Merge branch 'for0.20.4' into 0.20.X

Author: jnothman

.circleci/config.yml

@@ -42,11 +42,11 @@ jobs:
       - MINICONDA_PATH: ~/miniconda
       - CONDA_ENV_NAME: testenv
       - PYTHON_VERSION: "2"
-      - NUMPY_VERSION: "1.10"
-      - SCIPY_VERSION: "0.16"
-      - MATPLOTLIB_VERSION: "1.4"
-      - SCIKIT_IMAGE_VERSION: "0.11"
-      - PANDAS_VERSION: "0.17.1"
+      - NUMPY_VERSION: "1.*"
+      - SCIPY_VERSION: "0.*"
+      - MATPLOTLIB_VERSION: "*"
+      - SCIKIT_IMAGE_VERSION: "0.*"
+      - PANDAS_VERSION: "0.*"
     steps:
       - checkout
       - run: ./build_tools/circle/checkout_merge_commit.sh

.travis.yml

@@ -35,12 +35,12 @@ matrix:
             - libatlas-dev
     # Python 3.4 build
     - env: DISTRIB="conda" PYTHON_VERSION="3.4" INSTALL_MKL="false"
-           NUMPY_VERSION="1.10.4" SCIPY_VERSION="0.16.1" CYTHON_VERSION="0.25.2"
-           PILLOW_VERSION="4.0.0" COVERAGE=true
+           NUMPY_VERSION="1.10.4" SCIPY_VERSION="0.17" CYTHON_VERSION="0.25.2"
+           PILLOW_VERSION="4.0.0" COVERAGE=
       if: type != cron
     # Python 3.5 build
     - env: DISTRIB="conda" PYTHON_VERSION="3.5" INSTALL_MKL="false"
-           NUMPY_VERSION="1.10.4" SCIPY_VERSION="0.16.1" CYTHON_VERSION="0.25.2"
+           NUMPY_VERSION="1.10.4" SCIPY_VERSION="0.17" CYTHON_VERSION="0.25.2"
            PILLOW_VERSION="4.0.0" COVERAGE=true
            SKLEARN_SITE_JOBLIB=1 JOBLIB_VERSION="0.11"
       if: type != cron

build_tools/circle/build_doc.sh

@@ -119,7 +119,7 @@ conda update --yes --quiet conda
 # provided versions
 conda create -n $CONDA_ENV_NAME --yes --quiet python="${PYTHON_VERSION:-*}" \
   numpy="${NUMPY_VERSION:-*}" scipy="${SCIPY_VERSION:-*}" cython \
-  pytest coverage matplotlib="${MATPLOTLIB_VERSION:-*}" sphinx=1.6.2 pillow \
+  pytest coverage matplotlib="${MATPLOTLIB_VERSION:-*}" sphinx=1.6.* pillow \
   scikit-image="${SCIKIT_IMAGE_VERSION:-*}" pandas="${PANDAS_VERSION:-*}" \
   joblib
 

build_tools/travis/install.sh

@@ -11,7 +11,7 @@
 # matrix entry) from which we pull from local Travis repository. This allows
 # us to keep build artefact for gcc + cython, and gain time
 
-set -e
+set -ex
 
 echo 'List files from cached directories'
 echo 'pip:'
@@ -38,12 +38,18 @@ make_conda() {
     export PATH=$MINICONDA_PATH/bin:$PATH
     conda update --yes conda
 
-    conda create -n testenv --yes $TO_INSTALL
+    conda create -c conda-forge -n testenv --yes $TO_INSTALL
     source activate testenv
 }
 
+if [[ "$COVERAGE" == "true" ]]; then
+    TEST_DEPS="pytest pytest-cov"
+else
+    TEST_DEPS="pytest"
+fi
+
 if [[ "$DISTRIB" == "conda" ]]; then
-    TO_INSTALL="python=$PYTHON_VERSION pip pytest pytest-cov \
+    TO_INSTALL="python=$PYTHON_VERSION pip $TEST_DEPS \
                 numpy=$NUMPY_VERSION scipy=$SCIPY_VERSION \
                 cython=$CYTHON_VERSION"
 
