fix a bug

rdagent/app/data_science/loop.py

@@ -1,6 +1,6 @@
+from pathlib import Path
 from typing import Any
 
-from pathlib import Path
 import fire
 
 from rdagent.app.data_science.conf import DS_RD_SETTING
@@ -100,7 +100,7 @@ def feedback(self, prev_out: dict[str, Any]):
         self.trace.hist.append((prev_out["direct_exp_gen"].hypothesis, prev_out["running"], feedback))
 
 
-def main(path=None, step_n=None, competition=None):
+def main(path=None, step_n=None, competition="bms-molecular-translation"):
     """
     Auto R&D Evolving loop for models in a kaggle{} scenario.
     You can continue running session by

rdagent/components/coder/CoSTEER/evolving_agent.py

@@ -4,6 +4,7 @@
 from rdagent.core.evolving_framework import EvolvableSubjects
 from rdagent.core.exception import CoderError
 
+
 class FilterFailedRAGEvoAgent(RAGEvoAgent):
     def filter_evolvable_subjects_by_feedback(
         self, evo: EvolvableSubjects, feedback: CoSTEERSingleFeedbackDeprecated
@@ -15,8 +16,8 @@ def filter_evolvable_subjects_by_feedback(
         for index in range(len(evo.sub_workspace_list)):
             if evo.sub_workspace_list[index] is not None and feedback[index] and not feedback[index].final_decision:
                 evo.sub_workspace_list[index].clear()
-        
+
         if all(not f.final_decision for f in feedback if f):
             raise CoderError("All feedbacks of sub tasks are negative.")
-        
+
         return evo

rdagent/components/coder/data_science/ensemble/exp.py

@@ -7,5 +7,4 @@
 from rdagent.components.coder.CoSTEER.task import CoSTEERTask
 from rdagent.core.utils import cache_with_pickle
 
-
 EnsembleTask = CoSTEERTask

rdagent/components/coder/data_science/feature/eval.py

@@ -52,7 +52,8 @@ def evaluate(
             test_code = f.read()
             implementation.inject_files(**{fname: test_code})
         stdout = implementation.execute(env=de, entry=f"python {fname}")
-
+        if stdout is None:
+            stdout = "The execution exceeded the time limit, and no stdout information has been generated yet."
         system_prompt = T(".prompts:feature_eval.system").r(
             test_code=test_code, code=implementation.file_dict["feat01.py"]
         )

rdagent/components/coder/data_science/feature/exp.py

@@ -7,5 +7,4 @@
 from rdagent.components.coder.CoSTEER.task import CoSTEERTask
 from rdagent.core.utils import cache_with_pickle
 
-
 FeatureTask = CoSTEERTask

rdagent/components/coder/data_science/model/eval.py

@@ -70,6 +70,8 @@ def evaluate(
             test_code = f.read()
             implementation.inject_files(**{fname: test_code})
         stdout = implementation.execute(env=de, entry=f"python {fname}")
+        if stdout is None:
+            stdout = "The execution exceeded the time limit, and no stdout information has been generated yet."
         system_prompt = T(".prompts:model_eval.system").r(
             test_code=test_code, scenario="No scenario information yet.", spec=implementation.file_dict["spec/model.md"]
         )

rdagent/components/coder/data_science/raw_data_loader/eval.py

@@ -57,7 +57,8 @@ def evaluate(
             test_code = f.read()
             implementation.inject_files(**{fname: test_code})
         stdout = implementation.execute(env=de, entry=f"python {fname}")
-
+        if stdout is None:
+            stdout = "The execution exceeded the time limit, and no stdout information has been generated yet."
         system_prompt = T(".prompts:data_loader_eval.system").r(
             test_code=test_code, code=implementation.file_dict["load_data.py"]
         )

rdagent/components/coder/data_science/raw_data_loader/exp.py

@@ -11,5 +11,4 @@
 from rdagent.utils.agent.tpl import T
 from rdagent.utils.env import DockerEnv, DSDockerConf
 
