scikit-survival 0.23.0__tar.gz → 0.23.1__tar.gz

{scikit_survival-0.23.0/scikit_survival.egg-info → scikit_survival-0.23.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: scikit-survival
-Version: 0.23.0
+Version: 0.23.1
 Summary: Survival analysis built on top of scikit-learn
 Author-email: Sebastian Pölsterl <sebp@k-d-w.org>
 License:                     GNU GENERAL PUBLIC LICENSE
@@ -708,7 +708,7 @@ Requires-Dist: joblib
 Requires-Dist: numexpr
 Requires-Dist: numpy
 Requires-Dist: osqp!=0.6.0,!=0.6.1
-Requires-Dist: pandas>=1.0.5
+Requires-Dist: pandas>=1.4.0
 Requires-Dist: scipy>=1.3.2
 Requires-Dist: scikit-learn<1.6,>=1.4.0
 Provides-Extra: dev
@@ -736,7 +736,7 @@ Requires-Dist: sphinx-copybutton~=0.5.2; extra == "docs"
 Provides-Extra: nbval
 Requires-Dist: ipykernel; extra == "nbval"
 Requires-Dist: ipython!=8.7.0; extra == "nbval"
-Requires-Dist: matplotlib~=3.6.2; extra == "nbval"
+Requires-Dist: matplotlib~=3.8.0; extra == "nbval"
 Requires-Dist: nbformat; extra == "nbval"
 Requires-Dist: nbval>=0.10.0; extra == "nbval"
 Requires-Dist: seaborn~=0.11.2; extra == "nbval"
@@ -781,7 +781,7 @@ Requirements
 - numexpr
 - numpy
 - osqp
-- pandas 1.0.5 or later
+- pandas 1.4.0 or later
 - scikit-learn 1.4 or 1.5
 - scipy
 - C/C++ compiler

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/README.rst

@@ -38,7 +38,7 @@ Requirements
 - numexpr
 - numpy
 - osqp
-- pandas 1.0.5 or later
+- pandas 1.4.0 or later
 - scikit-learn 1.4 or 1.5
 - scipy
 - C/C++ compiler

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/doc/install.rst

@@ -1,21 +1,30 @@
 Installing scikit-survival
 ==========================
 
-This is the recommended and easiest to install scikit-survival is to use
-:ref:`install-conda`.
+The recommended and easiest way to install scikit-survival is to use
+:ref:`install-conda` or :ref:`install-pip`.
+Pre-built binary packages for scikit-survival are available for Linux, macOS, and Windows.
 Alternatively, you can install scikit-survival :ref:`install-from-source`.
 
 .. _install-conda:
 
-Anaconda
---------
+conda
+-----
 
-Pre-built binary packages for Linux, MacOS, and Windows are available
-for `Anaconda <https://www.anaconda.com/distribution/>`_.
-If you have Anaconda installed, run::
+If you have `conda <https://docs.anaconda.com/>`_ installed, you can
+install scikit-survival from the ``conda-forge`` channel by running::
 
   conda install -c conda-forge scikit-survival
 
+.. _install-pip:
+
+pip
+---
+
+If you use ``pip``, install the latest release of scikit-survival with::
+
+  pip install scikit-survival
+
 
 .. _install-from-source:
 
@@ -31,9 +40,9 @@ On Linux, you need to install *gcc*, which in most cases is available
 via your distribution's packaging system.
 Please follow your distribution's instructions on how to install packages.
 
-**MacOS**
+**macOS**
 
-On MacOS, you need to install *clang*, which is available from
+On macOS, you need to install *clang*, which is available from
 the *Command Line Tools* package. Open a terminal and execute::
 
   xcode-select --install
@@ -55,7 +64,7 @@ Latest Release
 
 To install the latest release of scikit-survival from source, run::
 
-  pip install scikit-survival
+  pip install scikit-survival --no-binary scikit-survival
 
 
 .. note::
@@ -90,7 +99,7 @@ The current minimum dependencies to run scikit-survival are:
 - numexpr
 - numpy
 - osqp
-- pandas 1.0.5 or later
+- pandas 1.4.0 or later
 - scikit-learn 1.4 or 1.5
 - scipy
 - C/C++ compiler

