Fix score log overwriting on re-initialization

- Modify project_score_save_class to check if the log file exists before writing headers.
- Prevent data loss when the logger is instantiated multiple times within the same experiment folder.
- Add regression test to verify that existing log entries are preserved during re-initialization.

Author: SamoraHunter

ml_grid/util/project_score_save.py

@@ -127,7 +127,8 @@ def __init__(self, experiment_dir: str):
         self.log_path.parent.mkdir(parents=True, exist_ok=True)
         self.models_dir.mkdir(parents=True, exist_ok=True)
 
-        df.to_csv(self.log_path, mode="w", header=True, index=False)
+        if not self.log_path.exists():
+            df.to_csv(self.log_path, mode="w", header=True, index=False)
 
     def update_score_log(
         self,

tests/test_project_score_save.py

@@ -130,5 +130,23 @@ def test_update_score_log_typo_and_missing_safety(self):
         df = pd.read_csv(log_path)
         self.assertEqual(len(df), 1)
 
+    def test_initialization_does_not_overwrite(self):
+        """Test that re-initializing the class does not wipe an existing log file."""
+        # First initialization
+        saver1 = project_score_save_class(str(self.experiment_dir))
+        log_path = self.experiment_dir / "final_grid_score_log.csv"
+        
+        # Simulate writing some data
+        with open(log_path, "a") as f:
+            f.write("test_data_entry\n")
+            
+        # Second initialization on same directory
+        saver2 = project_score_save_class(str(self.experiment_dir))
+        
+        # Verify data persists
+        with open(log_path, "r") as f:
+            content = f.read()
+        self.assertIn("test_data_entry", content)
+
 if __name__ == "__main__":
     unittest.main()