Fix ruff rule F401: remove unused imports (#31)

* Remove unused imports

* Move fixture to conftest.py

* Enable rule F401

* Remove one more import

docs/conf.py

@@ -22,7 +22,6 @@
 import os
 import sys
 import shutil
-import sphinx_rtd_theme
 print(os.path.abspath('./shap'))
 sys.path.insert(0, os.path.abspath('..'))
 

pyproject.toml

@@ -11,7 +11,10 @@ testpaths = ["tests"]
 select = ["F"]
 # For now, deselect rules that do not pass
 ignore = [
-  "F401", # unused imports
   "F811", # Redefinition of unused variable
   "F841", # Local variables assigned but unused
 ]
+
+[tool.ruff.per-file-ignores]
+# Unused imports often intentional in top-level __init__.py 
+"__init__.py" = ["F401"]

shap/_explanation.py

@@ -2,8 +2,6 @@
 import pandas as pd
 import numpy as np
 import scipy as sp
-import sys
-import warnings
 import copy
 import operator
 import sklearn

shap/actions/_optimizer.py

@@ -1,5 +1,4 @@
 import queue
-import numpy as np
 import warnings
 import copy
 from ._action import Action

shap/benchmark/_sequential.py

@@ -1,5 +1,5 @@
 from shap.utils import safe_isinstance, MaskedModel
-from shap.maskers import Independent, Partition, Impute, Text, Image, FixedComposite
+from shap.maskers import Text, Image, FixedComposite
 from shap import Explanation, links
 import matplotlib.pyplot as pl
 import sklearn

shap/benchmark/experiments.py

@@ -2,10 +2,7 @@
 from .. import datasets
 from . import metrics
 from . import models
-from . import methods
 from .. import __version__
-import numpy as np
-import sklearn
 import os
 import pickle
 import sys

shap/benchmark/framework.py

@@ -1,11 +1,7 @@
 import matplotlib.pyplot as plt 
-import pandas as pd 
 import numpy as np 
 import itertools as it 
-import sklearn 
-import shap 
-from sklearn.model_selection import train_test_split
-from shap.utils import safe_isinstance, MaskedModel
+from shap.utils import safe_isinstance
 from . import perturbation
 
 def update(model, attributions, X, y, masker, sort_order, perturbation_method, scores):

shap/benchmark/measures.py

@@ -1,6 +1,5 @@
 import numpy as np
 from tqdm.autonotebook import tqdm
-import gc
 import warnings
 import sklearn.utils
 

shap/benchmark/metrics.py

@@ -1,14 +1,8 @@
-from .. import LinearExplainer
-from .. import KernelExplainer
-from .. import SamplingExplainer
-from ..explainers import other
 from .. import __version__
 from . import measures
 from . import methods
 import sklearn
 import numpy as np
-import copy
-import functools
 import time
 import hashlib
 import os

shap/benchmark/models.py

@@ -1,6 +1,5 @@
 import sklearn
 import sklearn.ensemble
-import gc
 from sklearn.preprocessing import StandardScaler
 import numpy as np
 

shap/explainers/_additive.py

@@ -1,9 +1,6 @@
 import numpy as np
-import scipy as sp
-import warnings
 from ._explainer import Explainer
 from ..utils import safe_isinstance, MaskedModel
-from .. import maskers
 
 
 class Additive(Explainer):

shap/explainers/_kernel.py

@@ -2,7 +2,6 @@
 from ..utils._legacy import convert_to_instance_with_index, convert_to_link, IdentityLink, convert_to_data, DenseData, SparseData
 from ..utils import safe_isinstance
 from scipy.special import binom
-from scipy.sparse import issparse
 import numpy as np
 import pandas as pd
 import scipy as sp

shap/explainers/_partition.py

@@ -1,20 +1,12 @@
-import types
-import copy
-import inspect
 from ..utils import MaskedModel
 import numpy as np
-import warnings
 import time
 from tqdm.auto import tqdm
 import queue
-from ..utils import assert_import, record_import_error, safe_isinstance, make_masks, OpChain
+from ..utils import safe_isinstance, make_masks, OpChain
 from .. import Explanation
-from .. import maskers
 from ._explainer import Explainer
 from .. import links
-import cloudpickle
-import pickle
-from ..maskers import Masker
 from ..models import Model
 from numba import njit
 

