diff --git a/examples/umbridge_tsunamitutorial/example_umbridge.py b/examples/umbridge_tsunamitutorial/example_umbridge.py
deleted file mode 100644
index b83175cf7f9ab40ab1800e42f625f65e96ede59d..0000000000000000000000000000000000000000
--- a/examples/umbridge_tsunamitutorial/example_umbridge.py
+++ /dev/null
@@ -1,183 +0,0 @@
-# -*- coding: utf-8 -*-
-"""
-Created on Mon Dec 11 14:03:23 2023
-
-Example and test for using umbridge with bayesvalidrox.
-Example model is the tsunami model by
-docker run -it -p 4242:4242 -v ~/tsunami_output:/output linusseelinger/model-exahype-tsunami
-
-@author: rkohlhaas
-"""
-
-import joblib
-import numpy as np
-import umbridge
-
-import bayesvalidrox as bv
-
-
-if __name__ == '__main__':
-
- # This model has 2 inputs and four outputs
- n_prior_sample = 1000
- priors = bv.Input()
- priors.add_marginals()
- priors.Marginals[0].name = 'x'
- priors.Marginals[0].input_data = np.random.uniform(50,150,n_prior_sample)
- priors.add_marginals()
- priors.Marginals[1].name = 'y'
- priors.Marginals[1].input_data = np.random.uniform(50,150,n_prior_sample)
-
- # Define the model - level 0
- model0 = bv.PyLinkForwardModel()
- #model.link_type = 'Function'
- model0.link_type = 'umbridge'
- model0.py_file = 'tsunami_model'
- model0.name = 'tsunami_model'
- model0.Output.names = ['T1', 'T2', 'H1', 'H2']
- #model.observations = data_dict1_times
-
- # Define the model - level 1
- model1 = bv.PyLinkForwardModel()
- #model.link_type = 'Function'
- model1.link_type = 'umbridge'
- model1.py_file = 'tsunami_model1'
- model1.name = 'tsunami_model1'
- model1.Output.names = ['T1', 'T2', 'H1', 'H2']
- #model.observations = data_dict1_times
-
- # Create the surrogate
- surrogate_opts0 = bv.MetaModel(priors, model0)
-
- # Select your metamodel method
- surrogate_opts0.meta_model_type = 'aPCE'
- surrogate_opts0.pce_reg_method = 'FastARD'
- surrogate_opts0.pce_deg = np.arange(1, 5)
- surrogate_opts0.pce_q_norm = 0.4#1.0
-
- # Select your metamodel method
- surrogate_opts0.add_ExpDesign()
- surrogate_opts0.ExpDesign.method = 'normal'
- surrogate_opts0.ExpDesign.n_init_samples = 50
- surrogate_opts0.ExpDesign.sampling_method = 'latin-hypercube'
- surrogate0 = surrogate_opts0.create_metamodel()
- print('Surrogate 0 completed')
- print('')
-
- # Create the surrogate
- surrogate_opts1 = bv.MetaModel(priors, model1)
-
- # Select your metamodel method
- surrogate_opts1.meta_model_type = 'aPCE'
- surrogate_opts1.pce_reg_method = 'FastARD'
- surrogate_opts1.pce_deg = np.arange(1, 5)
- surrogate_opts1.pce_q_norm = 0.4#1.0
-
- # Select your metamodel method
- surrogate_opts1.add_ExpDesign()
- surrogate_opts1.ExpDesign.method = 'normal'
- surrogate_opts1.ExpDesign.n_init_samples = 50
- surrogate_opts1.ExpDesign.sampling_method = 'latin-hypercube'
- surrogate1 = surrogate_opts1.create_metamodel()
- print('Surrogate 1 completed')
- print('')
-
- # Save surrogate
- with open('tsunami0.pk1', 'wb') as output:
- joblib.dump(surrogate0, output, 2)
-
- # Save surrogate
- with open('tsunami1.pk1', 'wb') as output:
- joblib.dump(surrogate1, output, 2)
-
- if 0:
- # Post processing
- L2_PostPCE = bv.PostProcessing(surrogate1)
- L2_PostPCE.plot_moments(plot_type='line')
- # Plot to check validation visually.
