Diamond Light Source - Annual Review 2022/23

84 85 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 2 / 2 3 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 2 2 / 2 3 In Physical Sciences the community has sought to focus more upon on tracking and optimising both measurements and experiments with projects such as the Autonomous Formulation Laboratory, (AFL), (DOI: 10.1021/acs. chemmater.2c03118), and gpCAM (DOI: 10.1038/s41598-019-48114-3) and Summit (DOI: 10.1002/cmtd.202000051). These projects aim not only to optimise experimental parameters but also interact with data acquisition systems enabling autonomous experimentation. The AFL project a platform that allows for the optimisation of samples made from stock solutions primarily from small angle scattering data but also other techniques such as spectroscopy and rheological measurements. gpCAM optimises the sample scanning by focussing scan point density on more ‘data rich’ areas of a target scan area and weighting away from scanning ‘data poor’ areas. (Figure 3 below ). Finally, Summit is a framework for optimising chemical reaction conditions, evaluating and comparing different reaction strategies against each other to present a broader picture to the user. As can be seen, all these packages afford benefits to scientists whether on a beamline, in the lab or simply highlighting a more optimal data collection strategy. A new EngineeringWeb User Interface Software provides the User Interface (UI) for the operation of Diamond; from the control systems, providing the interfaces: to the underlying hardware and equipment, to the data acquisition systems for experiment control, and to the tools for analysis and cataloguing of the collected data. Facility users rely on UIs to make effective use of Diamond. As part of the Diamond-II software and computing programme, there is a plan to enhance UIs to provide a more integrated experience across the whole suite of Diamond software. This will be achieved by using web technologies. This approach will: make use of modern technologies and benefit from the latest developments in user interfaces, improve the user and developer experience of working with the UIs, and enable the software to be operated remotely, which becomes increasingly important. The first areawhere this approachwill be applied is the EngineeringUI, which is the interface to the control system, used by experts to configure and operate the accelerators and beamlines. This will be needed in the commissioning and operation of Diamond-II and needs to address obsolescence in the existing engineering UI technologies. The new engineering UI will be based on established web frameworks. Similarly, to the web applications used every day, it is made up of two parts, a client and a server. The front-end, running in a browser, acts as a client and is responsible for the final rendering of the interface, while the server is responsible for delivering correct, timely and up to date data to be rendered. The use of web technologies is ubiquitous in society now, and there is familiarity with its function and behaviour. It is also the maturity and performance of modern web frameworks and toolkits that make this an obvious choice to deliver the next generation of control interfaces. The UI front-end is written in TypeScript and makes use of the React framework for rendering and Redux for management of state. The server is written in Python and communication between them is overWebsockets. The scope of this project is to deliver browser-based client software to render the engineering UIs to develop the backend server software that can deliver data to the clients in a performant manner. This means that the server software needs to be able to manage many users connecting and demanding a wide range of data to be displayed. This variability of demand is ideally met by deploying the application in containers orchestrated in Kubernetes (See above: The Journey to Cloud Native). A working prototype has been developed and have been used for some initial performance testing and analysis. The prototype has been reviewed by external experts to confirmthat its implementation is based on solid future-proof technologies that will deliver a high-performance solution that will be fit for purpose for many years. The first roll out of a pilot project delivers Storage Ring Status displays around Diamond (Figure 4 above). Over the coming year further preliminary work will be completed including: additional functionality in the pilot project, documenting a full set of requirements from users, and improving knowledge in key areas. Over the next three to four years, the software will be developed further to deliver a reliable and highly performant interface across the accelerator and beamline control systems. This development will underpin the user interfaces for the higher-level control and data-analysis software. The benefit for Diamond will be one common code base supporting all applications. Figure 2: Output of a deep learning network trained to detect the location of drops and crystals in images. Figure 3: An overview of a data density driven scan from the gpCAM system (Image from Noack et al. – DOI: 10.1038/s41598-019-48114-3, work is under a CC BY 4.0 Licence). Figure 4: Storage Ring Status Display realised as a Web UI.