@@ -84,7 +90,7 @@ elif [[ "$DISTRIB" == "ubuntu" ]]; then
     # and scipy
     virtualenv --system-site-packages testvenv
     source testvenv/bin/activate
-    pip install pytest pytest-cov cython==$CYTHON_VERSION
+    pip install $TEST_DEPS cython==$CYTHON_VERSION
 
 elif [[ "$DISTRIB" == "scipy-dev" ]]; then
     make_conda python=3.7
@@ -96,7 +102,7 @@ elif [[ "$DISTRIB" == "scipy-dev" ]]; then
     echo "Installing joblib master"
     pip install https://github.com/joblib/joblib/archive/master.zip
     export SKLEARN_SITE_JOBLIB=1
-    pip install pytest pytest-cov
+    pip install $TEST_DEPS
 fi
 
 if [[ "$COVERAGE" == "true" ]]; then

build_tools/travis/test_pytest_soft_dependency.sh

@@ -7,6 +7,7 @@ if [[ "$CHECK_PYTEST_SOFT_DEPENDENCY" == "true" ]]; then
     if [[ "$COVERAGE" == "true" ]]; then
         # Need to append the coverage to the existing .coverage generated by
         # running the tests
+        pip install coverage
         CMD="coverage run --append"
     else
         CMD="python"

doc/index.rst

@@ -207,20 +207,20 @@
                     <li><em>On-going development:</em>
                     <a href="/dev/whats_new.html"><em>What's new</em> (Changelog)</a>
                     </li>
-                    <li><strong>Scikit-learn 0.21 will drop support for Python 2.7 and Python 3.4.</strong>
+                    <li><strong>Scikit-learn from 0.21 requires Python 3.5 or greater.</strong>
                     </li>
-                    <li><em>March 2019.</em> scikit-learn 0.20.3 is available for download (<a href="whats_new.html#version-0-20-3">Changelog</a>).
+                    <li><em>July 2019.</em> scikit-learn 0.21.3 (<a href="whats_new.html#version-0-21-3">Changelog</a>) and 0.20.4 (<a href="whats_new.html#version-0-20-4">Changelog</a>) are available for download.
+                    </li>
+                    <li><em>May 2019.</em> scikit-learn 0.21.0 to 0.21.2 are available for download (<a href="whats_new.html#version-0-21">Changelog</a>).
                     </li>
-                    <li><em>December 2018.</em> scikit-learn 0.20.2 is available for download (<a href="whats_new.html#version-0-20-2">Changelog</a>)
+                    <li><em>March 2019.</em> scikit-learn 0.20.3 is available for download (<a href="whats_new.html#version-0-20-3">Changelog</a>).
                     </li>
                     <li><em>September 2018.</em> scikit-learn 0.20.0 is available for download (<a href="whats_new.html#version-0-20-0">Changelog</a>).
                     </li>
                     <li><em>July 2018.</em> scikit-learn 0.19.2 is available for download (<a href="whats_new.html#version-0-19">Changelog</a>).
                     </li>
                     <li><em>July 2017.</em> scikit-learn 0.19.0 is available for download (<a href="whats_new/v0.19.html#version-0-19">Changelog</a>).
                     </li>
-                    <li><em>June 2017.</em> scikit-learn 0.18.2 is available for download (<a href="whats_new/v0.18.html#version-0-18-2">Changelog</a>).
-                    </li>
                     </ul>
                 </div>
 

