How to use test_ons method in keyboard

Best Python code snippet using keyboard

lightning_task.py

Source:lightning_task.py

1import torch2import numpy as np3from torch.utils.data import Dataset, DataLoader4import matplotlib.pyplot as plt5import pytorch_lightning as pl6import os7AVAIL_GPUS = min(1, torch.cuda.device_count())8BATCH_SIZE = 256 if AVAIL_GPUS else 649NUM_WORKERS = int(os.cpu_count() / 2)10class DMTS_Dataset(torch.utils.data.Dataset):11    'Characterizes a delay-match to sample task'12    def __init__(self, inps,out_des,list_IDs, labels,test_ons, distracted_bools):13        'Initialization'14        self.labels = labels15        self.list_IDs = list_IDs16        self.inps = inps17        self.out_des = out_des18        self.test_ons = test_ons19        self.distracted_bools = distracted_bools20    def __len__(self):21        'Denotes the total number of samples'22        return len(self.list_IDs)23    def __getitem__(self, index):24        'Generates one sample of data'25        # Select sample26        ID = self.list_IDs[index]27        # Load data and get label28        inp = self.inps[ID]29        y = self.labels[ID].long()30        out = self.out_des[ID]31        test_on = self.test_ons[ID]32        distracted_bool = self.distracted_bools[ID]33        34        return inp,out,y,test_on,distracted_bool35def generate_one_DMTS_IO(sample_mat, samp, noise_level, dt,alpha, time_limits = [-1,5.5], possible_delays = [1.  , 1.41, 2.  , 2.83, 4.  ], use_distractor = 1):36    37    num_samples = len(sample_mat)38    39    distractor1 = np.random.choice(np.delete(np.arange(num_samples),samp))40    41    #by convention the 8th and 9th samples are possible mid-delay distractors42    mid_delay_distractor_ind = np.random.choice([7,8])43    delay_length = 1000*np.random.choice(possible_delays)44    45    tvec = np.arange(time_limits[0],time_limits[1],1/1000)46    TIME_STEPS = len(tvec)47    48    49    inp = torch.zeros((TIME_STEPS,num_samples+1))   50    out_des = torch.zeros((TIME_STEPS,num_samples+1))     51    samp_on = np.argmin(np.abs(tvec-0))52    samp_off = samp_on + 50053    54    dis_on = int(delay_length/2) + samp_off55    dis_off  = dis_on + 25056    test_on = samp_off + int(delay_length)57    test_off = test_on + 500  58    59    #present sample60    inp[samp_on:samp_off,:-1] = sample_mat[samp]61    62    63    64    if np.heaviside(np.random.rand()-0.5,0):65        #present distractor on 50% of trials66        inp[dis_on:dis_off,:-1] = sample_mat[mid_delay_distractor_ind]67        distracted_bool = 168    else:69        distracted_bool = 070    71    #present test and sample72    inp[test_on:test_off,:-1] = sample_mat[samp] + sample_mat[distractor1]73    74    #fixate signal75    inp[0:test_on,-1] = 1 # fixation signal, answer when it goes off    76    77    #desired output78    #out_des[int(3000/dt):int(3500/dt),samp] = 179    out_des[test_on:,samp] = 180    out_des[0:test_on,-1] = 081    82    inp += np.sqrt(2/alpha)*noise_level*torch.randn(inp.shape)83    84    85    return inp[::dt] , out_des[::dt], int(test_on/dt), distracted_bool86def generate_DMTS(dt = 100, tau = 100, time_limits = [-1,5.5], num_samples = 8+2, variable_delay = True, mid_delay_distractor = True):87    88    """"89   Generates one delayed-match to sample dataset. 