Test your AI Agents with the all-new Agent to Agent Testing Platform.Learn More

How to use update_samples method in autotest

Best Python code snippet using autotest_python

solver.py

Source:solver.py

1import torch2import numpy as np3import os4import json5import matplotlib.pyplot as plt6from utils.dataloader import torcs_dataset,synthetic_example7from utils.utils import print_n_txt8from core.net import Decoder,Policy,Encoder9from core.transformer.encoder import Transformer_Encoder10from core.loss import INFONCE_loss,BT_loss,mdn_loss,recon_loss11import torch.nn.functional as F12class SOLVER():13    def __init__(self,args):14        self.device = 'cuda'15        self.args= args16        self.state_only=True17        self.path = './res/{}/'.format(args.id)18        try:19            os.mkdir(self.path)20        except:21            pass22        self.lr = args.lr23        self.load_dataset()24        self.load_model()25    def load_dataset(self):26        if self.args.data == 'torcs':27            self.state_samples = 5028            dataset = torcs_dataset(num_traj=self.args.num_traj,state_samples=self.state_samples)29            self.traj_dim = self.args.num_traj*3130            self.a_dim = 231            self.s_dim = 2932            self.dim = 3133            self.z_dim = 534            self.update_samples = 535        elif self.args.data == 'syn':36            self.state_samples = 4037            dataset = synthetic_example(num_traj=self.args.num_traj,fixed_len=self.args.fixed_len,38                            state_samples=self.state_samples)39            self.a_dim = 240            self.s_dim = 241            self.z_dim = 442            if self.state_only:43                self.dim = 244            else:45                self.dim=446            self.traj_dim = self.args.num_traj*self.dim47            self.update_samples = 548        self.dataloader = torch.utils.data.DataLoader(dataset, batch_size=self.args.batch_size,shuffle=True)49    def load_model(self):50        if self.args.encoder_base == 'mlp':51            self.encoder = Encoder(x_dim=int(self.args.num_traj*self.dim),z_dim=self.z_dim).to(self.device)52        elif self.args.encoder_base == 'transformer':53            self.encoder = Transformer_Encoder(x_dim=self.dim,z_dim=self.z_dim).to(self.device)54        self.policy = Policy(s_dim=self.s_dim,a_dim=self.a_dim,z_dim=self.z_dim).to(self.device)55        # self.model = MODEL(self.args,x_dim=int(self.args.num_traj*self.dim),s_dim=self.s_dim,a_dim=self.a_dim,56        #                         z_dim=self.z_dim,k=5,sig_max=None).to(self.device)57        if self.args.policy == 'mlp':58            self.weight = [1,1]59        elif self.args.policy == 'mdn':60            self.weight = [1,0.01]61        self.policy.init_param()62        self.encoder.init_param()63        if self.args.recon_loss:64            self.decoder = Decoder(x_dim=int(self.args.num_traj*self.dim),z_dim=5).to(self.device)65    def train(self):66        with open(self.path+'config.json','w') as jf:67            json.dump(self.args.__dict__, jf, indent=2)68        f =  open(self.path+'logs.txt', 'w') 69        self.loss = []70        self.ploss = []71        self.eloss = []72        if self.args.loss=='simclr':73            ecriterion = INFONCE_loss74        elif self.args.loss == 'BT':75            ecriterion = BT_loss76        if self.args.policy == 'mlp':77            pcriterion = F.mse_loss78        elif self.args.policy == 'mdn':79            pcriterion = mdn_loss80        flag=081        #optimizer = torch.optim.Adam(self.model.parameters(), self.lr,weight_decay=1e-4)82        poptimizer = torch.optim.Adam(self.policy.parameters(), self.lr,weight_decay=1e-4)83        eoptimizer = torch.optim.Adam(self.encoder.parameters(), self.lr,weight_decay=1e-4)84        for e in range(200):85            total_loss = 086            ploss = 087            eloss = 088            for traj_1,traj_2,state_1,action_1,state_2,action_2,_,traj_len in self.dataloader:89                if self.args.encoder_base == 'mlp':90                    traj_1 = traj_1.view(-1,self.traj_dim)91                    traj_2 = traj_2.view(-1,self.traj_dim)92                    sampled_z_1 = self.encoder(traj_1.to(self.device))93                    sampled_z_2 = self.encoder(traj_2.to(self.device))94                else:95                    traj_1 = traj_1.view(-1,self.args.num_traj,self.dim)96                    traj_2 = traj_2.view(-1,self.args.num_traj,self.dim)97                    # traj_1 = torch.cat((torch.zeros(traj_1.size(0),1,self.dim),traj_1),dim=1)98                    # traj_2 = torch.cat((torch.zeros(traj_1.size(0),1,self.dim),traj_2),dim=1)99                    sampled_z_1 = self.encoder(traj_1.to(self.device))#[:,0,:]100                    sampled_z_2 = self.encoder(traj_2.to(self.device))#[:,0,:]101                    index = traj_len.unsqueeze(-1).unsqueeze(-1).repeat(1,self.args.num_traj,self.z_dim).to(self.device)102                    sampled_z_1 = torch.gather(sampled_z_1,dim=1,index=index)[:,0,:]103                    sampled_z_2 = torch.gather(sampled_z_2,dim=1,index=index)[:,0,:]104                encoder_loss = ecriterion(sampled_z_1,sampled_z_2) # encoder loss105                if torch.sum(torch.isnan(sampled_z_1)).item()>0:106                    flag=1107                    break108                eoptimizer.zero_grad()109                encoder_loss.backward()110                eoptimizer.step()111                112                piter = self.state_samples//self.update_samples113                for p in range(piter):114                    sampled_z_1_s = sampled_z_1.detach().unsqueeze(1).repeat(1,self.update_samples,1).to(self.device)115                    sampled_z_2_s = sampled_z_2.detach().unsqueeze(1).repeat(1,self.update_samples,1).to(self.device)116                    state_1_s = state_1[:,p*self.update_samples:(p+1)*self.update_samples,:].to(self.device)117                    state_2_s = state_2[:,p*self.update_samples:(p+1)*self.update_samples,:].to(self.device)118                    action_1_s = action_1[:,p*self.update_samples:(p+1)*self.update_samples,:].to(self.device)119                    #action_1_s = action_1_s.view(-1,self.s_dim)120                    action_2_s = action_2[:,p*self.update_samples:(p+1)*self.update_samples,:].to(self.device)121                    input_1 = torch.cat((sampled_z_1_s,state_1_s),dim=-1).view(-1,self.s_dim+self.z_dim)122                    pred_1 = self.policy(input_1)123                    # print(pred_1[0,:],action_1_s[0,0,:])124                    pred_2 = self.policy(torch.cat((sampled_z_2_s,state_2_s),dim=-1).view(-1,self.s_dim+self.z_dim))125                    policy_loss = pcriterion(pred_1,action_1_s.view(-1,self.a_dim).to(self.device)) \126                                            + pcriterion(pred_2,action_2_s.view(-1,self.a_dim).to(self.device)) # Policy Loss127                    poptimizer.zero_grad()128                    policy_loss.backward()129                    poptimizer.step()130                loss = self.weight[0]*encoder_loss + self.weight[1]*policy_loss/self.update_samples131                if self.args.recon_loss:132                    recon_1 = self.decoder(sampled_z_1)133                    recon_2 = self.decoder(sampled_z_2)134                    loss += 0.1*(recon_loss(traj_1,recon_1) + recon_loss(traj_2,recon_2))135                total_loss += loss136                eloss += encoder_loss137                ploss += policy_loss/self.update_samples138            if flag==1:139                break140            total_loss /= len(self.dataloader)141            eloss /= len(self.dataloader)142            ploss /= len(self.dataloader)143            strtemp = ("Epoch: %d loss: %.3f encoder loss: %.3f policy loss: %.3f")%(e,total_loss,eloss,ploss)144            print_n_txt(_chars=strtemp, _f=f)145            self.loss.append(total_loss)146            self.eloss.append(eloss)147            self.ploss.append(ploss)148        torch.save(self.policy.state_dict(),self.path+'policy_final.pt')149        torch.save(self.encoder.state_dict(),self.path+'encoder_final.pt')150    def plot_loss(self):151        plt.figure(figsize=(10,4))152        plt.subplot(1,3,1)153        plt.title("Total Loss")154        plt.xlabel("Epochs")155        plt.plot(self.loss)156        plt.subplot(1,3,2)157        plt.title("Encoder Loss")158        plt.xlabel("Epochs")159        plt.plot(self.eloss)160        plt.subplot(1,3,3)161        plt.title("Policy Loss")162        plt.xlabel("Epochs")163        plt.plot(self.ploss)...

