PyRDF 0.1.0

Executive release notes

Initial functional release of a python wrapper
around ROOT's RDataFrame with supported for the
Spark distributed backend.

New

• Fully compatible with Python 2.7 and 3.7
• Run in a python notebook on SWAN connected to a Spark cluster
• New tutorials, synchronized with ROOT RDataFrame tutorials
• New documentation to show the usage of PyRDF on SWAN
• Users can send C++ headers and shared libraries needed for their
  analysis to the Spark executors and use them during distributed
  execution
• Documentation available on GitHub Pages

Improvements

• Improve logic for the management of the computational graph. Now
  it is Python version independent and sends to the distributed
  workers only the minimal information required for the execution
  of the operations on the RDataFrame

Bugfixes

• Python 3 bugs:
  • Import statements now use paths relative to the main folder of
    the project
  • Integers previously declared as long now are only integers
  • Division of integers is now correctly declared as floor() division