KaneAI - World’s First E2E Software Testing Agent.Book a Demo Join TestMu Conference

How to use lower_than method in Sure

Best Python code snippet using sure_python

day_09.py

Source:day_09.py

1# https://adventofcode.com/2021/day/92import functools3def get_lower_than(heightmap):4    rows, cols = len(heightmap), len(heightmap[0])5    lower_than = [[set() for _ in range(cols)] for _ in range(rows)]6    directions = ((1, 0), (-1, 0), (0, 1), (0, -1))7    for r in range(rows):8        for c in range(cols):9            if heightmap[r][c] == 9:10                continue11            elif heightmap[r][c] == 0:12                lower_than[r][c].update(13                    tuple(sum(x) for x in zip((r, c), d))14                    for d in directions)15            else:16                for (d_r, d_c) in directions:17                    if (18                        (r + d_r) not in range(rows) or19                        (c + d_c) not in range(cols) or20                        heightmap[r][c] < heightmap[r + d_r][c + d_c]21                    ):22                        lower_than[r][c].add((r + d_r, c + d_c))23    return lower_than24def get_basin(heightmap, low_point, lower_than):25    rows, cols = len(heightmap), len(heightmap[0])26    lower_than = get_lower_than(heightmap)27    (lp_r, lp_c) = low_point28    basin = set((low_point,))29    frontier = list(lower_than[lp_r][lp_c])30    visited = set()31    while frontier:32        (p_r, p_c) = frontier.pop()33        if (p_r, p_c) in visited:34            continue35        elif (36            (p_r, p_c) in basin or37            p_r not in range(rows) or38            p_c not in range(cols) or39            heightmap[p_r][p_c] == 940        ):41            visited.add((p_r, p_c))42            continue43        else:44            basin.add((p_r, p_c))45            visited.add((p_r, p_c))46            frontier += list(lower_than[p_r][p_c])47    return basin48def get_risk_level(heightmap):49    rows, cols = len(heightmap), len(heightmap[0])50    lower_than = get_lower_than(heightmap)51    return sum(heightmap[r][c] + 152               for r in range(rows)53               for c in range(cols)54               if len(lower_than[r][c]) == 455               )56def find_basins(heightmap):57    rows, cols = len(heightmap), len(heightmap[0])58    lower_than = get_lower_than(heightmap)59    low_points = [(r, c)60                  for r in range(rows)61                  for c in range(cols)62                  if len(lower_than[r][c]) == 463                  ]64    basin_sizes = [len(get_basin(heightmap, lp, lower_than))65                   for lp in low_points]66    return functools.reduce(67        lambda x, y: x * y,68        sorted(basin_sizes, reverse=True)[:3]69    )70if __name__ == '__main__':71    with open('input/day_09.txt', 'r') as f:72        heightmap = tuple(...