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/doc/release_notes/v0.23.rst

@@ -3,6 +3,28 @@
 What's new in 0.23
 ==================
 
+scikit-survival 0.23.1 (2024-11-04)
+-----------------------------------
+
+This release adds support for Python 3.13.
+The minimum required version for pandas has been bumped to pandas 1.4.0.
+
+Bug fixes
+^^^^^^^^^
+- Add `SurvivalAnalysisMixin` base class to :class:`sksurv.ensemble.ExtraSurvivalTrees`
+  to enable the :meth:`sksurv.ensemble.ExtraSurvivalTrees.score` method that has been
+  unintentially removed in 0.23.0 (:issue:`488`).
+
+Enhancements
+^^^^^^^^^^^^
+- Improve performance of :func:`sksurv.metrics.concordance_index_censored` and
+  :func:`sksurv.metrics.concordance_index_ipcw` (:issue:`465`).
+
+Backwards incompatible changes
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+- Support for pandas versions before 1.4.0 has been dropped.
+
+
 scikit-survival 0.23.0 (2024-06-30)
 -----------------------------------
 

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/doc/user_guide/00-introduction.ipynb

@@ -296,6 +296,7 @@
    "source": [
     "%matplotlib inline\n",
     "import matplotlib.pyplot as plt\n",
+    "\n",
     "from sksurv.nonparametric import kaplan_meier_estimator\n",
     "\n",
     "time, survival_prob, conf_int = kaplan_meier_estimator(\n",
@@ -637,6 +638,7 @@
    ],
    "source": [
     "from sklearn import set_config\n",
+    "\n",
     "from sksurv.linear_model import CoxPHSurvivalAnalysis\n",
     "\n",
     "set_config(display=\"text\")  # displays text representation of estimators\n",

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/doc/user_guide/boosting.ipynb

@@ -45,16 +45,15 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
     "import pandas as pd\n",
+    "from sklearn.model_selection import train_test_split\n",
     "\n",
     "%matplotlib inline\n",
     "\n",
-    "from sklearn.model_selection import train_test_split\n",
     "from sksurv.datasets import load_breast_cancer\n",
-    "from sksurv.ensemble import ComponentwiseGradientBoostingSurvivalAnalysis\n",
-    "from sksurv.ensemble import GradientBoostingSurvivalAnalysis\n",
+    "from sksurv.ensemble import ComponentwiseGradientBoostingSurvivalAnalysis, GradientBoostingSurvivalAnalysis\n",
     "from sksurv.preprocessing import OneHotEncoder"
    ]
   },

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/doc/user_guide/coxnet.ipynb

@@ -25,21 +25,21 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import pandas as pd\n",
-    "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
     "\n",
     "%matplotlib inline\n",
     "\n",
-    "from sksurv.datasets import load_breast_cancer\n",
-    "from sksurv.linear_model import CoxPHSurvivalAnalysis, CoxnetSurvivalAnalysis\n",
-    "from sksurv.preprocessing import OneHotEncoder\n",
-    "\n",
     "from sklearn import set_config\n",
     "from sklearn.model_selection import GridSearchCV, KFold\n",
     "from sklearn.pipeline import make_pipeline\n",
     "from sklearn.preprocessing import StandardScaler\n",
     "\n",
+    "from sksurv.datasets import load_breast_cancer\n",
+    "from sksurv.linear_model import CoxnetSurvivalAnalysis, CoxPHSurvivalAnalysis\n",
+    "from sksurv.preprocessing import OneHotEncoder\n",
+    "\n",
     "set_config(display=\"text\")  # displays text representation of estimators"
    ]
   },
@@ -311,7 +311,6 @@
    "source": [
     "def plot_coefficients(coefs, n_highlight):\n",
     "    _, ax = plt.subplots(figsize=(9, 6))\n",
-    "    n_features = coefs.shape[0]\n",
     "    alphas = coefs.columns\n",
     "    for row in coefs.itertuples():\n",
     "        ax.semilogx(alphas, row[1:], \".-\", label=row.Index)\n",
@@ -524,9 +523,8 @@
    ],
    "source": [
     "import warnings\n",
+    "\n",
     "from sklearn.exceptions import FitFailedWarning\n",
-    "from sklearn.pipeline import make_pipeline\n",
-    "from sklearn.preprocessing import StandardScaler\n",
     "\n",
     "coxnet_pipe = make_pipeline(StandardScaler(), CoxnetSurvivalAnalysis(l1_ratio=0.9, alpha_min_ratio=0.01, max_iter=100))\n",
     "warnings.simplefilter(\"ignore\", UserWarning)\n",
@@ -758,7 +756,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.5"
+   "version": "3.12.4"
   }
  },
  "nbformat": 4,

