fix ci

rdagent/app/data_science/loop.py

@@ -2,14 +2,18 @@
 from typing import Any, Literal
 
 import fire
-from rdagent.scenarios.data_science.experiment.experiment import DSExperiment
 
 from rdagent.app.data_science.conf import DS_RD_SETTING
-from rdagent.components.coder.data_science.raw_data_loader import DataLoaderCoSTEER
+from rdagent.components.coder.data_science.ensemble import EnsembleCoSTEER
+from rdagent.components.coder.data_science.ensemble.exp import EnsembleTask
 from rdagent.components.coder.data_science.feature import FeatureCoSTEER
+from rdagent.components.coder.data_science.feature.exp import FeatureTask
 from rdagent.components.coder.data_science.model import ModelCoSTEER
-from rdagent.components.coder.data_science.ensemble import EnsembleCoSTEER
+from rdagent.components.coder.data_science.model.exp import ModelTask
+from rdagent.components.coder.data_science.raw_data_loader import DataLoaderCoSTEER
+from rdagent.components.coder.data_science.raw_data_loader.exp import DataLoaderTask
 from rdagent.components.coder.data_science.workflow import WorkflowCoSTEER
+from rdagent.components.coder.data_science.workflow.exp import WorkflowTask
 from rdagent.components.workflow.conf import BasePropSetting
 from rdagent.components.workflow.rd_loop import NextLoopException, RDLoop
 from rdagent.core.exception import FactorEmptyError, ModelEmptyError
@@ -23,15 +27,13 @@
 from rdagent.core.scenario import Scenario
 from rdagent.core.utils import import_class
 from rdagent.log import rdagent_logger as logger
-from rdagent.scenarios.data_science.proposal.exp_gen import DSTrace, DSExpGen
-from rdagent.scenarios.kaggle.kaggle_crawler import download_data
-from rdagent.components.coder.data_science.ensemble.exp import EnsembleTask
-from rdagent.components.coder.data_science.feature.exp import FeatureTask
-from rdagent.components.coder.data_science.model.exp import ModelTask
-from rdagent.components.coder.data_science.raw_data_loader.exp import DataLoaderTask
-from rdagent.components.coder.data_science.workflow.exp import WorkflowTask
-from rdagent.scenarios.data_science.dev.runner import DSRunner
 from rdagent.scenarios.data_science.dev.feedback import DSExperiment2Feedback
+from rdagent.scenarios.data_science.dev.runner import DSRunner
+from rdagent.scenarios.data_science.experiment.experiment import DSExperiment
+from rdagent.scenarios.data_science.proposal.exp_gen import DSExpGen, DSTrace
+from rdagent.scenarios.kaggle.kaggle_crawler import download_data
+
+
 class DataScienceRDLoop(RDLoop):
     skip_loop_error = (NextLoopException,)
 
@@ -103,7 +105,7 @@ def feedback(self, prev_out: dict[str, Any]):
             prev_out["running"], prev_out["direct_exp_gen"].hypothesis, self.trace
         )
         self.trace.hist.append((prev_out["direct_exp_gen"].hypothesis, prev_out["running"], feedback))
-        
+
 
 def main(path=None, step_n=None, competition=None):
     """
@@ -129,8 +131,3 @@ def main(path=None, step_n=None, competition=None):
 
 if __name__ == "__main__":
     fire.Fire(main)
-
-
-
-
-

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

@@ -22,16 +22,13 @@
 from rdagent.components.coder.CoSTEER.knowledge_management import (
     CoSTEERQueriedKnowledge,
 )
-from rdagent.components.coder.data_science.ensemble.eval import (
-    EnsembleCoSTEEREvaluator,
-)
+from rdagent.components.coder.data_science.ensemble.eval import EnsembleCoSTEEREvaluator
 from rdagent.components.coder.data_science.ensemble.exp import EnsembleTask
+from rdagent.core.experiment import FBWorkspace
 from rdagent.core.scenario import Scenario
 from rdagent.oai.llm_utils import APIBackend
 from rdagent.utils.agent.tpl import T
 
