Merge pull request #3 from DeflateAwning/test-sample-db

* test: add sample file to read, create basic tests

* several CI fixes

* fix: ensure `\r` not in table_names on Windows

Author: DeflateAwning

.github/workflows/ci.yml

@@ -12,11 +12,12 @@ permissions:
 
 jobs:
   build:
-    timeout-minutes: 10
+    timeout-minutes: 5
     strategy:
       matrix:
         os: ["ubuntu-latest", "macos-latest", "windows-latest"]
         python-version: ["3.11", "3.12", "3.13"]
+        polars-version: ["1.20", "1.30", "1.31"]
         include:
           # Methodology: Use Ubuntu for basically everything, then target
           # specific cases for the OSes and Python versions.
@@ -26,8 +27,6 @@ jobs:
             python-version: "3.12"
 
           # Polars version tests (Python 3.12 only).
-          - python-version: "3.12"
-            polars-version: "0.20"
           - python-version: "3.12"
             polars-version: "1.20"
           - python-version: "3.12"
@@ -50,7 +49,7 @@ jobs:
       - name: Set up Python
         run: uv python install
 
-      - name: Install uv dependencies
+      - name: Install uv dependencies (Linux/macOS)
         if: runner.os == 'Linux' || runner.os == 'macOS'
         run: |
           if [ -n "${{ matrix.polars-version || '' }}" ]; then
@@ -62,6 +61,7 @@ jobs:
           fi
       - name: Install uv dependencies (Windows)
         if: runner.os == 'Windows'
+        shell: pwsh
         run: |
           if ($env:MATRIX_POLARS_VERSION) {
             Write-Host "Overriding Polars version to $env:MATRIX_POLARS_VERSION"
@@ -70,14 +70,16 @@ jobs:
           } else {
             uv sync --all-extras
           }
-        shell: pwsh
 
       - name: Run linting
         run: |
+          uv run python -c "import polars ; polars.show_versions()"
           uv run ruff format --check .
           uv run ruff check .
 
       - name: Run type checking
+        # Skip type checking on many old Polars versions.
+        if: matrix.polars-version != '0.20'
         run: uv run pyright .
 
       # Linux (Ubuntu)
@@ -102,20 +104,34 @@ jobs:
         if: runner.os == 'Windows'
         run: |
           $url = "https://github.com/lsgunth/mdbtools-win/archive/refs/tags/v1.0.0.zip"
-          $zipPath = "$env:RUNNER_TEMP\mdbtools-win.zip"
-          $extractPath = "$env:GITHUB_WORKSPACE\path"
+          $zipPath = "$env:RUNNER_TEMP\mdbtools-win-1.0.0.zip"
+          $extractRoot = "$env:RUNNER_TEMP\mdbtools"
+          $installPath = "$env:GITHUB_WORKSPACE\mdbtools"
 
           # Download the zip file
           Invoke-WebRequest -Uri $url -OutFile $zipPath
 
-          # Create the target directory if it doesn't exist
-          New-Item -ItemType Directory -Force -Path $extractPath | Out-Null
+          # Create the extraction root and install directories
+          New-Item -ItemType Directory -Force -Path $extractRoot | Out-Null
+          New-Item -ItemType Directory -Force -Path $installPath | Out-Null
 
           # Extract the zip file
-          Expand-Archive -Path $zipPath -DestinationPath $extractPath -Force
+          Expand-Archive -Path $zipPath -DestinationPath $extractRoot -Force
+
+          # Move contents from the subfolder (mdbtools-win-1.0.0) to install path
+          $innerDir = Join-Path $extractRoot "mdbtools-win-1.0.0"
+          Copy-Item -Path (Join-Path $innerDir "*") -Destination $installPath -Recurse -Force
 
-          Write-Host "Extracted to $extractPath"
+          # Add install path to PATH
+          Write-Host "Adding $installPath to PATH"
+          echo "$installPath" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
+
+          Write-Host "MDBTools installed at $installPath"
         shell: pwsh
 
+      - name: Verify MDBTools installation
+        run: |
+          mdb-ver --version
+
       - name: Run tests
         run: uv run pytest tests/

pyproject.toml

@@ -10,7 +10,7 @@ classifiers = [
   "License :: OSI Approved :: MIT License",
   "Operating System :: OS Independent",
 ]
-dependencies = ["polars>=0.20"]
+dependencies = ["polars>=1.20"]
 keywords = ["polars", "mdbtools", "access", "database", "dataframe"]
 
 [project.urls]
@@ -23,7 +23,12 @@ requires = ["hatchling"]
 build-backend = "hatchling.build"
 
 [dependency-groups]
-dev = ["pyright>=1.1.406", "pytest>=8.4.2", "ruff>=0.14.1"]
+dev = [
+    "pyright>=1.1.406",
+    "pytest>=8.4.2",
+    "requests>=2.32.5",
+    "ruff>=0.14.1",
+]
 