- L2_PostPCE.valid_metamodel(n_samples=1)
- # Compute and print RMSE error
- L2_PostPCE.check_accuracy(n_samples=300)
- total_sobol = L2_PostPCE.sobol_indices()
-
- # Test BayesInference
- import copy
- from tsunami_model import tsunami_model
- true_data = tsunami_model(np.array([[90.0,60.0]]))
- model0.observations = true_data
- model1.observations = true_data
- true_data_nox = copy.deepcopy(true_data)
- true_data_nox.pop('x_values')
-
- # Test surrogate output shape
- mean, std = surrogate1.eval_metamodel(np.array([[90.0,60.0]]))
-
- # Bayesian Inference
- if 0:
- # Set uncertainty
- import pandas as pd
- obsData = pd.DataFrame(true_data_nox, columns=model0.Output.names)
- DiscrepancyOpts = bv.Discrepancy('')
- DiscrepancyOpts.type = 'Gaussian'
- DiscrepancyOpts.parameters = (obsData*0.15)**2
-
- # Parameter estimation / single model validation via TOM
- BayesOpts = bv.BayesInference(surrogate1)
- BayesOpts.emulator= True
- BayesOpts.plot_post_pred = True
- #BayesOpts.inference_method = 'rejection'
- BayesOpts.Discrepancy = DiscrepancyOpts
- BayesOpts.bootstrap = True
- BayesOpts.n_bootstrap_itrs = 10
- BayesOpts.bootstrap_noise = 0.05
- BayesOpts.out_dir = ''
-
- # Set the MCMC parameters
- import emcee
- BayesOpts.inference_method = "MCMC"
- BayesOpts.mcmc_params = {
- 'n_steps': 1e3,
- 'n_walkers': 30,
- 'moves': emcee.moves.KDEMove(),
- 'multiprocessing': False,
- 'verbose': False
- }
-
- Bayes = BayesOpts.create_inference()
- # Pkl the inference results
- with open(f'Bayes_{model0.name}.pkl', 'wb') as output:
- joblib.dump(Bayes, output, 2)
-
- # Model Comparison
- if 1:
- # Set the models
- meta_models = {
- "M0": surrogate0,
- "M1": surrogate1,
- }
-
- # BME Bootstrap options
- opts_bootstrap = {
- "bootstrap": True,
- "n_samples": 1000,
- "Discrepancy": DiscrepancyOpts,
- "emulator": True,
- "plot_post_pred": False
- }
-
- # Run model comparison
- BayesOpts = bv.BayesModelComparison(
- justifiability=True,
- n_bootstarp=10,
- just_n_meas=2
- )
- output_dict = BayesOpts.create_model_comparison(
- meta_models,
- opts_bootstrap
- )
-
- import pickle as pkl
- # save model comparison results
- with open('ModelComparison.pkl', 'wb') as f:
- pkl.dump(output_dict, f)
-
\ No newline at end of file
diff --git a/examples/umbridge_tsunamitutorial/example_umbridge_testmodel.py b/examples/umbridge_tsunamitutorial/example_umbridge_testmodel.py
index 95ec964fa674a7ff9525c3557f65d7038fd2200a..22a0f750e96c4b7bf14913eb97f0040fb360a257 100644
--- a/examples/umbridge_tsunamitutorial/example_umbridge_testmodel.py
+++ b/examples/umbridge_tsunamitutorial/example_umbridge_testmodel.py
@@ -1,12 +1,12 @@
# -*- coding: utf-8 -*-
"""
-Created on Mon Dec 11 14:03:23 2023
+Example and test for using UM-Bridge with bayesvalidrox.
+This example uses the `testmodel` and sets the model type explicitly as
+UM-Bridge for the pylink model.
