New Structure Demo (#120)

better demo
---------

Co-authored-by: Young <[email protected]>
Co-authored-by: Taozhi Wang <[email protected]>
Co-authored-by: you-n-g <[email protected]>
Co-authored-by: cyncyw <[email protected]>

rdagent/app/model_extraction_and_code/GeneralModel.py

@@ -1,13 +1,12 @@
 from pathlib import Path
 
-from rdagent.components.coder.model_coder.model import (
-    ModelExperiment,
-)
+from rdagent.components.coder.model_coder.model import ModelExperiment
 from rdagent.core.prompts import Prompts
 from rdagent.core.scenario import Scenario
 
 prompt_dict = Prompts(file_path=Path(__file__).parent / "prompts.yaml")
 
+
 class GeneralModelScenario(Scenario):
     @property
     def background(self) -> str:
@@ -28,10 +27,10 @@ def interface(self) -> str:
     @property
     def simulator(self) -> str:
         return prompt_dict["general_model_simulator"]
-    
+
     @property
-    def rich_style_description(self)->str:
-        return '''
+    def rich_style_description(self) -> str:
+        return """
 # General Model Scenario
 
 ## Overview
@@ -61,7 +60,7 @@ def rich_style_description(self)->str:
 - **Time-Series Data:** Sequential data points indexed in time order, useful for forecasting and temporal pattern recognition.
 - **Graph Data:** Data structured as nodes and edges, suitable for network analysis and relational tasks.
 
-        '''
+        """
 
     def get_scenario_all_desc(self) -> str:
         return f"""Background of the scenario:

rdagent/app/model_extraction_and_code/model_extraction_and_implementation.py

@@ -3,18 +3,24 @@
 
 load_dotenv(override=True)
 
+import fire
+
+from rdagent.app.model_extraction_and_code.GeneralModel import GeneralModelScenario
 from rdagent.components.coder.model_coder.task_loader import (
     ModelExperimentLoaderFromPDFfiles,
 )
-from rdagent.scenarios.qlib.developer.model_coder import QlibModelCoSTEER
-from rdagent.app.model_extraction_and_code.GeneralModel import GeneralModelScenario
-from rdagent.components.document_reader.document_reader import extract_first_page_screenshot_from_pdf
+from rdagent.components.document_reader.document_reader import (
+    extract_first_page_screenshot_from_pdf,
+)
 from rdagent.log import rdagent_logger as logger
-import fire
+from rdagent.scenarios.qlib.developer.model_coder import QlibModelCoSTEER
 
-def extract_models_and_implement(report_file_path: str = "/home/v-xisenwang/RD-Agent/rdagent/app/model_extraction_and_code/test_doc1.pdf") -> None:
+
+def extract_models_and_implement(
+    report_file_path: str = "/home/v-xisenwang/RD-Agent/rdagent/app/model_extraction_and_code/test_doc1.pdf",
+) -> None:
     with logger.tag("r"):
-    # Save Relevant Images 
+        # Save Relevant Images
         img = extract_first_page_screenshot_from_pdf(report_file_path)
         logger.log_object(img, tag="pdf_image")
         scenario = GeneralModelScenario()
@@ -26,5 +32,6 @@ def extract_models_and_implement(report_file_path: str = "/home/v-xisenwang/RD-A
         logger.log_object(exp, tag="developed_experiment")
     return exp
 
+
 if __name__ == "__main__":
     fire.Fire(extract_models_and_implement)

rdagent/app/quant_factor_benchmark/analysis.py

@@ -41,6 +41,7 @@ def process_results(self, results):
             processed_data = self.analyze_data(summarized_data)
             final_res[experiment] = processed_data.iloc[-1, :]
         return final_res
+
     def reformat_succ_rate(self, display_df):
         new_idx = []
         display_df = display_df[display_df.index.isin(self.index_map.keys())]
@@ -166,6 +167,7 @@ def plot_data(data, file_name):
         plt.title("Comparison of Different Methods")
         plt.savefig(file_name)
 
+
 if __name__ == "__main__":
     settings = BenchmarkSettings()
     benchmark = BenchmarkAnalyzer(settings)

rdagent/app/quant_factor_benchmark/eval.py

@@ -15,12 +15,6 @@
     FactorTestCaseLoaderFromJsonFile,
 )
 
-from rdagent.core.utils import import_class
-from rdagent.core.scenario import Scenario
-from rdagent.scenarios.qlib.experiment.factor_experiment import QlibFactorScenario
-
-from pprint import pprint
-
 # 1.read the settings
 bs = BenchmarkSettings()
 

rdagent/components/benchmark/eval_method.py

@@ -35,6 +35,7 @@
     List[Tuple[FactorEvaluator, Union[object, RunnerException]]],
 ]
 
+
 class TestCase:
     def __init__(
         self,

rdagent/core/proposal.py

@@ -112,7 +112,7 @@ def gen(self, trace: Trace) -> Hypothesis:
 
 class Hypothesis2Experiment(ABC, Generic[ASpecificExp]):
     """
-    [Abstract description => concrete description] => Code implementation Card 
+    [Abstract description => concrete description] => Code implementation Card
     """
 
     @abstractmethod

rdagent/log/storage.py

@@ -90,22 +90,28 @@ def iter_msg(self, watch: bool = False) -> Generator[Message, None, None]:
                 message_end = next_match.start() if next_match else len(content)
                 message_content = content[message_start:message_end].strip()
 
+                if "Logging object in" in message_content:
+                    continue
+
                 m = Message(
                     tag=tag, level=level, timestamp=timestamp, caller=caller, pid_trace=pid, content=message_content
                 )
 
-                if isinstance(m.content, str) and "Logging object in" in m.content:
-                    absolute_p = m.content.split("Logging object in ")[1]
-                    relative_p = "." + absolute_p.split(self.path.name)[1]
-                    pkl_path = self.path / relative_p
-                    try:
-                        with pkl_path.open("rb") as f:
-                            m.content = pickle.load(f)
-                    except:
-                        continue
-
                 msg_l.append(m)
 
+        for file in self.path.glob("**/*.pkl"):
+            tag = ".".join(str(file.relative_to(self.path)).replace("/", ".").split(".")[:-3])
+            pid = file.parent.name
+
+            with file.open("rb") as f:
+                content = pickle.load(f)
+
+            timestamp = datetime.strptime(file.stem, "%Y-%m-%d_%H-%M-%S-%f").replace(tzinfo=timezone.utc)
+
+            m = Message(tag=tag, level="INFO", timestamp=timestamp, caller="", pid_trace=pid, content=content)
+
+            msg_l.append(m)
+
         msg_l.sort(key=lambda x: x.timestamp)
         for m in msg_l:
             yield m