can find the count_value returned by the function I used:

def read():
return task.read()
t=threadig.Timer(2,read)
t.start()

From this code I have no output. There is probably something I miss.

[seb5g]

Well for sure your async function won't return directly the counts, but the timer callback (self._ready) will be the one reading the counts and you can notify the user using two methods: polling the self.ready boolean or asking it to provide a method signature (callback) to notify a external program (for instance a pymodaq plugin).

class blabla:
    def __init__(self):
        self.task = Task(...)
        self.ready = False
        self.counts: int = None
        self.callback = None

    def _ready(self):
        self._ready = True
        self.counts = self.task.read()
        if self.callback is not None:
            self.callback(self.counts)

    def get_counts(self, dtime: float, callback = None):
        self.ready = False
        self.callback = callback
        t = threading.Timer(dtime, self._ready)
        t.start()

If you want a blocking function, then just write what you did before maybe inserting a sleep time in the while loop

[loicguilmard]

hello @jerlfan
You should check the last PR : nidaqmx PR
There is no counter but there is task management with the NI python librairy and scalable GroupParameter, and importing toml to GroupParameter in the pluging ParameterTree.