v0.1.1

README.md

@@ -8,8 +8,6 @@
 
 The geopins package provides geospatial support for the [Python pins package](https://github.com/rstudio/pins-python). The package publishes data, models, and other Python objects, making it easy to share them across projects and with your colleagues. With `geopins`, there is support for geospatial datatypes (e.g. `geopandas.GeoDatFrame`, and `rastr.Raster`) and filetypes (e.g. `GeoPackage`, and `GeoTIFF`), fully compatible with your existing pins boards.
 
-The pins package provides drivers for standard Python tabular datatypes (e.g. `pandas.DataFrame`) and filetypes (e.g. CSV, Parquet, etc.). The geopins package brings the same experience to geospatial datatypes (e.g. `geopandas.GeoDatFrame`, and `rastr.Raster` and filetypes, fully compatible with your existing pins boards.
-
 ## Installation
 
 ```bash

src/geopins/drivers/gdf/filetypes/gpkg.py

@@ -3,6 +3,7 @@
 import tempfile
 import warnings
 from pathlib import Path
+from sqlite3 import connect
 from typing import TYPE_CHECKING
 
 import geopandas as gpd
@@ -112,6 +113,9 @@ def pin_write_gdf_gpkg(  # noqa: PLR0913
         path = Path(tmpdir_path) / f"{name}.gpkg"
         x.to_file(path, driver="GPKG")
 
+        # Overwrite the modification time to keep hashing stable and release locks.
+        _snapshot_last_change(path=path)
+
         with warnings.catch_warnings():
             # Upstream issue relating to opening files without context managers
             warnings.simplefilter("ignore", category=ResourceWarning)
@@ -123,3 +127,21 @@ def pin_write_gdf_gpkg(  # noqa: PLR0913
                 description=description,
                 metadata=metadata,
             )
+
+
+def _snapshot_last_change(path: Path) -> None:
+    """Set the last_change timestamp to Unix epoch to keep GeoPackage hashing stable."""
+
+    # Avoid `with connect(...)` because the context manager delays handle release on
+    # Windows, which keeps the temporary GeoPackage locked during cleanup.
+    conn = connect(path.as_posix())
+    try:
+        conn.execute(
+            """
+            UPDATE gpkg_contents
+            SET last_change = '1970-01-01T00:00:00Z';
+            """
+        )
+        conn.commit()
+    finally:
+        conn.close()

tests/geopins/drivers/gdf/filetypes/test_gpkg.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+from time import sleep
 from typing import TYPE_CHECKING
 
 import geopandas as gpd
@@ -27,3 +28,20 @@ def test_round_trip(tmp_geoboard: GeoBaseBoard):
     # Assert
     assert gdf.equals(retrieved)
     assert gdf.crs == retrieved.crs
+
+
+def test_hash_is_not_dependent_on_file_write_time(tmp_geoboard: GeoBaseBoard):
+    # Arrange
+    gdf = gpd.GeoDataFrame(
+        {"id": [1, 2, 3]},
+        geometry=gpd.points_from_xy([0, 1, 2], [0, 1, 2]),
+        crs="EPSG:2193",  # NZGD2000 / New Zealand Transverse Mercator 2000
+    )
+
+    # Act
+    meta1 = tmp_geoboard.pin_write(gdf, name="test-gdf-hash", type="gpkg")
+    sleep(1)
+    meta2 = tmp_geoboard.pin_write(gdf, name="test-gdf-hash", type="gpkg")
+
+    # Assert
+    assert meta1.pin_hash == meta2.pin_hash