How to feed callback_data parameter to the callback passed as parameter of register_every_n_samples_acquired_into_buffer_event (in nidaqmx-python)

[ft_06]

Hello,

I wish to pass some data to my callback through callback_data, which seems to be custom info that you can pass (I used to do similar things in C 20 years ago). But I did not get how you give it to the API

In the doc: "@def callback(task_handle, every_n_samples_event_type, number_of_samples, callback_data): ... The callback_data parameter contains the value you passed in the callback_data parameter of this function"

Well, not a native english speaker but I am quite puzzled by the sentence. I see no "callback_data" parameter in any nidaqmx API.

Within the generated bindings to C, there are some "register_every_n_samples_event" where you could pass a callback and callback_data but this is internal. Is it just that the API is available in C and not in python ? 

OK, I have a fallback solution where I put every info as global but well, if I could get a bit cleaner...

[ft_06]

I had a look at C API

nt32 DAQmxRegisterEveryNSamplesEvent (TaskHandle taskHandle, int32 everyNsamplesEventType, uInt32 nSamples, uInt32 options, DAQmxEveryNSamplesEventCallbackPtr callbackFunction, void *callbackData);

In python: register_every_n_samples_acquired_into_buffer_event(sample_interval, callback_method)

As parameter, there is the callback... and the Callback_data stuff that you want to be given as parameter of your callback.

I don't see how this is translated in python

[ft_06]

Looking at internal code of ndaqmx-python:

def register_every_n_samples_acquired_into_buffer_event(self, sample_interval, callback_method):

    if callback_method is not None:

        event_handler = self._interpreter.register_every_n_samples_event(self._handle, EveryNSamplesEventType.ACQUIRED_INTO_BUFFER.value, sample_interval, 0, callback_method, None)

Cough, cough, .... callback_data gets discarded. I fear I have my answer

[pieleric · Accepted solution]

I also need to pass some info to the callback, in Python. Eventually, I've used "partial" to do that. Something like:

on_data_consumed = functools.partial(self._on_data_consumed, continuous=continuous)
ao_task.register_every_n_samples_transferred_from_buffer_event(ao_chunk_size, on_data_consumed)

[ft_06]

Right, I had seen something similar in NI code in their acceptance test for multi-threading. pyqtchart uses lambda functions to also "customize" their callbacks, I guess this would also work.

This is kind of "backup" solution but if NI is also doing this, let's consider this is the solution

code from NI:

for i in range(len(tasks)):

  def _every_n_samples_callback(
    i: int,
    task_handle: object,
...

) -> int:...

tasks[i].register_every_n_samples_acquired_into_buffer_event(

100, functools.partial(_every_n_samples_callback, i)
)