Make sure opened files are closed

datasets/apps_dataset.py

@@ -136,15 +136,17 @@ def initialize(self):
                 answer_type = "\nUse Standard Input format\n"
                 starter_code = ""
 
-            sols_str_list = json.load(open(sols_fname, 'r'))
+            with open(sols_fname, 'r') as f:
+                sols_str_list = json.load(f)
             gt_samples = self.load_gt_samples(sols_str_list, answer_type, starter_code, question_str)
             all_samples += gt_samples
 
             # Read all the solutions
             if self.tuning_mode in ['critic']:
                 for fname in gen_sols_fname:
                     if os.path.exists(fname):
-                        gen_sols = json.load(open(fname, 'r'))
+                        with open(fname, 'r') as f:
+                            gen_sols = json.load(f)
                         samples, info = self.load_gen_samples(gen_sols, answer_type, starter_code, question_str)
                         self.update_error_stat(info)
                         gen_samples += samples
@@ -158,14 +160,16 @@ def initialize(self):
             elif self.tuning_mode in ['rl']:
 
                 if self.relative_returns:
-                    baseline_sample = json.load(open(baseline_fname, 'r'))
+                    with open(baseline_fname, 'r') as f:
+                        baseline_sample = json.load(f)
                     baseline_error_type = self.get_baseline_error_type(baseline_sample)
                 else:
                     baseline_error_type = -1
 
                 for fname in gen_sols_fname:
                     if os.path.exists(fname):
-                        gen_sols = pkl.load(open(fname, 'rb'))
+                        with open(fname, 'rb') as f:
+                            gen_sols = pkl.load(f)
                         samples, info = self.load_rl_samples(gen_sols, baseline_error_type)
                         self.update_error_stat_rl(info)
                         gen_samples += samples

generate.py

@@ -61,7 +61,8 @@ def generate_critic_inputs(args, test_case_path, prompt_path, solutions_path, to
     in_tokens[:len(q_tokens)] = q_tokens
     in_tokens = in_tokens[:args.source_len]
 
-    solutions = json.load(open(solutions_path, 'r'))
+    with open(solutions_path, 'r') as f:
+        solutions = json.load(f)
 
     all_texts = []
     gt_errors = []
@@ -183,7 +184,8 @@ def main(args):
                 else:
                     scores_loc = os.path.join(prob_path, "gen_solutions_critic_scores.pkl")
 
-                pkl.dump(saved_critic_scores, open(scores_loc, 'wb'))
+                with open(scores_loc, 'wb') as f:
+                    pkl.dump(saved_critic_scores, f)
 
             else:
                 input_ids = torch.LongTensor(tokenizer.encode(input_text,

test_one_solution.py

@@ -46,7 +46,8 @@ def eval_and_save_problems(args):
         gen_codes = json.load(file)[str(real_index)]['code']
 
     test_file = os.path.join(problem, "input_output.json")
-    tests = json.load(open(test_file, 'r'))
+    with open(test_file, 'r') as file:
+        tests = json.load(file)
     nb_tests = len(tests['inputs'])
     if args.max_tests != -1 and nb_tests > args.max_tests:
         print(f"{test_file} contains more tests than --max_tests")
@@ -102,7 +103,8 @@ def eval_and_save_problems(args):
     '''
 
     save_results = {real_index: {'results': all_results, 'errors': all_errors, 'sols': all_sols}}
-    pkl.dump(save_results, open(outputs_loc, "wb"))
+    with open(outputs_loc, "wb") as file:
+        pkl.dump(save_results, file)
 
 
 def main(args):

train.py

@@ -132,7 +132,8 @@ def main(args):
     train_data = get_dataset(args)
 
     # Save args to file
-    json.dump(argsdict, open(os.path.join(args.save_dir, "args.json"), 'w'))
+    with open(os.path.join(args.save_dir, "args.json"), 'w') as file:
+        json.dump(argsdict, file)
 
     # Load and train model; save model checkpoints 
     run_training(args, train_data)