How to use unpad method in localstack

Best Python code snippet using localstack_python

detect.py

Source:detect.py

1#!/usr/bin/env python32import os3from this import d4from cv2 import KeyPoint5from matplotlib import image6os.environ['KMP_DUPLICATE_LIB_OK'] = 'True'7import triangulation as tri8import calibration9# Python imports10import numpy as np11import cv212import math13import message_filters14from geometry_msgs.msg import TransformStamped15import tf2_ros16import tf17# ROS imports18import rospy19import scipy.io as sio20import std_msgs.msg21# Deep Learning imports22import torch23import yaml24from cv_bridge import CvBridge, CvBridgeError25from geometry_msgs.msg import Point32, Polygon26from numpy import random27from rospkg import RosPack28from sensor_msgs.msg import Image ,LaserScan29from skimage.transform import resize30from std_msgs.msg import UInt831from torch.autograd import Variable32from yolov5_pytorch_ros.msg import BoundingBox, BoundingBoxes33from math import nan34from models.experimental import attempt_load35from utils.datasets import LoadImages, LoadStreams36from utils.general import (apply_classifier, check_img_size,37                           check_requirements, increment_path,38                           non_max_suppression, scale_coords, set_logging,39                           strip_optimizer)40from utils.plots import plot_one_box41# util + model imports42from utils.torch_utils import load_classifier, select_device, time_synchronized43package = RosPack()44package_path = package.get_path('yolov5_pytorch_ros')45topic_tf_child = "/object"46topic_tf_perent = "/base_link"47P = np.array([[0,0,0,0],48            [0,0,0,0],49            [0,0,0,0],50            [0,0,0,0]])51x_hat=None52prev=[0,0]53intila=054t = TransformStamped()55tf2_br = tf2_ros.TransformBroadcaster()56class Detector:57    def __init__(self):58        # Load weights parameter59        self.weights_path = rospy.get_param('~weights_path')60        rospy.loginfo("Found weights, loading %s", self.weights_path)61        # Raise error if it cannot find the model62        if not os.path.isfile(self.weights_path):63            raise IOError(('{:s} not found.').format(self.weights_path))64        # Load image parameter and confidence threshold65        self.image_topic = rospy.get_param(66            '~image_topic', '/camera/rgb/image_raw')67        log_str = "Subscribing to image topic: %s" % self.image_topic68        rospy.loginfo(log_str)69        self.conf_thres = rospy.get_param('~confidence', 0.25)70        self.C=np.array([[1,0,1,0],71                            [0,1,0,1],72                            [0,0,0,0],73                            [0,0,0,0]])74        self.B=np.array([0,0,0,0])75        dt=0.276        self.A=np.array([[1.0 ,0, dt,0],77                        [0,1.0,0,dt],78                        [0,0,1,0],79                        [0,0,0,1]])80        self.Q=np.array([[1.0,0,0,0],81                        [0,1.0,0,0],82                        [0,0,0.1,0],83                        [0,0,0,0.1]])84        self.R=np.array([[0.1,0,1,0],85                            [0,0.1,0,1],86                            [0,0,0.1,0],87                            [0,0,0,0.1]])88        self.I=np.array([[1,0,0,0],89                        [0,1,0,0],90                        [0,0,1,0],91                        [0,0,0,1]])92        # Load other parameters93        self.device_name = ''94        self.device = select_device(self.device_name)95        self.gpu_id = rospy.get_param('~gpu_id', 0)96        self.network_img_size = rospy.get_param('~img_size', 416)97        self.publish_image = rospy.get_param('~publish_image')98        self.iou_thres = 0.4599        self.augment = True100        self.classes = None101        self.agnostic_nms = False102        self.w = 0103        self.h = 0104        # Second-stage classifier105        self.classify = False106        # Initialize107        self.half = self.device.type != 'cpu'  # half precision only supported on CUDA108        # Load model109        self.model = attempt_load(110            self.weights_path, map_location=self.device)  # load FP32 model111        self.stride = int(self.model.stride.max())      # model stride112        self.network_img_size = check_img_size(113            self.network_img_size, s=self.stride)  # check img_size114        if self.half:115            self.model.half()  # to FP16116        if self.classify:117            self.modelc = load_classifier(name='resnet101', n=2)  # initialize118            self.modelc.load_state_dict(torch.load(119                