def cma_iter_plot_scatter3d_candidates(all_variables, parameters=None):
candidates = parameters['candidates']
fitness = parameters['fitness']
# Take candidates and fitness and append them
candidates.extend(all_variables['parameters_candidates_array'])
fitness.extend(all_variables['fitness_results'].tolist())
candidates_arr = np.array(candidates)
fitness_arr = np.array(fitness)
parameter_names_sorted = all_variables['parameter_names_sorted']
# Save data
if parameters['dataio'] is not None:
parameters['dataio'].save_variables_default(locals(), ['candidates', 'fitness', 'parameter_names_sorted'])
parameters['dataio'].make_link_output_to_dropbox(dropbox_current_experiment_folder='cmaes_fit_experiment')
# Do plots
# Param 1 vs 2 vs 3
parameters['ax'] = utils.scatter3d(candidates_arr[:, 0], candidates_arr[:, 1], candidates_arr[:, 2], c=np.log(fitness_arr), xlabel=parameter_names_sorted[0], ylabel=parameter_names_sorted[1], zlabel=parameter_names_sorted[2], ax_handle=parameters['ax'])
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_landscape_123_scatter3d_{label}_{unique_id}.pdf')
# Param 1 vs 2 vs 4
parameters['ax'] = utils.scatter3d(candidates_arr[:, 0], candidates_arr[:, 1], candidates_arr[:, 3], c=np.log(fitness_arr), xlabel=parameter_names_sorted[0], ylabel=parameter_names_sorted[1], zlabel=parameter_names_sorted[3], ax_handle=parameters['ax'])
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_landscape_124_scatter3d_{label}_{unique_id}.pdf')
# Param 2 vs 3 vs 4
parameters['ax'] = utils.scatter3d(candidates_arr[:, 1], candidates_arr[:, 2], candidates_arr[:, 3], c=np.log(fitness_arr), xlabel=parameter_names_sorted[1], ylabel=parameter_names_sorted[2], zlabel=parameter_names_sorted[3], ax_handle=parameters['ax'])
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_landscape_234_scatter3d_{label}_{unique_id}.pdf')
def plot_scatter(all_vars, result_dist_to_use_name, title='', log_color=True, downsampling=1, label_file=''):
fig = plt.figure()
ax = Axes3D(fig)
result_dist_to_use = all_vars[result_dist_to_use_name]
if not log_color:
result_dist_to_use = np.exp(result_dist_to_use)
utils.scatter3d(result_parameters_flat[:, 0], result_parameters_flat[:, 1], result_parameters_flat[:, 2], s=size_normal_points, c=np.log(result_dist_to_use), xlabel=parameter_names_sorted[0], ylabel=parameter_names_sorted[1], zlabel=parameter_names_sorted[2], title=title, ax_handle=ax)
best_points_result_dist_to_use = np.argsort(result_dist_to_use)[:nb_best_points]
utils.scatter3d(result_parameters_flat[best_points_result_dist_to_use, 0], result_parameters_flat[best_points_result_dist_to_use, 1], result_parameters_flat[best_points_result_dist_to_use, 2], c='r', s=size_best_points, ax_handle=ax)
print "Best points, %s:" % title
print '\n'.join(['sigma output %.2f, ratio %.2f, sigmax %.2f: %f' % (result_parameters_flat[i, 0], result_parameters_flat[i, 1], result_parameters_flat[i, 2], result_dist_to_use[i]) for i in best_points_result_dist_to_use])
if savefigs:
dataio.save_current_figure('scatter3d_%s%s_{label}_{unique_id}.pdf' % (result_dist_to_use_name, label_file))
if savemovies:
try:
utils.rotate_plot3d(ax, dataio.create_formatted_filename('scatter3d_%s%s_{label}_{unique_id}.mp4' % (result_dist_to_use_name, label_file)), bitrate=8000, min_duration=8)
utils.rotate_plot3d(ax, dataio.create_formatted_filename('scatter3d_%s%s_{label}_{unique_id}.gif' % (result_dist_to_use_name, label_file)), nb_frames=30, min_duration=8)
except Exception:
# Most likely wrong aggregator...
print "failed when creating movies for ", result_dist_to_use_name
ax.view_init(azim=90, elev=10)
dataio.save_current_figure('scatter3d_view2_%s%s_{label}_{unique_id}.pdf' % (result_dist_to_use_name, label_file))
return ax
def _show3d(self,
bmin=None,
bmax=None,
margin=0.2,
N_grid=20,
levels=None,
scatter_rule=None):
if (bmin is None) or (bmax is None):
bmin, bmax = self.gp.get_boundary(margin=margin)
fig = plt.figure()
ax = Axes3D(fig)
fax = (fig, ax)
raise NotImplementedError("under construction: delete rule")
def rule(Y):
idxes_1 = []
idxes_2 = []
for idx in range(len(Y)):
y = Y[idx]
if y > 0.0:
idxes_1.append(idx)
else:
idxes_2.append(idx)
return [(idxes_1, "blue"), (idxes_2, "red")]
utils.scatter3d(rule, self.gp.X, self.gp.Y, fax)
def _show3d(self,
bmin=None,
bmax=None,
margin=0.2,
N_grid=20,
levels=None,
scatter_rule=None):
if (bmin is None) or (bmax is None):
bmin, bmax = self.get_boundary(margin=margin)
fig = plt.figure()
ax = Axes3D(fig)
fax = (fig, ax)
if scatter_rule is None:
## Y -> [(idxes_1, c1), ..., (idxes_n, cn)]
def f(Y):
idxes = range(len(Y))
color = "red"
return [(idxes, color)]
scatter_rule = f
utils.scatter3d(scatter_rule, self.X, self.Y, fax)
def plot_scatter(all_vars, result_dist_to_use_name, title='', log_color=True, downsampling=1, label_file='', mask_outliers=True):
result_dist_to_use = all_vars[result_dist_to_use_name]
result_parameters_flat = all_vars['result_parameters_flat']
# Filter if downsampling
filter_downsampling = np.arange(0, result_dist_to_use.size, downsampling)
result_dist_to_use = result_dist_to_use[filter_downsampling]
result_parameters_flat = result_parameters_flat[filter_downsampling]
if mask_outliers:
result_dist_to_use = mask_outliers_array(result_dist_to_use)
best_points_result_dist_to_use = np.argsort(result_dist_to_use)[:nb_best_points]
# Construct all permutations of 3 parameters, for 3D scatters
params_permutations = set([tuple(np.sort(np.random.choice(result_parameters_flat.shape[-1], 3, replace=False)).tolist()) for i in xrange(1000)])
for param_permut in params_permutations:
fig = plt.figure()
ax = Axes3D(fig)
# One plot per parameter permutation
if log_color:
color_points = np.log(result_dist_to_use)
else:
color_points = result_dist_to_use
utils.scatter3d(result_parameters_flat[:, param_permut[0]], result_parameters_flat[:, param_permut[1]], result_parameters_flat[:, param_permut[2]], s=size_normal_points, c=color_points, xlabel=parameter_names_sorted[param_permut[0]], ylabel=parameter_names_sorted[param_permut[1]], zlabel=parameter_names_sorted[param_permut[2]], title=title, ax_handle=ax)
utils.scatter3d(result_parameters_flat[best_points_result_dist_to_use, param_permut[0]], result_parameters_flat[best_points_result_dist_to_use, param_permut[1]], result_parameters_flat[best_points_result_dist_to_use, param_permut[2]], c='r', s=size_best_points, ax_handle=ax)
if savefigs:
dataio.save_current_figure('scatter3d_%s_%s%s_{label}_{unique_id}.pdf' % (result_dist_to_use_name, '_'.join([parameter_names_sorted[i] for i in param_permut]), label_file))
if savemovies:
try:
utils.rotate_plot3d(ax, dataio.create_formatted_filename('scatter3d_%s_%s%s_{label}_{unique_id}.mp4' % (result_dist_to_use_name, '_'.join([parameter_names_sorted[i] for i in param_permut]), label_file)), bitrate=8000, min_duration=8)
utils.rotate_plot3d(ax, dataio.create_formatted_filename('scatter3d_%s_%s%s_{label}_{unique_id}.gif' % (result_dist_to_use_name, '_'.join([parameter_names_sorted[i] for i in param_permut]), label_file)), nb_frames=30, min_duration=8)
except Exception:
# Most likely wrong aggregator...
print "failed when creating movies for ", result_dist_to_use_name
if False and savefigs:
ax.view_init(azim=90, elev=10)
dataio.save_current_figure('scatter3d_view2_%s_%s%s_{label}_{unique_id}.pdf' % (result_dist_to_use_name, '_'.join([parameter_names_sorted[i] for i in param_permut]), label_file))
# plt.close('all')
print "Parameters: %s" % ', '.join(parameter_names_sorted)
print "Best points, %s:" % title
print '\n'.join([str_best_params(best_i, result_dist_to_use) for best_i in best_points_result_dist_to_use])
def cma_iter_plot_scatter3d_candidates(all_variables, parameters=None):
candidates = parameters['candidates']
fitness = parameters['fitness']
# Take candidates and fitness and append them
candidates.extend(all_variables['parameters_candidates_array'])
fitness.extend(all_variables['fitness_results'].tolist())
candidates_arr = np.array(candidates)
fitness_arr = np.ma.masked_greater(np.array(fitness), 1e8)
parameter_names_sorted = all_variables['parameter_names_sorted']
# Save data
if parameters['dataio'] is not None:
parameters['dataio'].save_variables_default(locals(), ['candidates', 'fitness', 'parameter_names_sorted'])
parameters['dataio'].make_link_output_to_dropbox(dropbox_current_experiment_folder='output_noise')
# Best median parameters for now
best_param_str = ' '.join(["%s: %s" % (param_val[0], np.round(param_val[1], 4)) for param_val in zip(parameter_names_sorted, np.median(candidates_arr[-100:], axis=0))])
print "Best parameters: %s" % best_param_str
# Do plots
ax = utils.scatter3d(candidates_arr[:, 0], candidates_arr[:, 1], candidates_arr[:, 2], c=np.log(fitness_arr), xlabel=all_variables['parameter_names_sorted'][0], ylabel=all_variables['parameter_names_sorted'][1], zlabel=all_variables['parameter_names_sorted'][2], title=best_param_str)
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_landscape_scatter3d_012_{label}_{unique_id}.pdf')
plt.close(ax.get_figure())
ax = utils.scatter3d(candidates_arr[:, 1], candidates_arr[:, 2], candidates_arr[:, 3], c=np.log(fitness_arr), xlabel=all_variables['parameter_names_sorted'][1], ylabel=all_variables['parameter_names_sorted'][2], zlabel=all_variables['parameter_names_sorted'][3], title=best_param_str)
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_landscape_scatter3d_123_{label}_{unique_id}.pdf')
plt.close(ax.get_figure())
ax = utils.scatter3d(candidates_arr[:, 0], candidates_arr[:, 1], candidates_arr[:, 3], c=np.log(fitness_arr), xlabel=all_variables['parameter_names_sorted'][0], ylabel=all_variables['parameter_names_sorted'][1], zlabel=all_variables['parameter_names_sorted'][3], title=best_param_str)
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_landscape_scatter3d_013_{label}_{unique_id}.pdf')
plt.close(ax.get_figure())
f, axes = plt.subplots(2, 1)
axes[0].plot(candidates_arr)
axes[0].set_xlabel('Time')
axes[0].set_ylabel('Variables')
axes[0].set_title('Best params: %s' % best_param_str)
axes[0].legend(all_variables['parameter_names_sorted'])
axes[1].plot(fitness_arr)
axes[1].set_xlabel('Time')
axes[1].set_ylabel('Fitness')
axes[1].set_title('Current fitness: %s' % np.round(np.median(fitness_arr[-100:]), 4))
f.canvas.draw()
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_time_{label}_{unique_id}.pdf')
plt.close(f)
def cma_iter_plot_scatter3d_candidates(all_variables, parameters=None):
candidates = parameters["candidates"]
fitness = parameters["fitness"]
# Take candidates and fitness and append them
candidates.extend(all_variables["parameters_candidates_array"])
fitness.extend(all_variables["fitness_results"].tolist())
candidates_arr = np.array(candidates)
fitness_arr = np.array(fitness)
parameter_names_sorted = all_variables["parameter_names_sorted"]
# Save data
if parameters["dataio"] is not None:
