def output_results(em):
def add_to_zip(mat, filename, zf):
numpy.savetxt('/tmp/' + filename,mat,delimiter=',')
zf.write('/tmp/' + filename, filename)
os.remove('/tmp/' + filename)
experiment_name = em.experiment_name
self.em.gm = em.initialize_generation_manager()
#Open a zip file to put results in
zip_file_name = '/var/www/eigenhand_project/' + experiment_name + '_csv.zip'
zf = zipfile.ZipFile(zip_file_name,'w')
#Create data structures for storing results
e_list = dict()
total_score_array = None
#Iterate over generations calculating the scores
for gen in xrange(em.db_interface.get_max_hand_gen() + 1):
print gen
#Set current generation
em.gm.generation = gen
#get updated grasps for genetaion
grasp_list = em.gm.get_all_grasps()
#get a list of only the top grasps by grasp force optimization score.
top_grasp_dict = experiment_manager.generation_manager.get_top_grasps_by_hand(grasp_list, 10)
top_grasp_list = []
[top_grasp_list.extend(gl) for gl in top_grasp_dict.values()]
#Load the hands for the relevant grasps. This gets only those hands for this generation that were valid enough to produce grasps
hands = em.db_interface.load_hands_for_grasps(grasp_list)
em.gm.hands = hands
#Get the numpy matrix of the grasps, top grasps, and hands
grasp_mat = eigenhand_db_tools.output_grasps_mat(grasp_list)
top_grasp_mat = eigenhand_db_tools.output_grasps_mat(top_grasp_list)
hand_mat = eigenhand_db_tools.output_hands_mat(em.gm.hands, gen)
#Re analyze the hands to get all of the relevant scores
e_list = examine_database.get_e_list(em.gm, [], em.eval_functor)
score_array = numpy.array([[e_list[k][0], k,e_list[k][1],numpy.sum(e_list[k][2]), numpy.sum(e_list[k][3]), e_list[k][4]] for k in e_list])
valid_score_array = score_array[numpy.nonzero((score_array[:,2] < 500) & (score_array[:,5] < 2))]
try:
if total_score_array is not None:
total_score_array = numpy.vstack([total_score_array, valid_score_array])
else:
total_score_array = valid_score_array
hand_mat = numpy.hstack([hand_mat, score_array[:,2:]])
except:
pdb.set_trace()
#output the matrices to a csv file.
add_to_zip(top_grasp_mat,'%s_top_grasp_gen_%i.csv'%(experiment_name,gen), zf)
add_to_zip(grasp_mat,'%s_grasp_gen_%i.csv'%(experiment_name,gen), zf)
add_to_zip(hand_mat,'%s_hand_gen_%i.csv'%(experiment_name,gen), zf)
#close the zip file
zf.close()
#Write current results to the web site
output_frontend_report(em,'5 object 100000 steps 5 tasks/object 1 atr/gen 10 gen, ', total_score_array)
def output_current_status(self):
filename = '/var/www/eigenhand_project/results'
self.e_list.update(examine_database.get_e_list(self.gm, [], self.eval_functor))
score_array = examine_database.e_list_to_score_array(self.e_list)
examine_database.plot_elist_vs_gen(score_array, filename)
def output_results(em):
def add_to_zip(mat, filename, zf):
numpy.savetxt('/tmp/' + filename, mat, delimiter=',')
zf.write('/tmp/' + filename, filename)
os.remove('/tmp/' + filename)
experiment_name = em.experiment_name
em.initialize_generation_manager()
#Open a zip file to put results in
zip_file_name = '/var/www/eigenhand_project/' + experiment_name + '_csv.zip'
zf = zipfile.ZipFile(zip_file_name, 'w')
#Create data structures for storing results
e_list = dict()
total_score_array = None
#Iterate over generations calculating the scores
for gen in xrange(em.db_interface.get_max_hand_gen() + 1):
print gen
#Set current generation
em.gm.generation = gen
#get updated grasps for genetaion
grasp_list = em.gm.get_all_grasps()
#get a list of only the top grasps by grasp force optimization score.
top_grasp_dict = experiment_manager.generation_manager.get_top_grasps_by_hand(
grasp_list, 10)
top_grasp_list = []
[top_grasp_list.extend(gl) for gl in top_grasp_dict.values()]
#Load the hands for the relevant grasps. This gets only those hands for this generation that were valid enough to produce grasps
hands = em.db_interface.load_hands_for_grasps(grasp_list)
em.gm.hands = hands
#Get the numpy matrix of the grasps, top grasps, and hands
grasp_mat = eigenhand_db_tools.output_grasps_mat(grasp_list)
top_grasp_mat = eigenhand_db_tools.output_grasps_mat(top_grasp_list)
hand_mat = eigenhand_db_tools.output_hands_mat(em.gm.hands, gen)
#Re analyze the hands to get all of the relevant scores
e_list = examine_database.get_e_list(em.gm, [], em.eval_functor)
score_array = numpy.array([[
e_list[k][0], k, e_list[k][1],
numpy.sum(e_list[k][2]),
numpy.sum(e_list[k][3]), e_list[k][4]
] for k in e_list])
valid_score_array = score_array[numpy.nonzero(
(score_array[:, 2] < 500) & (score_array[:, 5] < 2))]
try:
if total_score_array is not None:
total_score_array = numpy.vstack(
[total_score_array, valid_score_array])
else:
total_score_array = valid_score_array
hand_mat = numpy.hstack([hand_mat, score_array[:, 2:]])
except:
pdb.set_trace()
#output the matrices to a csv file.
add_to_zip(top_grasp_mat,
'%s_top_grasp_gen_%i.csv' % (experiment_name, gen), zf)
add_to_zip(grasp_mat, '%s_grasp_gen_%i.csv' % (experiment_name, gen),
zf)
add_to_zip(hand_mat, '%s_hand_gen_%i.csv' % (experiment_name, gen), zf)
#close the zip file
zf.close()
#Write current results to the web site
output_frontend_report(
em, 'weighted threshold ate with one object (drill) 50 generations',
total_score_array)
def output_current_status(self):
filename = '/var/www/eigenhand_project/results'
self.e_list.update(examine_database.get_e_list(self.gm, [], self.eval_functor))
score_array = examine_database.e_list_to_score_array(self.e_list)
examine_database.plot_elist_vs_gen(score_array, filename)
