Implementation of Histmaker

[jeyserma]

Implemented the histmaker functionality in FCCAnalysis. It can run over edm4hep files and/or mini-trees from previous stages and directly produces the histograms using the RunGraps() method in RDataframe.

The analyzer.py must have a build_graph() function that contains the analysis logic and histograms to be saved. The plotting script is expanded accordingly to accept the plotting of single histograms (without the convoluted selections).

The usage is similar to the other functions: fccanalysis run path/to/file.py

Other functionalities implemented:

• Added functionality to add cross-section to the processList. If provided, it will assign the cross-section for that process and normalize to it (regardless the value or existence in the procDict).
• Possibility to scale directly to the lumi in the histmaker (and cross-section).
• Added aliases in defines.h (Vec_f, Vec_i, Vec_rp, Vec_mc, Vec_tlv, …) in defines.h
• Easy way to include user-defined header files (multiple), provided as a list of header files in the input file
• Added ncpus = -1 to use all available cores
• Revised the getchunkList() function: usage of numpy functions to make the chunks
• Created an initialize() function in the FCCAnalysisRun.py that loads the custom analyzers, dicts, and sets the threads (used in both stage and histmaker)
• Simplified the code for tracking the number of events processed in previous stages

Todo:

• No condor support yet (will implement soon).
• Check the newest functionalities with the "final" stage run option (check the normalization, npus, etc). This part of the code is not touched.

[kjvbrt]

is p8_ee_ZH_ecm240 part of the winter2023/IDEA campaign?

[jeyserma]

No, it isn't. So in case the process does not exist, the user can provide a custom cross-section in the processList. Otherwise, a default of 1 is used