Page tree

Present: Lars Froehlich @Suren Karabekyan Martin Killenberg Maximilian Schuette @Sergey Tomin Josef Wilgen

Undulator controls

Olaf has prepared a new version of the wavelength control server that can handle up to three undulator groups with separate photon energies. The new version is installed on the Virtual XFEL. Raimund is working on new jddd panels.

Python data logger & machine abstraction framework

Sergey gives a brief report on his data acquisition and machine abstraction modules for Python. With this framework, he can easily read the machine state into a Pandas data frame and access orbit, screens, and other instrumentation. The code was used successfully during the machine shifts last Sunday. Sergey will give a more detailed overview when the framework is more fleshed out.

Python for control room applications

Tim is planning to upgrade the Anaconda installation on the BKR macs and on our Linux user machines. The new Python will probably be from the 3.7 version line. It is planned not to upgrade this base installation anymore in the foreseeable future. For our applications, we'll set up a virtual environment called hlc in which we can install packages and even upgrade the Python version if necessary. There is some discussion on control system bindings:

  • So far, we have no PyDOOCS module for Python 3.7 in the DOOCS release.
  • There is a PyDOOCS module for Python 3.7 in Christopher Behrens' Anaconda repository which could be installed under Ubuntu 16 using the conda package manager. (Addendum after the meeting: Tim does not recommend using the conda package because it cannot be supplied reliably for all our platforms.)
  • PyTINE works with Python 3.7.
  • Martin complains that the PyDOOCS module from the DOOCS release can only be used with Anaconda, not with the standard Python3 that comes packaged with Debian. This makes it impossible to combine it with other Python modules compiled against that Python version.
  • Maximilian points out that the zmqget functionality in PyDOOCS is currently broken due to a memory management bug.

To gather experience before the upgrade, Lars is going to set up a virtual environment hlc for xfeloper with the current Python 3.6. It will be stored under /home/xfeloper/.conda/envs/hlc and can be activated via conda activate hlc. In the beginning, the following modules will be installed (in addition to Anaconda's default set):

conda cx_oracle matplotlib numpy pymongo pyqt pyqtgraph scipy

Please inform Lars if you would like to have other modules preinstalled.


On February 11, this year's "Automatisierungsforum" takes place at FH Westküste: