Command line args for logging level

README.md

@@ -50,6 +50,11 @@ Many of the principles and best practices we acquired in this work, we believe,
 
 ## Installation
 
+### Pre-requisites
+
+1. Python3
+2. Rust compiler
+
 (1) setup a virtual environment：
 ```bash
 python3 -m venv venv

alpha_codium/constants.py

@@ -0,0 +1,3 @@
+"""Common constants used across the codebase."""
+
+LOG_LEVEL_CHOICES = ['DEBUG', 'INFO', 'WARNING', 'ERROR', 'CRITICAL']

alpha_codium/solve_dataset.py

@@ -2,16 +2,20 @@
 
 from alpha_codium.gen.dataset_solver import solve_dataset
 from alpha_codium.log import get_logger, setup_logger
+from alpha_codium.constants import LOG_LEVEL_CHOICES
 
 logger = get_logger(__name__)
 
 parser = argparse.ArgumentParser()
 parser.add_argument("--dataset_name", type=str, default="valid_and_test_processed")
 parser.add_argument("--split_name", type=str, default="valid")
 parser.add_argument("--database_solution_path", type=str, default="")
+parser.add_argument('--log-level', type=str, default='INFO',
+                    choices=LOG_LEVEL_CHOICES)
+
 if __name__ == "__main__":
     args = parser.parse_args()
-    setup_logger()
+    setup_logger(level=args.log_level)
 
     # set default database_solution_path
     args.database_solution_path = args.database_solution_path

alpha_codium/solve_my_problem.py

@@ -3,14 +3,18 @@
 
 from alpha_codium.gen.coding_competitor import solve_problem, solve_my_problem
 from alpha_codium.log import setup_logger
+from alpha_codium.constants import LOG_LEVEL_CHOICES
 from alpha_codium.settings.config_loader import get_settings
 
 parser = argparse.ArgumentParser()
 parser.add_argument("--my_problem_json_file", type=str, default="my_problem_example.json")
+parser.add_argument('--log-level', type=str, default='INFO',
+                    choices=LOG_LEVEL_CHOICES)
+
 
 if __name__ == "__main__":
     args = parser.parse_args()
-    setup_logger()
+    setup_logger(level=args.log_level)
 
     with open(args.my_problem_json_file, "r") as my_problem:
         solve_my_problem(json.load(my_problem))

alpha_codium/solve_problem.py

@@ -2,17 +2,20 @@
 
 from alpha_codium.gen.coding_competitor import solve_problem
 from alpha_codium.log import setup_logger
+from alpha_codium.constants import LOG_LEVEL_CHOICES
 from alpha_codium.settings.config_loader import get_settings
 
 parser = argparse.ArgumentParser()
 parser.add_argument("--dataset_name", type=str, default="valid_and_test_processed")
 parser.add_argument("--split_name", type=str, default="valid")
 parser.add_argument("--problem_number", type=int, default=0)
 parser.add_argument("--problem_name", type=str, default="")
+parser.add_argument('--log-level', type=str, default='INFO',
+                    choices=LOG_LEVEL_CHOICES)
 
 if __name__ == "__main__":
     args = parser.parse_args()
-    setup_logger()
+    setup_logger(level=args.log_level)
     solve_problem(dataset_name=args.dataset_name,
                   split_name=args.split_name,
                   problem_number=args.problem_number,

requirements.txt

@@ -1,10 +1,10 @@
 dynaconf==3.1.12
-fastapi==0.99.0
+fastapi==0.115.5
 PyGithub==1.59.*
 retry==0.9.2
 Jinja2==3.1.2
-tiktoken==0.5.2
-uvicorn==0.22.0
+tiktoken==0.8.0
+uvicorn==0.24.0
 pytest==7.4.0
 aiohttp==3.9.3
 atlassian-python-api==3.39.0
@@ -19,7 +19,7 @@ msrest==0.7.1
 ##
 openai
 litellm
-duckdb==0.9.2
+duckdb==1.0.0
 datasets
 notebook
 black