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/doc/user_guide/evaluating-survival-models.ipynb

@@ -29,12 +29,11 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
     "\n",
     "%matplotlib inline\n",
     "import pandas as pd\n",
-    "\n",
     "from sklearn import set_config\n",
     "from sklearn.impute import SimpleImputer\n",
     "from sklearn.model_selection import train_test_split\n",
@@ -42,7 +41,7 @@
     "\n",
     "from sksurv.datasets import load_flchain, load_gbsg2\n",
     "from sksurv.functions import StepFunction\n",
-    "from sksurv.linear_model import CoxPHSurvivalAnalysis, CoxnetSurvivalAnalysis\n",
+    "from sksurv.linear_model import CoxnetSurvivalAnalysis, CoxPHSurvivalAnalysis\n",
     "from sksurv.metrics import (\n",
     "    concordance_index_censored,\n",
     "    concordance_index_ipcw,\n",
@@ -1034,7 +1033,6 @@
     "    as_concordance_index_ipcw_scorer(rsf_gbsg, tau=gbsg_times[-1]),\n",
     "    param_grid=cv_param_grid,\n",
     "    cv=cv,\n",
-    "    n_jobs=4,\n",
     ").fit(gbsg_X, gbsg_y)"
    ]
   },
@@ -1056,7 +1054,6 @@
     "    as_cumulative_dynamic_auc_scorer(rsf_gbsg, times=gbsg_times),\n",
     "    param_grid=cv_param_grid,\n",
     "    cv=cv,\n",
-    "    n_jobs=4,\n",
     ").fit(gbsg_X, gbsg_y)"
    ]
   },
@@ -1080,7 +1077,6 @@
     "    as_integrated_brier_score_scorer(rsf_gbsg, times=gbsg_times),\n",
     "    param_grid=cv_param_grid,\n",
     "    cv=cv,\n",
-    "    n_jobs=4,\n",
     ").fit(gbsg_X, gbsg_y)"
    ]
   },

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/doc/user_guide/survival-svm.ipynb

@@ -58,15 +58,15 @@
    "outputs": [],
    "source": [
     "%matplotlib inline\n",
-    "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
     "import pandas\n",
     "import seaborn as sns\n",
     "from sklearn import set_config\n",
-    "from sklearn.model_selection import ShuffleSplit, GridSearchCV\n",
+    "from sklearn.model_selection import GridSearchCV, ShuffleSplit\n",
     "\n",
-    "from sksurv.datasets import load_veterans_lung_cancer\n",
     "from sksurv.column import encode_categorical\n",
+    "from sksurv.datasets import load_veterans_lung_cancer\n",
     "from sksurv.metrics import concordance_index_censored\n",
     "from sksurv.svm import FastSurvivalSVM\n",
     "\n",

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/pyproject.toml

@@ -10,7 +10,8 @@ requires = [
 
     # scikit-learn requirements
     "scikit-learn~=1.4.0; python_version<='3.12'",
-    "scikit-learn; python_version>'3.12'",
+    "scikit-learn~=1.5.0; python_version=='3.13'",
+    "scikit-learn; python_version>'3.13'",
 ]
 build-backend = "setuptools.build_meta"
 
@@ -48,7 +49,7 @@ dependencies = [
     "numexpr",
     "numpy",
     "osqp !=0.6.0,!=0.6.1",
-    "pandas >=1.0.5",
+    "pandas >=1.4.0",
     "scipy >=1.3.2",
     "scikit-learn >=1.4.0,<1.6",
 ]
@@ -93,7 +94,7 @@ docs = [
 nbval = [
     "ipykernel",
     "ipython!=8.7.0",
-    "matplotlib~=3.6.2",
+    "matplotlib~=3.8.0",
     "nbformat",
     "nbval>=0.10.0",
     "seaborn~=0.11.2",
@@ -197,6 +198,8 @@ select = [
 ]
 