shap/explainers/_permutation.py

@@ -1,17 +1,8 @@
-import functools
-import types
 from ..utils import partition_tree_shuffle, MaskedModel
-from .._explanation import Explanation
 from ._explainer import Explainer
 import numpy as np
-import pandas as pd
-import scipy as sp
-import pickle
-import cloudpickle
 import warnings
 from .. import links
-from .. import maskers
-from ..maskers import Masker
 from ..models import Model
 
 class Permutation(Explainer):

shap/explainers/_sampling.py

@@ -1,5 +1,4 @@
-from ..utils._legacy import convert_to_instance, convert_to_model, match_instance_to_data, match_model_to_data
-from ..utils._legacy import convert_to_instance_with_index, convert_to_link, IdentityLink, convert_to_data, DenseData
+from ..utils._legacy import convert_to_instance, match_instance_to_data
 from ..utils import safe_isinstance
 from .._explanation import Explanation
 from ._kernel import Kernel

shap/explainers/_tree.py

@@ -1,12 +1,8 @@
 import time
 import numpy as np
 import scipy.special
-import multiprocessing
-import sys
 import json
-import os
 import struct
-import itertools
 from packaging import version
 from ._explainer import Explainer
 from ..utils import assert_import, record_import_error, safe_isinstance
@@ -27,7 +23,7 @@
     record_import_error("cext", "C extension was not built during install!", e)
 
 try:
-    import pyspark
+    import pyspark  # noqa
 except ImportError as e:
     record_import_error("pyspark", "PySpark could not be imported!", e)
 
@@ -829,7 +825,6 @@ def __init__(self, model, data=None, data_missing=None, model_output=None):
             else:
                 assert False, "Unsupported Spark model type: " + str(type(model))
         elif safe_isinstance(model, "xgboost.core.Booster"):
-            import xgboost
             self.original_model = model
             self.model_type = "xgboost"
             xgb_loader = XGBTreeModelLoader(self.original_model)
@@ -840,7 +835,6 @@ def __init__(self, model, data=None, data_missing=None, model_output=None):
             if xgb_loader.num_class > 0:
                 self.num_stacked_models = xgb_loader.num_class
         elif safe_isinstance(model, "xgboost.sklearn.XGBClassifier"):
-            import xgboost
             self.input_dtype = np.float32
             self.model_type = "xgboost"
             self.original_model = model.get_booster()
@@ -864,7 +858,6 @@ def __init__(self, model, data=None, data_missing=None, model_output=None):
                 else:
                     self.model_output = "probability"
         elif safe_isinstance(model, "xgboost.sklearn.XGBRegressor"):
-            import xgboost
             self.original_model = model.get_booster()
             self.model_type = "xgboost"
             xgb_loader = XGBTreeModelLoader(self.original_model)
@@ -876,7 +869,6 @@ def __init__(self, model, data=None, data_missing=None, model_output=None):
             if xgb_loader.num_class > 0:
                 self.num_stacked_models = xgb_loader.num_class
         elif safe_isinstance(model, "xgboost.sklearn.XGBRanker"):
-            import xgboost
             self.original_model = model.get_booster()
             self.model_type = "xgboost"
             xgb_loader = XGBTreeModelLoader(self.original_model)

shap/explainers/pytree.py

@@ -5,7 +5,6 @@
 """
 import numpy as np
 #import numba
-from .explainer import Explainer
 from ..utils._exceptions import ExplainerError
 
 # class TreeExplainer(Explainer):

shap/maskers/_image.py

@@ -6,7 +6,7 @@
 import heapq
 from numba import njit
 try:
-    import torch
+    import torch  # noqa: F401
 except ImportError as e:
     record_import_error("torch", "torch could not be imported!", e)
 

shap/models/_model.py

@@ -2,7 +2,7 @@
 from ..utils import record_import_error, safe_isinstance
 from .._serializable import Serializable, Serializer, Deserializer
 try:
-    import torch
+    import torch  # noqa: F401
 except ImportError as e:
     record_import_error("torch", "torch could not be imported!", e)
 

shap/plots/_bar.py

@@ -6,11 +6,10 @@
     pass
 from ._labels import labels
 from ..utils import format_value, ordinal_str
-from ._utils import convert_ordering, convert_color, merge_nodes, get_sort_order, sort_inds, dendrogram_coords
+from ._utils import convert_ordering, merge_nodes, get_sort_order, sort_inds, dendrogram_coords
 from . import colors
 import numpy as np
 import scipy
-import copy
 from .. import Explanation, Cohorts
 
