MAINT: change conda dependency to libgdal-core + remove dependency on pyproj (#452)

Author: theroggy

CHANGES.md

@@ -2,6 +2,15 @@
 
 ## 0.9.1 (yyyy-mm-dd)
 
+### Packaging
+
+-   For the conda-forge package, change the dependency from `libgdal` to
+    `libgdal-core`. This package is significantly smaller as it doesn't contain
+    some large GDAL plugins. Extra plugins can be installed as seperate conda
+    packages if needed: more info [here](https://gdal.org/download.html#conda).
+    This also leads to `pyproj` becoming an optional dependency; you will need
+    to install `pyproj` in order to support spatial reference systems (#452).
+
 ### Bug fixes
 
 -   Silence warning from `write_dataframe` with `GeoSeries.notna()` (#435).

ci/envs/latest.yml

@@ -3,9 +3,8 @@ channels:
   - conda-forge
 dependencies:
   - numpy
-  - libgdal
+  - libgdal-core
   - pytest
   - shapely>=2
   - geopandas-base
   - pyarrow
-

ci/envs/minimal.yml

@@ -3,5 +3,5 @@ channels:
   - conda-forge
 dependencies:
   - numpy
-  - libgdal
+  - libgdal-core
   - pytest

ci/envs/nightly-deps.yml

@@ -2,7 +2,7 @@ name: test
 channels:
   - conda-forge
 dependencies:
-  - libgdal
+  - libgdal-core
   - pytest
   - geopandas-base
   - pip

environment-dev.yml

@@ -2,14 +2,17 @@ name: pyogrio-dev
 channels:
   - conda-forge
 dependencies:
+  # Required
   - numpy
+  - libgdal-core
+  - shapely>=2
+  # Optional
   - geopandas-base
-  - libgdal==3.8.3
+  - pyproj
   - pyarrow
-  - shapely>=2
   # Specific for dev
   - cython
   - pre-commit
   - pytest
-  - versioneer
   - ruff
+  - versioneer

pyogrio/_compat.py

@@ -8,6 +8,11 @@
 except ImportError:
     pyarrow = None
 
+try:
+    import pyproj
+except ImportError:
+    pyproj = None
+
 try:
     import shapely
 except ImportError:
@@ -27,6 +32,7 @@
 HAS_ARROW_API = __gdal_version__ >= (3, 6, 0)
 HAS_ARROW_WRITE_API = __gdal_version__ >= (3, 8, 0)
 HAS_PYARROW = pyarrow is not None
+HAS_PYPROJ = pyproj is not None
 
 HAS_GEOPANDAS = geopandas is not None
 

pyogrio/geopandas.py

@@ -447,7 +447,6 @@ def write_dataframe(
 
     from geopandas.array import to_wkb
     import pandas as pd
-    from pyproj.enums import WktVersion  # if geopandas is available so is pyproj
 
     if not isinstance(df, pd.DataFrame):
         raise ValueError("'df' must be a DataFrame or GeoDataFrame")
@@ -579,7 +578,7 @@ def write_dataframe(
         if epsg:
             crs = f"EPSG:{epsg}"  # noqa: E231
         else:
-            crs = geometry.crs.to_wkt(WktVersion.WKT1_GDAL)
+            crs = geometry.crs.to_wkt("WKT1_GDAL")
 
     if use_arrow:
         import pyarrow as pa

pyogrio/tests/conftest.py

@@ -14,6 +14,7 @@
     HAS_ARROW_WRITE_API,
     HAS_GDAL_GEOS,
     HAS_PYARROW,
+    HAS_PYPROJ,
     HAS_SHAPELY,
 )
 from pyogrio.raw import read, write
@@ -54,6 +55,8 @@ def pytest_report_header(config):
     not HAS_ARROW_API or not HAS_PYARROW, reason="GDAL>=3.6 and pyarrow required"
 )
 
+requires_pyproj = pytest.mark.skipif(not HAS_PYPROJ, reason="pyproj required")
+
 requires_arrow_write_api = pytest.mark.skipif(
     not HAS_ARROW_WRITE_API or not HAS_PYARROW,
     reason="GDAL>=3.8 required for Arrow write API",

pyogrio/tests/test_geopandas_io.py

@@ -20,9 +20,10 @@
     DRIVERS,
     requires_pyarrow_api,
     requires_arrow_write_api,
+    requires_pyproj,
     requires_gdal_geos,
 )
-from pyogrio._compat import PANDAS_GE_15, HAS_ARROW_WRITE_API
+from pyogrio._compat import HAS_PYPROJ, PANDAS_GE_15, HAS_ARROW_WRITE_API
 
 try:
     import pandas as pd
@@ -126,7 +127,8 @@ def test_read_csv_platform_encoding(tmp_path):
 def test_read_dataframe(naturalearth_lowres_all_ext):
     df = read_dataframe(naturalearth_lowres_all_ext)
 
-    assert df.crs == "EPSG:4326"
+    if HAS_PYPROJ:
+        assert df.crs == "EPSG:4326"
     assert len(df) == 177
     assert df.columns.tolist() == [
         "pop_est",
@@ -1071,6 +1073,8 @@ def test_write_empty_geometry(tmp_path):
     # Check that no warning is raised with GeoSeries.notna()
     with warnings.catch_warnings():
         warnings.simplefilter("error", UserWarning)
+        if not HAS_PYPROJ:
+            warnings.filterwarnings("ignore", message="'crs' was not provided.")
         write_dataframe(expected, filename)
     assert filename.exists()
 
@@ -1460,6 +1464,7 @@ def test_write_dataframe_infer_geometry_with_nulls(tmp_path, geoms, ext, use_arr
     "ignore: You will likely lose important projection information"
 )
 @pytest.mark.requires_arrow_write_api
+@requires_pyproj
 def test_custom_crs_io(tmp_path, naturalearth_lowres_all_ext, use_arrow):
     df = read_dataframe(naturalearth_lowres_all_ext)
     # project Belgium to a custom Albers Equal Area projection

pyogrio/tests/test_path.py

@@ -6,6 +6,7 @@
 import pytest
 
 import pyogrio
+from pyogrio._compat import HAS_PYPROJ
 import pyogrio.raw
 from pyogrio.util import vsi_path, get_vsi_path_or_buffer
 
@@ -238,6 +239,9 @@ def test_detect_zip_path(tmp_path, naturalearth_lowres):
     path = tmp_path / "test.zip"
     with ZipFile(path, mode="w", compression=ZIP_DEFLATED, compresslevel=5) as out:
         for ext in ["dbf", "prj", "shp", "shx"]:
+            if not HAS_PYPROJ and ext == "prj":
+                continue
+
             filename = f"test1.{ext}"
             out.write(tmp_path / filename, filename)