def generate_research():
for i in range(20):
researchers.append({
"name": myGenerator.name(),
"mwet_id": str(myGenerator.uuid4())
})
for i in range(5):
groups.append(myGenerator.last_name())
for i in range(10):
adsorptions.append(myGenerator.random_uppercase_letter() +
myGenerator.random_lowercase_letter())
for i in range(20):
polymers.append(myGenerator.random_uppercase_letter() +
str(myGenerator.random_digit_not_null()) +
str(myGenerator.random_digit_not_null()))
for i in range(100):
x = Experiment()
research_projects.append(x)
#Avoid that whole nightmare again
check_for_unprofessional_lang(research_projects[i])
researcher = random.choice(researchers)
research_projects[i].data["researcher"]["name"] = researcher["name"]
research_projects[i].data["researcher"]["mwet_id"] = researcher[
"mwet_id"]
research_projects[i].data["researcher"]["group"] = random.choice(
groups)
research_projects[i].data["researcher"]["institution"] = random.choice(
institutions)
research_projects[i].data["experiment_metadata"][
"gap"] = random.choice(gaps)
research_projects[i].set_polymer(random.choice(polymers))
research_projects[i].set_adsorbing(random.choice(adsorptions))
research_projects[i].data["uid"] = str(uuid.uuid4())
Critic,
'critic_kwargs':
critic_kwargs,
'critic_optim':
torch.optim.Adam,
'critic_optim_kwargs':
critic_optim_kwargs,
'num_iters':
200,
'ep_per_iter':
5,
'log_file':
'./kl/logs/{}/{}/{}_log.npz'.format(env_str, alg.__name__, seed)
}
experiment = Experiment(**experiment_parameters)
experiment.run()
kls = torch.tensor(experiment.results['kl'])
mean = torch.mean(kls).item()
std = torch.std(kls).item()
cvs.append(std / mean)
plt.plot(kls, alpha=0.2, color="blue")
#experiment.plot('./kl/plots/cartpole/{}/{}_'.format(alg.__name__, seed))
# show one example episode as a sanity check
#sample_memory(experiment.env, experiment.actor, 1, True)
print(cvs)
from experiment_class import Experiment
max_doses = [200]
curve_types = ['linear']
experiment_type = 'rate-survival'
n_sims = 10
slopes = [.1,.5,1.5]
# scale = 5 # scale time by two
mut_rate = 0.00005
death_rate = 0.3
# mut_rate = mut_rate/scale
# death_rate = death_rate/scale
options = {'mut_rate':mut_rate,
'death_rate':death_rate,
# 'x_lim':100,
# 'y_lim':2*10**4,
# 'counts_log_scale':True,
'plot':True}
e = Experiment(max_doses=max_doses,
slopes=slopes,
curve_types = curve_types,
experiment_type = experiment_type,
n_sims=n_sims,
population_options=options)
e.run_experiment()
e.plot_barchart()
kl_list.append(max_kl)
print('\nNow running wiht max_kl = {}'.format(max_kl))
# set up experiment
ac_alg_kwargs['max_kl'] = max_kl
experiment_parameters['ac_alg_kwargs'] = ac_alg_kwargs
trial_log_dir = log_dir + 'kl_{0:1.4f}/'.format(max_kl)
try:
os.mkdir(trial_log_dir)
os.mkdir(trial_log_dir + 'plots/')
except:
pass
experiment = Experiment(**experiment_parameters)
# run mutliple experiments with different seeds
mult_exp = mult_seed_exp(experiment_parameters, seeds, trial_log_dir)
mult_exp.run()
try:
mult_exp.plot(trial_log_dir + 'plots/')
except:
pass
# get results
mean_results, std_results = mult_exp.get_mean_results()
print(
'Mean return at termination with max_kl {0:1.5f} was {1:1.2f}'.format(
from experiment_class import Experiment
import seaborn as sns
import matplotlib.pyplot as plt
# max_doses = [100,300,400,500]
max_doses = [350]
experiment_type = 'dose-entropy'
n_sims = 100
e1 = Experiment(max_doses=max_doses,
experiment_type=experiment_type,
n_sims=n_sims)
e1.run_experiment()
e = e1.entropy_results
# fig,ax = plt.subplots()
# sns.swarmplot(x='dose',y='max entropy',data=e,ax=ax,hue='survive condition',dodge=True,color='black')
# sns.boxplot(x='dose',y='max entropy',data=e,ax=ax,hue='survive condition',dodge=True,palette='Set2')
# handles, labels = ax.get_legend_handles_labels()
# ax.legend(handles[:2], labels[:2])
# population_options=options)
# e.run_experiment()
# e.plot_barchart()
# experiment_type = 'inoculant-survival'
# options = {'n_gen':1000,'v2':True,'max_dose':400,'max_cells':10**7}
# inoculants = [10**3,10**4,10**5,10**6]
# curve_types=['constant','linear','pharm']
# n_sims = 10
# e = Experiment(experiment_type = experiment_type,
# inoculants=inoculants,
# curve_types=curve_types,
# n_sims=n_sims,
# population_options=options)
# e.run_experiment()
# e.plot_barchart()
experiment_type = 'inoculant-survival'
options = {'n_gen':1000,'v2':True,'max_dose':800,'max_cells':10**7}
inoculants = [10**3,10**4,10**5,10**6]
curve_types=['constant','linear','pharm']
n_sims = 10
e = Experiment(experiment_type = experiment_type,
inoculants=inoculants,
curve_types=curve_types,
n_sims=n_sims,
population_options=options)
e.run_experiment()
e.plot_barchart()
from experiment_class import Experiment
# import matplotlib.pyplot as plt
options = {
'n_impulse': 20,
'k_abs': 0.04,
'k_elim': 0.03,
'max_dose': 150,
'n_gen': 1000,
'mut_rate': 0.00005,
'death_rate': 0.3
}
e = Experiment(experiment_type='drug-regimen',
n_sims=10,
prob_drops=[0, .4, .7],
population_options=options,
debug=True)
e.run_experiment()
e.plot_barchart()
e.save_images()
# import matplotlib.pyplot as plt
# fig,ax = plt.subplots()
# ax.bar([0,1,2],[10,20,30])
# ax.set_xticks([0,1,2])
# ax.set_xticklabels(['a','b','c'])