Overview

This Python package provides utilities to work with National Instruments DAQs on top of the PyDAQmx package. Utilities include scanning the system for attached DAQs and the ability to start and control an analog input acquisition in another python process (server) on the local or a remote machine and return the acquired data (this is useful when one wants to work in a 64 bit environment because the NI drivers can only be interfaced from a 32 bit environment, which in this package is isolated to another process) to the client.

Warning

When running the server and clients on different machines, the communication connection between the two over a socket is not encrypted nor is any sort of authentication used. Thus, this feature should only be used on a trusted private network.

The package's source code is found at https://github.com/frejanordsiek/pynidaqutils

Installation

This package is supported for Python version 2.6 and later.

This package requires the numpy package for both the server and the client, the PyDAQmx package on the server, and the paramiko package on the client if the client and the server are on different machines (remote acquisition).

Warning

The server cannot operate on a machine that doesn't have niDAQmxBase installed and the lsdaq in the executable path. niDAQmx, the default on Windows, will not work since it does not have the lsdaq utility. Note that the PyDAQmx package must be one with niDAQmxBase support, so it must be a version greater than or equal to 1.3. This is not listed in the dependencies because only the acquisition server in pynidaqutils.analog_input needs it. The client doesn't.

To install pynidaqutils, download the package and run the command:

pythonX setup.py install

where X is the particular python you want to install it for.

Getting Started

First, import the modules:

>>> import pynidaqutils
>>> import pynidaqutils.analog_input

The :py:mod:`pynidaqutils` module had hardware information for various National Instrument DAQ's as well as a function that returns which DAQ's are attached to the computer:

>>> pynidaqutils.list_daqs()
{'Dev1': {'hw': 'USB0::0x3923::0x7270::0186F411::RAW', 'type': 'NI USB-6211'}}

The easiest way to do analog input is to make a :py:class:`pynidaqutils.analog_input.DaqInterface`, which can start a server (:py:class:`pynidaqutils.analog_input.DaqServer`) and a client (:py:class:`pynidaqutils.analog_input.DaqClient`) connected to it as well as stopping them:

>>> di = pynidaqutils.analog_input.DaqInterface()

When starting a server, the location, or host, of where it is to run and what is the name of the command to run python called must be specified. To run on the same computer and say run it on python3, one would do:

>>> di.start_server(host='localhost', python_command='python3')
{'PyDAQmx': '1.2.5.2', 'paramiko': '1.14.0', 'pynidaqutils': '0.2'}

and it then returns the package versions of pynidaqutils, PyDAQmx, and paramiko on the server. To instead run it on another machine, host must be set to the IP address of the other machine. That machine must be running an SSH server on port 22 and the paramiko package is needed to connect to it to start the server. The username and password to log into the server by SSH must also be provided:

>>> di.start_server(host='192.168.1.100', python_command='python3'
...                 username='daq', password=passwd)
{'PyDAQmx': '1.2.5.2', 'paramiko': '1.14.0', 'pynidaqutils': '0.2'}

where passwd holds the password.

A convenience function is provided that will make a client (:py:class:`pynidaqutils.analog_input.DaqClient`), connect it to the server, and put store it at di.client:

>>> di.start_client()
True

where the bool returned indicates whether the client was created and connected successfully or not.

di.client.daq_list contains the output of pynidaqutils.list_daqs() on the server. di.client.scan_daqs() will cause the server to re-scan for DAQ's and di.client.daq_list to be updated.

In order to acquire, the acquisition must be setup. If we want to acquire on device 'Dev1', get the value at each channel at a rate of 100 Hz by averaging 6 measurements together, acquire continuously till we explicitly tell the server to stop (denoted by -1), and return the measured values as single precision floating point numbers (numpy.float32) on differential channels 0 and 2 with input ranges of -10 to +10 V; one would do:

>>> channels = [{'channel': 0, 'voltage': 10.0, 'termination': b'Diff'},
...             {'channel': 2, 'voltage': 10.0, 'termination': b'Diff'}]
>>> success, = di.client.setup_daq(b'Dev1', frequency=100.0,
...                                averaged=6, count=-1,
...                                tp='single', channels=channels)
True

The other outputs of the function indicate what the actual command and configuration sent to the server looked like in the event that they are needed. Then, to start acquisition:

>>> di.client.start_daq()
True

Whether it was successfully started or not is returned. While it is acquiring, di.client.is_acquiring is True. While acquisition is occurring, the server is transferring it to the client over a socket in blocks. All the blocks acquired so far are obtained by:

>>> data, lg = di.client.get_new_data()

data is None if no blocks have been acquired, and a list of numpy.ndarray if there have been. Each bock is a numpy.ndarray where the columns are the different channels in the order given to setup_daq and the rows are successive time steps. lg is a list of tuple with a tuple for each block. The tuple specify the zero-indexed starting and ending sample number for the respective block. In the very off chance that the blocks get out of order or a block is lost, lg can be used to figure that out and reorder if necessary.

When done acquiring, call:

>>> di.client.stop_daq()
True

to stop the DAQ and transmit the last data blocks. Whether stopping it was successful or not is returned. At this point, the DAQ can be reconfigured and acquisition started again.

The client and server are closed by:

>>> di.stop_client()
True
>>> di.stop_server()
True

Though, calling stop_server will automatically call stop_client if the client was started.

Versions

0.2.1. Updated documentation to reflect new version of PyDAQmx.

0.2. Major changes and bugfixes. The highlights are

• Better support for the server and client being on different machines using paramiko.
• Added ability to do temporal binning of acquired samples server side.
• Changed communication protocol betweent the server and the client.
• Backported to Python 2.6 and 2.7
• Completed documentation.
• Many bugfixes.

0.1.1. Fixed bug in reporting version from pynidaqutils.analog_input

0.1. Initial version.