intermediate save

PaperPlotTable.py

@@ -0,0 +1,74 @@
+import os
+import pandas as pd
+import orloge as ol
+import copy
+import yaml
+from utils import config_decoder
+
+
+def best_parameter_distribution(dir_path, file_condition='best_params_', output_file_name='best_params.csv',
+                                save_output=True, noise_level=0):
+    file_list = [f_name for f_name in os.listdir(dir_path) if f_name.startswith(file_condition)]
+    path_list = [[os.path.join(dir_path, f_name), f_name] for f_name in file_list
+                 if os.path.isfile(os.path.join(dir_path, f_name))]
+    combined_csv = []
+    for f in path_list:
+        temp_csv = pd.read_csv(f[0], index_col=0)
+        temp_info = pd.DataFrame(dict(zip(['N', 'group_size', 'm', 's', 'seed'], f[1].strip('.csv').split('_')[2:])), index=[0])
+        temp_info['noise_level'] = noise_level
+        temp_csv = pd.concat([temp_info, temp_csv], axis=1)
+        combined_csv.append(copy.copy(temp_csv))
+    combined_csv = pd.concat(combined_csv).reset_index().drop(['index'], axis=1)
+
+    if save_output:
+        combined_csv.to_csv(os.path.join(dir_path, output_file_name), index=False, encoding='utf-8-sig')
+    else:
+        return combined_csv
+
+def cplex_log_files(dir_path, file_condition='log_', output_file_name='parsed_log.csv',
+                                save_output=True):
+    file_list = [f_name for f_name in os.listdir(dir_path) if f_name.startswith(file_condition)]
+    path_list = [[os.path.join(dir_path, f_name), f_name] for f_name in file_list
+                 if os.path.isfile(os.path.join(dir_path, f_name))]
+    combined_csv = []
+    for f in path_list:
+        temp_info = pd.DataFrame(dict(zip(['N', 'group_size', 'm', 's', 'seed'], f[1].strip('.txt').split('_')[1:])), index=[0])
+        flist = open(os.path.join(f[0])).readlines()
+        print(flist[-1])
+        temp_info['time(sec)'] = flist[-1].split(' ')[-4].strip('(')
+        temp_info['ticks'] = flist[-1].split(' ')[-2].strip('(')
+        combined_csv.append(copy.copy(temp_info))
+    combined_csv = pd.concat(combined_csv).reset_index().drop(['index'], axis=1)
+    if save_output:
+        combined_csv.to_csv(os.path.join(dir_path, output_file_name), index=False, encoding='utf-8-sig')
+    else:
+        return combined_csv
+
+# dir_path = '/Users/hoomanzabeti/Desktop/GTResults/Noisy/Threshold'
+# dir_list = [i for i in os.walk(dir_path)][0][1]
+#
+# # print(pd.concat([best_parameter_distribution(os.path.join(dir_path,i,'Logs'),save_output=False) for i in dir_list]).reset_index())
+# pd.concat([best_parameter_distribution(os.path.join(dir_path, i, 'Logs'), save_output=False)
+#            for i in dir_list]).reset_index().to_csv(os.path.join(dir_path, 'best_params.csv'),
+#                                                     index=False, encoding='utf-8-sig')
+
+# dir_path = '/Users/hoomanzabeti/Desktop/GTResults/LP_IT'
+# dir_list = [i for i in os.walk(dir_path)][0][1]
+# pd.concat([cplex_log_files(os.path.join(dir_path, i, 'Logs'), save_output=False)
+#            for i in dir_list]).reset_index().to_csv(os.path.join(dir_path, 'parsed_lp_logs.csv'),
+#                                                     index=False, encoding='utf-8-sig')
+
+dir_path = '/Users/hoomanzabeti/Desktop/GTResults/Noisy/Permutation/'
+dir_list = [i for i in os.walk(dir_path)][0][1]
+noise_level = []
+for f in dir_list:
+    with open(os.path.join(dir_path,f,'config.yml'), 'r') as config_file:
+        config_dict = yaml.load(config_file, Loader=yaml.FullLoader)
+    opts = config_decoder(config_dict['opts'])[0]
+    noise_level.append(opts['permutation_noise_prob'])
+    print(noise_level)
+    print(dir_list)
+# print(pd.concat([best_parameter_distribution(os.path.join(dir_path,i,'Logs'),save_output=False) for i in dir_list]).reset_index())
+pd.concat([best_parameter_distribution(os.path.join(dir_path, i, 'Logs'), save_output=False,noise_level=noise_level[idx])
+           for idx, i in enumerate(dir_list)]).reset_index().to_csv(os.path.join(dir_path, 'best_params.csv'),
+                                                    index=False, encoding='utf-8-sig')

