TinghuiWang / pyActLearn

    f.write('\\bottomrule\n')

def generate_timeliness_avg_table(methods, classes, result_array,

                                  background_class, filename=None):

    bg_class_id = _get_bg_class_id(classes, background_class)

    timeliness_values = []

    for i, method in enumerate(methods):

        start_mismatch, stop_mismatch = _get_timeliness_measures(classes, result_array[i][0], result_array[i][1],

                                                                 result_array[i][4])

        cur_timeliness = [start_mismatch[j] + stop_mismatch[j] for j in range(len(classes))]

        timeliness_values.append([np.abs(np.array(cur_timeliness[j])) for j in range(len(classes))])

    # Average, <60, >60

    if filename is None:

        f = sys.stdout

    else:

        f = open(filename, 'w')

    f.write('\\textbf{Activities} ')

    for method in methods:

        f.write('& \\textbf{%s} ' % method.replace('_', ' '))

    f.write('\\\\ \\midrule \n')

    for i, activity in enumerate(classes):

        if i != bg_class_id:

            f.write('%s ' % activity.replace('_', ' '))

            # Find maximum and store index

            for j, method in enumerate(methods):

                if len(timeliness_values[j][i]) == 0:

                    average_time = 0.

                else:

                    average_time = np.average(timeliness_values[j][i])

                f.write('& %.1f' % average_time)

            f.write('\\\\ \n')

    f.write('\\bottomrule \n')

def generate_offset_per_table(methods, classes, result_array,

    f.write('\\bottomrule \n')

def generate_offset_per_table(methods, classes, result_array,

                              background_class, filename=None):

    bg_class_id = _get_bg_class_id(classes, background_class)

    timeliness_values = []

    for i, method in enumerate(methods):

        start_mismatch, stop_mismatch = _get_timeoffset_measures(classes, result_array[i][0], result_array[i][1],

                                                                 result_array[i][4])

        cur_timeliness = [start_mismatch[j] + stop_mismatch[j] for j in range(len(classes))]

        timeliness_values.append([np.abs(np.array(cur_timeliness[j])) for j in range(len(classes))])

    # Average, <60, >60

    if filename is None:

        f = sys.stdout

    else:

        f = open(filename, 'w')

    f.write('\\textbf{Activities} ')

    for method in methods:

        f.write('& \\textbf{%s} ' % method.replace('_', ' '))

    f.write('\\\\ \\midrule \n')

    for i, activity in enumerate(classes):

        if i != bg_class_id:

            f.write('%s ' % activity.replace('_', ' '))

            # Find maximum and store index

            for j, method in enumerate(methods):

                total_num = len(timeliness_values[j][i])/2

                nonzero_num = np.count_nonzero(timeliness_values[j][i])

                f.write('& %d/%d' % (nonzero_num, total_num))

            f.write('\\\\ \n')

    f.write('\\bottomrule \n')