Abstract text (incl. figure legends and references)

Recent developments of add-on equipment in transmission electron microscopy cause a demand for modular and flexible data acquisition software. Fully integrated commercial solutions exist, but these are typically tailored towards a certain combination of proprietary hardware components from one manufacturer. For off-line data analysis and very specific high-throughput workflows open source software packages became available during the last years. Nevertheless, a lack of highly interactive software directly usable during the operation of the instrument with its different components is obvious. To close this gap, CEOS Panta Rhei is designed as an interactive platform for data acquisition, processing and visualization in electron microscopy written in Python using the Qt, ZMQ and numpy libraries.

The primary purpose of Panta Rhei is to enable simple steering and supervision of workflows combining the control of the microscope and its accessories with data acquisition, online data analysis, and direct visual feedback. The capabilities for online data evaluation are progressing. We currently concentrate on functionality for quick and meaningful assessment of data quality and online session planning like image filters, data statistics, diffractograms, live 4D-STEM evaluation, and elemental mapping and quantification for EELS as well as EFTEM. Data can be interchanged with other software using common file formats like npz, hspy, and tiff. The acquisition (via cameras or scan detectors) typically generates high volume data which has to be transferred and processed efficiently. Therefore, a central component of Panta Rhei is a separate server process called Repository providing access to a managed shared memory. As soon as an acquisition device stores data in the Repository other clients are notified and can directly access the data with minimal CPU load and memory consumption. Clients themselves may also use the Repository to store processed data.

The Panta Rhei GUI is an application to acquire, display and process data and control hardware components which connects as client to the Repository. It displays Views of data from the Repository via a multiple document interface (MDI). Views are used to live display certain aspects of the data and the numerous available DataTools continuously calculate dependent data from updated input. The name Panta Rhei (πάντα ῥεῖ -- everything flows) is motivated by these chains of transformations that may even run in separate processes. For custom extensions, a scripting interface provides control of data processing and display tools as well as hardware devices. For easy external access to hardware control, an RPC-interface is available.

Currently, device interfaces for the TEMs of three manufacturers, five families of detectors, two different scan generators, and the CEOS Imaging Energy Filter (CEFID [1]) exist. We expect that the number of compatible devices will continuously grow over time.

References:

[1] F. Kahl et al. (2019) AIEP 212 including Proceedings CPO-10.

<fig1>

Figure 1: Software architecture of the CEOS Panta Rhei platform.

<fig2>

Figure 2: Screenshot of CEOS Panta Rhei with Views for different imaging and spectroscopy workflows.