Merge pull request #422 from sebp/py-3-12

Add support for Python 3.12

.github/workflows/tests-workflow.yaml

@@ -12,10 +12,10 @@ jobs:
       fail-fast: false
       matrix:
         config:
-          - py38_pandas11_numpy121_sklearn13
           - py39_pandas14_numpy122_sklearn13
           - py310_pandas15_numpy124_sklearn13
           - py311_pandas20_numpy124_sklearn13
+          - py312_pandas21_numpy126_sklearn13
         runner:
           - ubuntu-latest
           - macos-latest

.github/workflows/wheels-workflow.yaml

@@ -14,7 +14,7 @@ jobs:
       fail-fast: false
       matrix:
         os: [ubuntu-22.04, windows-2019, macos-11]
-        python: [38, 39, 310, 311]
+        python: [38, 39, 310, 311, 312]
 
     steps:
       - uses: actions/checkout@v4

appveyor.yml

@@ -5,14 +5,14 @@ image: Visual Studio 2019
 environment:
   # see https://www.appveyor.com/docs/build-environment/#python
   matrix:
-    - TARGET_ARCH: "x64"
-      CONDA_PY: "38"
     - TARGET_ARCH: "x64"
       CONDA_PY: "39"
     - TARGET_ARCH: "x64"
       CONDA_PY: "310"
     - TARGET_ARCH: "x64"
       CONDA_PY: "311"
+    - TARGET_ARCH: "x64"
+      CONDA_PY: "312"
 
 
 # We always use a 64-bit machine.

ci/appveyor/py312.ps1

@@ -0,0 +1,4 @@
+$env:CI_PYTHON_VERSION="3.12.*"
+$env:CI_PANDAS_VERSION="2.1.*"
+$env:CI_NUMPY_VERSION="1.26.*"
+$env:CI_SKLEARN_VERSION="1.3.*"

ci/deps/py311_pandas20_numpy124_sklearn13.sh

@@ -3,4 +3,4 @@ export CI_PYTHON_VERSION='3.11.*'
 export CI_PANDAS_VERSION='2.0.*'
 export CI_NUMPY_VERSION='1.24.*'
 export CI_SKLEARN_VERSION='1.3.*'
-export CI_NO_SLOW=false
+export CI_NO_SLOW=true

ci/deps/py312_pandas21_numpy126_sklearn13.sh

@@ -0,0 +1,6 @@
+# shellcheck shell=sh
+export CI_PYTHON_VERSION='3.12.*'
+export CI_PANDAS_VERSION='2.1.*'
+export CI_NUMPY_VERSION='1.26.*'
+export CI_SKLEARN_VERSION='1.3.*'
+export CI_NO_SLOW=false

ci/deps/py39_pandas14_numpy122_sklearn13.sh

@@ -3,4 +3,4 @@ export CI_PYTHON_VERSION='3.9.*'
 export CI_PANDAS_VERSION='1.4.*'
 export CI_NUMPY_VERSION='1.22.*'
 export CI_SKLEARN_VERSION='1.3.*'
-export CI_NO_SLOW=true
+export CI_NO_SLOW=false

ci/deps/requirements.yaml.tmpl

@@ -6,11 +6,8 @@ dependencies:
   - coverage
   - cython>=0.29
   - ecos
-  - ipykernel
-  - ipython!=8.7.0
   - joblib
-  - matplotlib>=3.6.2,<3.7
-  - nbformat
+  - matplotlib>=3.8.0,<3.9
   - numexpr
   - numpy={CI_NUMPY_VERSION}
   - osqp!=0.6.0,!=0.6.1
@@ -29,4 +26,4 @@ dependencies:
     - build
     - importlib-resources>=1.3
     - nbval>=0.10.0
-    - tomli
+    - tomli

pyproject.toml

@@ -16,13 +16,15 @@ requires = [
     "numpy==1.19.3; python_version=='3.9' and platform_machine != 'arm64'",
     "numpy==1.21.6; python_version=='3.10'",
     "numpy==1.23.2; python_version=='3.11'",
-    "numpy; python_version>'3.11'",
+    "numpy==1.26.1; python_version=='3.12'",
+    "numpy; python_version>'3.12'",
 
     # scikit-learn requirements
     "scikit-learn~=1.3.0; python_version<='3.9'",
     "scikit-learn~=1.3.0; python_version=='3.10'",
     "scikit-learn~=1.3.0; python_version=='3.11'",
-    "scikit-learn; python_version>'3.11'",
+    "scikit-learn~=1.3.0; python_version=='3.12'",
+    "scikit-learn; python_version>'3.12'",
 ]
 build-backend = "setuptools.build_meta"
 
@@ -140,9 +142,11 @@ filterwarnings = [
     "ignore:distutils Version classes are deprecated.*:DeprecationWarning",
     "ignore:Jupyter is migrating its paths to use standard platformdirs:DeprecationWarning",
     # added with pandas 2.1
-    "ignore:is_categorical_dtype is deprecated and will be removed in a future version:FutureWarning",
-    # added with pandas 2.1
     "ignore:DataFrame.applymap has been deprecated:FutureWarning",
+    # deprecated since Python 3.12
+    "ignore:datetime\\.datetime\\.utcfromtimestamp\\(\\) is deprecated and scheduled for removal in a future version.*:DeprecationWarning",
+    "ignore:ast\\.Num is deprecated and will be removed in Python 3\\.14.*:DeprecationWarning",
+    "ignore:Attribute n is deprecated and will be removed in Python 3\\.14.*:DeprecationWarning",
 ]
 