config.yml

@@ -1,57 +1,68 @@
 ---
-opts:
-  'run_ID': 'debugging'
-  'verbose': False
-  'plotting': False
-  'saving': True
-  'graph_gen_method': 'no_multiple'
-  'test_noise_methods': ['permutation']
-  'permutation_noise_prob':
-    'mode': 'scalar'
-    'values': [0.01]
-  'theta_l':
-    'mode': 'scalar'
-    'values': [0.00]
-  'theta_u':
-    'mode': 'scalar'
-    'values': [0.0625]
-  'binary_symmetric_noise_prob':
-    'mode': 'scalar'
-    'values': [0.26]
-  'seed':
-    'mode': 'range'
-    'values': [0,10,1]
-  'N':
-    'mode': 'scalar'
-    'values': [1000]
-  's':
-    'mode': 'scalar'
-    'values': [50]
-  'group_size':
-    'mode': 'scalar'
-    'values': [16]
-  'm':
-    'mode': 'range'
-    'values': [10,1010,10]
-  'max_tests_per_individual':
-    'mode': 'scalar'
-    'values': [16]
+design_param:
+  'general':
+    'seed':
+      'mode': 'range'
+      'values': [0,1,1]
+    'N':
+      'mode': 'scalar'
+      'values': [1000]
+    'm':
+      'mode': 'range'
+      'values': [10,30,10]
+    'verbose': False
+  'generate_groups':
+    'run_ID': 'debugging'
+    'plotting': False
+    'saving': True
+    'graph_gen_method': 'no_multiple'
+    'group_size':
+      'mode': 'scalar'
+      'values': [16]
+    'max_tests_per_individual':
+      'mode': 'scalar'
+      'values': [16]
+  'generate_individual_status':
+    's':
+      'mode': 'scalar'
+      'values': [50]
+  'generate_test_results':
+    'test_noise_methods': ['permutation']
+    'permutation_noise_prob':
+      'mode': 'scalar'
+      'values': [0.01]
+    'theta_l':
+      'mode': 'scalar'
+      'values': [0.00]
+    'theta_u':
+      'mode': 'scalar'
+      'values': [0.0625]
+    'binary_symmetric_noise_prob':
+      'mode': 'scalar'
+      'values': [0.26]
+
+decoder_param:
+  'decode':
+    'lambda_w': 1
+    'lambda_e':
+      'mode': 'scalar'
+      'values': [0.125,0.25,0.5,1,2,4,8]
+    'defective_num_lower_bound': null
+    'sensitivity_threshold': null
+    'specificity_threshold': null
+    'is_it_noiseless': False
+    'lp_relaxation': False
+    'solver_name': 'CPLEX_PY'
+    'solver_options':
+      'mode': 'exact'
+      'timeLimit': 1800
+  'lambda_selection':
+    'cross_validation': False
+    'number_of_folds': 2
+    'cv_param':
+      'mode': 'exact'
+      'lambda_e': [0.125,0.25,0.5,1,2,4,8]
+    #'mip':True
+  'evaluation':
+    'eval_metric': 'balanced_accuracy'
 
-lambda_selection:
-  'cross_validation': False
-  'number_of_folds': 5
-  'param':
-    'lambda_e': [0.125,0.25,0.5,1,2,4,8]
-param:
-  'lambda_w': 1
-  'lambda_p': 100
-  'lambda_n': 100
-  'defective_num_lower_bound': null
-  'sensitivity_threshold': null
-  'specificity_threshold': null
-  'is_it_noiseless': False
-  'lp_relaxation': False
-  'solver_name': 'CPLEX_PY'
-  'solver_options':
-    'mode': 'exact'
-    'timeLimit': 1800

generate_test_results.py

@@ -88,7 +88,7 @@ def gen_test_vector(A, u, opts):
         elif method == 'threshold':
 
             # find the group sizes
-            Asum = np.sum(A, axis = 1)
+            Asum = np.sum(A, axis=1)
 
             print('sum is ' + str(Asum.shape))