How to use __calc_percentiles method in yandex-tank

Best Python code snippet using yandex-tank

screen.py

Source:screen.py

...563        for second in tuple(self.last_min.keys()):564            if ts - second > 60:565                self.last_min.pop(second)566        self.last_min[ts] = dist567    def __calc_percentiles(self):568        def hist_to_quant(histogram, quant):569            cumulative = histogram.cumsum()570            total = cumulative.max()571            positions = cumulative.searchsorted([float(i) / 100 * total for i in quant])572            quant_times = [cumulative.index[i] / 1000. for i in positions]573            return quant_times574        all_times = hist_to_quant(self.overall, self.quantiles)575        last_data = self.last_min[max(self.last_min.keys())]576        last_times = hist_to_quant(last_data, self.quantiles)577        # Check if we have precise data for last second quantiles instead of binned histogram578        for position, q in enumerate(self.quantiles):579            if q in self.precise_quantiles:580                last_times[position] = self.precise_quantiles[q]581        # Replace binned values with precise, if lower quantile bin happens to be582        # greater than upper quantile precise values583        for position in reversed(range(1, len(last_times))):584            if last_times[position - 1] > last_times[position]:585                last_times[position - 1] = last_times[position]586        last_1m = pd.Series()587        for ts, data in self.last_min.items():588            if last_1m.empty:589                last_1m = data590            else:591                last_1m = last_1m.add(data, fill_value=0)592        last_1m_times = hist_to_quant(last_1m, self.quantiles)593        quant_times = reversed(594            list(zip(self.quantiles, all_times, last_1m_times, last_times))595        )596        data = []597        for q, all_time, last_1m, last_time in quant_times:598            data.append({599                'quantile': q,600                'all': all_time,601                'last_1m': last_1m,602                'last': last_time603            })604        return data605    def render(self, expected_width=None):606        prepared = [(self.screen.markup.WHITE, self.title)]607        if self.overall is None:608            prepared.append(('',))609        else:610            data = self.__calc_percentiles()611            prepared += self.formatter.render_table(data, ['quantile', 'all', 'last_1m', 'last'])612        self.width, self.lines = self.fill_rectangle(prepared)613class CurrentHTTPBlock(AbstractBlock):614    ''' Http codes with highlight'''615    def __init__(self, screen):616        AbstractBlock.__init__(self, screen)617        self.overall_dist = defaultdict(int)618        self.title = 'HTTP codes: '619        self.total_count = 0620        template = {621            'count':       {'tpl': '{:>,}'},  # noqa: E241622            'last':        {'tpl': '+{:>,}'},  # noqa: E241623            'percent':     {'tpl': '{:>.2f}%'},  # noqa: E241624            'code':        {'tpl': '{:>3}', 'final': '{{{:}:<{len}}}'},  # noqa: E241...

generation.py

Source:generation.py

...31        else:32            for (i, j), x in np.ndenumerate(self.heightmap):33                self.heightmap[i][j] = x / mx34        return self35    def __calc_percentiles(self):36        """Converts each value in the heightmap to the largest percentile in which it37        still lies. If values are the same, they are assigned the same percentile.38        """39        arr_as_srtd_lst = sorted(list(np.ndenumerate(self.heightmap)),40                                 key=lambda x: x[1])41        last_percentile = None42        last_value = None43        for (k, ((i, j), v)) in enumerate(arr_as_srtd_lst):44            curr_percentile = (k + 1.0) / len(arr_as_srtd_lst)45            if v == last_value:46                curr_percentile = last_percentile47            self.heightmap[i][j] = curr_percentile48            last_percentile = curr_percentile49            last_value = v50        return self51    def __calc_groups(self):52        """Maps normalized tile values to groups. A group is a number indicating between53        which thresholds a value lies. The groups can later easily be mapped to54        color maps by a visualization strategy.55        """56        thresholds = self.config['thresholds']57        if thresholds is None:58            raise Exception("The config parameter 'thresholds' must be set.")59        for ((i, j), x) in np.ndenumerate(self.heightmap):60            groupFound = False61            for k, t in enumerate(thresholds):62                if x < t:63                    self.heightmap[i][j] = k64                    groupFound = True65                    break66            if not groupFound:67                # assign to last group68                self.heightmap[i][j] = len(thresholds)69        return self70    @staticmethod71    def __get_neighbors(pt, shape):72        """Get all neighbours for the given point pt. This function is used by the73        erosion filter. At most 8 neighbors are returned. The second parameter74        denotes the shape of the heightmap (i.e. width and height).75        """76        x, y = pt77        w, h = shape78        neighbors = [((x - 1, y), 1), ((x - 1, y - 1), 1), ((x - 1, y + 1), 1),79                     ((x, y), 2), ((x, y - 1), 1), ((x, y + 1), 1),80                     ((x + 1, y), 1), ((x + 1, y - 1), 1), ((x + 1, y + 1), 1)]81        return [((a, b), wt) for ((a, b), wt) in neighbors if 0 <= a < w and b >= 0 and b < h]82    def __smooth(self):83        """Erosion filter. This filter reduces some of the noise from the generated data.84        After applying it, it is unlikely for tiles of high and low height to be85        adjacent.86        """87        if 'erosion' in self.config:88            for m in range(self.config['erosion']):89                hm_copy = np.copy(self.heightmap)90                for i in range(self.width):91                    for j in range(self.height):92                        neighbors = self.__get_neighbors((i, j),93                                                         (self.width, self.height))94                        vals = [self.heightmap[k][l] * w for ((k, l), w) in neighbors]95                        sum_weights = sum([w for ((k, l), w) in neighbors])96                        hm_copy[i][j] = round(sum(vals) / (sum_weights))97                self.heightmap = hm_copy98        return self99    def generate(self):100        """Generates height data by calling generate_raw() which should be overridden by101        all subclasses. Afterwards, a normalization is performed and finally each102        normalized value is mapped to a group.103        """104        return self._generate_raw().__calc_percentiles().__calc_groups().__smooth()105class RandomStrategy(GenerationStrategy):106    """Simple height map generation strategy: Uses uniformly distributed random values107    for each tile.108    """109    def _generate_raw(self):110        """Creates height map of random values.111        """112        self.heightmap = np.random.rand(self.width, self.height)113        return self114class LinearFaultStrategy(GenerationStrategy):115    """Generates a random height map by using the "fault" algorithm. This algorithm116    iteratively creates random straights and increases the height for all117    points either below or above the straight. Which side is increased is also118    decided randomly....

Automation Testing Tutorials

Learn to execute automation testing from scratch with LambdaTest Learning Hub. Right from setting up the prerequisites to run your first automation test, to following best practices and diving deeper into advanced test scenarios. LambdaTest Learning Hubs compile a list of step-by-step guides to help you be proficient with different test automation frameworks i.e. Selenium, Cypress, TestNG etc.