'weights/resnet101.pt', map_location=self.device)['model']).to(self.device).eval()120        # Get names and colors121        self.names = self.model.module.names if hasattr(122            self.model, 'module') else self.model.names123        self.colors = [[random.randint(0, 255)124                        for _ in range(3)] for _ in self.names]125        # Run inference126        if self.device.type != 'cpu':127            self.model(torch.zeros(1, 3, self.network_img_size, self.network_img_size).to(128                self.device).type_as(next(self.model.parameters())))  # run once129        self.K=[203.42144086256206, 0.0, 206.12517266886093, 130                0.0, 203.55319738398958, 146.4392791209304, 131                0.0, 0.0, 1.0]132        t = 0133        self.camera_parameters=np.array([[0,0,0],[0,0,0],[0,0,0]])134        for i in range(3):135            for k in range(3):136                self.camera_parameters[i][k] = self.K[i+k]137                t+=1138        self.D=[-0.18296735250090237, 0.49168852367941696, -0.6266727991904993, 0.2636407064533411,0]139        self.camera_distortion_param = np.array([-0.426801, 0.249082, -0.002705, -0.001600, 0.000000])140        # Load CvBridge141        self.bridge = CvBridge()142        # Load publisher topic143        self.detected_objects_topic = rospy.get_param(144            '~detected_objects_topic')145        self.published_image_topic = rospy.get_param('~detections_image_topic')146        # Define subscribers147        self.image_sub = message_filters.Subscriber(148            self.image_topic, Image)149        self.image_bub2 = message_filters.Subscriber(150            "/scan", LaserScan)151        self.image_bub = message_filters.Subscriber(152            "/camera2/image_raw", Image)153        ts = message_filters.ApproximateTimeSynchronizer([self.image_sub, self.image_bub2,self.image_bub],  1, 1)154        ts.registerCallback(self.image_cb)155        # Define publishers156        self.pub_ = rospy.Publisher(157            self.detected_objects_topic, BoundingBoxes, queue_size=10)158        self.pub_viz_ = rospy.Publisher(159            self.published_image_topic, Image, queue_size=10)160        rospy.loginfo("Launched node for object detection")161        # Spin162        rospy.spin()163    def lidar_cb(self,data):164        self.ranges=data165    def pubTf(self,position, orientation):166        """167        publish find object to tf2168        :param position:169        :param orientation:170        :return:171        """172        global t173        t.header.stamp = rospy.Time.now()174        t.header.frame_id = "/base_link"175        t.child_frame_id = "/object"176        t.transform.translation.x = position[0]177        t.transform.translation.y = position[1]178        t.transform.translation.z = position[2]179        quaternion = tf.transformations.quaternion_from_euler(orientation[0], orientation[1], orientation[2])180        t.transform.rotation.x = quaternion[0]181        t.transform.rotation.y = quaternion[1]182        t.transform.rotation.z = quaternion[2]183        t.transform.rotation.w = quaternion[3]184        tf2_br.sendTransform(t)185    def kalamnFilter(self, v):186        global intila187        global P188        189        190        v=np.array(v)191        if intila==0:192            x_hat = v193            prev = v[0:1]194            intila=1195        else:196            197            K = P*self.C.transpose()*(self.C*P*self.C.transpose() + self.R).inverse()198            x_hat += K * (v - self.C*x_hat)199            x_hat = self.A*x_hat200            P = self.A*P*self.A.transpose() + self.Q201            P = (self.I - K*self.C)*P202            return x_hat203        return x_hat204    def image_cb(self, data,data2,data3):205        global prev206        # Convert the image to OpenCV207        self.ranges=data2.ranges208        try:209            self.cv_img = self.bridge.imgmsg_to_cv2(data, "rgb8")210        except CvBridgeError as e:211            print(e)212        try:213            self.cv_img2 = self.bridge.imgmsg_to_cv2(data3, "rgb8")214        except CvBridgeError as e:215            print(e)216        # Initialize detection results217        B = 5218        f=10 219        detection_results = BoundingBoxes()220        detection_results.header = data.header221        detection_results.image_header = data.header222        223        input_img = self.preprocess(self.cv_img)224        input_img2 = self.preprocess(self.cv_img2)225        #input_img = Variable(input_img.type(torch.FloatTensor))226        #input_img2 = Variable(input_img2.type(torch.FloatTensor))227        # Get detections from network228        with torch.no_grad():229            input_img = torch.from_numpy(input_img).to(self.device,dtype=torch.half)230            input_img2 = torch.from_numpy(input_img2).to(self.device,dtype=torch.half)231            detections = self.model(input_img)[0]232            detections2 = self.model(input_img2)[0]233            detections = non_max_suppression(detections, self.conf_thres, self.iou_thres,234                                             classes=self.classes, agnostic=self.agnostic_nms)235            detections2 = non_max_suppression(detections2, self.conf_thres, self.iou_thres,236                                             classes=self.classes, agnostic=self.agnostic_nms)237        alpha = 127.22339616    238        # Parse detections239        i=0240        strq='a'241        if detections[0] is not None:242            for detection in detections[0]:243                kp=0.2244                try:245                    xmin2, ymin2, xmax2, ymax2, conf2, det_class2=detections2[0][i]246                except:247                    pass248                # Get xmin, ymin, xmax, ymax, confidence and class249                xmin, ymin, xmax, ymax, conf, det_class = detection250        251                if self.names[int(det_class)] not in ['bin_rect','bin_side','bin_cir']:252                    continue253                pad_x = max(self.h - self.w, 0) * \254                    (self.network_img_size/max(self.h, self.w))255                pad_y = max(self.w - self.h, 0) * \256                    (self.network_img_size/max(self.h, self.w))257                unpad_h = self.network_img_size-pad_y258                unpad_w = self.network_img_size-pad_x259                xmin_unpad = ((xmin-pad_x//2)/unpad_w)*self.w260                xmax_unpad = ((xmax-xmin)/unpad_w)*self.w + xmin_unpad261                ymin_unpad = ((ymin-pad_y//2)/unpad_h)*self.h262                ymax_unpad = ((ymax-ymin)/unpad_h)*self.h + ymin_unpad263                xmin_unpad = xmin_unpad.cpu().detach().numpy()264                xmax_unpad = xmax_unpad.cpu().detach().numpy()265                ymin_unpad =  ymin_unpad.cpu().detach().numpy()266                ymax_unpad = ymax_unpad.cpu().detach().numpy()267                pad_x2 = max(self.h - self.w, 0) * \268                    (self.network_img_size/max(self.h, self.w))269                pad_y2 = max(self.w - self.h, 0) * \270                   (self.network_img_size/max(self.h, self.w))271                unpad_h2 = self.network_img_size-pad_y2272                unpad_w2 = self.network_img_size-pad_x2273                try:274                    xmin_unpad2 = ((xmin2-pad_x2//2)/unpad_w2)*self.w275                    xmax_unpad2 = ((xmax2-xmin2)/unpad_w2)*self.w + xmin_unpad2276                    ymin_unpad2 = ((ymin2-pad_y2//2)/unpad_h2)*self.h277                    ymax_unpad2 = ((ymax2-ymin2)/unpad_h2)*self.h + ymin_unpad2278                    xmin_unpad2 = xmin_unpad2.cpu().detach().numpy()279                    xmax_unpad2 = xmax_unpad2.cpu().detach().numpy()280                    ymin_unpad2 =  ymin_unpad2.cpu().detach().numpy()281                    ymax_unpad2 = ymax_unpad2.cpu().detach().numpy()282                    center_point_right=((xmin_unpad2-xmax_unpad2//2),(ymin_unpad2-ymax_unpad2//2))283                    center_point=((xmin_unpad-xmax_unpad//2),(ymin_unpad-ymax_unpad//2))284                    frame_right, frame_left = calibration.undistortRectify(self.cv_img, self.cv_img2)285                    depth2 = tri.find_depth(center_point_right, center_point, frame_right, frame_left, B, f, alpha)286                    depth2=depth2/10287                except:288                    kp=0289                    depth2=0290                    pass291                # Populate darknet message292                detection_msg = BoundingBox()293                detection_msg.xmin = int(xmin_unpad)294                detection_msg.xmax = int(xmax_unpad)295                detection_msg.ymin = int(ymin_unpad)296                detection_msg.ymax = int(ymax_unpad)297                detection_msg.probability = float(conf)298                detection_msg.Class = self.names[int(det_class)]299                300                angle=(2.22/408)*((xmin_unpad-xmax_unpad)//2)301                index=int((angle+0.462)//0.014032435603439808)302                depth1=self.ranges[index]303                304                if