doc/whats_new/v0.20.rst

@@ -2,6 +2,50 @@
 
 .. currentmodule:: sklearn
 
+ .. _changes_0_20_4:
+
+Version 0.20.4
+==============
+
+**July 30, 2019**
+
+This is a bug-fix release with some bug fixes applied to version 0.20.3.
+
+Changelog
+---------
+
+The bundled version of joblib was upgraded from 0.13.0 to 0.13.2.
+
+:mod:`sklearn.cluster`
+..............................
+
+- |Fix| Fixed a bug in :class:`cluster.KMeans` where KMeans++ initialisation
+  could rarely result in an IndexError. :issue:`11756` by `Joel Nothman`_.
+
+:mod:`sklearn.compose`
+.....................
+
+- |Fix| Fixed an issue in :class:`compose.ColumnTransformer` where using
+  DataFrames whose column order differs between :func:``fit`` and
+  :func:``transform`` could lead to silently passing incorrect columns to the
+  ``remainder`` transformer.
+  :pr:`14237` by `Andreas Schuderer <schuderer>`.
+
+:mod:`sklearn.model_selection`
+..............................
+
+- |Fix| Fixed a bug where :class:`model_selection.StratifiedKFold`
+  shuffles each class's samples with the same ``random_state``,
+  making ``shuffle=True`` ineffective.
+  :issue:`13124` by :user:`Hanmin Qin <qinhanmin2014>`.
+
+:mod:`sklearn.neighbors`
+......................
+
+- |Fix| Fixed a bug in :class:`neighbors.KernelDensity` which could not be
+  restored from a pickle if ``sample_weight`` had been used.
+  :issue:`13772` by :user:`Aditya Vyas <aditya1702>`.
+
  .. _changes_0_20_3:
 
 Version 0.20.3
@@ -30,7 +74,7 @@ Changelog
   :issue:`12946` by :user:`Pierre Tallotte <pierretallotte>`.
 
 :mod:`sklearn.covariance`
-......................
+.........................
 
 - |Fix| Fixed a regression in :func:`covariance.graphical_lasso` so that
   the case `n_features=2` is handled correctly. :issue:`13276` by

sklearn/__init__.py

@@ -44,7 +44,7 @@
 # Dev branch marker is: 'X.Y.dev' or 'X.Y.devN' where N is an integer.
 # 'X.Y.dev0' is the canonical version of 'X.Y.dev'
 #
-__version__ = '0.20.3'
+__version__ = '0.20.4'
 
 
 try:

sklearn/cluster/dbscan_.py

@@ -10,11 +10,11 @@
 # License: BSD 3 clause
 
 import numpy as np
+import warnings
 from scipy import sparse
 
 from ..base import BaseEstimator, ClusterMixin
 from ..utils import check_array, check_consistent_length
-from ..utils.testing import ignore_warnings
 from ..neighbors import NearestNeighbors
 
 from ._dbscan_inner import dbscan_inner
@@ -139,7 +139,8 @@ def dbscan(X, eps=0.5, min_samples=5, metric='minkowski', metric_params=None,
         X.sum_duplicates()  # XXX: modifies X's internals in-place
 
         # set the diagonal to explicit values, as a point is its own neighbor
-        with ignore_warnings():
+        with warnings.catch_warnings():
+            warnings.simplefilter('ignore', sparse.SparseEfficiencyWarning)
             X.setdiag(X.diagonal())  # XXX: modifies X's internals in-place
 
         X_mask = X.data <= eps

sklearn/cluster/k_means_.py

@@ -111,6 +111,9 @@ def _k_init(X, n_clusters, x_squared_norms, random_state, n_local_trials=None):
         rand_vals = random_state.random_sample(n_local_trials) * current_pot
         candidate_ids = np.searchsorted(stable_cumsum(closest_dist_sq),
                                         rand_vals)
+        # XXX: numerical imprecision can result in a candidate_id out of range
+        np.clip(candidate_ids, None, closest_dist_sq.size - 1,
+                out=candidate_ids)
 
         # Compute distances to center candidates
         distance_to_candidates = euclidean_distances(