Davout 0.1.1.dev136__tar.gz → 0.1.1.dev137__tar.gz

{davout-0.1.1.dev136/source/Davout.egg-info → davout-0.1.1.dev137}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: Davout
-Version: 0.1.1.dev136
+Version: 0.1.1.dev137
 Summary: Scientific computing utilities combining FEM, ANN and tooling
 Author: Matheus Janczkowski
 Author-email: Matheus Janczkowski <matheusj2009@hotmail.com>

{davout-0.1.1.dev136 → davout-0.1.1.dev137}/pyproject.toml

@@ -63,7 +63,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "Davout"
-version = "0.1.1.dev136"
+version = "0.1.1.dev137"
 description = "Scientific computing utilities combining FEM, ANN and tooling"
 readme = "README.md"
 authors = [

{davout-0.1.1.dev136 → davout-0.1.1.dev137}/source/Davout/DeepMech/aa_tests/mecsol_2026_quotient_space/box_with_imposed_displacement_gradient.py

@@ -10,7 +10,8 @@ from .....Davout.GraphUtilities.paraview_tools import frozen_snapshots
 
 def solve_BVP(results_path, displacement_file_name, young_modulus_file,
 mesh_file_name, displacement_gradient_components, 
-lagrange_multiplier_file_name, save_snapshot=False):
+lagrange_multiplier_file_name, save_snapshot=False, snapshot_file_name=
+"RVE_displacement_imposed_gradient.png"):
 
     ####################################################################
     ####################################################################
@@ -33,7 +34,7 @@ lagrange_multiplier_file_name, save_snapshot=False):
 
     post_processes[0][1]["SaveField"] = {"directory path": results_path, 
     "file name": displacement_file_name, "saving method": "binary", 
-    "visualization copy for readable xdmf": False}
+    "visualization copy for readable xdmf": True}
 
     post_processes[1][1]["SaveField"] = {"directory path": results_path, 
     "file name": lagrange_multiplier_file_name, "saving method": "bina"+
@@ -149,18 +150,18 @@ lagrange_multiplier_file_name, save_snapshot=False):
 
         frozen_snapshots(displacement_file_name+"_visualization_copy", 
         "Displacement", input_path=results_path, 
-        camera_position=[3.295421262287293, -4.617612102822565, 3.7314785394692436],
-        camera_focal_point=[0.3868267767658733, -0.07264176288584451, 0.4047648861287915],
-        camera_up_direction=[-0.14883962531451048, 0.5202719201071424, 0.8409303746947565],
-        camera_parallel_scale=1.6406698504546489,
+        camera_position=[4.451171571468425, -3.797472604571211, 4.430154873929185],
+        camera_focal_point=[0.7984414053662643, 0.5769704993902152, 0.33475865213051903],
+        camera_up_direction=[-0.3606934051472985, 0.45896443721366265, 0.8119432941138149],
+        camera_parallel_scale=1.8163576931289542,
         camera_rotation=[0.0, 0.0, 0.0],
-        legend_bar_position=[0.785724255357283, 0.10503211991434691],
-        legend_bar_length=0.7389935760171309,
-        size_in_pixels={'aspect ratio': 0.6768115942028986, 'pixels in width': 690},
+        legend_bar_position=[0.7406134390037274, 0.14364425162689798],
+        legend_bar_length=0.640195227765729,
+        size_in_pixels={'aspect ratio': 0.6739766081871345, 'pixels in width': 684},
         axes_color=[0.0, 0.0, 0.0], get_attributes_render=False, 
-        output_imageFileName="RVE_displacement_imposed_gradient.png", resolution_ratio=5, 
+        output_imageFileName=snapshot_file_name, resolution_ratio=5, 
         warp_by_vector=True, representation_type="Surface With Edges", 
-        set_camera_interactively=False, time=1.0)
+        set_camera_interactively=False, time=1.0, transparent_background=True)
 
 # Testing block
 
@@ -174,6 +175,8 @@ if __name__=="__main__":
 
     lagrange_multiplier_file_name = "lagrange_multiplier_displacement_gradient"
 
+    snapshot_file_name = "RVE_shear.png"
+
     # Defines the mesh file name and the name of the file of the field 
     # of Young modulus
 
@@ -186,9 +189,13 @@ if __name__=="__main__":
 
     displacement_gradient_components = [[2.5, 0.0, 0.0], [0.0, 2.5, 0.0], 
     [0.0, 0.0, 2.5]]
+    
+    displacement_gradient_components = [[3.0, 0.5, 0.0], [0.5, 2.5, 0.0], 
+    [0.0, 0.0, 2.0]]
 
