External libraries and dependencies

From SweepMe! Wiki
(Redirected from Device Class Dependencies)
Jump to: navigation, search

Not all files that you need to run your Device Class are part of SweepMe!, e.g.

  • Python packages that are not part of SweepMe!
  • C dll files
  • .NET files

Still, you can add these packages and files to your Device Class by loading them from a subfolder.

Python packages

Several important python packages are already shipped with SweepMe!, e.g.

  • numpy
  • scipy
  • matplotlib
  • pyvisa
  • pyserial
  • pythonnet
  • ... and many more, see credits.html

You can directly load them using "import ..." or "from ... import ..."

In case a python package is missing, you have to add it your your Device Class. Instead of copying the package into the Device Class folder, we recommend to use the Library Builder.

Please note that you can only redistribute python packages if the licenses of all additinally shipped packages allow it.

Library Builder

If you would simply copy a python package to your Device Class it might be that

  • you are missing further sub-dependencies
  • you are shipping sub-dependencies twice that are already included in SweepMe!

For that reason, we provide a tool called 'Library Builder' which adds all missing packages to a folder 'libs' which you can copy into your Device Class folder. The Library Builder is available through your sweep-me.net user account. Please contact us to get one (contact@sweep-me.net).


Install Python 3.6 (i.e. same version as used by SweepMe!) with:

  • all packages installed you need for your Device Class
  • pyinstaller 3.3.1 [1] (newer versions will not work)

Extracting the missing packages

  1. Goto LibraryBuilder folder
  2. Open build_library.py and change sweepme_path to your installation path of SweepMe!
  3. Open required_modules.py and import all packages which are missing in SweepMe!
  4. Run build_library.py
  5. Copy the folder libs_required_modules to your DC and rename as desired
  6. Do not forget to add licence files for each packages you added yourself


Device Classes that load additional python packages are:

  • Camera-PC_Webcam
  • Logger-PC_Midi
  • Logger-PC_Screenshot
  • Switch-PC_Winsound

Environment variables PATH and PYTHONPATH

Often, you have to add the folder and subfolders of a DeviceClass to the environment variable PATH or to PYTHONPATH to allow for importing libraries and dlls (dynamic link libraries). We created a convenience function that adds all folders and subfolders of your DeviceClass script to both environment variables, so that all further imports should work immediately.

Convenience function

The 'FolderManager' of SweepMe! provides a convenience function that automatically adds the folder and all subfolders of the script to PATH and PYTHONPATH. Just add the following two lines of code to your DeviceClass before you import external libraries:

from FolderManager import addFolderToPATH

Adding a folder to PYTHONPATH

Of course, you can add folders yourself, e.g. to load a python package that comes with your DeviceClass being in the subfolder "libs", you can use

# Add the path of your libs within your Device Class to the variable sys.path:
import os
import sys
libpath = os.path.dirname(__file__) + os.sep + "libs"
if not libpath in sys.path:
# Now import your libs as usual
import *name of the lib*

Adding a folder to PATH

In case you like to add a folder to environment variable PATH, use the following lines of code

# Add the path of your libs within your Device Class to the variable sys.path:
import os
import sys
libpath = os.path.dirname(__file__) + os.sep + "libs"
if not libpath in os.environ["PATH"].split(os.pathsep):
    os.environ["PATH"] = libpath + os.pathsep + os.environ["PATH"]

Dynamic Link libraries (DLL)



.Net DLL

version: 1.5.3


Before you can load a .NET dll, one has to make sure that SweepMe! can find it. Either put the dll into the folder 'ExternalLibraries' of the public SweepMe! folder or put it into the folder of your Device Class. In the latter case, the path of the folder needs to be added to the PATH variable, which can automatically be done with the following two lines.

import FolderManager

If you add the dll file to the folder of your Device Class and you like to share it, please make sure that you have the right to distribute the dll file.

A .NET assembly can be loaded using the pythonnet/clr package that already comes with SweepMe! with

import clr

at the top of the Device Class file.

With the following line, you can add the dll file to the namespace of python:

# use 'RgbDriverKit' if the file has the name 'RgbDriverKit.dll'

If the PATH was not updated correctly in the first step, you get an error message like 'Unable to find assembly'.

Now, you can use the name of the added reference like a python package:

from RgbDriverKit import SimulatedSpectrometer, RgbSpectrometer, SpectrometerStatus

I some cases, the name of the dll and the name of the package that is provided by 'clr.AddReference' are not identical. Check the namespace of clr by


in order to see the correct key/name that has to be used during import.

For example, a dll provided by Newport called 'Cornerstone.dll' to control monochromators can be used with the following two lines

import CornerstoneDll as cs

In this case, the name of the imported dll ('CornerstoneDll') is different from the name of the .NET assembly ('Cornerstone').

In order to see the available function of the .NET dll, you can use the tool 'ildasm.exe' that e.g. comes with Microsoft Visual Studio. Please find further information here: https://docs.microsoft.com/en-us/dotnet/framework/tools/ildasm-exe-il-disassembler


The following Device Classes are based on .NET dll and might help you to get startet with your Device class

  • Spectrometer-RGBphotonics_Qwave
  • Spectrometer-Labsphere_CDS6x0
  • Camera-Basler_ace