Removes dry unused variables

.github/workflows/test_gedi.yml

@@ -71,11 +71,11 @@ jobs:
 
     - name: Compare output 3
       run: 
-        diff data/validation/genELexperiment3_04.json output/features/grid_feat/2_enself_rt20v/genELexperiment3_04.json
+        diff data/validation/genELexperiment3_04.json output/features/grid_feat/2_enself_rt20v/genELexperiment3_04_nan.json
 
     - name: Compare output 4
       run: 
-        diff data/validation/genELexperiment4_02.json output/features/grid_feat/2_enself_rt20v/genELexperiment4_02.json
+        diff data/validation/genELexperiment4_02.json output/features/grid_feat/2_enself_rt20v/genELexperiment4_nan_02.json
 
   test_benchmark:
     runs-on: ubuntu-latest

gedi/generator.py

@@ -165,7 +165,7 @@ class GenerateEventLogs():
             with multiprocessing.Pool(num_cores) as p:
                 print(f"INFO: Generator starting at {start.strftime('%H:%M:%S')} using {num_cores} cores for {len(tasks)} tasks...")
                 random.seed(RANDOM_SEED)
-                log_config = p.map(self.generator_wrapper, [(index, row.dropna()) for index, row in tasks.iterrows()])
+                log_config = p.map(self.generator_wrapper, [(index, row) for index, row in tasks.iterrows()])
             self.log_config = log_config
 
         else:
@@ -194,7 +194,7 @@ class GenerateEventLogs():
         except IndexError:
             identifier = task[0]+1
         task = task[1].loc[lambda x, identifier=identifier: x!=identifier]
-        self.objectives = task.to_dict()
+        self.objectives = task.dropna().to_dict()
         random.seed(RANDOM_SEED)
         self.configs = self.optimize()
 
@@ -209,7 +209,7 @@ class GenerateEventLogs():
         if self.objectives.get('ratio_unique_traces_per_trace'):#HOTFIX
             self.objectives['ratio_variants_per_number_of_traces']=self.objectives.pop('ratio_unique_traces_per_trace')
 
-        save_path = get_output_key_value_location(self.objectives,
+        save_path = get_output_key_value_location(task.to_dict(),
                                          self.output_path, identifier, self.feature_keys)+".xes"
 
         write_xes(log_config['log'], save_path)
@@ -221,7 +221,7 @@ class GenerateEventLogs():
         if features_to_dump.get('ratio_unique_traces_per_trace'):#HOTFIX
             features_to_dump['ratio_variants_per_number_of_traces']=features_to_dump.pop('ratio_unique_traces_per_trace')
         features_to_dump['log'] = identifier.replace('genEL', '')
-        dump_features_json(features_to_dump, self.output_path, identifier, objectives=self.objectives, obj_keys=self.feature_keys)
+        dump_features_json(features_to_dump, save_path)
         return log_config
 
     def generate_optimized_log(self, config):

gedi/utils/io_helpers.py

@@ -72,15 +72,11 @@ def get_output_key_value_location(obj, output_path, identifier, obj_keys=None):
     save_path = os.path.join(folder_path, generated_file_name)
     return save_path
 
-def dump_features_json(features: dict, output_path, identifier, objectives=None, content_type="features", obj_keys=None):
-    output_parts = PurePath(output_path).parts
-    feature_dir = os.path.join(output_parts[0], content_type,
+def dump_features_json(features: dict, output_path, content_type="features"):
+    output_parts = PurePath(output_path.split(".xes")[0]).parts
+    features_path = os.path.join(output_parts[0], content_type,
                                    *output_parts[1:])
-    if objectives is not None:
-        json_path = get_output_key_value_location(objectives,
-                                                feature_dir, identifier, obj_keys)+".json"
-    else:
-        json_path = os.path.join(feature_dir, identifier)+".json"
+    json_path = features_path+'.json'
 
     os.makedirs(os.path.split(json_path)[0], exist_ok=True)
     with open(json_path, 'w') as fp: