Sklearn 1.0

.github/workflows/lint.yaml

@@ -0,0 +1,10 @@
+name: Linting
+on: [pull_request, push]
+
+jobs:
+  lint:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v2
+      - uses: actions/setup-python@v2
+      - uses: pre-commit/[email protected]

.github/workflows/tests.yaml

@@ -0,0 +1,44 @@
+name: Tests
+
+on: [push, pull_request]
+
+jobs:
+  test:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      fail-fast: false
+      matrix:
+        # os: ["windows-latest", "ubuntu-latest", "macos-latest"]
+        os: ["ubuntu-latest"]
+        python-version: ["3.7", "3.8", "3.9"]
+
+    env:
+      PYTHON_VERSION: ${{ matrix.python-version }}
+      PARALLEL: "true"
+      COVERAGE: "true"
+
+    steps:
+      - name: Checkout source
+        uses: actions/checkout@v2
+        with:
+          fetch-depth: 0  # Needed by codecov.io
+
+      - name: Setup Conda Environment
+        uses: conda-incubator/setup-miniconda@v2
+        with:
+          miniforge-variant: Mambaforge
+          miniforge-version: latest
+          use-mamba: true
+          channel-priority: strict
+          python-version: ${{ matrix.python-version }}
+          environment-file: ci/environment-${{ matrix.python-version }}.yaml
+          activate-environment: test-environment
+          auto-activate-base: false
+
+      - name: Install
+        shell: bash -l {0}
+        run: source ci/install.sh
+
+      - name: Run tests
+        shell: bash -l {0}
+        run: pytest -v

ci/environment-3.7.yaml

@@ -1,23 +1,14 @@
-name: dask-ml-test
+name: dask-ml-3.7
 channels:
   - conda-forge
   - defaults
 dependencies:
-  - black==19.10b0
-  - coverage
-  - codecov
-  # dask 2021.3.0 introduced a regression which causes tests to fail.
-  # The issue has been resolved upstream in dask and will be included
-  # in the next release. We temporarily apply a dask version contraint
-  # to allow CI to pass
-  - dask !=2021.3.0
-  - dask-glm >=0.2.0
-  - flake8
-  - isort==4.3.21
+  - dask
+  - dask-glm
   - multipledispatch >=0.4.9
   - mypy
   - numba
-  - numpy >=1.16.3
+  - numpy
   - numpydoc
   - packaging
   - pandas
@@ -26,10 +17,7 @@ dependencies:
   - pytest-cov
   - pytest-mock
   - python=3.7.*
-  - scikit-learn>=0.23.0
+  - scikit-learn>=1.0.0
   - scipy
   - sparse
-  - toolz
-  - pip
-  - pip:
-    - pytest-azurepipelines
+  - toolz

ci/environment-3.8.yaml

@@ -1,23 +1,14 @@
-name: dask-ml-test
+name: dask-ml-3.8
 channels:
   - conda-forge
   - defaults
 dependencies:
-  - black==19.10b0
-  - coverage
-  - codecov
-  # dask 2021.3.0 introduced a regression which causes tests to fail.
-  # The issue has been resolved upstream in dask and will be included
-  # in the next release. We temporarily apply a dask version contraint
-  # to allow CI to pass
-  - dask !=2021.3.0
-  - dask-glm >=0.2.0
-  - flake8
-  - isort==4.3.21
+  - dask
+  - dask-glm
   - multipledispatch >=0.4.9
   - mypy
   - numba
-  - numpy >=1.16.3
+  - numpy
   - numpydoc
   - packaging
   - pandas
@@ -26,10 +17,7 @@ dependencies:
   - pytest-cov
   - pytest-mock
   - python=3.8.*
-  - scikit-learn>=0.23.0
+  - scikit-learn>=1.0.0
   - scipy
   - sparse
-  - toolz
-  - pip
-  - pip:
-    - pytest-azurepipelines
+  - toolz

ci/environment-3.9.yaml

@@ -0,0 +1,23 @@
+name: dask-ml-3.9
+channels:
+  - conda-forge
+  - defaults
+dependencies:
+  - dask
+  - dask-glm
+  - multipledispatch >=0.4.9
+  - mypy
+  - numba
+  - numpy
+  - numpydoc
+  - packaging
+  - pandas
+  - psutil
+  - pytest
+  - pytest-cov
+  - pytest-mock
+  - python=3.9.*
+  - scikit-learn>=1.0.0
+  - scipy
+  - sparse
+  - toolz