-
 DataLoaderTask = CoSTEERTask

rdagent/components/coder/data_science/workflow/eval.py

@@ -55,6 +55,8 @@ def evaluate(
         de = DockerEnv(conf=ds_docker_conf)
         fname = "main.py"
         stdout = implementation.execute(env=de, entry=f"python {fname}")
+        if stdout is None:
+            stdout = "The execution exceeded the time limit, and no stdout information has been generated yet."
         system_prompt = T(".prompts:workflow_eval.system").r(
             scenario="No scenario information yet.", spec=implementation.file_dict["spec/workflow.md"]
         )

rdagent/components/coder/data_science/workflow/exp.py

@@ -7,5 +7,4 @@
 from rdagent.components.coder.CoSTEER.task import CoSTEERTask
 from rdagent.core.utils import cache_with_pickle
 
-
-WorkflowTask = CoSTEERTask
+WorkflowTask = CoSTEERTask

rdagent/core/experiment.py

@@ -50,6 +50,7 @@ def get_task_information(self) -> str:
     def __repr__(self) -> str:
         return f"<{self.__class__.__name__} {self.name}>"
 
+
 ASpecificTask = TypeVar("ASpecificTask", bound=Task)
 
 

rdagent/log/logger.py

@@ -1,7 +1,7 @@
 import json
 import os
-import sys
 import pickle
+import sys
 from contextlib import contextmanager
 from datetime import datetime, timezone
 from functools import partial

rdagent/scenarios/data_science/scen/__init__.py

@@ -1,4 +1,4 @@
-from .scen import DataScienceScen
 from .kaggle import KaggleScen
+from .scen import DataScienceScen
 
 __all__ = ["DataScienceScen", "KaggleScen"]

rdagent/scenarios/data_science/scen/kaggle.py

@@ -19,6 +19,7 @@ class KaggleScen(DataScienceScen):
           But we found that too much scenario unrelated code in kaggle scenario and hard to reuse.
           So we start from a simple one....
     """
+
     def _get_description(self):
         return crawl_descriptions(self.competition, DS_RD_SETTING.local_data_path)
 
@@ -32,4 +33,3 @@ def rich_style_description(self) -> str:
             name="Kaggle",
             competition=f"[{self.competition}](https://www.kaggle.com/competitions/{self.competition})",
         )
-

rdagent/scenarios/data_science/scen/scen.py

@@ -3,14 +3,13 @@
 
 from rdagent.app.data_science.conf import DS_RD_SETTING
 from rdagent.core.scenario import Scenario
+from rdagent.log import rdagent_logger as logger
 from rdagent.oai.llm_utils import APIBackend
 from rdagent.utils.agent.tpl import T
-from rdagent.log import rdagent_logger as logger
 
 
 class DataScienceScen(Scenario):
-    """Data Science Scenario
-    """
+    """Data Science Scenario"""
 
     def __init__(self, competition: str) -> None:
         self.competition = competition
@@ -24,7 +23,9 @@ def _get_description(self):
             with fp.open("r") as f:
                 return json.load(f)
         else:
-            logger.error(f"Cannot find {self.competition}.json in {DS_RD_SETTING.local_data_path}, please check the file.")
+            logger.error(
+                f"Cannot find {self.competition}.json in {DS_RD_SETTING.local_data_path}, please check the file."
+            )
 
     def _get_direction(self):
         return self.raw_description.get("metric_direction", "minimize")

test/utils/coder/test_CoSTEER.py

@@ -1,5 +1,6 @@
 import unittest
 
+
 class CoSTEERTest(unittest.TestCase):
 
     def setUp(self):
@@ -23,26 +24,31 @@ def test_feature(self):
         from rdagent.components.coder.data_science.feature.test import (
             develop_one_competition,
         )
+
         exp = develop_one_competition(self.test_competition)
 
     def test_model(self):
         from rdagent.components.coder.data_science.model.test import (
             develop_one_competition,
         )
+
         exp = develop_one_competition(self.test_competition)
 
     def test_ensemble(self):
         from rdagent.components.coder.data_science.ensemble.test import (
             develop_one_competition,
         )
+
         exp = develop_one_competition(self.test_competition)
-    
+
     def test_workflow(self):
         from rdagent.components.coder.data_science.workflow.test import (
             develop_one_competition,
         )
+
         exp = develop_one_competition(self.test_competition)
 
+
 if __name__ == "__main__":
     unittest.main()
     # pytest test/utils/coder/test_CoSTEER.py