The first stage of any data analysis is to unpack raw event files into a representation that is used internally by that analysis framework. In SpecTcl, for example, this unpacking typically fills in structs of treeparamteers. For Root, root trees are built and filled.
Without a lot of careful forethought and planning, this has often resulted in experimenters writing completely different unpacking software for Root and SpecTcl. The result is not only extra work but a maintenance problem as event structures evolve.
genx is a system that allows you to specify the structure of your unpacked data in a way that it can be converted into the data structures and executable code needed to support a framework independent unpacker that works regardless of the data analysis environment.
genx provides a data structure compiler that operates in two stages. In the first stage, your data structure is compiled into an internal representation. The second, target specific stage, translates that internal representation into headers and C++ executable code that supports a specific analysis framework. At this time two backends are suported, SpecTcl and CERN/Root.
This documentation is divided into two parts:
User manual is a user guide. For most people, this is the only part you need to read. It describes how to write data structure declaration files, how to compile those files for specific targets and how to write an a single unpacker for the files.
Reference information is also given in this part for the data structure definition language and the genx compiler command.
Programming manual provides sufficient
internals information to allow program maintainers to extend the
set of backend target code generators. This part describes the
strategy of compilation, the internal representation. It walks through
how a target generator was created and finally describes what you
need to do to extend the set of values the genx
command's --target understands.