parameters["dataio"].save_variables_default(locals(), ["candidates", "fitness", "parameter_names_sorted"])
parameters["dataio"].make_link_output_to_dropbox(dropbox_current_experiment_folder="output_noise")
# Do plots
parameters["ax"] = utils.scatter3d(
candidates_arr[:, 0],
candidates_arr[:, 1],
candidates_arr[:, 2],
c=np.log(fitness_arr),
xlabel=all_variables["parameter_names_sorted"][0],
ylabel=all_variables["parameter_names_sorted"][1],
zlabel=all_variables["parameter_names_sorted"][2],
ax_handle=parameters["ax"],
)
if parameters["dataio"] is not None:
parameters["dataio"].save_current_figure("cmaes_optim_landscape_scatter3d_{label}_{unique_id}.pdf")
def cma_iter_plot_scatter3d_candidates(all_variables, parameters=None):
candidates = parameters['candidates']
fitness = parameters['fitness']
# Take candidates and fitness and append them
candidates.extend(all_variables['parameters_candidates_array'])
fitness.extend(all_variables['fitness_results'].tolist())
candidates_arr = np.array(candidates)
fitness_arr = np.array(fitness)
parameter_names_sorted = all_variables['parameter_names_sorted']
# Save data
if parameters['dataio'] is not None:
parameters['dataio'].save_variables_default(locals(), ['candidates', 'fitness', 'parameter_names_sorted'])
parameters['dataio'].make_link_output_to_dropbox(dropbox_current_experiment_folder='output_noise')
# Do plots
parameters['ax'] = utils.scatter3d(candidates_arr[:, 0], candidates_arr[:, 1], candidates_arr[:, 2], c=np.log(fitness_arr), xlabel=all_variables['parameter_names_sorted'][0], ylabel=all_variables['parameter_names_sorted'][1], zlabel=all_variables['parameter_names_sorted'][2], ax_handle=parameters['ax'])
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_landscape_scatter3d_{label}_{unique_id}.pdf')
f, ax = plt.subplots()
ax.plot(candidates_arr)
ax.set_xlabel('Time')
ax.set_ylabel('Variables')
ax.set_title('Parameter optimisation')
ax.legend(all_variables['parameter_names_sorted'])
f.canvas.draw()
if parameters['dataio'] is not None:
parameters['dataio'].save_current_figure('cmaes_optim_time_{label}_{unique_id}.pdf')
plt.close(f)
def cma_iter_plot_contourf_candidates(all_variables, parameters=None):
candidates = parameters['candidates']
fitness = parameters['fitness']
# Take candidates and fitness and append them
candidates.extend(all_variables['parameters_candidates_array'])
fitness.extend(all_variables['fitness_results'].tolist())
candidates_arr = np.array(candidates)
fitness_arr = np.array(fitness)
# Do a plot
utils.scatter3d(candidates_arr[:, 0], candidates_arr[:, 1], candidates_arr[:, 2], c=np.log(fitness_arr), xlabel=all_variables['parameter_names_sorted'][0], ylabel=all_variables['parameter_names_sorted'][1], zlabel=all_variables['parameter_names_sorted'][2])
if parameters['dataio'] is not None:
parameters['dataio'].save_variables_default(locals(), ['candidates', 'fitness'])
parameters['dataio'].save_current_figure('cmaes_optim_landscape_scatter3d_{label}_{unique_id}.pdf')
def cma_iter_plot_scatter3d_candidates(all_variables, parameters=None):
candidates = parameters["candidates"]
fitness = parameters["fitness"]
# Take candidates and fitness and append them
candidates.extend(all_variables["parameters_candidates_array"])
fitness.extend(all_variables["fitness_results"].tolist())
candidates_arr = np.array(candidates)
fitness_arr = np.ma.masked_greater(np.array(fitness), 1e8)
parameter_names_sorted = all_variables["parameter_names_sorted"]
# Save data
if parameters["dataio"] is not None:
parameters["dataio"].save_variables_default(locals(), ["candidates", "fitness", "parameter_names_sorted"])
parameters["dataio"].make_link_output_to_dropbox(
dropbox_current_experiment_folder="sigmaoutput_normalisedsigmax_cmaes"
)
# Best median parameters for now
best_param_str = " ".join(
[
"%s: %s" % (param_val[0], np.round(param_val[1], 4))
