fix restart bug (#609)

rdagent/app/data_science/loop.py

@@ -125,9 +125,14 @@ class DataScienceRDLoop(RDLoop):
             )
             if self.trace.sota_experiment() is None and len(self.trace.hist) >= DS_RD_SETTING.consecutive_errors:
                 trace_exp_next_component_list = [
-                    exp.next_component_required() for exp, _ in self.trace.hist[-DS_RD_SETTING.consecutive_errors :]
+                    type(exp.pending_tasks_list[0][0])
+                    for exp, _ in self.trace.hist[-DS_RD_SETTING.consecutive_errors :]
                 ]
-                if None not in trace_exp_next_component_list and len(set(trace_exp_next_component_list)) == 1:
+                last_successful_exp = self.trace.last_successful_exp()
+                if (
+                    last_successful_exp not in [exp for exp, _ in self.trace.hist[-DS_RD_SETTING.consecutive_errors :]]
+                    and len(set(trace_exp_next_component_list)) == 1
+                ):
                     logger.error("Consecutive errors reached the limit. Dumping trace.")
                     logger.log_object(self.trace, tag="trace before restart")
                     self.trace = DSTrace(scen=self.trace.scen, knowledge_base=self.trace.knowledge_base)

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

@@ -11,8 +11,6 @@ from rdagent.oai.llm_utils import md5_hash
 from rdagent.utils.agent.tpl import T
 from rdagent.utils.env import DockerEnv, DSDockerConf
 
-DataLoaderTask = CoSTEERTask
-
 
 # Because we use isinstance to distinguish between different types of tasks, we need to use sub classes to represent different types of tasks
 class DataLoaderTask(CoSTEERTask):