-from rdagent.core.experiment import FBWorkspace
-
 
 class EnsembleMultiProcessEvolvingStrategy(MultiProcessEvolvingStrategy):
     def implement_one_task(
@@ -42,7 +39,7 @@ def implement_one_task(
     ) -> dict[str, str]:
         # Get task information for knowledge querying
         ensemble_information_str = target_task.get_task_information()
-        
+
         # Query knowledge
         queried_similar_successful_knowledge = (
             queried_knowledge.task_to_similar_task_successful_knowledge[ensemble_information_str]
@@ -60,24 +57,22 @@ def implement_one_task(
         system_prompt = T(".prompts:ensemble_coder.system").r(
             competition_info=competition_info,
             queried_similar_successful_knowledge=queried_similar_successful_knowledge,
-            queried_former_failed_knowledge=queried_former_failed_knowledge[0] if queried_former_failed_knowledge else None
-        )
-        user_prompt = T(".prompts:ensemble_coder.user").r(
-            ensemble_spec=workspace.code_dict["spec/ensemble.md"]
+            queried_former_failed_knowledge=(
+                queried_former_failed_knowledge[0] if queried_former_failed_knowledge else None
+            ),
         )
+        user_prompt = T(".prompts:ensemble_coder.user").r(ensemble_spec=workspace.code_dict["spec/ensemble.md"])
 
         ensemble_code = json.loads(
             APIBackend().build_messages_and_create_chat_completion(
-                user_prompt=user_prompt,
-                system_prompt=system_prompt,
-                json_mode=True
+                user_prompt=user_prompt, system_prompt=system_prompt, json_mode=True
             )
         )["code"]
 
         return {
             "ensemble.py": ensemble_code,
         }
-    
+
     def assign_code_list_to_evo(self, code_list: list[dict[str, str]], evo):
         """
         Assign the code list to the evolving item.
@@ -102,19 +97,7 @@ def __init__(
         *args,
         **kwargs,
     ) -> None:
-        eva = CoSTEERMultiEvaluator(
-            EnsembleCoSTEEREvaluator(scen=scen), scen=scen
-        )
+        eva = CoSTEERMultiEvaluator(EnsembleCoSTEEREvaluator(scen=scen), scen=scen)
         es = EnsembleMultiProcessEvolvingStrategy(scen=scen, settings=CoSTEER_SETTINGS)
 
-        super().__init__(
-            *args,
-            settings=CoSTEER_SETTINGS,
-            eva=eva,
-            es=es,
-            evolving_version=2,
-            scen=scen,
-            **kwargs
-        )
-
-
+        super().__init__(*args, settings=CoSTEER_SETTINGS, eva=eva, es=es, evolving_version=2, scen=scen, **kwargs)

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

@@ -1,2 +1,2 @@
 # Configuration file for ensemble component
-# Currently empty as no specific configuration is needed 
+# Currently empty as no specific configuration is needed

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

@@ -1,6 +1,7 @@
 import json
 from dataclasses import dataclass
 from pathlib import Path
+
 import numpy as np
 
 from rdagent.components.coder.CoSTEER.evaluators import (
@@ -12,12 +13,13 @@
 from rdagent.core.experiment import FBWorkspace, Task
 from rdagent.oai.llm_utils import APIBackend
 from rdagent.utils.agent.tpl import T
-from rdagent.utils.env import DSDockerConf, DockerEnv
+from rdagent.utils.env import DockerEnv, DSDockerConf
 
 DIRNAME = Path(__file__).absolute().resolve().parent
 
 EnsembleEvalFeedback = CoSTEERSingleFeedback
 
+
 class EnsembleCoSTEEREvaluator(CoSTEEREvaluator):
     def evaluate(
         self,
@@ -27,10 +29,12 @@ def evaluate(
         queried_knowledge: QueriedKnowledge = None,
         **kwargs,
     ) -> EnsembleEvalFeedback:
-        
+
         target_task_information = target_task.get_task_information()
-        if (queried_knowledge is not None and
-                target_task_information in queried_knowledge.success_task_to_knowledge_dict):
+        if (
+            queried_knowledge is not None
+            and target_task_information in queried_knowledge.success_task_to_knowledge_dict
+        ):
             return queried_knowledge.success_task_to_knowledge_dict[target_task_information].feedback
         elif queried_knowledge is not None and target_task_information in queried_knowledge.failed_task_info_set:
             return EnsembleEvalFeedback(
@@ -51,7 +55,5 @@ def evaluate(
         system_prompt = T(".prompts:ensemble_eval.system").r(test_code=test_code)
         user_prompt = T(".prompts:ensemble_eval.user").r(stdout=stdout)
 
