torchcell.viz

Plotting helpers used during training and analysis: predicted-versus-measured fitness and genetic-interaction plots, dataset split visualizations, graph-regularization and edge-recovery plots, transformer diagnostics, oversmoothing and oversquashing measures, and regression diagnostics logged to Weights & Biases.

`fitness`

Box-plot visualization of predicted vs measured growth/fitness.

Functions 

`fitness.box_plot`	Bin predictions into growth ranges and box-plot the measured values.
`fitness.generate_simulated_data`	Generate correlated synthetic true/predicted growth values with WT samples.
`fitness.generate_simulated_data_with_nan`	Generate synthetic growth data seeded with NaN, inf, and identical-value bins.

`genetic_interaction_score`

Visualizations of genetic interaction score predictions versus truth.

Functions 

`genetic_interaction_score.box_plot`	Build a binned box-plot figure comparing predictions to true scores.
`genetic_interaction_score.generate_simulated_data`	Generate synthetic true values and noisy predictions for plot testing.
`genetic_interaction_score.generate_simulated_data_with_nan`	Generate synthetic data with injected NaN values for robustness testing.