''' utility functions for bob.measure ''' import numpy import scipy.stats import bob.core def remove_nan(scores): """remove_nan Remove NaN(s) in the given array Parameters ---------- scores : :py:class:`numpy.ndarray` : array Returns ------- :py:class:`numpy.ndarray` : array without NaN(s) :py:class:`int` : number of NaN(s) in the input array :py:class:`int` : length of the input array """ nans = numpy.isnan(scores) sum_nans = sum(nans) total = len(scores) if sum_nans > 0: logger = bob.core.log.setup("bob.measure") logger.warning('Found {} NaNs in {} scores'.format(sum_nans, total)) return scores[numpy.where(~nans)], sum_nans, total def get_fta(scores): """get_fta calculates the Failure To Acquire (FtA) rate, i.e. proportion of NaN(s) in the input scores Parameters ---------- scores : Tuple of (``positive``, ``negative``) :py:class:`numpy.ndarray`. Returns ------- (:py:class:`numpy.ndarray`, :py:class:`numpy.ndarray`): scores without NaN(s) :py:class:`float` : failure to acquire rate """ fta_sum, fta_total = 0.0, 0.0 neg, sum_nans, total = remove_nan(scores[0]) fta_sum += sum_nans fta_total += total pos, sum_nans, total = remove_nan(scores[1]) fta_sum += sum_nans fta_total += total return ((neg, pos), fta_sum / fta_total) def get_fta_list(scores): """ Get FTAs for a list of scores Parameters ---------- scores: :any:`list` list of scores Returns ------- neg_list: :any:`list` list of negatives pos_list: :any:`list` list of positives fta_list: :any:`list` list of FTAs """ neg_list = [] pos_list = [] fta_list = [] for score in scores: neg = pos = fta = None if score is not None: (neg, pos), fta = get_fta(score) if neg is None: raise ValueError("While loading dev-score file") neg_list.append(neg) pos_list.append(pos) fta_list.append(fta) return (neg_list, pos_list, fta_list) def get_thres(criter, neg, pos, far=None): """Get threshold for the given positive/negatives scores and criterion Parameters ---------- criter : Criterion (`eer` or `hter` or `far`) neg : :py:class:`numpy.ndarray`: array of negative scores pos : :py:class:`numpy.ndarray`:: array of positive scores Returns ------- :py:obj:`float` threshold """ if criter == 'eer': from . import eer_threshold return eer_threshold(neg, pos) elif criter == 'min-hter': from . import min_hter_threshold return min_hter_threshold(neg, pos) elif criter == 'far': if far is None: raise ValueError("FAR value must be provided through " "``--far-value`` option.") from . import far_threshold return far_threshold(neg, pos, far) else: raise ValueError("Incorrect plotting criterion: ``%s``" % criter) def get_colors(n): """get_colors Get a list of matplotlib colors Parameters ---------- n : :obj:`int` Number of colors to output Returns ------- :any:`list` list of colors """ if n > 10: from matplotlib import pyplot cmap = pyplot.cm.get_cmap(name='magma') return [cmap(i) for i in numpy.linspace(0, 1.0, n + 1)] return ['C0', 'C1', 'C2', 'C3', 'C4', 'C5', 'C6', 'C7', 'C8', 'C9'] def get_linestyles(n, on=True): """Get a list of matplotlib linestyles Parameters ---------- n : :obj:`int` Number of linestyles to output Returns ------- :any:`list` list of linestyles """ if not on: return [None] * n list_linestyles = [ (0, ()), # solid (0, (1, 1)), # densely dotted (0, (5, 5)), # dashed (0, (5, 1)), # densely dashed (0, (3, 1, 1, 1, 1, 1)), # densely dashdotdotted (0, (3, 10, 1, 10, 1, 10)), # loosely dashdotdotted (0, (3, 5, 1, 5, 1, 5)), # dashdotdotted (0, (3, 1, 1, 1)), # densely dashdotted (0, (1, 5)), # dotted (0, (3, 5, 1, 5)), # dashdotted (0, (5, 10)), # loosely dashed (0, (3, 10, 1, 10)), # loosely dashdotted (0, (1, 10)) # loosely dotted ] while n > len(list_linestyles): list_linestyles += list_linestyles return list_linestyles def confidence_for_indicator_variable(x, n, alpha=0.05): '''Calculates the confidence interval for proportion estimates The Clopper-Pearson interval method is used for estimating the confidence intervals. Parameters ---------- x : int The number of successes. n : int The number of trials. alpha : :obj:`float`, optional The 1-confidence value that you want. For example, alpha should be 0.05 to obtain 95% confidence intervals. Returns ------- (:obj:`float`, :obj:`float`) a tuple of (lower_bound, upper_bound) which shows the limit of your success rate: lower_bound < x/n < upper_bound ''' lower_bound = scipy.stats.beta.ppf(alpha / 2.0, x, n - x + 1) upper_bound = scipy.stats.beta.ppf(1 - alpha / 2.0, x + 1, n - x) if numpy.isnan(lower_bound): lower_bound = 0 if numpy.isnan(upper_bound): upper_bound = 1 return (lower_bound, upper_bound)