90    91    ARGS:92        -dt: timestep to use93        -time_limits: beginning and end of trial. sample On = 0s. Units of s. 94        -num_samples: size of sample pool95        96    97    98    RETURNS:99        -inps: inputs into the rnn, size batch by time by num_samples + 1100        -out_des: desired outputs from the rnn, size batch by time by num_samples + 1101        -partition: training and testing IDs102        -labels: sample label for each element in dataset103    """   104    #use binary encoding of sample images105    sample_mat = torch.eye( num_samples )106    107    TIME_STEPS = len(np.arange(time_limits[0],time_limits[1], dt/1000))108        109    noise_level = 0.01110    111    num_train = int(2**14)#int(0.6*0.5*(2**13))112    num_test = int(2**12)#int(0.4*0.5*(2**13))113    num_val = int(2**12)114    num_examples = num_train + num_test + num_val #int(0.5*(2**10))115    inps = torch.empty((num_examples,TIME_STEPS, num_samples+1), requires_grad = False)116    out_des = torch.empty((num_examples,TIME_STEPS, num_samples+1),requires_grad = False)117    test_ons = torch.empty((num_examples),requires_grad = False)118    distracted_bools = torch.empty((num_examples),requires_grad = False)119    120    labels = torch.empty(num_examples)121    list_IDs = []122    for i in range(num_examples):123        samp = np.random.choice(np.arange(num_samples-2))   124        inps[i],out_des[i],test_ons[i],distracted_bools[i] = generate_one_DMTS_IO(sample_mat = sample_mat, samp = samp, noise_level = noise_level, dt = dt, alpha = dt/tau,time_limits = time_limits)125        labels[i] = samp126        list_IDs.append(i)127    partition = {'train': list_IDs[:num_train], 'test': list_IDs[num_train: num_train+ num_test],'val':list_IDs[num_train + num_test: num_train+ num_test+ num_val]}128    129    return inps,out_des,partition,labels,test_ons,distracted_bools130def get_DMTS_training_test_val_sets(dt_ann,train_batch = 256,test_batch = 256,val_batch = 2**12):131    132    inps,out_des,partition,labels,test_ons,distracted_bools = generate_DMTS(dt = dt_ann)133    training_set = DMTS_Dataset(inps,out_des,partition['train'],labels,test_ons,distracted_bools)134    test_set = DMTS_Dataset(inps,out_des,partition['test'],labels,test_ons, distracted_bools)135    validation_set = DMTS_Dataset(inps,out_des,partition['val'],labels,test_ons, distracted_bools)136    137    138    return training_set,test_set,validation_set139class dDMTSDataModule(pl.LightningDataModule):140    def __init__(self, dt_ann: int = 15, batch_size: int = 256):141        super().__init__()142        self.dt_ann = dt_ann  143        self.batch_size = batch_size144            145    def setup(self,stage = None):146        self.training_data,self.test_data,self.validation_data = get_DMTS_training_test_val_sets(self.dt_ann)147    def train_dataloader(self):148        return DataLoader(self.training_data, batch_size=BATCH_SIZE, shuffle=True,num_workers = NUM_WORKERS,drop_last = True,pin_memory=False)149    def val_dataloader(self):150        return DataLoader(self.test_data, batch_size=2**11,num_workers = NUM_WORKERS, drop_last = True, pin_memory=False)151    def test_dataloader(self):...

