Writing a custom app

Use a custom C++ app when configuration-only selection or histogramming cannot express the analysis. tea/create.py creates a skeleton in the analysis repository and updates the CMake cache state.

Create an app

Run from the analysis root:

python3 tea/create.py --type app --name my_analysis

Available app templates are:

• app: general event loop;
• printer: event loop intended for inspection;
• histogrammer: event loop with histogram helpers;
• task4_histogrammer: the DAS LLP tutorial template.

The command creates apps/my_analysis.cpp and configs/my_analysis_config.py.

Read configuration and events

A typical app parses --config, initializes ConfigManager, then reads events:

vector<string> requiredArgs = {"config"};
vector<string> optionalArgs = {"input_path", "output_hists_path"};

auto args = make_unique<ArgsManager>(argc, argv, requiredArgs, optionalArgs);
ConfigManager::Initialize(args);

auto& config = ConfigManager::GetInstance();

auto eventReader = std::make_shared<EventReader>();
for (int iEvent = 0; iEvent < eventReader->GetNevents(); ++iEvent) {
  auto event = eventReader->GetEvent(iEvent);
  // Analysis logic
}

Use the current generated template as the source of truth for constructor signatures; they evolve with the framework.

Build and run

source tea/build.sh
cd bin
./my_analysis --config my_analysis_config.py

Keep reusable domain behavior in a custom object or event class rather than growing one application indefinitely.