Skip to content
forked from a4/a4

A4 - An Analysis Tool for High-Energy Physics

License

Notifications You must be signed in to change notification settings

JohannesEbke/a4

 
 

Repository files navigation

A4 - An Analysis Tool for High-Energy Physics

A4 is a data processing format, tool-set and library which provides fast I/O of structured data using the Google protocol buffer library: http://code.google.com/p/protobuf/

This is currently an experimental project, but has already been successfully applied to physics analysis.

For instructions on how to install A4, see INSTALL. More documentation is in the doc/ directory. See also the FAQ. An overview of the structure and a preliminary look at the performance is available at arxiv:1208.1600.

If you are impatient, try:

git clone https://github.com/a4/a4.git
cd a4
./waf go
source install/bin/this_a4.sh
cd tutorial/

For a quick start install A4 and take a look at the examples in the tutorial directory.

Inspired and building on Samvel Khalatyans RTvsPBThreads

Overview

Modular organization

  • a4io provides fast I/O of protobuf events to a custom file format
  • a4process contains processor classes with overridable event loops
  • a4hist contains fast and easy to use histogramming classes
  • a4store provides a way to define and fill histograms on one line in the loop
  • a4root contains conversion routines and tools from a4 to ROOT and back
  • a4plot is a container module for ROOT plotting tools
  • a4atlas contains ATLAS-specific event and metadata definitions and tools

Key features

  • Store protobuf messages of arbitrary types
  • Stores the description of messages, making the format self-describing
  • Transparent compression using different algorithms (zlib, gzip and snappy)
  • Store metadata messages for blocks of events
  • Binary concatenation of A4 files yields a valid A4 file with all metadata
  • Support linear no-seeking mode of operation, suitable for network streaming
  • Tries to minimize manual book-keeping

About

A4 - An Analysis Tool for High-Energy Physics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • Python 62.7%
  • C++ 30.9%
  • C 2.9%
  • Shell 1.6%
  • JavaScript 1.5%
  • Objective-C 0.4%