Integrated Main.py into Snakefile.

Main.py

-# -*- coding: utf-8 -*-
-"""
-@author: Laura C. Kühle
-
-Code-Style: E226, W503
-Docstring-Style: D200, D400
-
-"""
-from DG_Approximation import DGScheme
-import timeit
-
-tic = timeit.default_timer()
-
-alpha = 1
-p = 2
-cfl = 0.2
-N = 32    # 40 elements work well for Condition 3
-finalTime = 1
-xL = -1
-xR = 1
-
-detector_config = {}
-init_config = {}
-limiter_config = {}
-quadrature_config = {}
-
-fold_len = 16
-whisker_len = 3
-
-detector_config['fold_len'] = fold_len
-detector_config['whisker_len'] = whisker_len
-
-init_config['factor'] = 4
-init_config['left_factor'] = 3
-
-limiter_config['mod_factor'] = 0
-limiter_config['erase_degree'] = 0
-
-quadrature_config['num_eval_points'] = 12  # 12
-
-# detector = 'Theoretical'
-detector = 'ArtificialNeuralNetwork'
-update_scheme = 'SSPRK3'
-init_cond = 'Sine'
-limiter = 'ModifiedMinMod'
-quadrature = 'Gauss'
-dg_scheme = DGScheme(detector, detector_config=detector_config, init_cond=init_cond,
-                     init_config=init_config, limiter=limiter, limiter_config=limiter_config,
-                     quadrature=quadrature, quadrature_config=quadrature_config,
-                     update_scheme=update_scheme, wave_speed=alpha, polynomial_degree=p,
-                     cfl_number=cfl, num_grid_cells=N, final_time=finalTime,
-                     verbose=False, left_bound=xL, right_bound=xR)
-
-dg_scheme.approximate()
-dg_scheme.save_plots()
-
-toc = timeit.default_timer()
-print('Time:', toc-tic)
-
-# if __name__ == '__main__':
-#     pass

Snakefile

 import sys
+import timeit
 
 import ANN_Data_Generator, Initial_Condition, ANN_Training
 from ANN_Training import evaluate_models
+from DG_Approximation import DGScheme
 import numpy as np
 
 configfile: 'config.yaml'
@@ -14,7 +16,54 @@ if config['random_seed'] is not None:
 rule all:
     input:
         expand(DIR+'/trained models/model__{model}.pt', model=MODELS),
-        DIR+'/model evaluation/classification_accuracy/' + '_'.join(MODELS.keys()) + '.pdf'
+        DIR+'/model evaluation/classification_accuracy/' + '_'.join(MODELS.keys()) + '.pdf',
+        config['dg_parameter']['plot_dir'] + '/error/' + 'Sine__ArtificialNeuralNetwork__' +
+            'ModifiedMinMod0__SSPRK3__Gauss12__final_time_1__wave_speed_1__' +
+            'number_of_cells_32__polynomial_degree_2.pdf'
+
+rule approximate_solution:
+    input:
+        config['dg_parameter']['detector_config']['model_state']
+        if config['dg_parameter']['detector'] == 'ArtificialNeuralNetwork' else ''
+    output:
+        error=config['dg_parameter']['plot_dir'] + '/error/' + 'Sine__ArtificialNeuralNetwork__' +
+              'ModifiedMinMod0__SSPRK3__Gauss12__final_time_1__wave_speed_1__' +
+              'number_of_cells_32__polynomial_degree_2.pdf'
+    params:
+        # plot_dir=config['plot_dir'],
+        # wave_speed=config['wave_speed'],
+        # polynomial_degree=config['polynomial_degree'],
+        # cfl_number=config['cfl_number'],
+        # num_grid_cells=config['num_grid_cells'],
+        # final_time=config['final_time'],
+        # left_bound=config['left_bound'],
+        # right_bound=config['right_bound'],
+        # verbose=config['verbose'],
+        # detector=config['detector'],
+        # detector_config=config['detector_config'],
+        # init_cond=config['init_cond'],
+        # init_config=config['init_config'],
+        # limiter=config['limiter'],
+        # limiter_config=config['limiter_config'],
+        # quadrature=config['quadrature'],
+        # quadrature_config=config['quadrature_config'],
+        # update_scheme=config['update_scheme']
+        dg_params=config['dg_parameter']
+    log:
+        DIR+'/log/approximate_solution.log'
+    run:
+        with open(str(log), 'w') as logfile:
+            tic = timeit.default_timer()
+
+            print(params.dg_params)
+            # dg_scheme = DGScheme(**params)
+            dg_scheme = DGScheme(**params.dg_params)
+
+            dg_scheme.approximate()
+            dg_scheme.save_plots()
+
+            toc = timeit.default_timer()
+            print('Time:',toc-tic)
 
 rule test_model:
     input:

config.yaml

 data_directory: "Snakemake-Test"
 random_seed: 1234
 
+# Parameter for Approximation with Troubled Cell Detection
+dg_parameter:
+    plot_dir: 'testing'
+    wave_speed: 1
+    polynomial_degree: 2
+    cfl_number: 0.2
+    num_grid_cells: 32    # 40 elements work well for Condition 3
+    final_time: 1
+    left_bound: -1
+    right_bound: 1
+    verbose: True
+
+    #detector: 'Theoretical'
+    detector: 'ArtificialNeuralNetwork'
+    detector_config:
+      fold_len: 16
+      whisker_len: 3
+      model_state: 'Snakemake-Test/trained models/model__Adam.pt'
+
+    init_cond: 'Sine'
+    init_config:
+      factor: 4
+      left_factor: 3
+
+    limiter: 'ModifiedMinMod'
+    limiter_config:
+      mod_factor: 0
+      erase_degree: 0
+
+    quadrature: 'Gauss'
+    quadrature_config:
+      num_eval_points: 12
+
+    update_scheme: 'SSPRK3'
+
 # Parameter for Training Data Generation
 sample_number: 100
 
@@ -25,6 +60,7 @@ functions:
 
 # Parameter for Model Training and Evaluation
 compare_normalization: True
+
 classification_colors:
   Accuracy: '#FF00FF'  # magenta
   Precision_Smooth: '#FF0000'  # red
@@ -34,6 +70,7 @@ classification_colors:
   F-Score_Smooth: '#00FF00'  # green
   F-Score_Troubled: '#006400'  # dark green
   AUROC: '#FFFF00'  # yellow
+
 models:
   Adam:
     num_epochs: 1000
@@ -52,3 +89,6 @@ models:
     loss_function: BCELoss
     optimizer: SGD
 
+# Loss Functions: BCELoss, BCEWithLogitsLoss,
+# CrossEntropyLoss (not working), MSELoss (with reduction='sum')
+# Optimizer: Adam, SGD
\ No newline at end of file