Source code for orion.analysis.regret_utils

# -*- coding: utf-8 -*-
Provide tools to calculate regret
import numpy
import pandas as pd

[docs]def regret(trials, names=("best", "best_id")): """ Calculates the regret for a collection of :class:`orion.core.worker.trial.Trial`. The regret is calculated sequentially from the order of the collection. Parameters ---------- trials: DataFrame or dict A dataframe of trials containing, at least, the columns 'objective' and 'id'. Or a dict equivalent. names: A tuple containing the names of the columns. Default is ('best', 'best-id'). Returns ------- A copy of the original dataframe with two new columns containing respectively the best value so far and its trial id. """ if len(names) != 2: raise ValueError( f"`names` requires a tuple with 2 elements. {len(names)} provided." ) df = pd.DataFrame(trials, copy=True) if df.empty: return df regrets_idx = get_regrets_idx(df["objective"]) df[names[0]] = df["objective"].to_numpy()[list(regrets_idx)] df[names[1]] = df["id"].to_numpy()[regrets_idx] return df
[docs]def get_regrets_idx(objectives): """Return the indices corresponding to the cumulative minimum""" minima = numpy.minimum.accumulate(objectives) diff = numpy.diff(minima) jumps = numpy.arange(len(objectives)) jumps[1:] *= diff != 0 jumps = numpy.maximum.accumulate(jumps) return jumps