refactor: improve project structure, update docs workflow, and add notebook

- CI: Add `torch` dependency to `docs.yml` workflow to support autodoc imports.
- Refactor: Reorganize imports in time-series model modules for better readability and consistency.
- Fix: Use `Exception` instead of bare `except` in `data_feature_methods.py` for safer error handling.
- Cleanup: Remove unused `NoFeaturesError` import in `test_data_pipeline.py`.
- Refactor: Standardize scikit-learn metric imports in `global_params.py` and `project_score_save.py`.

Author: SamoraHunter

.github/workflows/docs.yml

@@ -70,6 +70,8 @@ jobs:
           # Ensure setuptools and wheel are present before installing the project
           pip install --upgrade pip setuptools wheel
           # psutil/ipython required for imports during autodoc (H2OBaseClassifier, pipeline.data)
+          # torch is needed for some model classes that are imported during autodoc
+          pip install torch
           pip install psutil ipython
           pip install .[docs]
 
@@ -133,6 +135,8 @@ jobs:
         run: |
           pip install --upgrade pip setuptools wheel
           # psutil/ipython required for imports during autodoc (H2OBaseClassifier, pipeline.data)
+          # torch is needed for some model classes that are imported during autodoc
+          pip install torch
           pip install psutil ipython
           pip install .[docs]
 

ml_grid/model_classes_time_series/TapNetClassifier_module.py

@@ -2,7 +2,14 @@
 
 import numpy as np
 import keras
-from sklearn.base import BaseEstimator, ClassifierMixin
+from sklearn.base import BaseEstimator, ClassifierMixin, TransformerMixin
+from sklearn.impute import SimpleImputer
+from sklearn.pipeline import Pipeline
+from sklearn.preprocessing import StandardScaler
+from skopt.space import Categorical
+
+from ml_grid.pipeline.data import pipe
+from ml_grid.util.param_space import ParamSpace
 
 
 class _DummyClassifier(BaseEstimator, ClassifierMixin):
@@ -44,16 +51,6 @@ def _predict_proba(self, X, **kwargs):
     except ImportError:
         TapNetClassifier = _DummyClassifier
 
-from skopt.space import Categorical
-from sklearn.pipeline import Pipeline
-from sklearn.impute import SimpleImputer
-from sklearn.preprocessing import StandardScaler
-from sklearn.base import BaseEstimator, TransformerMixin
-
-
-from ml_grid.pipeline.data import pipe
-from ml_grid.util.param_space import ParamSpace
-
 
 class TimeSeriesStandardScaler(BaseEstimator, TransformerMixin):
     def __init__(self, epsilon=1e-6):

ml_grid/model_classes_time_series/shapeDTWClassifier_module.py

@@ -1,6 +1,9 @@
 from typing import Any, Dict, List
 import numpy as np
 from sklearn.base import BaseEstimator, ClassifierMixin
+from skopt.space import Categorical
+
+from ml_grid.pipeline.data import pipe
 
 
 class _DummyClassifier(BaseEstimator, ClassifierMixin):
@@ -25,8 +28,6 @@ def predict_proba(self, X):
 except ImportError:
     # ShapeDTW was removed in aeon v0.11.0. Use a dummy placeholder.
     ShapeDTW = _DummyClassifier
-from skopt.space import Categorical
-from ml_grid.pipeline.data import pipe
 
 
 class ShapeDTW_class:

ml_grid/pipeline/data_feature_methods.py

@@ -215,7 +215,7 @@ def getNFeaturesMarkovBlanket(
                         os.close(devnull_fd)
                     os.close(old_stdout_fd)
                     os.close(old_stderr_fd)
-                except:
+                except Exception:
                     pass  # Silently fail if restoration doesn't work
 
         # Get the feature indices from the Markov blanket (MB)

ml_grid/pipeline/test_data_pipeline.py

@@ -16,7 +16,7 @@
 
 # Ensure the project root is in the Python path to allow for module imports
 try:
-    from ml_grid.pipeline.data import NoFeaturesError, pipe
+    from ml_grid.pipeline.data import pipe
     from ml_grid.util.global_params import global_parameters
 except ImportError:
     # This allows the test to be run from the project root directory

ml_grid/util/global_params.py

@@ -10,6 +10,15 @@
 from typing import Any, Callable, Dict, List, Union
 import logging
 
+import numpy as np
+from sklearn.metrics import (
+    accuracy_score,
+    f1_score,
+    make_scorer,
+    recall_score,
+    roc_auc_score,
+)
+
 # Disable aeon telemetry to prevent SSL errors in restricted environments
 os.environ["AEON_TRACKING"] = "False"
 
@@ -61,15 +70,6 @@ def _configure_tensorflow_gpu_env():
 _configure_tensorflow_gpu_env()
 # --- END FIX ---
 
-import numpy as np
-from sklearn.metrics import (
-    accuracy_score,
-    f1_score,
-    make_scorer,
-    recall_score,
-    roc_auc_score,
-)
-
 
 def custom_roc_auc_score(y_true: np.ndarray, y_pred: np.ndarray) -> float:
     """Calculates ROC AUC score, handling cases with only one class in y_true.

ml_grid/util/project_score_save.py

@@ -4,7 +4,14 @@
 import pandas as pd
 from ml_grid.util.global_params import global_parameters
 from sklearn import metrics
-from sklearn.metrics import *
+from sklearn.metrics import (
+    accuracy_score,
+    f1_score,
+    matthews_corrcoef,
+    precision_recall_fscore_support,
+    precision_score,
+    recall_score,
+)
 import pickle
 import logging
 import warnings

notebooks/unit_test_synthetic.ipynb

@@ -332,7 +332,7 @@
     "            try:\n",
     "                # Try to get a meaningful name for the model\n",
     "                model_name = error_info[0].__class__.__name__\n",
-    "            except:\n",
+    "            except Exception:\n",
     "                model_name = \"Unknown Model\"\n",
     "            error_exception = error_info[1]\n",
     "            print(f\"    {i+1}. Model: {model_name}, Error: {error_exception}\")\n",

tests/test_tabpfn_classifier_class.py

@@ -3,6 +3,7 @@
 import os
 import pandas as pd
 from unittest.mock import MagicMock, patch
+import importlib.util
 from ml_grid.model_classes.tabpfn_classifier_class import (
     TabPFNClassifierClass,
 )
@@ -27,9 +28,7 @@
 
 # Conditionally mock tabpfn. If it's installed, we might want to use it for integration tests.
 # If not installed, we mock it to allow importing the wrapper class.
-try:
-    import tabpfn
-except ImportError:
+if importlib.util.find_spec("tabpfn") is None:
     mock_tabpfn_module = MagicMock()
     sys.modules["tabpfn"] = mock_tabpfn_module
     sys.modules["tabpfn.constants"] = MagicMock()