-Example and test for using umbridge with bayesvalidrox.
-Example model is the tsunami model by
-docker run -it -p 4242:4242 -v ~/tsunami_output:/output linusseelinger/model-exahype-tsunami
+An example for using UM-Bridge with an explicitly defined model can be found in
+`example_umbridge_tsunamimodel.py`
-@author: rkohlhaas
"""
import joblib
@@ -31,32 +31,30 @@ if __name__ == '__main__':
priors.Marginals[0].name = 'x'
priors.Marginals[0].input_data = np.random.uniform(50,150,n_prior_sample)
- # Define the model - level 0
+ # Define the model directly via Pylink
model = PyLinkForwardModel()
- #model.link_type = 'Function'
- #model.py_file = 'testmodel'
- #model.name = 'testmodel'
model.link_type = 'umbridge'
model.host = 'http://testmodel.linusseelinger.de'
model.modelparams = {}
model.name = 'testmodel'
model.Output.names = ['y']
- model.output_type = 'single-valued'
+ model.x_values = np.array([0])
# Create the surrogate
surrogate_opts = MetaModel(priors)
- # Select your metamodel method
+ # Select the surrogate type and properties
surrogate_opts.meta_model_type = 'aPCE'
surrogate_opts.pce_reg_method = 'FastARD'
surrogate_opts.pce_deg = np.arange(1, 5)
surrogate_opts.pce_q_norm = 0.4
- # Select your metamodel method
+ # Define the experimental design and sampling methods
ExpDesign = ExpDesigns(priors)
ExpDesign.n_init_samples = 50
ExpDesign.sampling_method = 'latin-hypercube'
+ # Start and run the engine
engine = Engine(surrogate_opts, model, ExpDesign)
engine.start_engine()
engine.train_normal(parallel = False)
@@ -70,8 +68,10 @@ if __name__ == '__main__':
# Post processing
L2_PostPCE = PostProcessing(engine)
L2_PostPCE.plot_moments(plot_type='line')
+
# Plot to check validation visually.
L2_PostPCE.valid_metamodel(n_samples=1)
+
# Compute and print RMSE error
L2_PostPCE.check_accuracy(n_samples=300)
total_sobol = L2_PostPCE.sobol_indices()
diff --git a/examples/umbridge_tsunamitutorial/example_umbridge_tsunamitutorial.py b/examples/umbridge_tsunamitutorial/example_umbridge_tsunamitutorial.py
new file mode 100644
index 0000000000000000000000000000000000000000..c456787d0c5d9d9714083b406895b8096a4d16ce
--- /dev/null
+++ b/examples/umbridge_tsunamitutorial/example_umbridge_tsunamitutorial.py
@@ -0,0 +1,183 @@
+# -*- coding: utf-8 -*-
+"""
+Example and test for using UM-Bridge with bayesvalidrox.
+Example model is the tsunami model described here:
+ https://um-bridge-benchmarks.readthedocs.io/en/docs/inverse-benchmarks/exahype-tsunami.html
+It needs to run separately while starting this example.
+ docker run -it -p 4242:4242 -v ~/tsunami_output:/output linusseelinger/model-exahype-tsunami
+
+This example uses models that are exmplicitly defined as functions outside of
+PyLink. An example for the link_type 'umbdridge' can be found in
+`example_umbridge_testmodel.py`.
