kd_tree_store.py 文件源码

def run_profiling(self, num_loops, num_neighbors, age_proximity):
        """Executes the k_nearest_neighbors algorithm for num_loops times and returns the average running time

        Args:
            num_loops: number of loops for which we query the server
            num_neighbors: number of neighbors to query for
            age_proximity: maximum difference between a candidate neighbor's age and the user


        Returns:

        """
        print('profiling over ', num_loops, ' times')
        random_latitudes = random.uniform(-90, 90, num_loops)
        random_longitudes = random.uniform(-180, 180, num_loops)
        time_list = []

        for i in tqdm(range(len(random_latitudes))):
            start_time = time.clock()
            kd_store.k_nearest_neighbors({'name': 'bla bla', 'age': 23, 'latitude': random_latitudes[i] / 2,
                                          'longitude': random_longitudes[i]}, num_neighbors, age_proximity)
            end_time = time.clock()
            time_list.append(end_time - start_time)

        # get the timing statistics
        stats_desc = stats.describe(time_list)
        frac_times_exceeded = len(np.where(np.array(time_list) >= 1)[0]) / len(time_list)
        print('\nfraction of times with delay > 1 is: ', frac_times_exceeded, '\n')
        print('\nStats:\n', stats_desc)
        return stats_desc