helpers

Shared computation helpers for MMM summary factories.

Functions

compute_channel_shares(channel_contributions)	Compute each channel's share of total channel contribution.
compute_residuals(data)	Compute residuals as target minus posterior predictions.
compute_summary_stats_with_hdi(data, hdi_probs)	Module-level wrapper around StatsHelper.compute_summary_stats_with_hdi().
compute_waterfall_components(data[, dims, ...])	Mean contributions over date per component for waterfall summaries.
convert_output(df[, output_format])	Module-level wrapper around StatsHelper.convert_output().
dataframe_to_json_records(df)	Convert a summary DataFrame to JSON-serializable records.
get_channel_x_data(data, apply_cost_per_unit)	Return channel spend or raw channel data based on cost-per-unit flag.
get_prior_for_plot(data, original_scale)	Return the correct idata group for prior predictive plotting.
prepare_sensitivity_data(sa_da, data[, ...])	Prepare sensitivity data for summary tables (steps 1-5 of sensitivity plots).

Classes

StatsHelper()

Lightweight helper for HDI summary stats and output conversion.