for param_val in zip(parameter_names_sorted, np.median(candidates_arr[-100:], axis=0))
]
) + " > %f" % np.median(fitness_arr[-100:])
print "Best parameters: %s" % best_param_str
# Do plots
ax = utils.scatter3d(
candidates_arr[:, 0],
candidates_arr[:, 1],
candidates_arr[:, 2],
c=np.log(fitness_arr),
xlabel=all_variables["parameter_names_sorted"][0],
ylabel=all_variables["parameter_names_sorted"][1],
zlabel=all_variables["parameter_names_sorted"][2],
title=best_param_str,
)
if parameters["dataio"] is not None:
parameters["dataio"].save_current_figure("cmaes_optim_landscape_scatter3d_012_{label}_{unique_id}.pdf")
plt.close(ax.get_figure())
ax = utils.scatter3d(
candidates_arr[:, 1],
candidates_arr[:, 2],
candidates_arr[:, 3],
c=np.log(fitness_arr),
xlabel=all_variables["parameter_names_sorted"][1],
ylabel=all_variables["parameter_names_sorted"][2],
zlabel=all_variables["parameter_names_sorted"][3],
title=best_param_str,
)
if parameters["dataio"] is not None:
parameters["dataio"].save_current_figure("cmaes_optim_landscape_scatter3d_123_{label}_{unique_id}.pdf")
plt.close(ax.get_figure())
ax = utils.scatter3d(
candidates_arr[:, 0],
candidates_arr[:, 1],
candidates_arr[:, 3],
c=np.log(fitness_arr),
xlabel=all_variables["parameter_names_sorted"][0],
ylabel=all_variables["parameter_names_sorted"][1],
zlabel=all_variables["parameter_names_sorted"][3],
title=best_param_str,
)
if parameters["dataio"] is not None:
parameters["dataio"].save_current_figure("cmaes_optim_landscape_scatter3d_013_{label}_{unique_id}.pdf")
plt.close(ax.get_figure())
f, axes = plt.subplots(2, 1)
axes[0].plot(candidates_arr)
axes[0].set_xlabel("Time")
axes[0].set_ylabel("Variables")
axes[0].set_title("Best params: %s" % best_param_str)
axes[0].legend(all_variables["parameter_names_sorted"])
axes[1].plot(fitness_arr)
axes[1].set_xlabel("Time")
axes[1].set_ylabel("Fitness")
axes[1].set_title("Current fitness: %s" % np.round(np.median(fitness_arr[-100:]), 4))
f.canvas.draw()
if parameters["dataio"] is not None:
parameters["dataio"].save_current_figure("cmaes_optim_time_{label}_{unique_id}.pdf")
plt.close(f)
# same plot but normalized
f, axes = plt.subplots(2, 1)
axes[0].plot(candidates_arr / np.max(candidates_arr, axis=0))
axes[0].set_xlabel("Time")
axes[0].set_ylabel("Variables")
axes[0].set_title("N Best params: %s" % best_param_str)
axes[0].legend(all_variables["parameter_names_sorted"])
axes[1].plot(fitness_arr)
axes[1].set_xlabel("Time")
axes[1].set_ylabel("Fitness")
axes[1].set_title("N Current fitness: %s" % np.round(utils.nanmedian(fitness_arr[-100:]), 4))
f.canvas.draw()
if parameters["dataio"] is not None:
parameters["dataio"].save_current_figure("cmaes_optim_time_normalized_{label}_{unique_id}.pdf")
plt.close(f)
c = [0.0]
for i, j in indx:
c.append(1.0)
write_spda_file(sdp_filename, m, nblocks, size, c, indx)
os.system("/usr/local/bin/csdp {0} {1}".format(sdp_filename, sol_filename))
y, Z, X = read_sol_file(sol_filename, size=49)
u, s, v = la.svd(X)
Y = u * np.sqrt(s)
pylab.clf()
pylab.plot(s[:10])
save_many('exp_comb/mvu_scree')
pylab.clf()
scatter3d(Y[:, 0], Y[:, 1], Y[:, 2], c=colors)
save_many('exp_comb/mvu')
#
# Isomap
#
X = expand(N)
Y, s = mds(X, 3)
pylab.clf()
pylab.plot(s[:10])
save_many('exp_comb/isomab_scree')
pylab.clf()
scatter3d(Y[:, 0], Y[:, 1], Y[:, 2], c=colors)
c.append(1.0)
write_spda_file(sdp_filename, m, nblocks, size, c, indx)
os.system("/usr/local/bin/csdp {0} {1}".format(sdp_filename,
sol_filename))
y, Z, X = read_sol_file(sol_filename, size=49)
u, s, v = la.svd(X)
Y = u * np.sqrt(s)
pylab.clf()
pylab.plot(s[:10])
save_many('exp_comb/mvu_scree')
pylab.clf()
scatter3d(Y[:, 0], Y[:, 1], Y[:, 2], c=colors)
save_many('exp_comb/mvu')
#
# Isomap
#
X = expand(N)
Y, s = mds(X, 3)
pylab.clf()
pylab.plot(s[:10])
save_many('exp_comb/isomab_scree')
pylab.clf()
scatter3d(Y[:, 0], Y[:, 1], Y[:, 2], c=colors)