Handle geopandas and shapely geometries via geo_interface link

.github/workflows/ci_tests.yaml

@@ -45,11 +45,14 @@ jobs:
           #   python-version: 3.7
           #   isDraft: true
         # Pair Python 3.7 with NumPy 1.17 and Python 3.9 with NumPy 1.20
+        # Only install geopandas on Python 3.9/NumPy 1.20
         include:
           - python-version: 3.7
             numpy-version: '1.17'
+            geopandas: ''
           - python-version: 3.9
             numpy-version: '1.20'
+            geopandas: 'geopandas'
     defaults:
       run:
         shell: bash -l {0}
@@ -87,7 +90,7 @@ jobs:
       - name: Install dependencies
         run: |
           conda install gmt=6.1.1 numpy=${{ matrix.numpy-version }} \
-                        pandas xarray netCDF4 packaging \
+                        pandas xarray netCDF4 packaging ${{ matrix.geopandas }} \
                         codecov coverage[toml] dvc ipython make \
                         pytest-cov pytest-mpl pytest>=6.0 \
                         sphinx-gallery

MAINTENANCE.md

@@ -78,9 +78,12 @@ There are 9 configuration files located in `.github/workflows`:
 
    This is run on every commit to the *master* and Pull Request branches.
    It is also scheduled to run daily on the *master* branch.
-   In draft Pull Requests, only two jobs on Linux (minimum NEP29 Python/NumPy versions
-   and latest Python/NumPy versions) are triggered to save on Continuous Integration
-   resources.
+   In draft Pull Requests, only two jobs on Linux are triggered to save on
+   Continuous Integration resources:
+
+   - Minimum [NEP29](https://numpy.org/neps/nep-0029-deprecation_policy)
+     Python/NumPy versions
+   - Latest Python/NumPy versions + GeoPandas
 
 3. `ci_docs.yml` (Build documentation on Linux/macOS/Windows)
 

doc/conf.py

@@ -48,6 +48,7 @@
 # intersphinx configuration
 intersphinx_mapping = {
     "python": ("https://docs.python.org/3/", None),
+    "geopandas": ("https://geopandas.org/", None),
     "numpy": ("https://numpy.org/doc/stable/", None),
     "pandas": ("https://pandas.pydata.org/pandas-docs/stable/", None),
     "xarray": ("https://xarray.pydata.org/en/stable/", None),

doc/install.rst

@@ -86,9 +86,10 @@ PyGMT requires the following libraries to be installed:
 * `netCDF4 <https://unidata.github.io/netcdf4-python>`__
 * `packaging <https://packaging.pypa.io>`__
 
-The following are optional (but recommended) dependencies:
+The following are optional dependencies:
 
-* `IPython <https://ipython.org>`__: For embedding the figures in Jupyter notebooks.
+* `IPython <https://ipython.org>`__: For embedding the figures in Jupyter notebooks (recommended).
+* `GeoPandas <https://geopandas.org>`__: For using and plotting GeoDataFrame objects.
 
 Installing GMT and other dependencies
 -------------------------------------

pygmt/clib/session.py

@@ -25,7 +25,7 @@
     GMTInvalidInput,
     GMTVersionError,
 )
-from pygmt.helpers import data_kind, dummy_context
+from pygmt.helpers import data_kind, dummy_context, tempfile_from_geojson
 
 FAMILIES = [
     "GMT_IS_DATASET",
@@ -1417,12 +1417,18 @@ def virtualfile_from_data(
 
         if check_kind == "raster" and kind not in ("file", "grid"):
             raise GMTInvalidInput(f"Unrecognized data type for grid: {type(data)}")
-        if check_kind == "vector" and kind not in ("file", "matrix", "vectors"):
-            raise GMTInvalidInput(f"Unrecognized data type: {type(data)}")
+        if check_kind == "vector" and kind not in (
+            "file",
+            "matrix",
+            "vectors",
+            "geojson",
+        ):
+            raise GMTInvalidInput(f"Unrecognized data type for vector: {type(data)}")
 
         # Decide which virtualfile_from_ function to use
         _virtualfile_from = {
             "file": dummy_context,
+            "geojson": tempfile_from_geojson,
             "grid": self.virtualfile_from_grid,
             # Note: virtualfile_from_matrix is not used because a matrix can be
             # converted to vectors instead, and using vectors allows for better
@@ -1432,7 +1438,7 @@ def virtualfile_from_data(
         }[kind]
 
         # Ensure the data is an iterable (Python list or tuple)
-        if kind in ("file", "grid"):
+        if kind in ("file", "geojson", "grid"):
             _data = (data,)
         elif kind == "vectors":
             _data = [np.atleast_1d(x), np.atleast_1d(y)]

pygmt/helpers/__init__.py

@@ -2,7 +2,7 @@
 Functions, classes, decorators, and context managers to help wrap GMT modules.
 """
 from pygmt.helpers.decorators import fmt_docstring, kwargs_to_strings, use_alias
-from pygmt.helpers.tempfile import GMTTempFile, unique_name
+from pygmt.helpers.tempfile import GMTTempFile, tempfile_from_geojson, unique_name
 from pygmt.helpers.utils import (
     args_in_kwargs,
     build_arg_string,

pygmt/helpers/tempfile.py

@@ -3,6 +3,7 @@
 """
 import os
 import uuid
+from contextlib import contextmanager
 from tempfile import NamedTemporaryFile
 
