Lint tabular (#3347)

.github/workflow_scripts/lint_check.sh

@@ -17,6 +17,7 @@ function lint_check_all {
     lint_check common 160
     lint_check core 160
     lint_check features 160
+    lint_check tabular 160
 }
 
 bandit -r multimodal/src -ll

tabular/setup.py

@@ -1,11 +1,13 @@
 #!/usr/bin/env python
 ###########################
 # This code block is a HACK (!), but is necessary to avoid code duplication. Do NOT alter these lines.
+import importlib.util
 import os
+
 from setuptools import setup
-import importlib.util
+
 filepath = os.path.abspath(os.path.dirname(__file__))
-filepath_import = os.path.join(filepath, '..', 'core', 'src', 'autogluon', 'core', '_setup_utils.py')
+filepath_import = os.path.join(filepath, "..", "core", "src", "autogluon", "core", "_setup_utils.py")
 spec = importlib.util.spec_from_file_location("ag_min_dependencies", filepath_import)
 ag = importlib.util.module_from_spec(spec)
 # Identical to `from autogluon.core import _setup_utils as ag`, but works without `autogluon.core` being installed.
@@ -17,82 +19,81 @@
 version = ag.load_version_file()
 version = ag.update_version(version)
 
-submodule = 'tabular'
+submodule = "tabular"
 install_requires = [
     # version ranges added in ag.get_dependency_version_ranges()
-    'numpy',  # version range defined in `core/_setup_utils.py`
-    'scipy',  # version range defined in `core/_setup_utils.py`
-    'pandas',  # version range defined in `core/_setup_utils.py`
-    'scikit-learn',  # version range defined in `core/_setup_utils.py`
-    'networkx',  # version range defined in `core/_setup_utils.py`
-    f'{ag.PACKAGE_NAME}.core=={version}',
-    f'{ag.PACKAGE_NAME}.features=={version}',
+    "numpy",  # version range defined in `core/_setup_utils.py`
+    "scipy",  # version range defined in `core/_setup_utils.py`
+    "pandas",  # version range defined in `core/_setup_utils.py`
+    "scikit-learn",  # version range defined in `core/_setup_utils.py`
+    "networkx",  # version range defined in `core/_setup_utils.py`
+    f"{ag.PACKAGE_NAME}.core=={version}",
+    f"{ag.PACKAGE_NAME}.features=={version}",
 ]
 
 extras_require = {
-    'lightgbm': [
-        'lightgbm>=3.3,<3.4',
+    "lightgbm": [
+        "lightgbm>=3.3,<3.4",
     ],
-    'catboost': [
+    "catboost": [
         # CatBoost wheel build is not working correctly on darwin for CatBoost 1.2, so use old version in this case.
         # https://github.com/autogluon/autogluon/pull/3190#issuecomment-1540599280
         'catboost>=1.1,<1.2 ; sys_platform == "darwin"',
-        'catboost>=1.1,<1.3',
+        "catboost>=1.1,<1.3",
     ],
     # FIXME: Debug why xgboost 1.6 has 4x+ slower inference on multiclass datasets compared to 1.4
     #  It is possibly only present on MacOS, haven't tested linux.
     # XGBoost made API breaking changes in 1.6 with custom metric and callback support, so we don't support older versions.
-    'xgboost': [
-        'xgboost>=1.6,<1.8',
+    "xgboost": [
+        "xgboost>=1.6,<1.8",
     ],
-    'fastai': [
-        'torch>=1.9,<1.14',
-        'fastai>=2.3.1,<2.8',
+    "fastai": [
+        "torch>=1.9,<1.14",
+        "fastai>=2.3.1,<2.8",
     ],
-    'tabpfn': [
-        'tabpfn>=0.1,<0.2',
+    "tabpfn": [
+        "tabpfn>=0.1,<0.2",
     ],
-    'ray': [
-        f'{ag.PACKAGE_NAME}.core[all]=={version}',
+    "ray": [
+        f"{ag.PACKAGE_NAME}.core[all]=={version}",
     ],
-    'skex': [
+    "skex": [
         # Note: 2021.7 released on Sep 2022, version 2022.x doesn't exist (went directly from 2021.7 to 2023.0)
-        'scikit-learn-intelex>=2021.7,<2023.2',
+        "scikit-learn-intelex>=2021.7,<2023.2",
     ],
-    'imodels': [
-        'imodels>=1.3.10,<1.4.0',  # 1.3.8/1.3.9 either remove/renamed attribute `complexity_` causing failures. https://github.com/csinva/imodels/issues/147
+    "imodels": [
+        "imodels>=1.3.10,<1.4.0",  # 1.3.8/1.3.9 either remove/renamed attribute `complexity_` causing failures. https://github.com/csinva/imodels/issues/147
     ],
-    'vowpalwabbit': [
+    "vowpalwabbit": [
         # FIXME: 9.5+ causes VW to save an empty model which always predicts 0. Confirmed on MacOS (Intel CPU). Unknown how to fix.
-        'vowpalwabbit>=9,<9.5',
+        "vowpalwabbit>=9,<9.5",
     ],
-    'skl2onnx': [
-        'skl2onnx>=1.13.0,<1.14.0',
+    "skl2onnx": [
+        "skl2onnx>=1.13.0,<1.14.0",
         # For macOS, there isn't a onnxruntime-gpu package installed with skl2onnx.
         # Therefore, we install onnxruntime explicitly here just for macOS.
-        'onnxruntime>=1.13.0,<1.14.0'
-    ] if sys.platform == 'darwin' else [
-        'skl2onnx>=1.13.0,<1.14.0',
-        'onnxruntime-gpu>=1.13.0,<1.14.0'
+        "onnxruntime>=1.13.0,<1.14.0",
     ]
+    if sys.platform == "darwin"
+    else ["skl2onnx>=1.13.0,<1.14.0", "onnxruntime-gpu>=1.13.0,<1.14.0"],
 }
 
 # TODO: v1.0: Rename `all` to `core`, make `all` contain everything.
 all_requires = []
 # TODO: Consider adding 'skex' to 'all'
-for extra_package in ['lightgbm', 'catboost', 'xgboost', 'fastai', 'ray']:
+for extra_package in ["lightgbm", "catboost", "xgboost", "fastai", "ray"]:
     all_requires += extras_require[extra_package]
 all_requires = list(set(all_requires))
-extras_require['all'] = all_requires
+extras_require["all"] = all_requires
 
 
 test_requires = []
-for test_package in ['tabpfn', 'imodels', 'vowpalwabbit', 'skl2onnx']:
+for test_package in ["tabpfn", "imodels", "vowpalwabbit", "skl2onnx"]:
     test_requires += extras_require[test_package]
-extras_require['tests'] = test_requires
+extras_require["tests"] = test_requires
 install_requires = ag.get_dependency_version_ranges(install_requires)
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     ag.create_version_file(version=version, submodule=submodule)
     setup_args = ag.default_setup_args(version=version, submodule=submodule)
     setup(

tabular/src/autogluon/tabular/__init__.py

@@ -1,5 +1,5 @@
 from autogluon.common.features.feature_metadata import FeatureMetadata
-from autogluon.common.utils.log_utils import _add_stream_handler 
+from autogluon.common.utils.log_utils import _add_stream_handler
 from autogluon.core.dataset import TabularDataset
 
 try: