use description from markdown

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

@@ -37,6 +37,9 @@ train_X, val_X, train_y, val_y = train_test_split(X, y, test_size=0.2, random_st
 
 # Print the types of train_y and val_y
 print(f"train_y type: {type(train_y)}, val_y type: {type(val_y)}")
+if isinstance(val_y, pd.Series | pd.DataFrame):
+    print(f"val_y shape: {val_y.shape} and first few rows of val_y:")
+    print(val_y.head())
 
 test_preds_dict = {}
 val_preds_dict = {}

rdagent/components/coder/data_science/model/eval_tests/model_test.txt

@@ -24,6 +24,8 @@ print(f"train_X.shape: {train_X.shape}")
 print(f"train_y.shape: {train_y.shape}" if not isinstance(train_y, list) else f"train_y(list)'s length: {len(train_y)}")
 print(f"val_X.shape: {val_X.shape}")
 print(f"val_y.shape: {val_y.shape}" if not isinstance(val_y, list) else f"val_y(list)'s length: {len(val_y)}")
+if isinstance(train_y, list) and not isinstance(val_y[0], int | str | float):
+    print(f"first 3 train_y: {train_y[:3]}")
 
 # First execution
 print("The first execution begins.\n")

rdagent/components/coder/data_science/raw_data_loader/eval_tests/data_loader_test.txt

@@ -30,6 +30,7 @@ assert X is not None, "Training data (X) is None."
 assert y is not None, "Training labels (y) are None."
 assert X_test is not None, "Test data (X_test) is None."
 assert test_ids is not None, "Test IDs (test_ids) are None."
+assert X.shape
 
 assert get_length(X_test) == get_length(
     test_ids

rdagent/scenarios/data_science/scen/__init__.py

@@ -229,7 +229,6 @@ class DataScienceScen(Scenario):
     def __init__(self, competition: str) -> None:
         self.competition = competition
         self.raw_description = self._get_description()
-        self.json_description = self._get_description_json()
         self.processed_data_folder_description = self._get_data_folder_description()
         self._analysis_competition_description()
         self.metric_direction = self._get_direction()
@@ -244,9 +243,6 @@ def _get_description(self):
                 f"Cannot find {self.competition}.json in {DS_RD_SETTING.local_data_path}, please check the file."
             )
 
-    def _get_description_json(self):
-        return self.raw_description
-
     def _get_direction(self):
         return self.metric_direction_guess if hasattr(self, "metric_direction_guess") else True
 
@@ -266,9 +262,9 @@ def _analysis_competition_description(self):
         response_json_analysis = json.loads(response_analysis)
         self.task_type = response_json_analysis.get("Task Type", "No type provided")
         self.data_type = response_json_analysis.get("Data Type", "No data type provided")
-        self.brief_description = self.json_description.get("Description", response_json_analysis.get("Brief Description", "No brief description provided"))
-        self.dataset_description = self.json_description.get("Data Description", response_json_analysis.get("Dataset Description", "No dataset description provided"))
-        self.target_description = self.json_description.get("Evaluation", response_json_analysis.get("Evaluation Description", "No target description provided"))
+        self.brief_description = response_json_analysis.get("Brief Description", "No brief description provided")
+        self.dataset_description = response_json_analysis.get("Dataset Description", "No dataset description provided")
+        self.target_description = response_json_analysis.get("Evaluation Description", "No target description provided")
         self.submission_specifications = response_json_analysis.get(
             "Submission Specifications", "No submission requirements provided"
         )
@@ -291,9 +287,7 @@ def background(self) -> str:
         background_prompt = background_template.r(
             task_type=self.task_type,
             data_type=self.data_type,
-            brief_description=self.brief_description,
-            dataset_description=self.dataset_description,
-            target_description=self.target_description,
+            raw_description=self.raw_description,
         )
         return background_prompt
 
@@ -340,9 +334,6 @@ class KaggleScen(DataScienceScen):
     def _get_description(self):
         return crawl_descriptions(self.competition, DS_RD_SETTING.local_data_path)
 
-    def _get_description_json(self):
-        return crawl_descriptions(self.competition, DS_RD_SETTING.local_data_path, force=True)
-
     def _get_direction(self):
         if DS_RD_SETTING.if_using_mle_data:
             return super()._get_direction()

rdagent/scenarios/data_science/scen/prompts.yaml

@@ -48,9 +48,7 @@ competition_background: |-
 
   The task type for this competition is {{ task_type }}.
   The data type used in this competition is {{ data_type }}.
-  Briefly, the competition involves: {{ brief_description }}.
-  The dataset used in this competition is: {{ dataset_description }}.
-  Your goal in this competition is to: {{target_description }}.
+  The description of the competition is: {{ raw_description }}
 
 rich_style_description: |-
   ### {{ name }} Agent: Automated Feature Engineering & Model Tuning Evolution

rdagent/scenarios/kaggle/kaggle_crawler.py

@@ -41,7 +41,7 @@ def crawl_descriptions(
         logger.info(f"Found {competition}/description.md, loading from it.")
         return fp.read_text()
 
-    if (fp := Path(f"{local_data_path}/{competition}.json")).exists():
+    if (fp := Path(f"{local_data_path}/{competition}.json")).exists() and not force:
         logger.info(f"Found {competition}.json, loading from local file.")
         with fp.open("r") as f:
             return json.load(f)