Join the 1-day Testing & QA Summit featuring 15+ Expert Speakers.Register for FREE! Join TestMu Conference

How to use t_nl method in avocado

Best Python code snippet using avocado_python

trajectory_plot.py

Source:trajectory_plot.py

1"""2Plot the state and input trajectory.3D. Malyuta -- ACL, University of Washington4B. Acikmese -- ACL, University of Washington5Copyright 2019 University of Washington. All rights reserved.6"""7import pickle8import numpy as np9import numpy.linalg as la10import matplotlib11import matplotlib.pyplot as plt12import progressbar13from csm import ApolloCSM14import tools15matplotlib.rc('font',**{'family':'serif','size':14})16matplotlib.rc('text', usetex=True)17matplotlib.rc('font',**{'family':'serif','size':14})18matplotlib.rc('text', usetex=True)19with open('data/tf_150_tc_2.pkl','rb') as f:20    data = pickle.load(f)21csm = ApolloCSM()22t_f = tools.get_values(data,'solver_input','t_f')[0]23q_f = tools.get_values(data,'solver_input','xf')[0]['q']24t_pulse_max = tools.get_values(data,'solver_input','t_pulse_max')[0]25r_a = tools.get_values(data,'solver_input','r_app')[0]26idx = 027t_sol = tools.get_values(data,'solver_output','optimizer','t')[idx]28p_sol = tools.get_values(data,'solver_output','optimizer','p')[idx]29v_sol = tools.get_values(data,'solver_output','optimizer','v')[idx]30q_sol = tools.get_values(data,'solver_output','optimizer','q')[idx]31omega_sol = tools.get_values(data,'solver_output','optimizer','w')[idx]32u_sol = tools.get_values(data,'solver_output','optimizer','u')[idx]33vc_sol = tools.get_values(data,'solver_output','history','vc')[idx][-1]34t_nl = tools.get_values(data,'solver_output','nl_prop','t')[idx]35p_nl = tools.get_values(data,'solver_output','nl_prop','p')[idx]36v_nl = tools.get_values(data,'solver_output','nl_prop','v')[idx]37q_nl = tools.get_values(data,'solver_output','nl_prop','q')[idx]38omega_nl = tools.get_values(data,'solver_output','nl_prop','w')[idx]39ang_err_sol = np.array([np.rad2deg(2.*np.arccos(max(-1,min(1,tools.qmult(40    tools.qconj(q_sol_k),q_f)[0])))) for q_sol_k in q_sol.T])41ang_err_nl = np.array([np.rad2deg(2.*np.arccos(max(-1,min(1,tools.qmult(42    tools.qconj(q_nl_k),q_f)[0])))) for q_nl_k in q_nl.T])43vc_sol_norm1 = np.array([la.norm(vc_sol[k],ord=1)44                         for k in range(vc_sol.shape[0])])45rpy_sol = np.column_stack([tools.q2rpy(q_sol.T[k])46                           for k in range(q_sol.shape[1])])47rpy_nl = np.column_stack([tools.q2rpy(q_nl.T[k])48                          for k in range(q_nl.shape[1])])49time_app = np.array([t_nl[k] for k in range(p_nl.shape[1])50                     if la.norm(p_nl[:,k]-p_nl[:,-1])<r_a])51# translation52fig = plt.figure(1)53plt.clf()54ax = fig.add_subplot(211)55ax.grid()56ax.plot(t_sol,p_sol[0],color='red',linestyle='none',marker='.',markersize=5)57ax.plot(t_sol,p_sol[1],color='green',linestyle='none',marker='.',markersize=5)58ax.plot(t_sol,p_sol[2],color='blue',linestyle='none',marker='.',markersize=5)59ax.plot(t_nl,p_nl[0],color='red')60ax.plot(t_nl,p_nl[1],color='green')61ax.plot(t_nl,p_nl[2],color='blue')62ax.set_xlabel('Time $t$ [s]')63ax.set_ylabel('Position [m]')64ax.autoscale(tight=True)65y_lim = ax.get_ylim()66x_lim = ax.get_xlim()67ax.fill_between(time_app,68                np.repeat(y_lim[0],time_app.size),69                np.repeat(y_lim[1],time_app.size),70                linewidth=0,color='black',alpha=0.2,zorder=1)71ax.set_xlim(x_lim)72ax.set_ylim(y_lim)73ax = fig.add_subplot(212)74ax.grid()75ax.plot(t_sol,v_sol[0],color='red',linestyle='none',marker='.',markersize=5)76ax.plot(t_sol,v_sol[1],color='green',linestyle='none',marker='.',markersize=5)77ax.plot(t_sol,v_sol[2],color='blue',linestyle='none',marker='.',markersize=5)78ax.plot(t_nl,v_nl[0],color='red')79ax.plot(t_nl,v_nl[1],color='green')80ax.plot(t_nl,v_nl[2],color='blue')81ax.set_xlabel('Time $t$ [s]')82ax.set_ylabel('Velocity [m/s]')83ax.autoscale(tight=True)84plt.tight_layout()85y_lim = ax.get_ylim()86x_lim = ax.get_xlim()87ax.fill_between(time_app,88                np.repeat(y_lim[0],time_app.size),89                np.repeat(y_lim[1],time_app.size),90                linewidth=0,color='black',alpha=0.2,zorder=1)91ax.set_xlim(x_lim)92ax.set_ylim(y_lim)93plt.show(block=False)94fig.savefig('./figures/translation.pdf',95            bbox_inches='tight',format='pdf',transparent=True)96# attitude97fig = plt.figure(2)98plt.clf()99ax = fig.add_subplot(211)100ax.grid()101ax.plot(t_sol,np.unwrap(rpy_sol[0],discont=359.)-360.,color='red',linestyle='none',marker='.',markersize=5)102ax.plot(t_sol,rpy_sol[1],color='green',linestyle='none',marker='.',markersize=5)103ax.plot(t_sol,rpy_sol[2],color='blue',linestyle='none',marker='.',markersize=5)104ax.plot(t_nl,np.unwrap(rpy_nl[0],discont=359.)-360.,color='red')105ax.plot(t_nl,rpy_nl[1],color='green')106ax.plot(t_nl,rpy_nl[2],color='blue')107ax.set_xlabel('Time $t$ [s]')108ax.set_ylabel('Roll, pitch, yaw [$^\circ$]')109ax.autoscale(tight=True)110y_lim = ax.get_ylim()111x_lim = ax.get_xlim()112ax.fill_between(time_app,113                np.repeat(y_lim[0],time_app.size),114                np.repeat(y_lim[1],time_app.size),115                linewidth=0,color='black',alpha=0.2,zorder=1)116ax.set_xlim(x_lim)117ax.set_ylim(y_lim)118ax = fig.add_subplot(212)119ax.grid()120ax.plot(t_sol,np.rad2deg(omega_sol[0]),color='red',linestyle='none',marker='.',markersize=5)121ax.plot(t_sol,np.rad2deg(omega_sol[1]),color='green',linestyle='none',marker='.',markersize=5)122ax.plot(t_sol,np.rad2deg(omega_sol[2]),color='blue',linestyle='none',marker='.',markersize=5)123ax.plot(t_nl,np.rad2deg(omega_nl[0]),color='red')124ax.plot(t_nl,np.rad2deg(omega_nl[1]),color='green')125ax.plot(t_nl,np.rad2deg(omega_nl[2]),color='blue')126ax.set_xlabel('Time $t$ [s]')127ax.set_ylabel('Angular velocity [$^\circ$/s]')128ax.autoscale(tight=True)129plt.tight_layout()130y_lim = ax.get_ylim()131x_lim = ax.get_xlim()132ax.fill_between(time_app,133                np.repeat(y_lim[0],time_app.size),134                np.repeat(y_lim[1],time_app.size),135                linewidth=0,color='black',alpha=0.2,zorder=1)136ax.set_xlim(x_lim)137ax.set_ylim(y_lim)138plt.show(block=False)139fig.savefig('./figures/attitude.pdf',140            bbox_inches='tight',format='pdf',transparent=True)141# RCS thrusts142fig = plt.figure(3,figsize=(9.6,9.86))143plt.clf()144for i in range(4): # quad number145    for j in range(4): # thruster of quad146        k = i*4+j147        ax = fig.add_subplot(4,4,k+1)148        ax.grid()149        ax.plot(t_sol[:-1],u_sol[k]*1e3,color='black',marker='.',150                linestyle='none',markersize=5)151        ax.axhline(y=csm.t_pulse_max*1e3,color='red',linestyle='--',linewidth=1)152        ax.axhline(y=csm.t_pulse_min*1e3,color='blue',linestyle='--',153                   linewidth=1)154        ax.set_xlabel('Time $t$ [s]')155        ax.set_ylabel('Pulse width $\Delta t_k^{%d}$ [ms]'%(k+1))156        ax.set_title('Thruster $i=%d$'%(k+1))157        ax.autoscale(tight=True)158        yticks = (np.ceil(np.linspace(0,0.6,7)*1e3)).astype(int)159        xticks = (np.linspace(0,t_f,3)).astype(int)160        plt.yticks(yticks,yticks)161        plt.xticks(xticks,xticks)162        ax.set_ylim([0,t_pulse_max*1.1*1e3])163        if 'p_f' in csm.i2thruster[4*i+j]:164            y_lim = ax.get_ylim()165            x_lim = ax.get_xlim()166            ax.fill_between(time_app,167                            np.repeat(y_lim[0],time_app.size),168                            np.repeat(y_lim[1],time_app.size),169                            linewidth=0,color='black',alpha=0.2,zorder=1)170            ax.set_xlim(x_lim)171            ax.set_ylim(y_lim)172        173plt.tight_layout()174plt.show(block=False)175fig.savefig('./figures/inputs.pdf',...

