def run_sim_job(batch_idx, in_dir, out_dir):
in_file = '%s/batch%d_in.csv' % (in_dir, batch_idx)
out_file = 'batch%d_out.csv' % batch_idx
in_data = pd.read_csv(in_file)
run = Samplerun(no_docker=True)
run.perform_sample(out_file=out_file,
out_dir=out_dir,
param_values=in_data)
def test_samplerun():
'''
Sets up uniform generation of data points and csv output.
'''
n_samples = 5
run1 = Samplerun()
run1.perform_sample(out_file='100uniform.csv',
n_samples=n_samples,
domain=Domain(),
sampling_strategy=UniformSamplingStrategy())
def test_runfromfile():
'''
Sets up uniform generation of data points and csv output.
'''
n_samples = 100
run1 = Samplerun()
run1.perform_sample(
out_file='100fix0000000out.csv',
n_samples=n_samples,
param_values=pd.read_csv("params/100params0000000.csv"),
domain=Domain(),
sampling_strategy=UniformSamplingStrategy())
def query_tbr(self):
run = Samplerun(no_docker=True)
sampled = run.perform_sample(
out_file=None,
param_values=self.tbr_params,
progress_handler=self.progress_handler)
param_names = [self.x_param_name, self.y_param_name]
sampled = pd.merge(self.tbr_params,
sampled,
how='left',
left_on=param_names,
right_on=param_names,
suffixes=('', '_dup_'))
sampled.drop([
column
for column in sampled.columns if column.endswith('_dup_')
],
axis=1,
inplace=True)
self.samples_available.emit(sampled)
self.finished.emit()
def theory_TBR(params, domain, n_samples=1):
run = Samplerun()
return run.perform_sample(domain=domain,
n_samples=n_samples,
param_values=params)