 

shap/plots/_beeswarm.py

@@ -14,7 +14,7 @@
     pass
 from ._labels import labels
 from . import colors
-from ..utils import safe_isinstance, OpChain, format_value
+from ..utils import safe_isinstance
 from ._utils import convert_ordering, convert_color, merge_nodes, get_sort_order, sort_inds
 from .. import Explanation
 

shap/plots/_decision.py

@@ -10,7 +10,6 @@
 try:
     import matplotlib.cm as cm
     import matplotlib.pyplot as pl
-    import matplotlib
 except ImportError:
     warnings.warn("matplotlib could not be loaded!")
     pass

shap/plots/_embedding.py

@@ -1,9 +1,7 @@
-import numpy as np
 import sklearn
 import warnings
 try:
     import matplotlib.pyplot as pl
-    import matplotlib
 except ImportError:
     warnings.warn("matplotlib could not be loaded!")
     pass

shap/plots/_group_difference.py

@@ -2,7 +2,6 @@
 import warnings
 try:
     import matplotlib.pyplot as pl
-    import matplotlib
 except ImportError:
     warnings.warn("matplotlib could not be loaded!")
     pass

shap/plots/_heatmap.py

@@ -1,13 +1,12 @@
 import numpy as np
 try:
     import matplotlib.pyplot as pl
-    import matplotlib
 except ImportError:
     pass
 from . import colors
 from .. import Explanation
 from ..utils import OpChain
-from ._utils import convert_ordering, convert_color
+from ._utils import convert_ordering
 
 def heatmap(shap_values, instance_order=Explanation.hclust(), feature_values=Explanation.abs.mean(0), 
             feature_order=None, max_display=10, cmap=colors.red_white_blue, show=True,

shap/plots/_monitoring.py

@@ -3,7 +3,6 @@
 import warnings
 try:
     import matplotlib.pyplot as pl
-    import matplotlib
 except ImportError:
     warnings.warn("matplotlib could not be loaded!")
     pass

shap/plots/_partial_dependence.py

@@ -7,7 +7,6 @@
 except ImportError:
     warnings.warn("matplotlib could not be loaded!")
     pass
-from mpl_toolkits.mplot3d import Axes3D
 import numpy as np
 import pandas as pd
 

shap/plots/_text.py

@@ -1,7 +1,5 @@
 import numpy as np
-import warnings
 from . import colors
-from ..utils import ordinal_str
 import random
 import string
 import json

shap/plots/colors/_colors.py

@@ -7,7 +7,7 @@
 from ._colorconv import lab2rgb, lch2lab
 
 try:
-    import matplotlib
+    import matplotlib  # noqa: F401
     from matplotlib.colors import LinearSegmentedColormap
 
     def lch2rgb(x):

shap/utils/_clustering.py

@@ -1,6 +1,5 @@
 import numpy as np
 import scipy as sp
-from scipy.spatial.distance import pdist
 from numba import njit
 import sklearn
 import warnings

shap/utils/_general.py

@@ -1,12 +1,9 @@
 import re
-import pandas as pd
 import numpy as np
 import scipy as sp
-from scipy.spatial.distance import pdist
 import sys
 import warnings
 import sklearn
-import importlib
 import copy
 from contextlib import contextmanager
 import sys, os

shap/utils/image.py

@@ -1,9 +1,7 @@
 import cv2
-import json
 import matplotlib.pyplot as plt 
 import numpy as np
 import os
-import requests
 import shap
 from textwrap import wrap
 

tests/benchmark/perturbation.py

@@ -1,5 +1,4 @@
 import numpy as np
-import shap 
 import shap.benchmark as benchmark 
 from transformers import AutoTokenizer
 from shap.maskers import Independent, Partition, Impute, Text, Image, FixedComposite