ci/environment-docs.yaml

@@ -5,12 +5,10 @@ channels:
 dependencies:
   - black
   - coverage
-  - flake8
   - graphviz
   - heapdict
   - ipykernel
   - ipython
-  - isort==4.3.21
   - multipledispatch
   - mypy
   - nbsphinx
@@ -21,18 +19,14 @@ dependencies:
   - numpydoc
   - pandas
   - psutil
-  - pytest
-  - pytest-cov
-  - pytest-mock
-  - python=3.7
+  - python=3.8
   - sortedcontainers
-  - scikit-learn>=0.23.1
+  - scikit-learn>=1.0.0
   - scipy
   - sparse
-  - sphinx==1.8.5
+  - sphinx
   - sphinx_rtd_theme
   - sphinx-gallery
-  - testpath<0.4
   - tornado
   - toolz
   - xgboost

ci/install.sh

@@ -0,0 +1,4 @@
+python -m pip install --quiet --no-deps -e .
+
+echo mamba list
+mamba list

dask_ml/_compat.py

@@ -16,8 +16,6 @@
 PANDAS_VERSION = packaging.version.parse(pandas.__version__)
 DISTRIBUTED_VERSION = packaging.version.parse(distributed.__version__)
 
-SK_0_23_2 = SK_VERSION >= packaging.version.parse("0.23.2")
-SK_024 = SK_VERSION >= packaging.version.parse("0.24.0.dev0")
 DASK_240 = DASK_VERSION >= packaging.version.parse("2.4.0")
 DASK_2130 = DASK_VERSION >= packaging.version.parse("2.13.0")
 DASK_2_20_0 = DASK_VERSION >= packaging.version.parse("2.20.0")
@@ -49,9 +47,7 @@ def _check_multimetric_scoring(estimator, scoring=None):
     from sklearn.metrics._scorer import _check_multimetric_scoring
     from sklearn.metrics import check_scoring
 
-    if SK_024:
-        if callable(scoring) or isinstance(scoring, (type(None), str)):
-            scorers = {"score": check_scoring(estimator, scoring=scoring)}
-            return scorers, False
-        return _check_multimetric_scoring(estimator, scoring), True
-    return _check_multimetric_scoring(estimator, scoring)
+    if callable(scoring) or isinstance(scoring, (type(None), str)):
+        scorers = {"score": check_scoring(estimator, scoring=scoring)}
+        return scorers, False
+    return _check_multimetric_scoring(estimator, scoring), True

dask_ml/_partial.py

@@ -160,7 +160,7 @@ def _blocks_and_name(obj):
 
 
 def _predict(model, x):
-    return model.predict(x)[:, None]
+    return model.predict(x)
 
 
 def predict(model, x):
@@ -173,15 +173,25 @@ def predict(model, x):
 
     See docstring for ``da.learn.fit``
     """
-    if not hasattr(x, "chunks") and hasattr(x, "to_dask_array"):
-        x = x.to_dask_array()
-    assert x.ndim == 2
-    if len(x.chunks[1]) > 1:
-        x = x.rechunk(chunks=(x.chunks[0], sum(x.chunks[1])))
     func = partial(_predict, model)
-    xx = np.zeros((1, x.shape[1]), dtype=x.dtype)
-    dt = model.predict(xx).dtype
-    return x.map_blocks(func, chunks=(x.chunks[0], (1,)), dtype=dt).squeeze()
+
+    if getattr(model, "feature_names_in_", None) is not None:
+        meta = model.predict(x._meta_nonempty)
+        return x.map_partitions(func, meta=meta)
+    else:
+        if len(x.chunks[1]) > 1:
+            x = x.rechunk(chunks=(x.chunks[0], sum(x.chunks[1])))
+
+        xx = np.zeros((1, x.shape[1]), dtype=x.dtype)
+        meta = model.predict(xx)
+
+        if meta.ndim > 1:
+            chunks = (x.chunks[0], (1,))
+            drop_axis = None
+        else:
+            chunks = (x.chunks[0],)
+            drop_axis = 1
+        return x.map_blocks(func, chunks=chunks, meta=meta, drop_axis=drop_axis)
 
 
 def _copy_partial_doc(cls):