test_databus.py

Source:test_databus.py

...10		meta = ChannelMeta(name='RPM')11		sample.channel_metas = [meta]12		sample.samples = [SampleValue(1234, meta)]13		14		dataBus.update_samples(sample)15		dataBus.notify_listeners(None)16		17		value = dataBus.getData('RPM')18		self.assertEqual(value, 1234)19	listenerVal0 = None20	def test_listener(self):21		22		def listener(value):23			self.listenerVal0 = value24		sample = Sample()25		26		meta = ChannelMeta(name='RPM')27		sample.channel_metas = [meta]28		sample.samples = [SampleValue(1111, meta)]29		dataBus = DataBus()30		dataBus.addChannelListener('RPM', listener)31		dataBus.update_samples(sample)32		dataBus.notify_listeners(None)33		self.assertEqual(self.listenerVal0, 1111)34	35	listenerVal1 = None36	listenerVal2 = None37	def test_multiple_listeners(self):38		def listener1(value):39			self.listenerVal1 = value40		def listener2(value):41			self.listenerVal2 = value42			43		dataBus = DataBus()44		dataBus.addChannelListener('RPM', listener1)45		dataBus.addChannelListener('RPM', listener2)46		47		sample = Sample()48		meta = ChannelMeta(name='RPM')49		sample.channel_metas = [meta]50		sample.samples = [SampleValue(1111, meta)]51		52		dataBus.update_samples(sample)53		dataBus.notify_listeners(None)54		self.assertEqual(self.listenerVal1, 1111)55		self.assertEqual(self.listenerVal2, 1111)56		57	listenerVal3 = None58	listenerVal4 = None59	def test_mixed_listeners(self):60		def listener3(value):61			self.listenerVal3 = value62		def listener4(value):63			self.listenerVal4 = value64			65			66		sample = Sample()67		metaRpm = ChannelMeta(name='RPM')68		metaEngineTemp = ChannelMeta(name='EngineTemp')69		sample.channel_metas = [metaRpm, metaEngineTemp]70		sample.samples = [SampleValue(1111, metaRpm)]71			72		dataBus = DataBus()73		dataBus.addChannelListener('RPM', listener3)74		dataBus.addChannelListener('EngineTemp', listener4)75		dataBus.update_samples(sample)76		dataBus.notify_listeners(None)77		#ensure we don't set the wrong listener78		self.assertEqual(self.listenerVal3, 1111)79		self.assertEqual(self.listenerVal4, None)80		81		sample.samples = [SampleValue(1111, metaRpm), SampleValue(199, metaEngineTemp)]82		83		dataBus.update_samples(sample)84		dataBus.notify_listeners(None)85		#ensure we don't affect unrelated channels86		self.assertEqual(self.listenerVal3, 1111)87		self.assertEqual(self.listenerVal4, 199)88		89	def test_no_listener(self):90		sample = Sample()91		meta = ChannelMeta(name='EngineTemp')92		sample.channel_metas = [meta]93		sample.samples = [SampleValue(200, meta)]94		95		dataBus = DataBus()96		dataBus.update_samples(sample)97		dataBus.notify_listeners(None)98		#no listener for this channel, should not cause an error99	100	channelMeta = None101	def test_meta_listener(self):102		dataBus = DataBus()103		104		def metaListener(channel):105			self.channelMeta = channel106		metas = ChannelMetaCollection()107		metas.channel_metas = [ChannelMeta(name='RPM')]108		dataBus.addMetaListener(metaListener)109		dataBus.update_channel_meta(metas)110		dataBus.notify_listeners(None)...

06zmq-pub-sub.py

Source:06zmq-pub-sub.py

1from handofcats import as_subcommand2import os3import sys4import zmq5import subprocess6from tinyrpc.protocols.jsonrpc import JSONRPCProtocol7@as_subcommand8def run(*, endpoint: str = "tcp://127.0.0.1:5001"):9    sp = subprocess.Popen([sys.executable, __file__, "server", "--endpoint", endpoint])10    cp = subprocess.Popen([sys.executable, __file__, "client", "--endpoint", endpoint])11    cp.wait()12    sp.terminate()13    print("ok")14@as_subcommand15def server(*, endpoint: str):16    import time17    from random import randrange18    context = zmq.Context()19    publisher = context.socket(zmq.PUB)20    publisher.bind(endpoint)21    message_id = 022    while True:23        zipcode = randrange(10001, 10010)24        temprature = randrange(0, 215) - 8025        relhumidity = randrange(0, 50) + 1026        update = "%05d %d %d %d" % (zipcode, temprature, relhumidity, message_id)27        message_id += 128        print(os.getpid(), update)29        time.sleep(0.5)30        publisher.send(update.encode("utf-8"))31@as_subcommand32def client(*, endpoint: str, zipfilter="10001 "):33    context = zmq.Context()34    subscriber = context.socket(zmq.SUB)35    subscriber.connect(endpoint)36    # receive only message with zipcode being 1000137    subscriber.setsockopt(zmq.SUBSCRIBE, zipfilter.encode("utf-8"))38    update_samples = 1039    for updates in range(update_samples):40        message = subscriber.recv()41        zipcode, temprature, relhumidity, message_id = message.split()42        print(43            "zip:%s, temp:%s, relh:%s, id:%s"44            % (zipcode, temprature, relhumidity, message_id)45        )46        total_temp = float(temprature)47    print(48        "average temprature for zipcode '%s' was '%f'"49        % (zipfilter, total_temp / update_samples)50    )...

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.