TestMu 2025 - The Largest Virtual Software Testing Conference is Back | 50,000+ Attendees

How to use test_run_list method in tempest

Best Python code snippet using tempest_python

kamitani_data_handler.py

1from scipy.io import loadmat2import numpy as np3import pandas as pd4import sklearn.preprocessing5from sklearn import preprocessing6class kamitani_data_handler():7    """Generate batches for FMRI prediction8    frames_back - how many video frames to take before FMRI frame9    frames_forward - how many video frames to take after FMRI frame10    """11    def __init__(self, matlab_file ,test_img_csv = 'KamitaniData/imageID_test.csv',train_img_csv = 'KamitaniData/imageID_training.csv',voxel_spacing =3,log = 0 ):12        mat = loadmat(matlab_file)13        self.data = mat['dataSet'][:,3:]14        self.sample_meta = mat['dataSet'][:,:3]15        meta = mat['metaData']16        self.meta_keys = list(l[0] for l in meta[0][0][0][0])17        self.meta_desc = list(l[0] for l in meta[0][0][1][0])18        self.voxel_meta = np.nan_to_num(meta[0][0][2][:,3:])19        test_img_df = pd.read_csv(test_img_csv, header=None)20        train_img_df =pd.read_csv(train_img_csv, header=None)21        self.test_img_id = test_img_df[0].values22        self.train_img_id = train_img_df[0].values23        self.sample_type = {'train':1 , 'test':2 , 'test_imagine' : 3}24        self.voxel_spacing = voxel_spacing25        self.log = log26    def get_meta_field(self,field = 'DataType'):27        index = self.meta_keys.index(field)28        if(index <3): # 3 first keys are sample meta29            return self.sample_meta[:,index]30        else:31            return self.voxel_meta[index]32    def print_meta_desc(self):33        print(self.meta_desc)34    def get_labels(self, imag_data = 0,test_run_list = None):35        le = preprocessing.LabelEncoder()36        img_ids = self.get_meta_field('Label')37        type = self.get_meta_field('DataType')38        train = (type == self.sample_type['train'])39        test = (type == self.sample_type['test'])40        imag = (type == self.sample_type['test_imagine'])41        img_ids_train = img_ids[train]42        img_ids_test = img_ids[test]43        img_ids_imag = img_ids[imag]44        train_labels  = []45        test_labels  =  []46        imag_labels = []47        for id in img_ids_test:48            idx = (np.abs(id - self.test_img_id)).argmin()49            test_labels.append(idx)50        for id in img_ids_train:51            idx = (np.abs(id - self.train_img_id)).argmin()52            train_labels.append(idx)53        for id in img_ids_imag:54            idx = (np.abs(id - self.test_img_id)).argmin()55            imag_labels.append(idx)56        if (test_run_list is not None):57            run = self.get_meta_field('Run')58            test = (self.get_meta_field('DataType') == 2).astype(bool)59            run = run[test]60            select = np.in1d(run, test_run_list)61            test_labels = test_labels[select]62        #imag_labels = le.fit_transform(img_ids_imag)63        if(imag_data):64            return np.array(train_labels), np.array(test_labels), np.array(imag_labels)65        else:66            return np.array(train_labels),np.array(test_labels)67    def get_data(self,normalize =1 ,roi = 'ROI_VC',imag_data = 0,test_run_list = None):   # normalize 0-no, 1- per run , 2- train/test seperatly68        type = self.get_meta_field('DataType')69        train = (type == self.sample_type['train'])70        test = (type == self.sample_type['test'])71        test_imag = (type == self.sample_type['test_imagine'])72        test_all  = np.logical_or(test,test_imag)73        roi_select = self.get_meta_field(roi).astype(bool)74        data = self.data[:,roi_select]75        if(self.log ==1):76            data = np.log(1+np.abs(data))*np.sign(data)77        if(normalize==1):78            run = self.get_meta_field('Run').astype('int')-179            num_runs = np.max(run)+180            data_norm = np.zeros(data.shape)81            for r in range(num_runs):82                data_norm[r==run] = sklearn.preprocessing.scale(data[r==run])83            train_data = data_norm[train]84            test_data  = data_norm[test]85            test_all = data_norm[test_all]86            test_imag = data_norm[test_imag]87        else:88            train_data = data[train]89            test_data  =  data[test]90            if(normalize==2):91                train_data = sklearn.preprocessing.scale(train_data)92                test_data = sklearn.preprocessing.scale(test_data)93        if(self.log ==2):94            train_data = np.log(1+np.abs(train_data))*np.sign(train_data)95            test_data = np.log(1+np.abs(test_data))*np.sign(test_data)96            train_data = sklearn.preprocessing.scale(train_data)97            test_data = sklearn.preprocessing.scale(test_data)98        test_labels =  self.get_labels()[1]99        imag_labels = self.get_labels(1)[2]100        num_labels = max(test_labels)+1101        test_data_avg = np.zeros([num_labels,test_data.shape[1]])102        test_imag_avg = np.zeros([num_labels,test_data.shape[1]])103        if(test_run_list is not None):104            run = self.get_meta_field('Run')105            test = (self.get_meta_field('DataType') == 2).astype(bool)106            run = run[test]107            select = np.in1d(run, test_run_list)108            test_data = test_data[select,:]109            test_labels = test_labels[select]110        for i in range(num_labels):111            test_data_avg[i] = np.mean(test_data[test_labels==i],axis=0)112            test_imag_avg[i] = np.mean(test_imag[imag_labels == i], axis=0)113        if(imag_data):114            return train_data, test_data, test_data_avg,test_imag,test_imag_avg115        else:116            return train_data, test_data, test_data_avg117    def get_voxel_loc(self):118        x = self.get_meta_field('voxel_x')119        y = self.get_meta_field('voxel_y')120        z = self.get_meta_field('voxel_z')121        dim = [int(x.max() -x.min()+1),int(y.max() -y.min()+1), int(z.max() -z.min()+1)]...

