def plot_helper(gt_data_frame, est_data_frame, id_gt, id_est, title, y_label):
''' Helper function for plotting forces/moments/cop.
'''
for i in range(3):
a = np.array(est_data_frame.iloc[:, i + id_est]) # estimate
b = np.array(gt_data_frame.iloc[:, i + id_gt]) # ground-truth
# compute error
err = round(
mean_squared_error(b[~np.isnan(a)], a[~np.isnan(a)],
squared=False), 2)
# plot
fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(5, 4))
ax.plot(gt_data_frame.time, b, label='Measured')
ax.plot(est_data_frame.time, a, '--', label='Predicted')
if gait_cycle:
ax.set_xlabel('gait cycle (%)')
else:
ax.set_xlabel('Time (seconds)')
ax.set_ylabel(y_label)
ax.set_title(title[i])
ax.legend(loc='lower right')
annotate_plot(ax, 'RMSE = ' + str(err))
fig.patch.set_alpha(1)
fig.tight_layout()
pdf.savefig(fig)
plt.close()
scale * reconstructed_markers.iloc[:, i],
label='Reconstructed',
linestyle='--')
ax.axvspan(occlusion_start_perc,
occlusion_stop_perc,
label="Occlusion Period",
color="tab:blue",
alpha=0.3)
if gait_cycle:
ax.set_xlabel('gait cycle (%)')
else:
ax.set_xlabel('time (s)')
ax.set_ylabel('position' + position_unit)
ax.set_title(original_markers.columns[j])
annotate_plot(ax, 'RMSE = ' + str(d_q))
ax.legend(loc='lower left')
fig.tight_layout()
pdf.savefig(fig)
plt.close()
print('d_q: μ = ', np.round(np.mean(d_q_total), 3), ' σ = ',
np.round(np.std(d_q_total, ddof=1), 3))
with open(results_dir + '_metrics.txt', 'w') as file_handle:
file_handle.write('RMSE\n')
file_handle.write('\td_q: μ = ' + str(np.round(np.mean(d_q_total), 3)) +
' σ = ' + str(np.round(np.std(d_q_total, ddof=1), 3)))
# %%
units = ' (N)'
fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(5, 4))
ax.plot(tau_reference.time, tau_reference.iloc[:, i],
label='OpenSim ID')
ax.plot(tau_rt.time, tau_rt.iloc[:, j], label='Real-time ID',
linestyle='--')
if gait_cycle:
ax.set_xlabel('gait cycle (%)')
else:
ax.set_xlabel('time (s)')
ax.set_ylabel('generalized forces' + units)
ax.set_title(tau_rt.columns[j])
annotate_plot(ax, 'RMSE = ' + str(d_tau))
ax.legend(loc='lower left')
fig.tight_layout()
pdf.savefig(fig)
plt.close()
print('d_tau: μ = ', np.round(np.mean(d_tau_total), 3),
' σ = ', np.round(np.std(d_tau_total, ddof=1), 3))
with open(output_dir + 'metrics.txt', 'w') as file_handle:
file_handle.write('RMSE\n')
file_handle.write('\td_q: μ = ' + str(np.round(np.mean(d_tau_total), 3)) +
' σ = ' + str(np.round(np.std(d_tau_total, ddof=1), 3)))
# %%
#plt.xticks(timepoints, timepoints, rotation=45)
if options.title:
plt.title("%s (%s)" % (options.title,metabolite))
else:
plt.title(metabolite)
plt.gca().xaxis.set_label_text(options.xlabel)
plt.gca().yaxis.set_label_text(options.ylabel)
# Add some padding either side of graphs
#plt.xlim( ind[0]-1, ind[-1]+1)
if options.annotate:
utils.annotate_plot(plt,options)
fig = plt.gcf()
# Horizontal axis through zero
#plt.axhline(0, color='k')
if options.multiplot: #Scale the multiplots a bit more reasonably
