Fix empty dict parquet issues and typing issues

src/dehb/optimizers/dehb.py

@@ -5,7 +5,7 @@
 from copy import deepcopy
 from pathlib import Path
 from threading import Timer
-from typing import Union
+from typing import List, Tuple, Union
 
 import ConfigSpace
 import numpy as np
@@ -144,7 +144,7 @@ def reset(self, *, reset_seeds: bool = True):
     def _init_population(self):
         raise NotImplementedError("Redefine!")
 
-    def _get_next_iteration(self, iteration: int) -> tuple[np.array, np.array]:
+    def _get_next_iteration(self, iteration: int) -> Tuple[np.array, np.array]:
         """Computes the Successive Halving spacing.
 
         Given the iteration index, computes the fidelity spacing to be used and
@@ -171,7 +171,7 @@ def _get_next_iteration(self, iteration: int) -> tuple[np.array, np.array]:
 
         return ns, fidelities
 
-    def get_incumbents(self) -> tuple[Union[dict, ConfigSpace.Configuration], float]:
+    def get_incumbents(self) -> Tuple[Union[dict, ConfigSpace.Configuration], float]:
         """Retrieve current incumbent configuration and score.
         
         Returns:
@@ -306,7 +306,7 @@ def _f_objective(self, job_info):
             run_info.update({"device_id": device_id})
         return run_info
 
-    def _create_cuda_visible_devices(self, available_gpus: list[int], start_id: int) -> str:
+    def _create_cuda_visible_devices(self, available_gpus: List[int], start_id: int) -> str:
         """Generates a string to set the CUDA_VISIBLE_DEVICES environment variable.
 
         Given a list of available GPU device IDs and a preferred ID (start_id), the environment
@@ -675,7 +675,7 @@ def _get_next_job(self):
                 break
         return job_info
 
-    def ask(self, n_configs: int=1) -> Union[dict, list[dict]]:
+    def ask(self, n_configs: int=1) -> Union[dict, List[dict]]:
         """Get the next configuration to run from the optimizer.
 
         The retrieved configuration can then be evaluated by the user.
@@ -846,6 +846,10 @@ def _save_history(self, name="history.parquet.gzip"):
             history_path = self.output_path / name
             history_df = pd.DataFrame(self.history, columns=["config_id", "config", "fitness",
                                                              "cost", "fidelity", "info"])
+            # Check if the 'info' column is empty or contains only None values
+            if history_df["info"].apply(lambda x: (isinstance(x, dict) and len(x) == 0)).all():
+                # Drop the 'info' column
+                history_df = history_df.drop(columns=["info"])
             history_df.to_parquet(history_path, compression="gzip")
         except Exception as e:
             self.logger.warning(f"History not saved: {e!r}")
@@ -925,7 +929,7 @@ def _load_checkpoint(self, run_dir: str):
             result = {
                 "fitness": row["fitness"],
                 "cost": row["cost"],
-                "info": row["info"],
+                "info": row.get("info", {}),
             }
 
             self.tell(job_info, result, replay=True)
@@ -984,7 +988,7 @@ def tell(self, job_info: dict, result: dict, replay: bool=False) -> None:
         self._tell_counter += 1
         # Update bracket information
         fitness, cost = float(result["fitness"]), float(result["cost"])
-        info = result["info"] if "info" in result else dict()
+        info = result["info"] if "info" in result else {}
         fidelity, parent_id = job_info["fidelity"], job_info["parent_id"]
         config, config_id = job_info["config"], job_info["config_id"]
         bracket_id = job_info["bracket_id"]
@@ -1025,7 +1029,7 @@ def tell(self, job_info: dict, result: dict, replay: bool=False) -> None:
 
     @logger.catch
     def run(self, fevals=None, brackets=None, total_cost=None, single_node_with_gpus=False,
-            verbose=False, debug=False, **kwargs) -> tuple[np.array, np.array, np.array]:
+            verbose=False, debug=False, **kwargs) -> Tuple[np.array, np.array, np.array]:
         """Main interface to run optimization by DEHB.
 
         This function waits on workers and if a worker is free, asks for a configuration and a