+
+"""
+
+import copy
+import joblib
+import numpy as np
+import pandas as pd
+import umbridge
+
+import bayesvalidrox as bv
+from tsunami_model import tsunami_model
+
+if __name__ == '__main__':
+ # This model has 2 inputs and four outputs
+ n_prior_sample = 1000
+ priors = bv.Input()
+ priors.add_marginals()
+ priors.Marginals[0].name = 'x'
+ priors.Marginals[0].input_data = np.random.uniform(50,150,n_prior_sample)
+ priors.add_marginals()
+ priors.Marginals[1].name = 'y'
+ priors.Marginals[1].input_data = np.random.uniform(50,150,n_prior_sample)
+
+ # Define the model - level 0
+ model0 = bv.PyLinkForwardModel()
+ model0.link_type = 'function'
+ model0.py_file = 'tsunami_model'
+ model0.name = 'tsunami_model'
+ model0.Output.names = ['T1', 'T2', 'H1', 'H2']
+ #model.observations = data_dict1_times
+
+ # Define the model - level 1
+ model1 = bv.PyLinkForwardModel()
+ model1.link_type = 'function'
+ model1.py_file = 'tsunami_model1'
+ model1.name = 'tsunami_model1'
+ model1.Output.names = ['T1', 'T2', 'H1', 'H2']
+ #model.observations = data_dict1_times
+
+ # Create the surrogate
+ surrogate_opts0 = bv.MetaModel(priors, model0)
+
+ # Select your metamodel method
+ surrogate_opts0.meta_model_type = 'aPCE'
+ surrogate_opts0.pce_reg_method = 'FastARD'
+ surrogate_opts0.pce_deg = np.arange(1, 5)
+ surrogate_opts0.pce_q_norm = 0.4#1.0
+
+ # Define ExpDesign - this is the same for both models
+ ExpDesign0 = bv.ExpDesigns(priors)
+ ExpDesign0.method = 'normal'
+ ExpDesign0.n_init_samples = 50
+ ExpDesign0.sampling_method = 'latin-hypercube'
+ ExpDesign1 = copy.deepcopy(ExpDesign0)
+
+ # Start and run the engine
+ engine0 = bv.Engine(surrogate_opts0, model0, ExpDesign0)
+ engine0.start_engine()
+ engine0.train_normal(parallel = False)
+ print('Surrogate 0 completed')
+ print('')
+
+ # Create the surrogate
+ surrogate_opts1 = bv.MetaModel(priors, model1)
+
+ # Select your metamodel method
+ surrogate_opts1.meta_model_type = 'aPCE'
+ surrogate_opts1.pce_reg_method = 'FastARD'
+ surrogate_opts1.pce_deg = np.arange(1, 5)
+ surrogate_opts1.pce_q_norm = 0.4#1.0
+
+ # Start and run the engine
+ engine1 = bv.Engine(surrogate_opts1, model1, ExpDesign1)
+ engine1.start_engine()
+ engine1.train_normal(parallel = False)
+ print('Surrogate 1 completed')
+ print('')
+
+ # Save surrogate
+ with open('tsunami0.pk1', 'wb') as output:
+ joblib.dump(engine0.MetaModel, output, 2)
+
+ # Save surrogate
+ with open('tsunami1.pk1', 'wb') as output:
+ joblib.dump(engine1.MetaModel, output, 2)
+
+ # Post processing on model 1
+ L2_PostPCE = bv.PostProcessing(engine1)
+ L2_PostPCE.plot_moments(plot_type='line')
+ # Plot to check validation visually.