fig.set_size_inches(5 + len(metabolites)*3,6)
else:
fig.set_size_inches(8,6)
# Get ylimits for significance bars
if options.ylim:
ylim=options.ylim.split(',')
plt.ylim( int(ylim[0]), int(ylim[1]) )
if args.title:
plt.title("%s (%s)" % (args.title, metabolite))
else:
plt.title(metabolite)
plt.gca().set_xticklabels(graph['timepoints'], minor=False, rotation=45)
plt.gca().xaxis.set_label_text(args.xlabel)
plt.gca().yaxis.set_label_text(args.ylabel)
# Add some padding either side of graphs
#plt.xlim( ind[0]-1, ind[-1]+1)
if args.annotate:
utils.annotate_plot(plt, args)
fig = plt.gcf()
# Horizontal axis through zero
#plt.axhline(0, color='k')
if args.multiplot: # Scale the multiplots a bit more reasonably
fig.set_size_inches(5 + len(metabolites) * 3, 6)
else:
fig.set_size_inches(8, 6)
# Get ylimits for significance bars
if args.ylim:
ylim = args.ylim.split(',')
plt.ylim(int(ylim[0]), int(ylim[1]))
fig, ax = plt.subplots(nrows=1, ncols=3, figsize=(13, 4))
ax[0].plot(q_reference.time, q_reference.iloc[:, i], label='OpenSim')
ax[0].plot(q_filtered.time,
scale * q_filtered.iloc[:, j],
label='Proposed filter',
linestyle='--')
ax[0].plot(q_filtered_sp.time,
scale * q_filtered_sp.iloc[:, j],
label='Spatial filter',
linestyle=':')
ax[0].set_xlabel(x_label)
ax[0].set_ylabel('coordinate' + position_unit)
ax[0].set_title(q_reference.columns[i])
annotate_plot(ax[0], 'RMSE = ' + str(d_q))
annotate_plot(ax[0], 'RMSE = ' + str(d_q_sp), 'upper right')
ax[0].legend(loc='lower left')
ax[1].plot(q_dot_reference.time,
q_dot_reference.iloc[:, i],
label='OpenSim')
ax[1].plot(q_dot_filtered.time,
scale * q_dot_filtered.iloc[:, j],
label='Proposed filter',
linestyle='--')
ax[1].plot(q_dot_filtered_sp.time,
scale * q_dot_filtered_sp.iloc[:, j],
label='Spatial filter',
linestyle=':')
ax[1].set_xlabel(x_label)
i])
d_a = rmse_metric(q_ddot_reference.iloc[:, i], q_ddot_filtered.iloc[:,
i])
if not np.isnan(d_q): # NaN when siganl is zero
d_q_total.append(d_q)
d_u_total.append(d_u)
d_a_total.append(d_a)
fig, ax = plt.subplots(nrows=1, ncols=3, figsize=(12, 4))
ax[0].plot(q_reference.time, q_reference.iloc[:, i], label='OpenSim')
ax[0].plot(q_filtered.time, q_filtered.iloc[:, i], label='filtered')
ax[0].set_xlabel('time (s)')
ax[0].set_ylabel('coordinate (deg | m)')
ax[0].set_title(q_reference.columns[i])
annotate_plot(ax[0], 'RMSE = ' + str(d_q))
ax[0].legend(loc='lower left')
ax[1].plot(q_dot_reference.time,
q_dot_reference.iloc[:, i],
label='OpenSim')
ax[1].plot(q_dot_filtered.time,
q_dot_filtered.iloc[:, i],
label='filtered')
ax[1].set_xlabel('time (s)')
ax[1].set_ylabel('speed (deg / s | m / s)')
ax[1].set_title(q_dot_reference.columns[i])
annotate_plot(ax[1], 'RMSE = ' + str(d_u))
# ax[1].legend()
ax[2].plot(q_ddot_reference.time,
plt.xticks(ind, graph['ticks']) #, rotation=45)
if args.title:
plt.title(args.title)
else:
plt.title(metabolite)
plt.gca().xaxis.set_label_text(args.xlabel)
plt.gca().yaxis.set_label_text(args.ylabel)