     # Solves the boundary value problem
 
     solve_BVP(results_path, displacement_file_name, young_modulus_file,
     mesh_file_name, displacement_gradient_components, 
-    lagrange_multiplier_file_name, save_snapshot=True)
+    lagrange_multiplier_file_name, save_snapshot=True, 
+    snapshot_file_name=snapshot_file_name)

{davout-0.1.1.dev136 → davout-0.1.1.dev137}/source/Davout/DeepMech/aa_tests/mecsol_2026_quotient_space/neural_network_trainer.py

@@ -17,7 +17,7 @@ from .....Davout.PythonicUtilities.path_tools import get_parent_path_of_file
 
 def train_surrogate_model(displacement_data_file, input_data_file, 
 saved_model_file, results_path, n_training_samples, 
-quotient_space_dimension):
+quotient_space_dimension, n_monte_carlo_realizations, n_best_models):
 
     # Reads the two files
 
@@ -83,11 +83,24 @@ quotient_space_dimension):
 
     print("\nTrains at "+str(elapsed_time)+" seconds")
 
+    # Tests Monte Carlo training
+
+    training_class.monte_carlo_training(n_realizations=
+    n_monte_carlo_realizations, best_models_rank_size=n_best_models, 
+    show_reinitialization_distance=True)
+
+    # Checks the loss again with the best model of the Monte Carlo
+    # training
+
+    print("\nThe loss function evaluated again over the set of trainin"+
+    "g data for the best model is "+str(training_class.loss_unseen_data(
+    training_true_values, training_data, output_as_numpy=True)))
+
 # Defines a function to test the model
 
 def test_surrogate_model(displacement_data_file, input_data_file, 
-saved_model_file, results_path, n_training_samples, 
-quotient_space_dimension):
+results_path, n_training_samples, quotient_space_dimension, 
+n_best_models):
 
     # Reads the two files
 
@@ -109,38 +122,44 @@ quotient_space_dimension):
 
     test_true_values = output_data[n_training_samples:,:]
 
-    # Loads it back
-
-    loaded_model = tf.keras.models.load_model(results_path+"//"+
-    saved_model_file)
-
     # Defines the loss function metric
 
     loss_metric = tf.keras.losses.MeanAbsoluteError()
 
-    # Gets the output of the loaded model
+    # Iterates through the best models
 
-    output_model = loaded_model(test_data)
+    for i in range(n_best_models):
 
-    # Gets the loss of the test data
+        # Loads the i-th best model back
 
-    test_loss = loss_metric(test_true_values, output_model)
+        loaded_model = tf.keras.models.load_model(results_path+"//"+str(
+        i+1)+"_best_model.keras")
 
-    print("\nLoss function on test set:", format(test_loss.numpy(),'.5'+
-    'e')+"\n")
+        # Gets the output of the loaded model
 
-    # Verifies with the maximum absolute error
+        output_model = loaded_model(test_data)
 
-    maximum_absolute_error = MaximumAbsoluteError()
+        # Gets the loss of the test data
 
-    maximum_absolute_value = maximum_absolute_error(test_true_values, 
-    output_model)
+        test_loss = loss_metric(test_true_values, output_model)
 
-    print("\nMaximum absolute error on test set:"+str(format(
-    maximum_absolute_value.numpy(),'.5e'))+"\nwhereas the minimum abso"+
-    "lute error is "+str(format(
-    maximum_absolute_error.minimum_absolute_error(test_true_values, 
-    output_model).numpy(), '.5e'))+"\n")
+        print("Loads the "+str(i+1)+"-th best model")
+
+        print("Loss function on test set:", format(test_loss.numpy(),
+        '.5e')+"\n")
+
+        # Verifies with the maximum absolute error
+
+        maximum_absolute_error = MaximumAbsoluteError()
+
+        maximum_absolute_value = maximum_absolute_error(test_true_values, 
+        output_model)
+
+        print("Maximum absolute error on test set:"+str(format(
+        maximum_absolute_value.numpy(),'.5e'))+"\nwhereas the minimum abso"+
+        "lute error is "+str(format(
+        maximum_absolute_error.minimum_absolute_error(test_true_values, 
+        output_model).numpy(), '.5e'))+"\n")
 
 # Testing block
 
@@ -160,14 +179,19 @@ if __name__=="__main__":
 
     # Trains a new model
 
+    n_monte_carlo_realizations = 10
+
+    n_best_models = 5
+
     training_flag = False 
 
     if training_flag:
 
         train_surrogate_model(displacement_data_file, input_data_file, 
         saved_model_file, results_path, n_training_samples,
-        quotient_space_dimension)
+        quotient_space_dimension, n_monte_carlo_realizations, 
+        n_best_models)
 
     test_surrogate_model(displacement_data_file, input_data_file,
-    saved_model_file, results_path, n_training_samples,
-    quotient_space_dimension)
+    results_path, n_training_samples, quotient_space_dimension, 
+    n_best_models)