 [tool.ruff.lint.flake8-pytest-style]
+fixture-parentheses = true
+mark-parentheses = true
 parametrize-names-type = "csv"
 
 [tool.ruff.lint.mccabe]

{scikit_survival-0.23.0 → scikit_survival-0.23.1/scikit_survival.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: scikit-survival
-Version: 0.23.0
+Version: 0.23.1
 Summary: Survival analysis built on top of scikit-learn
 Author-email: Sebastian Pölsterl <sebp@k-d-w.org>
 License:                     GNU GENERAL PUBLIC LICENSE
@@ -708,7 +708,7 @@ Requires-Dist: joblib
 Requires-Dist: numexpr
 Requires-Dist: numpy
 Requires-Dist: osqp!=0.6.0,!=0.6.1
-Requires-Dist: pandas>=1.0.5
+Requires-Dist: pandas>=1.4.0
 Requires-Dist: scipy>=1.3.2
 Requires-Dist: scikit-learn<1.6,>=1.4.0
 Provides-Extra: dev
@@ -736,7 +736,7 @@ Requires-Dist: sphinx-copybutton~=0.5.2; extra == "docs"
 Provides-Extra: nbval
 Requires-Dist: ipykernel; extra == "nbval"
 Requires-Dist: ipython!=8.7.0; extra == "nbval"
-Requires-Dist: matplotlib~=3.6.2; extra == "nbval"
+Requires-Dist: matplotlib~=3.8.0; extra == "nbval"
 Requires-Dist: nbformat; extra == "nbval"
 Requires-Dist: nbval>=0.10.0; extra == "nbval"
 Requires-Dist: seaborn~=0.11.2; extra == "nbval"
@@ -781,7 +781,7 @@ Requirements
 - numexpr
 - numpy
 - osqp
-- pandas 1.0.5 or later
+- pandas 1.4.0 or later
 - scikit-learn 1.4 or 1.5
 - scipy
 - C/C++ compiler

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/scikit_survival.egg-info/SOURCES.txt

@@ -474,6 +474,7 @@ tests/test_binarytrees.py
 tests/test_boosting.py
 tests/test_clinical_kernel.py
 tests/test_column.py
+tests/test_common.py
 tests/test_compare.py
 tests/test_coxnet.py
 tests/test_coxph.py

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/scikit_survival.egg-info/requires.txt

@@ -3,7 +3,7 @@ joblib
 numexpr
 numpy
 osqp!=0.6.0,!=0.6.1
-pandas>=1.0.5
+pandas>=1.4.0
 scipy>=1.3.2
 scikit-learn<1.6,>=1.4.0
 
@@ -34,7 +34,7 @@ sphinx-copybutton~=0.5.2
 [nbval]
 ipykernel
 ipython!=8.7.0
-matplotlib~=3.6.2
+matplotlib~=3.8.0
 nbformat
 nbval>=0.10.0
 seaborn~=0.11.2

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/sksurv/column.py

@@ -64,18 +64,8 @@ def standardize(table, with_std=True):
         Table with numeric columns normalized.
         Categorical columns in the input table remain unchanged.
     """
-    if isinstance(table, pd.DataFrame):
-        cat_columns = table.select_dtypes(include=["category"]).columns
-    else:
-        cat_columns = []
-
     new_frame = _apply_along_column(table, standardize_column, with_std=with_std)
 
-    # work around for apply converting category dtype to object
-    # https://github.com/pydata/pandas/issues/9573
-    for col in cat_columns:
-        new_frame[col] = table[col].copy()
-
     return new_frame
 
 

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/sksurv/ensemble/boosting.py

@@ -44,7 +44,7 @@ def _sample_binomial_plus_one(p, size, random_state):
     return drop_model, n_dropped
 
 
-class ComponentwiseLeastSquares(BaseEstimator):
+class _ComponentwiseLeastSquares(BaseEstimator):
     def __init__(self, component):
         self.component = component
 