# Add some padding either side of graphs
plt.xlim(ind[0] - 1, ind[-1] + 1)
if args.annotate:
utils.annotate_plot(plt, options)
fig = plt.gcf()
# Horizontal axis through zero
plt.axhline(0, color='k')
if args.multiplot: # Scale the multiplots a bit more reasonably
fig.set_size_inches(5 + len(metabolites) * 3, 6)
else:
fig.set_size_inches(8, 6)
# Get ylimits for significance bars
ymin, ymax = graph['ylim']
sigs = list()
if args.tests:
if args.title:
plt.title("%s (%s)" % (args.title, metabolite))
else:
plt.title(metabolite)
plt.gca().set_xticklabels(graph['timepoints'], minor=False, rotation=45)
plt.gca().xaxis.set_label_text(args.xlabel)
plt.gca().yaxis.set_label_text(args.ylabel)
# Add some padding either side of graphs
#plt.xlim( ind[0]-1, ind[-1]+1)
if args.annotate:
utils.annotate_plot(plt, args)
fig = plt.gcf()
# Horizontal axis through zero
#plt.axhline(0, color='k')
if args.multiplot: # Scale the multiplots a bit more reasonably
fig.set_size_inches(5 + len(metabolites) * 3, 6)
else:
fig.set_size_inches(8, 6)
# Get ylimits for significance bars
if args.ylim:
ylim = args.ylim.split(',')
plt.ylim(int(ylim[0]), int(ylim[1]))
"Build from source (via GitHub)",
"Windows binaries (.msi file), downloaded from GitHub",
"macOS installer (.pkg file), downloaded from GitHub",
"Ubuntu Personal Package Archive (PPA)",
"Other",
]
frequency = CategoricalDtype(["Often", "Sometimes", "Rarely", "Never"])
# %%
interface_data = data[range(8, 13)]
interface_data.columns = interfaces
n_responses = len(interface_data.index)
ax = (interface_data.count().sort_index() / n_responses).plot.barh()
annotate_plot(
ax,
f"Which interface(s) do you use to access Cantera? {n_responses} Responses",
n_responses,
)
# %%
os_data = data[range(1, 8)]
os_data.columns = operating_systems
n_responses = len(os_data.index)
ax = (os_data.count().sort_index() / n_responses).plot.barh()
annotate_plot(ax,
f"What operating system(s) do you use? {n_responses} Responses",
n_responses)
# %%
install_data = data[range(13, 21)]
install_data.columns = install_methods
fig, ax = plt.subplots(nrows=1, ncols=1, figsize=(5, 4))
ax.plot(tau_reference.time, tau_reference.iloc[:, i],
label='OpenSim')
ax.plot(tau_rt.time, tau_rt.iloc[:, j], label='Proposed filter',
linestyle='--')
ax.plot(tau_rt_sp.time, tau_rt_sp.iloc[:, j], label='Spatial filter',
linestyle=':')
if gait_cycle:
ax.set_xlabel('gait cycle (%)')
else:
ax.set_xlabel('time (s)')
ax.set_ylabel('generalized force' + units)
ax.set_title(tau_rt.columns[j])
annotate_plot(ax, 'RMSE = ' + str(d_tau))
annotate_plot(ax, 'RMSE = ' + str(d_tau_sp), 'upper right')
ax.legend(loc='lower right')
fig.tight_layout()
pdf.savefig(fig)
plt.close()
print('d_tau: μ = ', np.round(np.mean(d_tau_total), 3),
' σ = ', np.round(np.std(d_tau_total, ddof=1), 3))
with open(output_dir + 'metrics.txt', 'w') as file_handle:
file_handle.write('RMSE\n')
file_handle.write('\td_q: μ = ' + str(np.round(np.mean(d_tau_total), 3)) +
' σ = ' + str(np.round(np.std(d_tau_total, ddof=1), 3)))