 import numpy as np
@@ -104,3 +105,54 @@ def loadtxt(self, **kwargs):
             Data read from the text file.
         """
         return np.loadtxt(self.name, **kwargs)
+
+
+@contextmanager
+def tempfile_from_geojson(geojson):
+    """
+    Saves any geo-like Python object which implements ``__geo_interface__``
+    (e.g. a geopandas GeoDataFrame) to a temporary OGR_GMT text file.
+
+    Parameters
+    ----------
+    geojson : geopandas.GeoDataFrame
+        A geopandas GeoDataFrame, or any geo-like Python object which
+        implements __geo_interface__, i.e. a GeoJSON
+
+    Yields
+    ------
+    tmpfilename : str
+        A temporary OGR_GMT format file holding the geographical data.
+        E.g. 'track-1a2b3c4.tsv'.
+    """
+    with GMTTempFile(suffix=".gmt") as tmpfile:
+        os.remove(tmpfile.name)  # ensure file is deleted first
+        ogrgmt_kwargs = dict(filename=tmpfile.name, driver="OGR_GMT", mode="w")
+        try:
+            # Using geopandas.to_file to directly export to OGR_GMT format
+            geojson.to_file(**ogrgmt_kwargs)
+        except AttributeError:
+            # pylint: disable=import-outside-toplevel
+            # Other 'geo' formats which implement __geo_interface__
+            import json
+
+            import fiona
+            import geopandas as gpd
+
+            with fiona.Env():
+                jsontext = json.dumps(geojson.__geo_interface__)
+                # Do Input/Output via Fiona virtual memory
+                with fiona.io.MemoryFile(file_or_bytes=jsontext.encode()) as memfile:
+                    geoseries = gpd.GeoSeries.from_file(filename=memfile)
+                    geoseries.to_file(**ogrgmt_kwargs)
+
+                #     with memfile.open(driver="GeoJSON") as collection:
+                #         # Get schema from GeoJSON
+                #         schema = collection.schema
+                # # Write to temporary OGR_GMT format file
+                # with fiona.open(
+                #    fp=tmpfile.name, mode="w", driver="OGR_GMT", schema=schema
+                # ) as ogrgmtfile:
+                #     ogrgmtfile.write(geojson.__geo_interface__)
+
+        yield tmpfile.name

pygmt/helpers/utils.py

@@ -67,6 +67,8 @@ def data_kind(data, x=None, y=None, z=None):
         kind = "file"
     elif isinstance(data, xr.DataArray):
         kind = "grid"
+    elif hasattr(data, "__geo_interface__"):
+        kind = "geojson"
     elif data is not None:
         kind = "matrix"
     else:

pygmt/src/info.py

@@ -46,7 +46,8 @@ def info(table, **kwargs):
 
     Parameters
     ----------
-    table : str or np.ndarray or pandas.DataFrame or xarray.Dataset
+    table : str or numpy.ndarray or pandas.DataFrame or xarray.Dataset or
+            geopandas.GeoDataFrame
         Pass in either a file name to an ASCII data table, a 1D/2D numpy array,
         a pandas dataframe, or an xarray dataset made up of 1D xarray.DataArray
         data variables.

pygmt/tests/test_geopandas.py

@@ -0,0 +1,66 @@
+"""
+Tests on integration with geopandas.
+"""
+import numpy.testing as npt
+import pytest
+from pygmt import info
+
+gpd = pytest.importorskip("geopandas")
+shapely = pytest.importorskip("shapely")
+
+
+@pytest.fixture(scope="module", name="gdf")
+def fixture_gdf():
+    """
+    Create a sample geopandas GeoDataFrame object with shapely geometries of
+    different types.
+    """
+    linestring = shapely.geometry.LineString([(20, 15), (30, 15)])
+    polygon = shapely.geometry.Polygon([(20, 10), (23, 10), (23, 14), (20, 14)])
+    multipolygon = shapely.geometry.shape(
+        {
+            "type": "MultiPolygon",
+            "coordinates": [
+                [
+                    [[0, 0], [20, 0], [10, 20], [0, 0]],  # Counter-clockwise
+                    [[3, 2], [10, 16], [17, 2], [3, 2]],  # Clockwise
+                ],
+                [[[6, 4], [14, 4], [10, 12], [6, 4]]],  # Counter-clockwise
+                [[[25, 5], [30, 10], [35, 5], [25, 5]]],
+            ],
+        }
+    )
+    # Multipolygon first so the OGR_GMT file has @GMULTIPOLYGON in the header
+    gdf = gpd.GeoDataFrame(
+        index=["multipolygon", "polygon", "linestring"],
+        geometry=[multipolygon, polygon, linestring],
+    )
+
+    return gdf
+
+
+def test_geopandas_info_geodataframe(gdf):
+    """
+    Check that info can return the bounding box region from a
+    geopandas.GeoDataFrame.
+    """
+    output = info(table=gdf, per_column=True)
+    npt.assert_allclose(actual=output, desired=[0.0, 35.0, 0.0, 20.0])
+
+
+@pytest.mark.parametrize(
+    "geomtype,desired",
+    [
+        ("multipolygon", [0.0, 35.0, 0.0, 20.0]),
+        ("polygon", [20.0, 23.0, 10.0, 14.0]),
+        ("linestring", [20.0, 30.0, 15.0, 15.0]),
+    ],
+)
+def test_geopandas_info_shapely(gdf, geomtype, desired):
+    """
+    Check that info can return the bounding box region from a shapely.geometry
+    object that has a __geo_interface__ property.
+    """
+    geom = gdf.loc[geomtype].geometry
+    output = info(table=geom, per_column=True)
+    npt.assert_allclose(actual=output, desired=desired)