-        resp = APIBackend().build_messages_and_create_chat_completion(
-            user_prompt, system_prompt, json_mode=True
-        )
-        return EnsembleEvalFeedback(**json.loads(resp)) 
+        resp = APIBackend().build_messages_and_create_chat_completion(user_prompt, system_prompt, json_mode=True)
+        return EnsembleEvalFeedback(**json.loads(resp))

rdagent/components/coder/data_science/ensemble/eval_tests/ensemble_test.py

@@ -5,12 +5,14 @@
 - Have correct shapes for inputs and outputs
 - Use validation data appropriately
 """
-import numpy as np
+
 import logging
+
+import numpy as np
 from ensemble import ens_and_decision
 
 # Setup logging
-logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
+logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
 
 # Create test data
 n_models = 3
@@ -24,17 +26,17 @@
 # Run ensemble
 try:
     final_predictions = ens_and_decision(test_pred_l, val_pred_l, val_label)
-    
+
     # Check shape
     assert final_predictions.shape == (n_samples, 1), "Wrong output shape"
-    
+
     # Check binary values
     assert np.all(np.isin(final_predictions, [0, 1])), "Predictions must be binary (0 or 1)"
-    
+
     logging.info("Ensemble test passed successfully.")
     logging.info(f"Output shape: {final_predictions.shape}")
     logging.info(f"Unique values in predictions: {np.unique(final_predictions)}")
-    
+
 except Exception as e:
     logging.error(f"Test failed: {str(e)}")
-    raise 
+    raise

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

@@ -1,23 +1,31 @@
 """
 Helper functions for testing the ensemble coder(CoSTEER-based) component.
 """
+
 import sys
 from pathlib import Path
 
 from rdagent.components.coder.data_science.ensemble import EnsembleCoSTEER
 from rdagent.components.coder.data_science.ensemble.exp import EnsembleTask
-from rdagent.scenarios.data_science.scen import DataScienceScen
 from rdagent.scenarios.data_science.experiment.experiment import DSExperiment
+from rdagent.scenarios.data_science.scen import DataScienceScen
 
 # Add the competition folder to path
-COMPETITION_PATH = Path(__file__).parent.parent.parent.parent.parent / "scenarios" / "kaggle" / "tpl_ex" / "aerial-cactus-identification"
+COMPETITION_PATH = (
+    Path(__file__).parent.parent.parent.parent.parent
+    / "scenarios"
+    / "kaggle"
+    / "tpl_ex"
+    / "aerial-cactus-identification"
+)
 sys.path.append(str(COMPETITION_PATH))
 
-EnsembleExperiment = DSExperiment 
+EnsembleExperiment = DSExperiment
+
 
 def load_ensemble_spec():
     spec_path = COMPETITION_PATH / "spec" / "ensemble.md"
-    with open(spec_path, 'r') as f:
+    with open(spec_path, "r") as f:
         return f.read()
 
 
@@ -31,15 +39,12 @@ def develop_ensemble():
     # Create the ensemble task with actual data context and specification
     task = EnsembleTask(
         name="EnsembleTask",
-        description=
-        """
+        description="""
         Implement ensemble and decision making for model predictions.
-        """
+        """,
     )
 
-    exp = EnsembleExperiment(
-        sub_tasks=[task]
-    )
+    exp = EnsembleExperiment(sub_tasks=[task])
 
     # Injecting the corresponding specification
     exp.experiment_workspace.inject_code(**{"spec/ensemble.md": ensemble_spec})
@@ -50,4 +55,4 @@ def develop_ensemble():
 
 
 if __name__ == "__main__":
-    develop_ensemble() 
+    develop_ensemble()

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

@@ -41,8 +41,10 @@ def implement_one_task(
         )
 
         # 2. code
-        system_prompt = T(".prompts:feature.system").r(queried_similar_successful_knowledge=queried_similar_successful_knowledge,
-                    queried_former_failed_knowledge=queried_former_failed_knowledge[0])
+        system_prompt = T(".prompts:feature.system").r(
+            queried_similar_successful_knowledge=queried_similar_successful_knowledge,
+            queried_former_failed_knowledge=queried_former_failed_knowledge[0],
+        )
         user_prompt = T(".prompts:feature.user").r(
             feature_spec=workspace.code_dict["spec/feature.md"],
             latest_code=workspace.code_dict.get("feat01.py"),