MSP.py

Source:MSP.py

1import matplotlib.pyplot as plt2import torch3import torch.nn as nn4import torch.nn.functional as F5class MultiSimilarity(nn.Module):6    def __init__(self):7        super().__init__()8    def forward(self, s_g, t_g):9        # --------------------------------------------10        # Shape of s_g : ((s_nl,), (bs, s_ch, s_h, s_w))11        # Shape of t_g : ((t_nl,), (bs, t_ch, t_h, t_w))12        # --------------------------------------------13        s_nl = len(s_g)14        t_nl = len(t_g)15        s_g_mtx = torch.stack([self.get_sim_matrix(s_f) for s_f in s_g])  # (s_nl, bs, bs)16        t_g_mtx = torch.stack([self.get_sim_matrix(t_f) for t_f in t_g])  # (t_nl, bs, bs)17        s_g_mtx = torch.unsqueeze(s_g_mtx, dim=1).repeat(1, t_nl, 1, 1)  # (s_nl, t_nl, bs, bs)18        t_g_mtx = torch.unsqueeze(t_g_mtx, dim=0).repeat(s_nl, 1, 1, 1)  # (s_nl, t_nl, bs, bs)19        loss = (s_g_mtx - t_g_mtx).pow(2).view(s_nl, -1).mean(1).mean(0)  # (1,)20        return loss21    def get_sim_matrix(self, f, is_at=True):22        # --------------------------------------------23        # Shape of f : (bs, ch, h, w)24        # --------------------------------------------25        if is_at:26            f = self.at(f)27        f = f.view(f.shape[0], -1)  # (bs, ch * h * w)28        f = F.normalize(f, dim=1)  # (bs, ch * h * w)29        mtx = torch.matmul(f, torch.t(f))  # (bs, bs)30        mtx = F.normalize(mtx, dim=1)  # (bs, bs)31        return mtx32    def at(self, f, is_flat=True):33        # --------------------------------------------34        # Shape of f : (bs, ch, h, w)35        # --------------------------------------------36        out = f.pow(2).mean(1)  # (bs, h, w)37        if is_flat:38            out =F.normalize(out.view(f.size(0), -1))  # (bs, h * w)39        return out40class MultiSimilarityPlotter(MultiSimilarity):41    def __init__(self):42        super().__init__()43    def plot(self, s_g, t_g, input, target, n_samp=10):44        # --------------------------------------------45        # Shape of s_g    : ((s_nl,), (bs, s_ch, s_h, s_w))46        # Shape of t_g    : ((t_nl,), (bs, t_ch, t_h, t_w))47        # Shape of input  : (bs, ch, h, w)48        # Shape of target : (bs,)49        # --------------------------------------------50        s_g = [s_f.detach() for s_f in s_g]51        t_g = [t_f.detach() for t_f in t_g]52        input = input.permute(0, 2, 3, 1)  # (bs, h, w, ch)53        s_attn_g = [self.at(s_f, is_flat=False) for s_f in s_g]  # ((s_nl,), (bs, s_h, s_w))54        t_attn_g = [self.at(t_f, is_flat=False) for t_f in t_g]  # ((t_nl,), (bs, t_h, t_w))55        s_sim_g = torch.stack([self.get_sim_matrix(s_f) for s_f in s_g])  # (s_nl, bs, bs)56        t_sim_g = torch.stack([self.get_sim_matrix(t_f) for t_f in t_g])  # (t_nl, bs, bs)57        s_pos_ind_g = torch.argsort(-s_sim_g, dim=2)  # (s_nl, bs, bs)58        t_pos_ind_g = torch.argsort(-t_sim_g, dim=2)  # (t_nl, bs, bs)59        s_neg_ind_g = torch.argsort(s_sim_g, dim=2)  # (s_nl, bs, bs)60        t_neg_ind_g = torch.argsort(t_sim_g, dim=2)  # (t_nl, bs, bs)61        pivot = 062        for i, (t_pos_ind_f, t_neg_ind_f, t_attn_f) in enumerate(zip(t_pos_ind_g, t_neg_ind_g, t_attn_g)):63            pos_ind = t_pos_ind_f[pivot, :n_samp]  # (n_samp,)64            neg_ind = t_neg_ind_f[pivot, :n_samp]  # (n_samp,)65            r_pos_imgs = input[pos_ind]  # (n_samp, h, w, ch). Sampled raw images66            r_neg_imgs = input[neg_ind]  # (n_samp, h, w, ch). Sampled raw images67            t_pos_attns = t_attn_f[pos_ind]  # (n_samp, t_h, t_w)68            t_neg_attns = t_attn_f[neg_ind]  # (n_samp, t_h, t_w)69            fig, axs = plt.subplots(4, n_samp)  # 4 rows and n_samp columns70            for j in range(n_samp):71                axs[0, j].imshow(r_pos_imgs[j], interpolation="bicubic")72                axs[1, j].imshow(t_pos_attns[j],  cmap="jet")73                axs[2, j].imshow(r_neg_imgs[j], interpolation="bicubic")74                axs[3, j].imshow(t_neg_attns[j],  cmap="jet")...

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.