sage.plotting.manager

Filename : manager.py Description : Short description of the file

Created on 2026-03-21 17:16:10

__author__ = Narenraju Nagarajan __copyright__ = Copyright 2026, ProjectName __license__ = MIT Licence __version__ = 0.0.1 __maintainer__ = Narenraju Nagarajan __affiliation__ = N/A __email__ = N/A __status__ = [‘inProgress’, ‘Archived’, ‘inUsage’, ‘Debugging’]

GitHub Repository: NULL

Documentation: NULL

Classes

ValidationPlotManager

Loads saved validation results from HDF5 and dispatches all diagnostic plots.

Module Contents

class ValidationPlotManager(validation_h5, losses_h5, export_dir=None)

Loads saved validation results from HDF5 and dispatches all diagnostic plots.

Reads the per-epoch validation HDF5 (network outputs, targets, signal parameters, signal injection indices) and the losses HDF5, then exposes a single plot_all() method that generates the full suite of training-diagnostics plots (ROC, loss curves, efficiency, parameter recovery, etc.) into export_dir.

Parameters:

• validation_h5 (str) – Path to the per-epoch validation output HDF5 file.

• losses_h5 (str) – Path to the epoch loss HDF5 file produced by HDF5LossLogger.

• export_dir (str or None) – Directory to save plots. Subdirectories are created per plot type.

validation_h5

losses_h5

export_dir = None

validation_data

training_loss = None

validation_loss = None

make_all_plots(save=True)

Dispatch the full suite of validation diagnostic plots.

Iterates over all saved epochs, generates per-epoch plots (ROC, loss curves, efficiency, parameter recovery, etc.), and produces cross-epoch summaries (separation trajectory, parameter evolution). All figures are written to self.export_dir.

Parameters:

save (bool) – If True (default), save all plots to disk; otherwise display.