INTRODUCTION
The CMS DBS data discovery is my current project. The aim of the project is efficiently lookup data in CMS DBS system. Here I'm outline main requirements of the system:
  • no a-priory knowledge of which data will be looked-up, although a common pattern exists;
  • easy data access and navigation;
  • intelligent data query;
  • hierarchical view (dataset⇒blocks⇒LFNs⇒Objects);
  • integration with other web services;
  • rapid web development to accommodate schema/use case/view changes;

IMPLEMENTATION
The Data Discovery web interface was developed purely in python. The CherryPy pythonic, object-oriented HTTP framework, was used as main engine. We heavily rely on Cheetah python template framework to build UI. Below you can find schematic view of Data Discovery interface.

Further reading
A breif introduction to the system can be found in this talk, while a full project description is spread around multiple twiki pages on CMS web site. The current version is located here, feel free to explore.