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

@@ -71,9 +71,7 @@ def evaluate(
             implementation.inject_code(**{fname: test_code})
         stdout = implementation.execute(env=de, entry=f"python {fname}")
         system_prompt = T(".prompts:model_eval.system").r(
-            test_code=test_code,
-            scenario="No scenario information yet.",
-            spec=implementation.code_dict["spec/model.md"]
+            test_code=test_code, scenario="No scenario information yet.", spec=implementation.code_dict["spec/model.md"]
         )
         user_prompt = T(".prompts:model_eval.user").r(
             stdout=stdout,

rdagent/components/coder/data_science/model/eval_tests/model_execute.py

@@ -24,7 +24,11 @@
 print("The first execution begins.\n")
 # Call model_workflow
 val_pred, test_pred, hypers = model_workflow(
-    X=train_X, y=train_y, val_X=val_X, val_y=val_y, test_X=None,
+    X=train_X,
+    y=train_y,
+    val_X=val_X,
+    val_y=val_y,
+    test_X=None,
 )
 # val_pred = np.random.rand(8, 1)
 # test_pred = np.random.rand(8, 1)

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

@@ -66,13 +66,17 @@ def implement_one_task(
             if queried_knowledge is not None
             else []
         )
-    
+
         # 1. specifications
         # TODO: Why is queried_former_failed_knowledge[0] used here?
-        system_prompt = T(".prompts:spec.system").r(competition_info=competition_info,
-                                                    queried_similar_successful_knowledge=queried_similar_successful_knowledge,
-                    queried_former_failed_knowledge=queried_former_failed_knowledge[0])
-        data_loader_prompt = T(".prompts:spec.user.data_loader").r(latest_spec=workspace.code_dict.get("spec/data_loader.md"))
+        system_prompt = T(".prompts:spec.system").r(
+            competition_info=competition_info,
+            queried_similar_successful_knowledge=queried_similar_successful_knowledge,
+            queried_former_failed_knowledge=queried_former_failed_knowledge[0],
+        )
+        data_loader_prompt = T(".prompts:spec.user.data_loader").r(
+            latest_spec=workspace.code_dict.get("spec/data_loader.md")
+        )
         feature_prompt = T(".prompts:spec.user.feature").r(latest_spec=workspace.code_dict.get("spec/feature.md"))
         model_prompt = T(".prompts:spec.user.model").r(latest_spec=workspace.code_dict.get("spec/model.md"))
         ensemble_prompt = T(".prompts:spec.user.ensemble").r(latest_spec=workspace.code_dict.get("spec/ensemble.md"))
@@ -97,10 +101,12 @@ def implement_one_task(
         # 2. code
         system_prompt = T(".prompts:data_loader_coder.system").r(
             queried_similar_successful_knowledge=queried_similar_successful_knowledge,
-                    queried_former_failed_knowledge=queried_former_failed_knowledge[0])
+            queried_former_failed_knowledge=queried_former_failed_knowledge[0],
+        )
         user_prompt = T(".prompts:data_loader_coder.user").r(
-            competition_info=competition_info, data_loader_spec=data_loader_spec,
-            latest_code=workspace.code_dict.get("load_data.py")
+            competition_info=competition_info,
+            data_loader_spec=data_loader_spec,
+            latest_code=workspace.code_dict.get("load_data.py"),
         )
 
         data_loader_code = json.loads(

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

@@ -8,7 +8,9 @@
     CoSTEEREvaluator,
     CoSTEERSingleFeedback,
 )
-from rdagent.components.coder.CoSTEER.knowledge_management import CoSTEERQueriedKnowledgeV2
+from rdagent.components.coder.CoSTEER.knowledge_management import (
+    CoSTEERQueriedKnowledgeV2,
+)
 from rdagent.core.evaluation import Feedback
 from rdagent.core.experiment import FBWorkspace, Task, Workspace
 from rdagent.oai.llm_utils import APIBackend
@@ -56,7 +58,9 @@ def evaluate(
             implementation.inject_code(**{fname: test_code})
         stdout = implementation.execute(env=de, entry=f"python {fname}")
 
-        system_prompt = T(".prompts:data_loader_eval.system").r(test_code=test_code, code=implementation.code_dict["load_data.py"])
+        system_prompt = T(".prompts:data_loader_eval.system").r(
+            test_code=test_code, code=implementation.code_dict["load_data.py"]
+        )
         user_prompt = T(".prompts:data_loader_eval.user").r(stdout=stdout)
 
         resp = APIBackend().build_messages_and_create_chat_completion(user_prompt, system_prompt, json_mode=True)