depth2<0:305                    depth2=depth2*-1306                307                if depth1==nan:308                    depth1=0309                fx=203.42144086256206310                fy=203.55319738398958311                rotation_rad=[0,0,0]312                cx=206.12517266886093313                cy=146.4392791209304314                kn=1-kp315                depth=kn*depth1+kp*depth2316                317                v=([depth,(depth) * (((410-(xmax_unpad-xmin_unpad//2))-cx)/fx),prev[0]-depth,prev[1]-(((410-(xmax_unpad-xmin_unpad//2))-cx)/fx)])318                prev=[depth,(depth) * (((410-(xmax_unpad-xmin_unpad//2))-cx)/fx)]319                v_out=(v)320                print("lidar:",depth1," stereo",depth2," fused:",depth,"kalman out:",v_out[1])321                t = TransformStamped()322                tf2_br = tf2_ros.TransformBroadcaster()323                t.header.stamp = rospy.Time.now()324                t.header.frame_id = "/base_link"325                t.child_frame_id = "/object"+strq326                strq+='a'327                t.transform.translation.x = v_out[0]328                t.transform.translation.y = v_out[1]329                t.transform.translation.z = 0330                quaternion = tf.transformations.quaternion_from_euler(0, 0, 0)331                t.transform.rotation.x = quaternion[0]332                t.transform.rotation.y = quaternion[1]333                t.transform.rotation.z = quaternion[2]334                t.transform.rotation.w = quaternion[3]335                tf2_br.sendTransform(t)336                # Append in overall detection message337                detection_results.bounding_boxes.append(detection_msg)338        # Publish detection results339        self.pub_.publish(detection_results)340        # Visualize detection results341        if (self.publish_image):342            self.visualize_and_publish(detection_results, self.cv_img)343        return True344    def preprocess(self, img):345        # Extract image and shape346        img = np.copy(img)347        img = img.astype(float)348        height, width, channels = img.shape349        if (height != self.h) or (width != self.w):350            self.h = height351            self.w = width352            # Determine image to be used353            self.padded_image = np.zeros(354                (max(self.h, self.w), max(self.h, self.w), channels)).astype(float)355        # Add padding356        if (self.w > self.h):357            self.padded_image[(self.w-self.h)//2: self.h +358                              (self.w-self.h)//2, :, :] = img359        else:360            self.padded_image[:, (self.h-self.w)//2: self.w +361                              (self.h-self.w)//2, :] = img362        # Resize and normalize363        input_img = resize(self.padded_image, (self.network_img_size, self.network_img_size, 3))/255.364        # Channels-first365        input_img = np.transpose(input_img, (2, 0, 1))366        # As pytorch tensor367        #input_img = torch.from_numpy(input_img).float()368        input_img = input_img[None]369        return input_img370    def visualize_and_publish(self, output, imgIn):371        # Copy image and visualize372        imgOut = imgIn.copy()373        font = cv2.FONT_HERSHEY_SIMPLEX374        fontScale = 0.8375        thickness = 2376        for index in range(len(output.bounding_boxes)):377            label = output.bounding_boxes[index].Class378            x_p1 = output.bounding_boxes[index].xmin379            y_p1 = output.bounding_boxes[index].ymin380            x_p3 = output.bounding_boxes[index].xmax381            y_p3 = output.bounding_boxes[index].ymax382            confidence = output.bounding_boxes[index].probability383            # Set class color384            color = self.colors[self.names.index(label)]385            # Create rectangle386            cv2.rectangle(imgOut, (int(x_p1), int(y_p1)), (int(x_p3), int(387                y_p3)), (color[0], color[1], color[2]), thickness)388            text = ('{:s}: {:.3f}').format(label, confidence)389            390            cv2.putText(imgOut, text, (int(x_p1), int(y_p1+20)), font,391                        fontScale, (255, 255, 255), thickness, cv2.LINE_AA)392        # Publish visualization image393        image_msg = self.bridge.cv2_to_imgmsg(imgOut, "rgb8")394        image_msg.header.frame_id = 'camera'395        image_msg.header.stamp = rospy.Time.now()396        self.pub_viz_.publish(image_msg)397if __name__ == '__main__':398    rospy.init_node('detector')399    # Define detector object...