 [tool.pyright]
 typeCheckingMode = "strict"

src/polars_access_mdbtools/__init__.py

@@ -36,10 +36,15 @@ def list_table_names(db_path: str | Path) -> list[str]:
     :param db_path: The MS Access database file.
     :return: A list of the tables in a given database.
     """
-    tables = subprocess.check_output(  # noqa: S603
-        ["mdb-tables", "--single-column", _path_to_cmd_str(db_path)],  # noqa: S607
-    ).decode()
-    return tables.strip().split("\n")
+    tables = (
+        subprocess.check_output(  # noqa: S603
+            ["mdb-tables", "--single-column", _path_to_cmd_str(db_path)],  # noqa: S607
+        )
+        .decode()
+        .replace("\r\n", "\n")
+        .strip()
+    )
+    return tables.split("\n")
 
 
 def _convert_data_type_from_access_to_polars(  # noqa: C901, PLR0911, PLR0912

src/polars_access_mdbtools/py.typed

@@ -0,0 +1,45 @@
+"""Fixtures (sample files) for tests."""
+
+import hashlib
+import tempfile
+from collections.abc import Iterator
+from pathlib import Path
+
+import pytest
+import requests
+
+
+def _sha256_checksum(file_path: Path) -> str:
+    """Compute the SHA-256 checksum of a file.
+
+    :param file_path: The path to the file.
+    :return: The SHA-256 checksum as a hexadecimal string.
+    """
+    sha256 = hashlib.sha256()
+    with file_path.open("rb") as f:
+        for chunk in iter(lambda: f.read(8192), b""):
+            sha256.update(chunk)
+    return sha256.hexdigest()
+
+
+@pytest.fixture
+def sample_db_1() -> Iterator[Path]:
+    """Give a sample access database file for tests.
+
+    Download the Access Example .accdb file to a temporary directory,
+    yield its path for tests, and delete it afterward.
+    """
+    url = "https://github.com/Access-projects/Access-examples/raw/refs/heads/master/Access_Example_VBA.accdb"
+    with tempfile.TemporaryDirectory() as temp_dir_str:
+        db_path = Path(temp_dir_str) / "Access_Example_VBA.accdb"
+        response = requests.get(url, stream=True, timeout=10)
+        response.raise_for_status()
+
+        db_path.write_bytes(response.content)
+        assert (
+            _sha256_checksum(db_path)
+            == "3cd1fc88d70bc93909ae6d190fa9692e709d5641b205fd2f66185c905e42cfbf"
+        ), "Downloaded file checksum does not match expected value."
+
+        yield db_path  # Provide file path to test.
+        # Tempdir automatically cleaned up on exit.

tests/conftest.py

@@ -0,0 +1,23 @@
+"""Test the `list_table_names()` function."""
+
+from pathlib import Path
+
+from polars_access_mdbtools import list_table_names
+
+
+def test_list_table_names_db_1(sample_db_1: Path) -> None:
+    table_names = list_table_names(sample_db_1)
+    assert isinstance(table_names, list)
+    assert table_names == [
+        "CMD_LINE_TB",
+        "CUM_VAL_TB",
+        "DATABASE_STRUCTURE_TB",
+        "DECUM_VAL_TB",
+        "ROULETTE_TB",
+        "TAG_GRP_TB",
+        "TAG_NME_TB",
+        "tblContacts",
+        "tblDefaults",
+        "tblFileList",
+        "USysRibbons",
+    ]

tests/test_list_table_names.py

@@ -0,0 +1,24 @@
+"""Test to ensure mdbtools is installed and accessible."""
+
+import subprocess
+
+import pytest
+
+
+def test_mdbtools_installed() -> None:
+    """Test that mdbtools commands are accessible."""
+    try:
+        result = subprocess.run(
+            ["mdb-ver", "--version"],  # noqa: S607
+            capture_output=True,
+            text=True,
+            check=True,
+        )
+        output = result.stdout.strip()
+        assert "mdbtools" in output, (
+            "mdb-ver output does not indicate MDB Tools is installed."
+        )
+    except FileNotFoundError:
+        pytest.fail("mdb-ver command not found. MDB Tools may not be installed.")
+    except subprocess.CalledProcessError as e:
+        pytest.fail(f"mdb-ver command failed with error: {e.stderr.strip()}")

tests/test_mdbtools_installed.py

@@ -1,6 +1,26 @@
 """Tests for reading tables from Access databases."""
 
+from pathlib import Path
 
-def test_placeholder() -> None:
-    """A placeholder test."""
-    assert True
+from polars_access_mdbtools import read_table
+
+
+def test_can_read_all_tables_in_sample_db_1(sample_db_1: Path) -> None:
+    """Test reading all tables in sample_db_1. Check row counts."""
+    for table_name, expected_rows in [
+        ("CMD_LINE_TB", 14),
+        ("CUM_VAL_TB", 59),
+        ("DATABASE_STRUCTURE_TB", 46),
+        ("DECUM_VAL_TB", 59),
+        ("ROULETTE_TB", 38),
+        ("TAG_GRP_TB", 19),
+        ("TAG_NME_TB", 933),
+        ("tblContacts", 0),
+        ("tblDefaults", 1),
+        ("tblFileList", 0),
+        ("USysRibbons", 2),
+    ]:
+        df = read_table(sample_db_1, table_name)
+        assert df.height == expected_rows, (
+            f"Table {table_name} should have {expected_rows} rows, but had {df.height}."
+        )