PDV – a PVSS Data Viewer Application

The four main LHC experiments and the accelerator department have agreed to use a unique commercial SCADA system, PVSS II from ETM (Austria) for detector and accelerator slow control purposes, in order to ease connection and exchange of data among all collaborators and to minimise central support effort for commonly used software and LHC specific libraries provided by the CERN IT department. Knowing this, the PVSS Data Viewer (PDV) had originally been developed to access environment and control data of the Pixel detector of the ATLAS experiment, with an effort to be sufficiently generic to provide access to data of other subdetectors and even data of other experiments, or PVSS systems in general. Other important keys for the design were independence from any existing PVSS installation and universality regarding computer operating systems or user environments. The universality condition has led to the choice of Java as programming language, which also allows to start applications with WebStart technology right from an HTML page, seen by the user as just a click away on the web. The application retrieves and decompresses PVSS data from the database, while optimisingil transferred data bandwidth for query periods of one hour as well as for one year. Data queries are displayed as 2D plot graphics on the screen with time on the abscissa by default, using the open source graphics chart library JChart2D. Different ordinates can be use d for groups of values; a default grouping is made according to the "physical unit" associated with a value in the database. Optionally the data can be shown as 1D "projection" or "value frequency histogram". Several values can be shown in a plot correlating them two by two. A data query can be saved as vector or bitmap graphics or be exported as a comma separated value list in plain text format for use by other applications. An export to the proprietary format of the ROOT analysis platform has been implemeted to satisfy the specific needs of the high energy physics community. As LHC experiments typically record millions of data values (for example 5�106 in the case of Atlas), the search and selection mechanism of data values for a query is not trivial, and efficient tools have been developed to assist the user with this task without compromising the general approach of representing data from all detectors in a common scheme. The user has several ways to select values from a sorted tree or regular expression interface. Furthermore, detector or user specific ways to select data values can be implemented by the user in form of code plugins. Plugin code can also be used to build up parts of the data value tree, to export or to select data. Plugin interfaces to implement detector specific, geographic representations of data values for selection as well as for display will be provided at a later stage. All data queries are kept i n a ring buffer similar to the well known "last documents" menues in commercial applications, and allow the user to easily access queries again from a previous work session. In addition a bookmark collection can be constituted, which has the advantage of being persistent and has the vocation of being exchanged with other users. The exchange of such query configurations, which are systematically stored in XML format, among a big collaboration like Atlas needs a special database, beyond traditional file exchange through email or web pages. Code management in the context of thousands of potential users is an important issue that had to be addressed during development of the application. We will briefly present the tools and methods that have been chosen for that. PDV is open source software. The generic approach of implementation of the PDV application has paid off, as it has grown from a detector specific tool to one of the standard tools for PVSS data viewing in Atlas. Three subdetectors have provided their own plugin codes. The number of its users is still growing and the project is in maintenance phase by now. Parts of the code have been reused as plotting interface in a newly developed web based data viewer by ATLAS.