main.py

Source: main.py

1from agent import *2training_eps=100003test_runs=54max_len=10005testing_steps=[]6ag=Agent()7for episode in range(training_eps):8  w=World(0,0)9  for step in range(max_len):10    a=ag.e_greedy(w.pos,w.vel)11    nextp,nextv,R=w.move(a)12    a_next=ag.e_greedy(nextp,nextv)13    ag.update(w.pos,w.vel,a,R,nextp,nextv,a_next)14    w.pos=nextp15    w.vel=nextv16    if(R==0):17      break18    elif(w.pos<=w.pmin):19      w.vel=0 # the car bumps to a stop 20  """21  running agent after each simulation22  multiple runs to get a better estimate of agent performance at this episode23  test_run_list stores the corresponding "test_run" number of values24  """25  test_run_list=[]26  for run in range(test_runs):27    w=World(0,0)28    for step in range(max_len):29      a=ag.greedy(w.pos,w.vel)30      nextp,nextv,R=w.move(a)31      a_next=ag.greedy(nextp,nextv)32      """33      The next command is required even in test case because the agent must learn to move in the environment every time it interacts with it. If it encounters a state upon which it is not trained well the agent would fail repeatedly across test runs unless we update it between runs.34      """35      ag.update(w.pos,w.vel,a,R,nextp,nextv,a_next) 36      w.pos=nextp37      w.vel=nextv38      if(R==0):39        break40      elif(w.pos<w.pmin):41        w.vel=042    test_run_list.append(step)43  testing_steps.append(mean(test_run_list))44  if(episode%1000==0):45    print('Done episode {}'.format(episode))    46# save agent47f='agent.sav'48file=open(f,'wb')49pickle.dump(ag,file)50file.close()51# plot peformance of agent with training episodes52plt.plot(range(1,1+len(testing_steps)),testing_steps,label='No. of steps')53plt.title('No. of steps to reach terminal states vs. training episodes')54plt.xlabel('Training Episodes')55plt.ylabel('No. of steps')56plt.legend()...

Blogs

Check out the latest blogs from LambdaTest on this topic:

Are Agile Self-Managing Teams Realistic with Layered Management?

Agile software development stems from a philosophy that being agile means creating and responding to change swiftly. Agile means having the ability to adapt and respond to change without dissolving into chaos. Being Agile involves teamwork built on diverse capabilities, skills, and talents. Team members include both the business and software development sides working together to produce working software that meets or exceeds customer expectations continuously.

How To Automate Mouse Clicks With Selenium Python

Sometimes, in our test code, we need to handle actions that apparently could not be done automatically. For example, some mouse actions such as context click, double click, drag and drop, mouse movements, and some special key down and key up actions. These specific actions could be crucial depending on the project context.

How To Write End-To-End Tests Using Cypress App Actions

When I started writing tests with Cypress, I was always going to use the user interface to interact and change the application’s state when running tests.

How To Handle Dynamic Dropdowns In Selenium WebDriver With Java

Joseph, who has been working as a Quality Engineer, was assigned to perform web automation for the company’s website.

Fluent Interface Design Pattern in Automation Testing

Recently, I was going through some of the design patterns in Java by reading the book Head First Design Patterns by Eric Freeman, Elisabeth Robson, Bert Bates, and Kathy Sierra.

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.