Merge branch 'develop'

.gitignore

@@ -221,10 +221,6 @@ tags
 
 ### VisualStudioCode ###
 .vscode/*
-!.vscode/settings.json
-!.vscode/tasks.json
-!.vscode/launch.json
-!.vscode/extensions.json
 
 ### VisualStudioCode Patch ###
 # Ignore all local history of files

doc/examples/algorithms/plot_algorithm_mestatic_IO.py

@@ -110,7 +110,6 @@ def deq(rho, t, alpha, beta, gamma):
                n_grid=[100, 100],
                coords=grid.coords,
                results=results,
-               fn_out=None,
                camera_pos=[45., 65])
 
 # On Windows subprocesses will import (i.e. execute) the main module at start.

doc/examples/algorithms/plot_algorithm_mestaticprojection.py

@@ -31,15 +31,15 @@
 # gPC options
 options = dict()
 options["method"] = "reg"
-options["solver"] = "Moore-Penrose"
+options["solver"] = "LarsLasso"
 options["settings"] = None
-options["order"] = [3, 3]
-options["order_max"] = 3
+options["order"] = [5, 5]
+options["order_max"] = 5
 options["interaction_order"] = 2
 options["n_cpu"] = 0
-options["gradient_enhanced"] = True
-options["gradient_calculation"] = "FD_fwd"
-options["gradient_calculation_options"] = {"dx": 0.001, "distance_weight": -2}
+# options["gradient_enhanced"] = False
+# options["gradient_calculation"] = "FD_fwd"
+# options["gradient_calculation_options"] = {"dx": 0.001, "distance_weight": -2}
 options["error_type"] = "nrmsd"
 options["n_samples_validation"] = 1e3
 options["qoi"] = "all"
@@ -52,7 +52,7 @@
 options["save_session_format"] = save_session_format
 options["grid"] = pygpc.Random
 options["grid_options"] = {"seed": 1}
-options["n_grid"] = 1000
+options["n_grid"] = 2000
 options["adaptive_sampling"] = False
 
 # define algorithm

doc/examples/algorithms/plot_algorithm_static_IO.py

@@ -59,6 +59,12 @@
 options["backend"] = "omp"
 options["verbose"] = True
 
+# determine number of gPC coefficients (hint: compare it with the amount of output data you have)
+n_coeffs = pygpc.get_num_coeffs_sparse(order_dim_max=options["order"],
+                                       order_glob_max=options["order_max"],
+                                       order_inter_max=options["interaction_order"],
+                                       dim=len(parameters))
+
 # define algorithm
 algorithm = pygpc.Static_IO(parameters=parameters, options=options, grid=grid, results=results)
 
@@ -87,15 +93,19 @@
                              calc_pdf=True,
                              algorithm="standard")
 
+# get a summary of the sensitivity coefficients
+sobol, gsens = pygpc.get_sens_summary(fn_results, parameters)
+print(sobol)
+print(gsens)
+
 # plot gPC approximation and IO data
 pygpc.plot_gpc(session=session,
                coeffs=coeffs,
                random_vars=["x1", "x3"],
-               output_idx= 0,
+               output_idx=0,
                n_grid=[100, 100],
                coords=grid.coords,
-               results=results,
-               fn_out=None)
+               results=results)
 
 # On Windows subprocesses will import (i.e. execute) the main module at start.
 # You need to insert an if __name__ == '__main__': guard in the main module to avoid

doc/examples/examples/plot_Lorenz.py

@@ -30,6 +30,8 @@
 import pygpc
 import numpy as np
 from collections import OrderedDict
+import matplotlib
+matplotlib.use("Qt5Agg")
 
 #%%
 # At first, we are loading the model:
@@ -93,23 +95,26 @@
                              output_idx=None,
                              calc_sobol=True,
                              calc_global_sens=True,
-                             calc_pdf=False,
-                             n_samples=int(1e4))
+                             calc_pdf=False)
 