__init__.py

Source:__init__.py

1from analyze.deeppoly.transform import Transformer2import torch3class ReluTransformer(Transformer):4    def __init__(self, heuristic):5        self._heuristic = heuristic6    def _transform(self, ad, input, ads=None):7        ad_relu = ad.clone()8        # 1. setting the equations to zero for nodes with upper_bound lower than zero9        set_to_zero = (ad.upper_bounds <= 0).flatten()10        ad_relu.lower_than.A[set_to_zero, :] = 0.011        ad_relu.lower_than.v[set_to_zero, :] = 0.012        ad_relu.greater_than.A[set_to_zero, :] = 0.013        ad_relu.greater_than.v[set_to_zero, :] = 0.014        # 2. for nodes where the lower_bound is bigger than 0 we don't do anything.15        # 3. for nodes with lower_bound < 0 and upper_bound > 016        relu_mask = ((ad.lower_bounds < 0) & (ad.upper_bounds > 0)).flatten()17        # 3.1 we set the upper bounds and relational contraints:18        # x^L_j <= u_j/(u_j-l_j) * (x^(L-1) - l_j) where l_j <= x^(L-1) <= u_j19        delta = (ad.upper_bounds[relu_mask] - ad.lower_bounds[relu_mask])20        lambda_up = ad.upper_bounds[relu_mask] / delta21        c_up = - ad.lower_bounds[relu_mask] * ad.upper_bounds[relu_mask] / delta22        ad_relu.greater_than.A[relu_mask, :] *= lambda_up23        ad_relu.greater_than.v[relu_mask] *= lambda_up24        ad_relu.greater_than.v[relu_mask] += c_up25        ad_relu.greater_than.A_no_sub = torch.eye(len(relu_mask))26        ad_relu.greater_than.A_no_sub[relu_mask, :] = (torch.eye(len(relu_mask))[relu_mask] * lambda_up)27        ad_relu.greater_than.v_no_sub = torch.zeros(ad_relu.greater_than.v_no_sub.shape)28        ad_relu.greater_than.v_no_sub.T[relu_mask] = c_up29        # 3.2 for the lower bound, the heuristic..30        # x^L_j >= lambda * x^(L-1)_j31        lambda_low = self._heuristic.compute_lambda(ad.lower_bounds[relu_mask], ad.upper_bounds[relu_mask])32        ad_relu.lower_than.A[relu_mask, :] *= lambda_low33        ad_relu.lower_than.v[relu_mask] *= lambda_low34        ad_relu.lower_than.A_no_sub = torch.eye(len(relu_mask))35        ad_relu.lower_than.A_no_sub[relu_mask, :] = (torch.eye(len(relu_mask))[relu_mask] * lambda_low)36        ad_relu.lower_than.v_no_sub = torch.zeros(ad_relu.lower_than.v_no_sub.shape)37        ad_relu.lower_than.v_no_sub.T[relu_mask] = lambda_low38        #if backprop:39        #    ad_relu.upper_bounds = ad_relu.greater_than.compute_bounds_backprop(ad_relu, ads)40        #    ad_relu.lower_bounds = ad_relu.lower_than.compute_bounds_backprop(ad_relu, ads)41        #else:42        ad_relu.lower_bounds = ad_relu.lower_than.compute_bounds(input.lower_bounds, input.upper_bounds)43        ad_relu.upper_bounds = ad_relu.greater_than.compute_bounds(input.lower_bounds, input.upper_bounds)...

day9.py

Source:day9.py

1import numpy as np2inp = []3with open("day9_input.txt") as file:4    inp = [[int(y) for y in x[:-1]] for x in file]5A = np.array(inp)6lower_than = np.full_like(A,0)7lower_than[1:,:] += A[1:,:] < A[:-1,:] #above8lower_than[:-1,:] += A[:-1,:] < A[1:,:] #below9lower_than[:,1:] += A[:,1:] < A[:,:-1] #left10lower_than[:,:-1] += A[:,:-1] < A[:,1:] #right11lower_than[0,:] += 112lower_than[-1,:] += 113lower_than[:,0] += 114lower_than[:,-1] += 115risk_level = 016low_points = []17for i in range(lower_than.shape[0]):18    for j in range(lower_than.shape[1]):19        if lower_than[i,j] == 4:20            risk_level += A[i,j] + 121            low_points.append((i,j))22print("part1", risk_level)23basin_sizes = []24for low_point in low_points:25    flood = np.full_like(A, False, dtype=np.bool_)26    flood[low_point] = True27    previous_sum = 028    while np.sum(flood) > previous_sum:29        previous_sum = np.sum(flood)30        for i in range(100):31            for j in range(100):32                if flood[i,j] == False:33                    continue34                if i-1 >= 0 and A[i-1,j] != 9:35                    flood[i-1,j] = True36                if i+1 < 100 and A[i+1,j] != 9:37                    flood[i+1,j] = True38                if j-1 >= 0 and A[i,j-1] != 9:39                    flood[i,j-1] = True40                if j+1 < 100 and A[i,j+1] != 9:41                    flood[i,j+1] = True42    basin_sizes.append(previous_sum)43basin_sizes.sort()...

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.