Analysis Module

analysis is the top-level pipeline class. It accepts user-defined event, graph, selection, model, and metric templates; reads ROOT and HDF5 files; builds all compiled objects in parallel; and launches the training/evaluation loop. It inherits from notification (logging) and tools (utilities).

Class: `analysis`

Header: <AnalysisG/analysis.h>

Inheritance: notification, tools

Configuration Field

Field	Type	Description
`m_settings`	`settings_t`	Aggregated settings struct controlling output paths, thread counts, batch size, kfolds, epochs, TrainSize, etc.
`meta_data`	`std::map<std::string, meta*>`	Dataset metadata objects indexed by file path.

Registration Methods

Signature	Description
`void add_samples(std::string path, std::string label)`	Registers a ROOT file or directory path under dataset label label.
`void add_event_template(event_template* ev, std::string label)`	Registers an event-class prototype ev for files with label label.
`void add_graph_template(graph_template* gr, std::string label)`	Registers a graph-class prototype gr for label label.
`void add_selection_template(selection_template* sel)`	Registers a selection prototype (applied to all events).
`void add_metric_template(metric_template* mx, model_template* mdl)`	Associates a metric mx with model mdl.
`void add_model(model_template* model, optimizer_params_t* op, std::string run_name)`	Registers a model with optimiser parameters and run name.
`void add_model(model_template* model, std::string run_name)`	Registers a model with default optimiser parameters.

Execution Methods

Signature	Description
`void attach_threads()`	Attaches the thread pool to the event/graph compilation workers.
`void start()`	Runs the complete pipeline: reads files → builds events → builds graphs → builds selections → trains models → evaluates metrics.
`std::map<std::string, std::vector<float>> progress()`	Returns per-model training progress as a map of model-name → list of (epoch, loss) pairs.
`std::map<std::string, std::string> progress_mode()`	Returns the current training mode string per model.
`std::map<std::string, std::string> progress_report()`	Returns a human-readable progress report per model.
`std::map<std::string, bool> is_complete()`	Returns whether each model has finished all epochs.