Last week I talked with my mentors about how the plotting needs to be done: instead of reading the TMVA output file, we need to get the data directly from TMVA.

To start with something easier we agreed on making a JavaScript visualization for correlation matrix. First, I created a C++ library, that was friended with ROOT::TMVA::DataLoader and ROOT::TMVA::Factory classes. I created the JPyInterface python module, which loads the C++ library, and if we call the init function of this module (we have to pass the TMVA::Factory object, because C++ library needs pointer to that class to access DataLoader) and attach the drawing method to factory object. This worked fine, but wasn’t very nice, because we needed to init the JPyInterface. For developing the JPyInterface I needed to learn a little about RequireJS because, Jupyter notebooks use this library for importing libraries (also JsROOT uses this library).

I will have skype call with my mentors about this week’s result on May 13.