 [tool.coverage.run]

sksurv/column.py

@@ -14,7 +14,7 @@
 
 import numpy as np
 import pandas as pd
-from pandas.api.types import is_categorical_dtype
+from pandas.api.types import CategoricalDtype, is_object_dtype
 
 __all__ = ["categorical_to_numeric", "encode_categorical", "standardize"]
 
@@ -125,12 +125,12 @@ def encode_categorical(table, columns=None, **kwargs):
         Numeric columns in the input table remain unchanged.
     """
     if isinstance(table, pd.Series):
-        if not is_categorical_dtype(table.dtype) and not table.dtype.char == "O":
+        if not isinstance(table.dtype, CategoricalDtype) and not is_object_dtype(table.dtype):
             raise TypeError(f"series must be of categorical dtype, but was {table.dtype}")
         return _encode_categorical_series(table, **kwargs)
 
     def _is_categorical_or_object(series):
-        return is_categorical_dtype(series.dtype) or series.dtype.char == "O"
+        return isinstance(series.dtype, CategoricalDtype) or is_object_dtype(series.dtype)
 
     if columns is None:
         # for columns containing categories
@@ -191,9 +191,9 @@ def categorical_to_numeric(table):
     """
 
     def transform(column):
-        if is_categorical_dtype(column.dtype):
+        if isinstance(column.dtype, CategoricalDtype):
             return column.cat.codes
-        if column.dtype.char == "O":
+        if is_object_dtype(column.dtype):
             try:
                 nc = column.astype(np.int64)
             except ValueError:

sksurv/datasets/base.py

@@ -3,7 +3,7 @@
 
 import numpy as np
 import pandas as pd
-from pandas.api.types import is_categorical_dtype
+from pandas.api.types import CategoricalDtype
 
 from ..column import categorical_to_numeric, standardize
 from ..io import loadarff
@@ -102,7 +102,7 @@ def get_x_y(data_frame, attr_labels, pos_label=None, survival=True):
 def _loadarff_with_index(filename):
     dataset = loadarff(filename)
     if "index" in dataset.columns:
-        if is_categorical_dtype(dataset["index"].dtype):
+        if isinstance(dataset["index"].dtype, CategoricalDtype):
             # concatenating categorical index may raise TypeError
             # see https://github.com/pandas-dev/pandas/issues/14586
             dataset["index"] = dataset["index"].astype(object)

sksurv/io/arffwrite.py

@@ -15,7 +15,7 @@
 
 import numpy as np
 import pandas as pd
-from pandas.api.types import is_categorical_dtype, is_object_dtype
+from pandas.api.types import CategoricalDtype, is_object_dtype
 
 _ILLEGAL_CHARACTER_PAT = re.compile(r"[^-_=\w\d\(\)<>\.]")
 
@@ -70,7 +70,7 @@ def _write_header(data, fp, relation_name, index):
         name = attribute_names[column]
         fp.write(f"@attribute {name}\t")
 
-        if is_categorical_dtype(series) or is_object_dtype(series):
+        if isinstance(series.dtype, CategoricalDtype) or is_object_dtype(series):
             _write_attribute_categorical(series, fp)
         elif np.issubdtype(series.dtype, np.floating):
             fp.write("real")
@@ -114,7 +114,7 @@ def _check_str_value(x):
 
 def _write_attribute_categorical(series, fp):
     """Write categories of a categorical/nominal attribute"""
-    if is_categorical_dtype(series.dtype):
+    if isinstance(series.dtype, CategoricalDtype):
         categories = series.cat.categories
         string_values = _check_str_array(categories)
     else:

sksurv/kernels/clinical.py

@@ -12,7 +12,7 @@
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 import numpy as np
 import pandas as pd
-from pandas.api.types import is_categorical_dtype, is_numeric_dtype
+from pandas.api.types import CategoricalDtype, is_numeric_dtype
 from sklearn.base import BaseEstimator, TransformerMixin
 from sklearn.utils.validation import check_is_fitted
 
@@ -178,7 +178,7 @@ def _prepare_by_column_dtype(self, X):
 
         for i, dt in enumerate(X.dtypes):
             col = X.iloc[:, i]
-            if is_categorical_dtype(dt):
+            if isinstance(dt, CategoricalDtype):
                 if col.cat.ordered:
                     numeric_ranges.append(col.cat.codes.max() - col.cat.codes.min())
                     numeric_columns.append(i)

sksurv/util.py

@@ -12,7 +12,7 @@
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 import numpy as np
 import pandas as pd
-from pandas.api.types import is_categorical_dtype
+from pandas.api.types import CategoricalDtype
 from sklearn.utils import check_array, check_consistent_length
 
 __all__ = ["check_array_survival", "check_y_survival", "safe_concat", "Surv"]
@@ -251,7 +251,7 @@ def safe_concat(objs, *args, **kwargs):
     categories = {}
     for df in objs:
         if isinstance(df, pd.Series):
-            if is_categorical_dtype(df.dtype):
+            if isinstance(df.dtype, CategoricalDtype):
                 categories[df.name] = {"categories": df.cat.categories, "ordered": df.cat.ordered}
         else:
             dfc = df.select_dtypes(include=["category"])