CI/TST: pip extras install (#49241)

Author: mroeschke

.github/workflows/package-checks.yml

@@ -0,0 +1,56 @@
+name: Package Checks
+
+on:
+  push:
+    branches:
+      - main
+      - 1.5.x
+  pull_request:
+    branches:
+      - main
+      - 1.5.x
+
+permissions:
+  contents: read
+
+jobs:
+  pip:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        extra: ["test", "performance", "timezone", "computation", "fss", "aws", "gcp", "excel", "parquet", "feather", "hdf5", "spss", "postgresql", "mysql", "sql-other", "html", "xml", "plot", "output_formatting", "clipboard", "compression", "all"]
+      fail-fast: false
+    name: Install Extras - ${{ matrix.extra }}
+    concurrency:
+      # https://github.community/t/concurrecy-not-work-for-push/183068/7
+      group: ${{ github.event_name == 'push' && github.run_number || github.ref }}-pip-extras-${{ matrix.extra }}
+      cancel-in-progress: true
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
+
+      - name: Setup Python
+        id: setup_python
+        uses: actions/setup-python@v3
+        with:
+          python-version: '3.8'
+
+      # Hacky patch to disable building cython extensions.
+      # This job should only check that the extras successfully install.
+      - name: Disable building ext_modules
+        run: |
+          sed -i '/ext_modules=/d' setup.py
+        shell: bash -el {0}
+
+      - name: Install required dependencies
+        run: |
+          python -m pip install --upgrade pip setuptools wheel python-dateutil pytz numpy cython
+        shell: bash -el {0}
+
+      - name: Pip install with extra
+        run: |
+          python -m pip install -e .[${{ matrix.extra }}] --no-build-isolation
+        shell: bash -el {0}

doc/source/getting_started/install.rst

@@ -415,7 +415,7 @@ depending on required compatibility.
 Dependency                Minimum Version    optional_extra   Notes
 ========================= ================== ================ =============================================================
 PyTables                  3.6.1              hdf5             HDF5-based reading / writing
-blosc                     1.21.0             hdf5             Compression for HDF5
+blosc                     1.21.0             hdf5             Compression for HDF5; only available on ``conda``
 zlib                                         hdf5             Compression for HDF5
 fastparquet               0.6.3              -                Parquet reading / writing (pyarrow is default)
 pyarrow                   6.0.0              parquet, feather Parquet, ORC, and feather reading / writing

scripts/validate_min_versions_in_sync.py

@@ -24,7 +24,7 @@
 )
 CODE_PATH = pathlib.Path("pandas/compat/_optional.py").resolve()
 SETUP_PATH = pathlib.Path("setup.cfg").resolve()
-EXCLUDE_DEPS = {"tzdata"}
+EXCLUDE_DEPS = {"tzdata", "blosc"}
 # pandas package is not available
 # in pre-commit environment
 sys.path.append("pandas/compat")
@@ -38,10 +38,11 @@
 
 
 def get_versions_from_code() -> dict[str, str]:
+    """Min versions for checking within pandas code."""
     install_map = _optional.INSTALL_MAPPING
     versions = _optional.VERSIONS
     for item in EXCLUDE_DEPS:
-        versions.pop(item)
+        versions.pop(item, None)
     return {
         install_map.get(k, k).casefold(): v
         for k, v in versions.items()
@@ -50,6 +51,7 @@ def get_versions_from_code() -> dict[str, str]:
 
 
 def get_versions_from_ci(content: list[str]) -> tuple[dict[str, str], dict[str, str]]:
+    """Min versions in CI job for testing all optional dependencies."""
     # Don't parse with pyyaml because it ignores comments we're looking for
     seen_required = False
     seen_optional = False
@@ -79,6 +81,7 @@ def get_versions_from_ci(content: list[str]) -> tuple[dict[str, str], dict[str,
 
 
 def get_versions_from_setup() -> dict[str, str]:
+    """Min versions in setup.cfg for pip install pandas[extra]."""
     install_map = _optional.INSTALL_MAPPING
     optional_dependencies = {}
 
@@ -99,7 +102,7 @@ def get_versions_from_setup() -> dict[str, str]:
         optional_dependencies[install_map.get(package, package).casefold()] = version
 
     for item in EXCLUDE_DEPS:
-        optional_dependencies.pop(item)
+        optional_dependencies.pop(item, None)
 
     return optional_dependencies
 

setup.cfg

@@ -85,7 +85,8 @@ parquet =
 feather =
     pyarrow>=6.0.0
 hdf5 =
-    blosc>=1.20.1
+    # blosc only available on conda (https://github.com/Blosc/python-blosc/issues/297)
+    # blosc>=1.20.1
     tables>=3.6.1
 spss =
     pyreadstat>=1.1.2
@@ -121,7 +122,8 @@ compression =
 # `all ` should be kept as the complete set of pandas optional dependencies for general use.
 all =
     beautifulsoup4>=4.9.3
-    blosc>=1.21.0
+    # blosc only available on conda (https://github.com/Blosc/python-blosc/issues/297)
+    # blosc>=1.21.0
     bottleneck>=1.3.2
     brotlipy>=0.7.0
     fastparquet>=0.6.3