main.py

Source:main.py

1def read_line(line):2    on = line.split()[0] == "on"  # True if on, False if off3    xyz = []4    for val in line.split()[1].split(","):5        xyz.append([int(i) for i in val.split("=")[1].split("..")])6    return on, xyz7def read_input(filename):8    ons = []9    xyzs = []10    with open(filename, "r") as fid:11        for line in fid.readlines():12            on, xyz = read_line(line.strip())13            ons.append(on)14            xyzs.append(xyz)15    return ons, xyzs16def on_cubes(ons, xyzs):17    output = set()18    for i in range(len(ons)):19        on = ons[i]20        xyz = xyzs[i]21        for x in range(max(xyz[0][0], -50), min(xyz[0][1], 50) + 1):22            for y in range(max(xyz[1][0], -50), min(xyz[1][1], 50) + 1):23                for z in range(max(xyz[2][0], -50), min(xyz[2][1], 50) + 1):24                    if on:25                        output.add((x, y, z))26                    else:27                        output.discard((x, y, z))28    return len(output)29def volume(xyz):30    return (31        (xyz[0][1] - xyz[0][0] + 1)32        * (xyz[1][1] - xyz[1][0] + 1)33        * (xyz[2][1] - xyz[2][0] + 1)34    )35def overlap(xyza, xyzb):36    xyzc = []37    for dim in range(3):38        if xyza[dim][0] <= xyzb[dim][0] <= xyza[dim][1]:39            xyzc.append([xyzb[dim][0], min(xyza[dim][1], xyzb[dim][1])])40        elif xyza[dim][0] <= xyzb[dim][1] <= xyza[dim][1]:41            xyzc.append([xyza[dim][0], min(xyza[dim][1], xyzb[dim][1])])42        elif xyzb[dim][0] < xyza[dim][0] and xyzb[dim][1] > xyza[dim][1]:43            xyzc.append(xyza[dim])44        else:45            return []46    return xyzc47def split(xyz, overlap):48    new_xyzs = []49    new_xlims = [xyz[0][0], xyz[0][1]]50    if xyz[0][0] < overlap[0][0]:51        new_xyzs.append([[xyz[0][0], overlap[0][0] - 1], xyz[1], xyz[2]])52        new_xlims[0] = overlap[0][0]53    if overlap[0][1] < xyz[0][1]:54        new_xyzs.append([[overlap[0][1] + 1, xyz[0][1]], xyz[1], xyz[2]])55        new_xlims[1] = overlap[0][1]56    new_ylims = [xyz[1][0], xyz[1][1]]57    if xyz[1][0] < overlap[1][0]:58        new_xyzs.append([new_xlims, [xyz[1][0], overlap[1][0] - 1], xyz[2]])59        new_ylims[0] = overlap[1][0]60    if overlap[1][1] < xyz[1][1]:61        new_xyzs.append([new_xlims, [overlap[1][1] + 1, xyz[1][1]], xyz[2]])62        new_ylims[1] = overlap[1][1]63    if xyz[2][0] < overlap[2][0]:64        new_xyzs.append([new_xlims, new_ylims, [xyz[2][0], overlap[2][0] - 1]])65    if overlap[2][1] < xyz[2][1]:66        new_xyzs.append([new_xlims, new_ylims, [overlap[2][1] + 1, xyz[2][1]]])67    return new_xyzs68def maybe_better_on_cubes(ons, xyzs):69    on_cubes = []70    for i in range(len(ons)):71        on = ons[i]72        xyz = xyzs[i]73        if on:74            on_cubes.append(xyz)75        else:76            add = []77            remove = []78            for cube in on_cubes:79                cube_overlap = overlap(xyz, cube)80                if cube_overlap:81                    remove.append(cube)82                    add += split(cube, cube_overlap)83            for cube in remove:84                on_cubes.remove(cube)85            on_cubes += add86    final_cubes = []87    while on_cubes:88        xyz = on_cubes[0]89        on_cubes = on_cubes[1:]90        for fc in final_cubes:91            cube_overlap = overlap(xyz, fc)92            if cube_overlap:93                on_cubes += split(xyz, fc)94                break95        else:96            final_cubes.append(xyz)97    return sum([volume(cube) for cube in final_cubes])98if __name__ == "__main__":99    test_ons, test_xyzs = read_input("test_input.txt")100    assert on_cubes(test_ons, test_xyzs) == 590784101    ons, xyzs = read_input("input.txt")102    print(on_cubes(ons, xyzs))103    assert maybe_better_on_cubes(test_ons, test_xyzs) == 590784104    test_ons, test_xyzs = read_input("test_input_2.txt")105    assert maybe_better_on_cubes(test_ons, test_xyzs) == 2758514936282235...

test_ons.py

Source:test_ons.py

1import tigercontrol2from tigercontrol.utils.optimizers.ons import ONS3from tigercontrol.utils.optimizers.losses import mse4import matplotlib.pyplot as plt5def test_ons(show=False):6    environment = tigercontrol.environment('LDS')7    x = environment.reset(p=2,q=0)8    controller = tigercontrol.controllers('LSTM')9    controller.initialize(n=1, m=1, l=5, h=10, optimizer=ONS) # initialize with class10    controller.predict(1.0) # call controllers to verify it works11    controller.update(1.0)12    optimizer = ONS(learning_rate=0.001)13    controller = tigercontrol.controllers('LSTM')14    controller.initialize(n=1, m=1, l=3, h=10, optimizer=optimizer) # reinitialize with instance15    loss = []16    for t in range(1000):17        y_pred = controller.predict(x)18        y_true = environment.step()19        loss.append(mse(y_pred, y_true))20        controller.update(y_true)21        x = y_true22    if show:23        plt.plot(loss)24        plt.show(block=False)25        plt.pause(3)26        plt.close()27    print("test_ons passed")28if __name__ == "__main__":...

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.