fix(utils): clean the configured tensorboard_path, not only the runs/ dir

src/spotoptim/SpotOptim.py

@@ -63,7 +63,8 @@ class SpotOptimConfig:
         ocba_delta (int): Delta for OCBA.
         tensorboard_log (bool): Whether to log to TensorBoard.
         tensorboard_path (Optional[str]): Path to TensorBoard logs.
-        tensorboard_clean (bool): Whether to clean TensorBoard logs.
+        tensorboard_clean (bool): Whether to clean old TensorBoard logs (the
+            configured tensorboard_path, or the 'runs' folder if no path is set).
         fun_mo2so (Optional[Callable]): Function to convert multi-objective to single-objective.
         seed (Optional[int]): Seed for random number generator.
         verbose (bool): Whether to print verbose output.
@@ -360,9 +361,12 @@ class SpotOptim(BaseEstimator):
             Path for TensorBoard log files. If None and tensorboard_log
             is True, creates a default path: runs/spotoptim_YYYYMMDD_HHMMSS. Defaults to None.
         tensorboard_clean (bool, optional):
-            If True, removes all old TensorBoard log directories from
-            the 'runs' folder before starting optimization. Use with caution as this permanently
-            deletes all subdirectories in 'runs'. Defaults to False.
+            If True, removes old TensorBoard logs before starting optimization
+            so every run begins with a fresh dashboard. With tensorboard_path
+            set, the configured directory itself is removed (and re-created
+            empty by the writer); without a path, all subdirectories of the
+            default 'runs' folder are removed. Use with caution as this
+            permanently deletes the affected log directories. Defaults to False.
         fun_mo2so (callable, optional):
             Function to convert multi-objective values to single-objective.
             Takes an array of shape (n_samples, n_objectives) and returns array of shape (n_samples,).

src/spotoptim/utils/tensorboard.py

@@ -21,14 +21,31 @@
 
 
 def clean_tensorboard_logs(optimizer: SpotOptimProtocol) -> None:
-    """Clean old TensorBoard log directories from the runs folder.
+    """Clean old TensorBoard logs if tensorboard_clean is True.
 
-    Removes all subdirectories in the 'runs' directory if tensorboard_clean is True.
+    When ``tensorboard_path`` is set, removes that directory (the writer
+    re-creates it empty afterwards), so every run starts with a fresh
+    dashboard. When ``tensorboard_path`` is None, falls back to the legacy
+    behavior of removing all subdirectories of the default ``runs``
+    directory.
 
     Args:
         optimizer: SpotOptim instance.
     """
     if optimizer.tensorboard_clean:
+        if optimizer.tensorboard_path is not None:
+            path = optimizer.tensorboard_path
+            if os.path.isdir(path):
+                try:
+                    shutil.rmtree(path)
+                    if optimizer.verbose:
+                        print(f"Removed old TensorBoard logs: {path}")
+                except Exception as e:
+                    if optimizer.verbose:
+                        print(f"Warning: Could not remove {path}: {e}")
+            elif optimizer.verbose:
+                print(f"No old TensorBoard logs to clean at '{path}'")
+            return
         runs_dir = "runs"
         if os.path.exists(runs_dir) and os.path.isdir(runs_dir):
             subdirs = [

tests/test_tensorboard_clean.py

@@ -230,12 +230,15 @@ def teardown_method(self):
             shutil.rmtree("runs")
 
     def test_clean_with_custom_tensorboard_path(self):
-        """Test that clean doesn't interfere with custom paths."""
-        # Create custom directory
+        """Test that clean targets the custom path and leaves 'runs' alone."""
+        # Create custom directory with a stale event file from a prior run
         custom_path = "my_logs/experiment_1"
         os.makedirs(custom_path, exist_ok=True)
+        stale_file = os.path.join(custom_path, "events.out.tfevents.stale")
+        with open(stale_file, "w") as f:
+            f.write("stale")
 
-        # Create some old logs in runs
+        # Create some old logs in runs (unrelated to the configured path)
         os.makedirs("runs/old_log", exist_ok=True)
 
         try:
@@ -251,22 +254,56 @@ def test_clean_with_custom_tensorboard_path(self):
                 verbose=False,
             )
 
-            # Runs directory should be cleaned
-            if os.path.exists("runs"):
-                subdirs = [
-                    d
-                    for d in os.listdir("runs")
-                    if os.path.isdir(os.path.join("runs", d))
-                ]
-                assert len(subdirs) == 0
+            # Custom directory exists (re-created by the writer) but the
+            # stale event file from the previous run is gone
+            assert os.path.isdir(custom_path)
+            assert not os.path.exists(stale_file)
 
-            # Custom directory should still exist
-            assert os.path.exists(custom_path)
+            # The unrelated 'runs' directory is untouched
+            assert os.path.isdir("runs/old_log")
 
         finally:
             if os.path.exists("my_logs"):
                 shutil.rmtree("my_logs")
 
+    def test_clean_with_custom_path_without_logging(self):
+        """Clean honors a custom path even when logging is disabled."""
+        custom_path = "my_logs/experiment_2"
+        os.makedirs(custom_path, exist_ok=True)
+        with open(os.path.join(custom_path, "events.out.tfevents.stale"), "w") as f:
+            f.write("stale")
+
+        try:
+            _ = SpotOptim(
+                fun=lambda X: np.sum(X**2, axis=1),
+                bounds=[(-5, 5), (-5, 5)],
+                max_iter=10,
+                n_initial=5,
+                tensorboard_path=custom_path,
+                tensorboard_clean=True,
+                verbose=False,
+            )
+
+            # Directory removed and not re-created (no writer without logging)
+            assert not os.path.exists(custom_path)
+
+        finally:
+            if os.path.exists("my_logs"):
+                shutil.rmtree("my_logs")
+
+    def test_clean_with_missing_custom_path(self):
+        """Clean handles a non-existent custom path gracefully."""
+        optimizer = SpotOptim(
+            fun=lambda X: np.sum(X**2, axis=1),
+            bounds=[(-5, 5), (-5, 5)],
+            max_iter=10,
+            n_initial=5,
+            tensorboard_path="my_logs/does_not_exist",
+            tensorboard_clean=True,
+            verbose=False,
+        )
+        assert optimizer.tensorboard_clean is True
+
     def test_clean_with_nested_directories(self):
         """Test that clean handles nested structures correctly."""
         # Create nested structure