MAINT add manifest check in circle ci config (scikit-learn#18128)

.github/workflows/check-manifest.yml

@@ -0,0 +1,21 @@
+name: "Check Manifest"
+
+on:
+  schedule:
+    - cron: '0 0 * * *'
+
+jobs:
+  check:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v2
+      - uses: actions/setup-python@v2
+        with:
+          python-version: '3.8'
+      - name: Install dependencies
+        # scipy and cython are required to build sdist
+        run: |
+          python -m pip install --upgrade pip
+          pip install check-manifest scipy cython
+      - run: |
+          check-manifest -v

MANIFEST.in

@@ -6,3 +6,27 @@ recursive-include sklearn/datasets *.csv *.csv.gz *.rst *.jpg *.txt *.arff.gz *.
 include COPYING
 include README.rst
 include pyproject.toml
+include sklearn/externals/README
+include sklearn/svm/src/liblinear/COPYRIGHT
+include sklearn/svm/src/libsvm/LIBSVM_CHANGES
+include conftest.py
+include Makefile
+include MANIFEST.in
+include .coveragerc
+
+# exclude from sdist
+recursive-exclude asv_benchmarks *
+recursive-exclude benchmarks *
+recursive-exclude build_tools *
+recursive-exclude maint_tools *
+recursive-exclude benchmarks *
+recursive-exclude .binder *
+recursive-exclude .circleci *
+exclude .codecov.yml
+exclude .mailmap
+exclude .pre-commit-config.yaml
+exclude azure-pipelines.yml
+exclude lgtm.yml
+exclude CODE_OF_CONDUCT.md
+exclude CONTRIBUTING.md
+exclude PULL_REQUEST_TEMPLATE.md

setup.cfg

@@ -32,3 +32,10 @@ ignore=E121,E123,E126,E226,E24,E704,W503,W504
 [mypy]
 ignore_missing_imports = True
 allow_redefinition = True
+
+[check-manifest]
+# ignore files missing in VCS
+ignore =
+    sklearn/linear_model/_sag_fast.pyx
+    sklearn/utils/_seq_dataset.pxd
+    sklearn/utils/_seq_dataset.pyx