Refactor tests to generate test and session scoped Zarr and Icechunk stores independently.

Cargo.toml

@@ -24,5 +24,9 @@ zarrs_object_store = "0.5.0"
 zarrs_storage = { version = "0.4.0", features = ["async"] }
 
 [dev-dependencies]
+chrono = "0.4.44"
+geo = "0.31"
 geoarrow-array = "0.7.0"
+tempfile = "3"
 tokio = { version = "1.48", features = ["macros", "rt-multi-thread"] }
+wkb = "0.9"

python/pyproject.toml

@@ -31,3 +31,24 @@ dev-dependencies = [
     "pytest",
     "shapely",
 ]
+
+
+[tool.ruff]
+line-length = 88
+target-version = "py311"
+
+[tool.ruff.lint]
+select = ["E", "F", "I"]
+per-file-ignores = {}
+
+[tool.ruff.format]
+quote-style = "double"
+indent-style = "space"
+skip-magic-trailing-comma = false
+line-ending = "auto"
+
+[tool.pytest.ini_options]
+# Directories to search for tests
+testpaths = [
+    "tests",
+]

python/python/zarr_datafusion_search/__init__.py

@@ -1,7 +1,6 @@
 from __future__ import annotations
 
-from ._rust import ZarrTable
-from ._rust import ___version
+from ._rust import ZarrTable, ___version
 
 __version__: str = ___version()
 

python/tests/conftest.py

@@ -0,0 +1,51 @@
+"""Pytest fixtures for zarr-datafusion tests."""
+import pytest
+
+from .fixtures import create_test_icechunk_data, create_test_zarr_data
+
+
+@pytest.fixture
+def zarr_store(tmp_path):
+    """Create a temporary zarr store with test data.
+
+    Yields the path to the zarr store, which is automatically cleaned up after the test.
+    """
+    store_path = tmp_path / "zarr_store.zarr"
+    create_test_zarr_data(store_path)
+    yield store_path
+
+
+@pytest.fixture
+def icechunk_store(tmp_path):
+    """Create a temporary icechunk store with test data.
+
+    Yields the path to the icechunk store, which is automatically cleaned up
+    after the test.
+    """
+    store_path = tmp_path / "icechunk"
+    create_test_icechunk_data(store_path)
+    yield store_path
+
+
+@pytest.fixture(scope="session")
+def session_zarr_store(tmp_path_factory):
+    """Create a session-scoped zarr store with test data.
+
+    This fixture is created once per test session and shared across all tests.
+    Useful for read-only tests that don't modify the data.
+    """
+    store_path = tmp_path_factory.mktemp("data") / "zarr_store.zarr"
+    create_test_zarr_data(store_path)
+    yield store_path
+
+
+@pytest.fixture(scope="session")
+def session_icechunk_store(tmp_path_factory):
+    """Create a session-scoped icechunk store with test data.
+
+    This fixture is created once per test session and shared across all tests.
+    Useful for read-only tests that don't modify the data.
+    """
+    store_path = tmp_path_factory.mktemp("data") / "icechunk"
+    create_test_icechunk_data(store_path)
+    yield store_path

python/tests/fixtures.py

@@ -0,0 +1,89 @@
+from pathlib import Path
+
+import icechunk
+import numpy as np
+import shapely
+import zarr
+from zarr.core.group import Group
+from zarr.dtype import VariableLengthBytes, VariableLengthUTF8
+
+
+def generate_test_data(root: Group) -> None:
+    """Create test zarr data
+
+    Adds to a zarr store:
+    - a /meta group which contains
+    - date: datetime64[ms] array with dates [2023-01-01, 2023-01-02,
+      2023-01-03]
+    - collection: variable length UTF8 array with ["collection_a",
+      "collection_b", "collection_c"]
+    - bbox: variable length bytes (WKB) array with boxes at (-10,-10,10,10),
+      (-20,-20,20,20), (-30,-30,30,30)
+
+    Args:
+        root: The root Group in the target Zarr or Icechunk store.
+    """
+    meta = root.create_group("meta")
+
+    # Create date array
+    date_data = np.array(
+        ["2023-01-01", "2023-01-02", "2023-01-03"], dtype="datetime64[ms]"
+    )
+    meta.create_array("date", data=date_data)
+
+    # Create collection array
+    collection_data = ["collection_a", "collection_b", "collection_c"]
+    collection_array = meta.create_array(
+        "collection",
+        shape=(len(collection_data),),
+        dtype=VariableLengthUTF8(),
+    )
+    collection_array[:] = collection_data
+
+    # Create bbox array with WKB-encoded geometries
+    bbox_data = shapely.to_wkb(
+        [
+            shapely.box(-10.0, -10.0, 10.0, 10.0),
+            shapely.box(-20.0, -20.0, 20.0, 20.0),
+            shapely.box(-30.0, -30.0, 30.0, 30.0),
+        ]
+    )
+
+    bbox_array = meta.create_array(
+        "bbox",
+        shape=(len(bbox_data),),
+        dtype=VariableLengthBytes(),
+    )
+    bbox_array[:] = bbox_data
+
+
+def create_test_zarr_data(store_path: Path | str) -> None:
+    """Create test store data at the specified path.
+
+    Args:
+        store_path: Path where the Zarr store will be created. Can be string
+            or Path object.
+
+    """
+    store_path = Path(store_path) if isinstance(store_path, str) else store_path
+    root = zarr.open_group(store_path, mode="w", zarr_format=3)
+    generate_test_data(root=root)
+
+
+def create_test_icechunk_data(store_path: Path | str) -> None:
+    """Create test icechunk store at the specified path.
+
+    Args:
+        store_path: Path where the Icechunk repository will be created.
+            Can be string or Path object.
+
+    """
+    store_path = Path(store_path) if isinstance(store_path, str) else store_path
+
+    storage = icechunk.local_filesystem_storage(str(store_path))
+    repo = icechunk.Repository.create(storage)
+    session = repo.writable_session("main")
+
+    root = zarr.open_group(session.store, mode="w", zarr_format=3)
+    generate_test_data(root=root)
+    session.commit("Initial test data")