|
High Energy Physics (HEP), tries to understand the real world, by explaining the physics model at the sub-atomic level. This entails to find new particles, derive their properties and study their interaction with the rest of the matter. In order to do that, big research projects are set and gather institutions and Universities most of the times from many different countries that contribute with scientists and technical people from different scientific areas. The main goal is to build giant machines to accelerate particles and make them collide. The product data from these experiments is stored and analyzed afterwards by physicists. These High tech projects, imply to push for the state of the art techniques in areas such as Networks, distributed computing, databases, real-time systems, etc.
Presently, in the new HEP experiments we are observing a growing complexity in the interfaces of the analysis tools. As a consequence, end users have to master programming, understand complex frameworks and data storage details before being able to work with the data from experiments. This reduces the efficiency of the analysis process significantly.
We have designed a declarative domain specific visual query language for HEP data analysis. It has been designed based on our experience dealing with Hera-B event data and query patterns. Its main goal is to allow the physicist to describe the decay selection queries by means of visual operators, to be run against the experiments' existing storage bases and analysis frameworks.
Our visual language aims to be a simple-to-use tool with which a user can express complex decay queries in an easy way with reduced programming efforts. Indeed, the user does not have to deal with intricacies like storage details.
Our framework to support the analysis process, Pheasant, consists of three layers. On the conceptual level it features the aforementioned domain specific visual query language. On the logical level we provide a more detailed representation of the data in form of a logical schema (still hiding implementation details). The visual
language queries are mapped on a logical algebra. Finally, on the lowest level, the physical level, different (existing) tools can be integrated into our framework via code generation modules (which translates the intermediate algebra expressions into the language of the respective tool). In this way, developers may introduce changes to their tools or extend the framework without affecting the upper layers.
In a first prototype we have implemented the interface (for the conceptual layer), the algebra (for the logical layer), and integrated the BEE framework, which is one of the analysis frameworks for the Hera-B experiment based in the ROOT histogramming tool (on the physical layer). A first assessment of Pheasant by real users has encouraged us to continue on improving our framework. The physicists trying out this prototype were quite enthusiastic, which leads us to believe that this framework will have an impact on the HEP analysis process.
Now that we have a prototype, we want to move on to the product release, improve the performance of our framework and increase the number of existing tools that are supported.
|