@@ -70,7 +70,7 @@ def _fit_stage_componentwise(X, residuals, sample_weight, **fit_params):  # pyli
     base_learners = []
     error = np.empty(n_features)
     for component in range(n_features):
-        learner = ComponentwiseLeastSquares(component).fit(X, residuals, sample_weight)
+        learner = _ComponentwiseLeastSquares(component).fit(X, residuals, sample_weight)
         l_pred = learner.predict(X)
         error[component] = squared_norm(residuals - l_pred)
         base_learners.append(learner)

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/sksurv/ensemble/forest.py

@@ -633,7 +633,7 @@ class RandomSurvivalForest(SurvivalAnalysisMixin, _BaseSurvivalForest):
         return super().predict_survival_function(X, return_array)
 
 
-class ExtraSurvivalTrees(_BaseSurvivalForest):
+class ExtraSurvivalTrees(SurvivalAnalysisMixin, _BaseSurvivalForest):
     """An extremely random survival forest.
 
     This class implements a meta estimator that fits a number of randomized

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/sksurv/metrics.py

@@ -96,12 +96,14 @@ def _iter_comparable(event_indicator, event_time, order):
         # check for tied event times
         event_at_same_time = event_indicator[order[i:end]]
         censored_at_same_time = ~event_at_same_time
+
+        mask = np.zeros(n_samples, dtype=bool)
+        mask[end:] = True
+        # an event is comparable to censored samples at same time point
+        mask[i:end] = censored_at_same_time
+
         for j in range(i, end):
             if event_indicator[order[j]]:
-                mask = np.zeros(n_samples, dtype=bool)
-                mask[end:] = True
-                # an event is comparable to censored samples at same time point
-                mask[i:end] = censored_at_same_time
                 tied_time += censored_at_same_time.sum()
                 yield (j, mask, tied_time)
         i = end

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/sksurv/testing.py

@@ -18,9 +18,9 @@ import pkgutil
 import numpy as np
 from numpy.testing import assert_almost_equal, assert_array_equal
 import pytest
+from sklearn.base import BaseEstimator, TransformerMixin
 
 import sksurv
-from sksurv.base import SurvivalAnalysisMixin
 from sksurv.metrics import concordance_index_censored
 
 
@@ -72,8 +72,15 @@ def assert_chf_properties(chf):
         raise AssertionError(f"most ({num_closer_to_one}) hazard rates at first time point are closer to 1 than 0")
 
 
-def _is_survival_mixin(x):
-    return inspect.isclass(x) and x is not SurvivalAnalysisMixin and issubclass(x, SurvivalAnalysisMixin)
+def _is_survival_estimator(x):
+    return (
+        inspect.isclass(x)
+        and issubclass(x, BaseEstimator)
+        and not issubclass(x, TransformerMixin)
+        and x.__module__.startswith("sksurv.")
+        and not x.__name__.startswith("_")
+        and x.__module__.split(".", 2)[1] not in {"metrics", "nonparametric"}
+    )
 
 
 def all_survival_estimators():
@@ -84,7 +91,7 @@ def all_survival_estimators():
         if modname.startswith("sksurv.meta"):
             continue
         module = import_module(modname)
-        for _name, cls in inspect.getmembers(module, _is_survival_mixin):
+        for _name, cls in inspect.getmembers(module, _is_survival_estimator):
             if inspect.isabstract(cls):
                 continue
             all_classes.append(cls)

scikit_survival-0.23.1/tests/test_common.py

@@ -0,0 +1,12 @@
+import pytest
+
+from sksurv.base import SurvivalAnalysisMixin
+from sksurv.testing import all_survival_estimators
+
+
+@pytest.mark.parametrize("estimator_cls", all_survival_estimators())
+def test_survival_analysis_base_clas(estimator_cls):
+    assert hasattr(estimator_cls, "fit")
+    assert hasattr(estimator_cls, "predict")
+    assert hasattr(estimator_cls, "score")
+    assert issubclass(estimator_cls, SurvivalAnalysisMixin)

{scikit_survival-0.23.0 → scikit_survival-0.23.1}/tox.ini

@@ -11,7 +11,7 @@ deps =
 description = Run linters
 skip_install = true
 deps =
-    ruff~=0.4.3
+    ruff~=0.7.1
 commands = ruff check sksurv/ tests/ setup.py
 pass_env = RUFF_*