 # extract sensitivity coefficients from results .hdf5 file
 sobol, gsens = pygpc.get_sens_summary(fn_gpc=fn_results,
                                       parameters_random=session.parameters_random,
                                       fn_out=fn_results + "_sens_summary.txt")
 
-# plot time course of sensitivity coefficients and mean and standard deviation of x(t)
+# plot time course of mean together with probability density, sobol sensitivity coefficients and global derivatives
 t = np.arange(0.0, parameters["t_end"], parameters["step_size"])
-pygpc.plot_sens_summary(sobol=sobol,
+pygpc.plot_sens_summary(session=session,
+                        coeffs=coeffs,
+                        sobol=sobol,
                         gsens=gsens,
-                        multiple_qoi=True,
+                        plot_pdf_over_output_idx=True,
                         qois=t,
-                        results=results,
+                        mean=pygpc.SGPC.get_mean(coeffs),
+                        std=pygpc.SGPC.get_std(coeffs),
                         x_label="t in s",
-                        y_label="x(t)")
+                        y_label="x(t)",
+                        zlim=[0, 0.4])
 
 #
 # On Windows subprocesses will import (i.e. execute) the main module at start.

doc/examples/examples/plot_Lorenz_julia.py

@@ -100,6 +100,8 @@ def simulate(self, process_id=None, matlab_engine=None):
     import pygpc
     import numpy as np
     from collections import OrderedDict
+    import matplotlib
+    matplotlib.use("Qt5Agg")
 
     # Windows users have to encapsulate the code into a main function to avoid multiprocessing errors.
     # def main():
@@ -162,14 +164,17 @@ def simulate(self, process_id=None, matlab_engine=None):
 
     # plot sobol indices over time and mean and standard deviation of x(t)
     t = np.arange(0.0, parameters["t_end"], parameters["step_size"])
-    pygpc.plot_sens_summary(sobol=sobol,
+    pygpc.plot_sens_summary(session=session,
+                            coeffs=coeffs,
+                            sobol=sobol,
                             gsens=gsens,
-                            multiple_qoi=True,
+                            plot_pdf_over_output_idx=True,
                             qois=t,
-                            results=results,
+                            mean=pygpc.SGPC.get_mean(coeffs),
+                            std=pygpc.SGPC.get_std(coeffs),
                             x_label="t in s",
-                            y_label="x(t)")
-
+                            y_label="x(t)",
+                            zlim=[0, 0.4])
 """
 import matplotlib.pyplot as plt
 

doc/examples/examples/plot_ishigami.py

@@ -18,6 +18,8 @@
 import pygpc
 import numpy as np
 from collections import OrderedDict
+import matplotlib
+matplotlib.use("Qt5Agg")
 
 fn_results = "tmp/example_ishigami"
 
@@ -112,8 +114,8 @@
 #%%
 # Sensitivity analysis
 # ^^^^^^^^^^^^^^^^^^^^
-sobol, gsens = pygpc.get_sens_summary(fn_results, parameters_random)
-pygpc.plot_sens_summary(sobol, gsens)
+sobol, gsens = pygpc.get_sens_summary(fn_results, parameters_random, fn_results + "_sens_summary.txt")
+pygpc.plot_sens_summary(sobol=sobol, gsens=gsens)
 
 #
 # On Windows subprocesses will import (i.e. execute) the main module at start.

doc/examples/examples/tmp/example_ishigami_sens_summary.txt

@@ -0,0 +1,14 @@
+Normalized Sobol indices:
+==========================
+['x2']:           0.442411
+['x1']:           0.313905
+['x1', 'x3']:     0.243684
+['x1', 'x2']:     0.000000
+['x2', 'x3']:     0.000000
+['x3']:           0.000000
+
+Average derivatives:
+==========================
+['x1']:           8.47e-06
+['x2']:          -3.25e-07
+['x3']:          -1.51e-06