test_Padding.py

Source:test_Padding.py

...40        padded = pad(b(""), 4)41        self.assertTrue(padded == uh(b("04040404")))42        padded = pad(b(""), 4, 'pkcs7')43        self.assertTrue(padded == uh(b("04040404")))44        back = unpad(padded, 4)45        self.assertTrue(back == b(""))46    def test2(self):47        padded = pad(uh(b("12345678")), 4)48        self.assertTrue(padded == uh(b("1234567804040404")))49        back = unpad(padded, 4)50        self.assertTrue(back == uh(b("12345678")))51    def test3(self):52        padded = pad(uh(b("123456")), 4)53        self.assertTrue(padded == uh(b("12345601")))54        back = unpad(padded, 4)55        self.assertTrue(back == uh(b("123456")))56    def test4(self):57        padded = pad(uh(b("1234567890")), 4)58        self.assertTrue(padded == uh(b("1234567890030303")))59        back = unpad(padded, 4)60        self.assertTrue(back == uh(b("1234567890")))61    def testn1(self):62        self.assertRaises(ValueError, pad, uh(b("12")), 4, 'pkcs8')63    def testn2(self):64        self.assertRaises(ValueError, unpad, b("\0\0\0"), 4)65        self.assertRaises(ValueError, unpad, b(""), 4)66    def testn3(self):67        self.assertRaises(ValueError, unpad, b("123456\x02"), 4)68        self.assertRaises(ValueError, unpad, b("123456\x00"), 4)69        self.assertRaises(ValueError, unpad, b("123456\x05\x05\x05\x05\x05"), 4)70class X923_Tests(unittest.TestCase):71    def test1(self):72        padded = pad(b(""), 4, 'x923')73        self.assertTrue(padded == uh(b("00000004")))74        back = unpad(padded, 4, 'x923')75        self.assertTrue(back == b(""))76    def test2(self):77        padded = pad(uh(b("12345678")), 4, 'x923')78        self.assertTrue(padded == uh(b("1234567800000004")))79        back = unpad(padded, 4, 'x923')80        self.assertTrue(back == uh(b("12345678")))81    def test3(self):82        padded = pad(uh(b("123456")), 4, 'x923')83        self.assertTrue(padded == uh(b("12345601")))84        back = unpad(padded, 4, 'x923')85        self.assertTrue(back == uh(b("123456")))86    def test4(self):87        padded = pad(uh(b("1234567890")), 4, 'x923')88        self.assertTrue(padded == uh(b("1234567890000003")))89        back = unpad(padded, 4, 'x923')90        self.assertTrue(back == uh(b("1234567890")))91    def testn1(self):92        self.assertRaises(ValueError, unpad, b("123456\x02"), 4, 'x923')93        self.assertRaises(ValueError, unpad, b("123456\x00"), 4, 'x923')94        self.assertRaises(ValueError, unpad, b("123456\x00\x00\x00\x00\x05"), 4, 'x923')95        self.assertRaises(ValueError, unpad, b(""), 4, 'x923')96class ISO7816_Tests(unittest.TestCase):97    def test1(self):98        padded = pad(b(""), 4, 'iso7816')99        self.assertTrue(padded == uh(b("80000000")))100        back = unpad(padded, 4, 'iso7816')101        self.assertTrue(back == b(""))102    def test2(self):103        padded = pad(uh(b("12345678")), 4, 'iso7816')104        self.assertTrue(padded == uh(b("1234567880000000")))105        back = unpad(padded, 4, 'iso7816')106        self.assertTrue(back == uh(b("12345678")))107    def test3(self):108        padded = pad(uh(b("123456")), 4, 'iso7816')109        self.assertTrue(padded == uh(b("12345680")))110        #import pdb; pdb.set_trace()111        back = unpad(padded, 4, 'iso7816')112        self.assertTrue(back == uh(b("123456")))113    def test4(self):114        padded = pad(uh(b("1234567890")), 4, 'iso7816')115        self.assertTrue(padded == uh(b("1234567890800000")))116        back = unpad(padded, 4, 'iso7816')117        self.assertTrue(back == uh(b("1234567890")))118    def testn1(self):119        self.assertRaises(ValueError, unpad, b("123456\x81"), 4, 'iso7816')120        self.assertRaises(ValueError, unpad, b(""), 4, 'iso7816')121def get_tests(config={}):122    tests = []123    tests += list_test_cases(PKCS7_Tests)124    tests += list_test_cases(X923_Tests)125    tests += list_test_cases(ISO7816_Tests)126    return tests127if __name__ == '__main__':128    suite = lambda: unittest.TestSuite(get_tests())...

test_utils_functional.py

Source:test_utils_functional.py

...31@pytest.mark.parametrize(32    ["shape", "padding"],33    [((1, 3, 221, 234), 32), ((1, 3, 256, 256), 32), ((1, 3, 512, 512), 16), ((1, 3, 512, 512), 7)],34)35def test_pad_unpad(shape, padding):36    x = torch.randn(shape)37    x_padded, pad_params = pad_image_tensor(x, pad_size=padding)38    assert x_padded.size(2) % padding == 039    assert x_padded.size(3) % padding == 040    y = unpad_image_tensor(x_padded, pad_params)41    assert (x == y).all()42@pytest.mark.parametrize(["shape", "padding"], [((1, 3, 512, 512), (7, 13))])43def test_pad_unpad_nonsymmetric(shape, padding):44    x = torch.randn(shape)45    x_padded, pad_params = pad_image_tensor(x, pad_size=padding)46    assert x_padded.size(2) % padding[0] == 047    assert x_padded.size(3) % padding[1] == 048    y = unpad_image_tensor(x_padded, pad_params)49    assert (x == y).all()

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.