+ L2_PostPCE.valid_metamodel(n_samples=1)
+
+ # Compute and print RMSE error
+ L2_PostPCE.check_accuracy(n_samples=300)
+ total_sobol = L2_PostPCE.sobol_indices()
+
+ # Get reference evaluation as 'true data'
+ true_data = tsunami_model(np.array([[90.0,60.0]]))
+ model0.observations = true_data
+ model1.observations = true_data
+ true_data_nox = copy.deepcopy(true_data)
+ true_data_nox.pop('x_values')
+
+ # Direct surrogate evaluation
+ mean, std = engine1.MetaModel.eval_metamodel(np.array([[90.0,60.0]]))
+
+ # Bayesian Inference
+ # Set uncertainty
+ obsData = pd.DataFrame(true_data_nox, columns=model0.Output.names)
+ DiscrepancyOpts = bv.Discrepancy('')
+ DiscrepancyOpts.type = 'Gaussian'
+ DiscrepancyOpts.parameters = (obsData*0.15)**2
+
+ # Parameter estimation / single model validation via TOM for model 1
+ BayesOpts = bv.BayesInference(engine1)
+ BayesOpts.emulator= True
+ BayesOpts.plot_post_pred = True
+ #BayesOpts.inference_method = 'rejection'
+ BayesOpts.Discrepancy = DiscrepancyOpts
+ BayesOpts.bootstrap = True
+ BayesOpts.n_bootstrap_itrs = 10
+ BayesOpts.bootstrap_noise = 0.05
+ BayesOpts.out_dir = ''
+
+ # Set the MCMC parameters
+ import emcee
+ BayesOpts.inference_method = "MCMC"
+ BayesOpts.mcmc_params = {
+ 'n_steps': 1e3,
+ 'n_walkers': 30,
+ 'moves': emcee.moves.KDEMove(),
+ 'multiprocessing': False,
+ 'verbose': False
+ }
+
+ Bayes = BayesOpts.create_inference()
+
+ # Pkl the inference results
+ with open(f'Bayes_{model0.name}.pkl', 'wb') as output:
+ joblib.dump(Bayes, output, 2)
+
+ # Model Comparison
+ # Set the models
+ meta_models = {
+ "M0": engine0,
+ "M1": engine1,
+ }
+
+ # BME Bootstrap options
+ opts_bootstrap = {
+ "bootstrap": True,
+ "n_samples": 1000,
+ "Discrepancy": DiscrepancyOpts,
+ "emulator": True,
+ "plot_post_pred": False
+ }
+
+ # Run model comparison
+ BayesOpts = bv.BayesModelComparison(
+ justifiability=True,
+ n_bootstarp=10,
+ just_n_meas=2
+ )
+ output_dict = BayesOpts.create_model_comparison(
+ meta_models,
+ opts_bootstrap
+ )
+
+ # save model comparison results
+ with open('ModelComparison.pkl', 'wb') as f:
+ joblib.dump(output_dict, f)
+
\ No newline at end of file
diff --git a/examples/umbridge_tsunamitutorial/testmodel.pk1 b/examples/umbridge_tsunamitutorial/testmodel.pk1
deleted file mode 100644
index 700c643fc23e631a23cb23dd0972954904268377..0000000000000000000000000000000000000000
Binary files a/examples/umbridge_tsunamitutorial/testmodel.pk1 and /dev/null differ
diff --git a/examples/umbridge_tsunamitutorial/testmodel.py b/examples/umbridge_tsunamitutorial/testmodel.py
deleted file mode 100644
index b213d2a3acc0b6840f7dd24b7f87670640575368..0000000000000000000000000000000000000000
--- a/examples/umbridge_tsunamitutorial/testmodel.py
+++ /dev/null
@@ -1,20 +0,0 @@
-# -*- coding: utf-8 -*-
-"""
-Runs the umbridge command for the tsunami model
-
-@author: Rebecca Kohlhaas
-"""
-import umbridge
-import numpy as np
-
-def testmodel(params):
- # Get the model
- model = umbridge.HTTPModel('http://testmodel.linusseelinger.de', 'forward')
- # Run the model
- params = np.ndarray.tolist(params)
- out = np.array(model(params))
- #print(out)
- return {'y':out[:,0],'x_values':[0]}
-
-
-
\ No newline at end of file
diff --git a/examples/umbridge_tsunamitutorial/tsunami_model.py b/examples/umbridge_tsunamitutorial/tsunami_model.py
index 7f8e54170b2013915bb2cbd3c44c2db43e45bad5..2d837a9ae4bb27e8ef2ca622011e3e67012f2e94 100644
--- a/examples/umbridge_tsunamitutorial/tsunami_model.py
+++ b/examples/umbridge_tsunamitutorial/tsunami_model.py
@@ -1,8 +1,6 @@
# -*- coding: utf-8 -*-
"""
-Runs the umbridge command for the tsunami model
-
-@author: Rebecca Kohlhaas
+Runs the umbridge command for the tsunami model.
"""
import umbridge
import numpy as np
diff --git a/examples/umbridge_tsunamitutorial/tsunami_model1.py b/examples/umbridge_tsunamitutorial/tsunami_model1.py
index 3e56801e5d009d3d8bda0f6c7b19af67d58b3417..239670773cf71e84b4ef0caec35c4e024ca1bd66 100644
--- a/examples/umbridge_tsunamitutorial/tsunami_model1.py
+++ b/examples/umbridge_tsunamitutorial/tsunami_model1.py
@@ -1,8 +1,6 @@
# -*- coding: utf-8 -*-
"""
-Runs the umbridge command for the tsunami model
-
-@author: Rebecca Kohlhaas
+Runs the umbridge command for the tsunami model on level 1.
"""
import umbridge
import numpy as np
diff --git a/src/bayesvalidrox/pylink/pylink.py b/src/bayesvalidrox/pylink/pylink.py
index 5084f1cf1450ea08d91d97b8a64cf1d4a12e494e..8899447c57bff01a2f4976d19df4e2033212f7bd 100644
--- a/src/bayesvalidrox/pylink/pylink.py
+++ b/src/bayesvalidrox/pylink/pylink.py
@@ -473,6 +473,8 @@ class PyLinkForwardModel(object):
# TODO: if umbridge, just run, no parallel from here
# If the link type is UM-Bridge, then no parallel needs to be started from here
if self.link_type.lower() == 'umbridge':
+ if not hasattr(self, 'x_values'):
+ raise AttributeError('For model type `umbridge` the attribute `x_values` needs to be set for the model!')
# Init model
#model = umbridge.HTTPModel('http://localhost:4242', 'forward')
self.model = umbridge.HTTPModel(self.host, 'forward') # TODO: is this always forward?
@@ -608,30 +610,31 @@ class PyLinkForwardModel(object):
out_dict = {}
cnt = 0
for key in self.Output.names:
- # If needed resort into single-value outputs
- if self.output_type == 'single-valued':
- if out.shape[1]>1: # TODO: this doesn't fully seem correct??
- for i in range(out[:,key]): # TODO: this doesn't fully seem correct??
- new_key = key+str(i)
- if new_key not in self.Output.names:
- self.Output.names.append(new_key)
- if i == 0:
- self.Ouptut.names.remove(key)
- out_dict[new_key] = out[:,cnt,i] # TODO: not sure about this, need to test
- else:
- out_dict[key] = out[:,cnt]
-
-
- else:
- out_dict[key] = out[:,cnt]
+ # # If needed resort into single-value outputs
+ # if self.output_type == 'single-valued':
+ # if out.shape[1]>1: # TODO: this doesn't fully seem correct??
+ # for i in range(out[:,key]): # TODO: this doesn't fully seem correct??
+ # new_key = key+str(i)
+ # if new_key not in self.Output.names:
+ # self.Output.names.append(new_key)
+ # if i == 0:
+ # self.Ouptut.names.remove(key)
+ # out_dict[new_key] = out[:,cnt,i] # TODO: not sure about this, need to test
+ # else:
+ # out_dict[key] = out[:,cnt]
+ #
+ #
+ # else:
+ out_dict[key] = out[:,cnt]
cnt += 1
- # TODO: how to deal with the x-values?
- if self.output_type == 'single-valued':
- out_dict['x_values'] = [0]
- else:
- out_dict['x_values'] = np.arange(0,out[:,0].shape[0],1)
+ ## TODO: how to deal with the x-values?
+ #if self.output_type == 'single-valued':
+ # out_dict['x_values'] = [0]
+ #else:
+ # out_dict['x_values'] = np.arange(0,out[:,0].shape[0],1)
+ out_dict['x_values'] = self.x_values
#return {'T1':out[:,0], 'T2':out[:,1], 'H1':out[:,2], 'H2':out[:,3],
# 'x_values':[0]}