updated a script to extract results

phy-q · Dec 31, 2021 · 0a0fd31 · 0a0fd31
1 parent e68730d
commit 0a0fd31
Showing 1 changed file with 61 additions and 0 deletions.
diff --git a/sciencebirdsagents/PhyreAgents/results/get_results.py b/sciencebirdsagents/PhyreAgents/results/get_results.py
@@ -0,0 +1,61 @@
+import os
+import argparse
+
+import numpy as np
+import pandas as pd
+def arg_parse():
+    parser = argparse.ArgumentParser(description='RPIN Parameters')
+    parser.add_argument('--folder', required=True, help='folder name to retrive results', type=str)
+
+    return parser.parse_args()
+
+
+def main():
+    '''
+    returns two csv files:
+    2. scenario/template, temp_num, avg., std
+    '''
+    args = arg_parse()
+    log_files = os.listdir(args.folder)
+    ret_1 = {}
+
+    for log_file in log_files:
+
+        with open(os.path.join(args.folder,log_file), 'r') as f:
+            temp_num = list(filter(lambda x: 'log' not in x, log_file.replace('.txt', '').split('_')))
+            if len(temp_num) == 3:
+                mode = 'template'
+            elif len(temp_num) == 2:
+                mode = 'scenario'
+            else:
+                raise ValueError(f'incorrect length of temp {log_file}')
+
+            temp_num = "_".join(temp_num)
+
+            content = f.readlines()
+            passing_rates = list(filter(lambda x : 'on val levels' in x and 'INFO: 004/' in x, content))
+            for rate in passing_rates:
+                rate = float(rate.split(':')[-1])
+                if temp_num in ret_1:
+                    num_fold = len(ret_1[temp_num])
+                    ret_1[temp_num][f'fold_{num_fold}'] = rate if str(rate) != 'nan' else 0
+                else:
+                    ret_1[temp_num] = {}
+                    ret_1[temp_num]['fold_0'] = rate if str(rate) != 'nan' else 0
+
+    out = {'template':[], 'average':[], 'std':[] }
+
+    for temp_num in ret_1:
+        avg = np.mean(list(ret_1[temp_num].values()))
+        std = np.std(list(ret_1[temp_num].values()))
+
+        out['template'].append(temp_num)
+        out['average'].append(avg)
+        out['std'].append(std)
+    out = pd.DataFrame(out)
+
+    out.to_csv(args.folder+'.csv', index=False)
+
+    return out
+if __name__ == '__main__':
+    results = main()