Merge pull request #22 from aaronspring/AS_intake_regionmask

intake_geopandas.regionmask

.travis.yml

@@ -20,7 +20,7 @@ script:
 - |
   flake8 .
   if [ "$TRAVIS_OS_NAME" = "linux" ]; then
-      travis_wait 30 conda build -c defaults -c conda-forge ./conda
+      travis_wait 40 conda build -c defaults -c conda-forge ./conda
   else
       # Workaround for Travis-CI bug #2: https://github.com/travis-ci/travis-ci/issues/7773
       conda build -c defaults -c conda-forge --no-test ./conda

intake_geopandas/__init__.py

@@ -1,18 +1,18 @@
 # -*- coding: utf-8 -*-
+import intake
+
 from ._version import get_versions
+from .geopandas import GeoJSONSource, PostGISSource, ShapefileSource, SpatiaLiteSource
+from .regionmask import RegionmaskSource
+
 __version__ = get_versions()['version']
 del get_versions
 
-import intake
-from .geopandas import (
-    GeoJSONSource,
-    PostGISSource,
-    ShapefileSource,
-    SpatiaLiteSource
-)
+
 __all__ = [
     'GeoJSONSource',
     'PostGISSource',
     'ShapefileSource',
-    'SpatiaLiteSource'
+    'SpatiaLiteSource',
+    'RegionmaskSource',
 ]

intake_geopandas/geopandas.py

@@ -5,13 +5,17 @@
 import geopandas
 from intake.source.base import DataSource, Schema
 
-from . import __version__
+from ._version import get_versions
+
+__version__ = get_versions()['version']
+del get_versions
 
 
 class GeoPandasSource(DataSource, ABC):
     """
     Base class intake source for loading GeoDataFrames.
     """
+
     version = __version__
     container = 'dataframe'
     partition_access = True
@@ -29,11 +33,13 @@ def _get_schema(self):
 
         dtypes = self._dataframe.dtypes.to_dict()
         dtypes = {n: str(t) for (n, t) in dtypes.items()}
-        return Schema(datashape=None,
-                      dtype=dtypes,
-                      shape=(None, len(dtypes)),
-                      npartitions=1,
-                      extra_metadata={})
+        return Schema(
+            datashape=None,
+            dtype=dtypes,
+            shape=(None, len(dtypes)),
+            npartitions=1,
+            extra_metadata={},
+        )
 
     def _get_partition(self, i):
         self._get_schema()
@@ -73,7 +79,8 @@ def __init__(
             Whether to use fsspec to open `urlpath`. By default, `urlpath` is passed
             directly to GeoPandas, which opens the file using `fiona`. However, for some
             use cases it may be beneficial to read the file using `fsspec` before
-            passing the resulting bytes to GeoPandas (e.g., when using `fsspec` caching).
+            passing the resulting bytes to GeoPandas (e.g., when using `fsspec`
+            caching).
             Note that fiona/GDAL and `fsspec` have mutually-incompatible URL chaining
             syntaxes, so the URLs passed to each may be significantly different.
 
@@ -90,11 +97,10 @@ def __init__(
         """
         self.urlpath = urlpath
         self._use_fsspec = use_fsspec
-        self._storage_options = storage_options or {}
+        self.storage_options = storage_options or {}
         self._bbox = bbox
         self._geopandas_kwargs = geopandas_kwargs or {}
         self._dataframe = None
-        self.storage_options = storage_options or {}
 
         super().__init__(metadata=metadata)
 
@@ -103,53 +109,50 @@ def _open_dataset(self):
         Open dataset using geopandas.
         """
         if self._use_fsspec:
-            with fsspec.open_files(self.urlpath, **self._storage_options) as f:
+            with fsspec.open_files(self.urlpath, **self.storage_options) as f:
                 f = self._resolve_single_file(f) if len(f) > 1 else f[0]
                 self._dataframe = geopandas.read_file(
-                    f,
-                    bbox=self._bbox,
-                    **self._geopandas_kwargs,
+                    f, bbox=self._bbox, **self._geopandas_kwargs,
                 )
         else:
             self._dataframe = geopandas.read_file(
-                self.urlpath,
-                bbox=self._bbox,
-                **self._geopandas_kwargs
+                self.urlpath, bbox=self._bbox, **self._geopandas_kwargs
             )
 
     def _resolve_single_file(self, filelist):
         """
         Given a list of fsspec OpenFiles, choose one to pass to geopandas.
         """
         raise NotImplementedError(
-            "Opening multiple files is not supported by this driver"
+            'Opening multiple files is not supported by this driver'
         )
 
 
 class GeoJSONSource(GeoPandasFileSource):
-    name = "geojson"
+    name = 'geojson'
 
 
 class ShapefileSource(GeoPandasFileSource):
-    name = "shapefile"
+    name = 'shapefile'
 
     def _resolve_single_file(self, filelist):
         """
         Given a list of fsspec OpenFiles, find a .shp file.
         """
         local_files = fsspec.open_local(self.urlpath, **self.storage_options)
         for f in local_files:
-            if f.endswith(".shp"):
+            if f.endswith('.shp'):
                 return f
         raise ValueError(
-            f"No shapefile found in {filelist}, if you are using fsspec caching"
-            " consider using same_names=True"
+            f'No shapefile found in {filelist}, if you are using fsspec caching'
+            ' consider using same_names=True'
         )
 
 
 class GeoPandasSQLSource(GeoPandasSource):
-    def __init__(self, uri, sql_expr=None, table=None,
-                 geopandas_kwargs=None, metadata=None):
+    def __init__(
+        self, uri, sql_expr=None, table=None, geopandas_kwargs=None, metadata=None
+    ):
         """
         Parameters
         ----------
@@ -168,9 +171,9 @@ def __init__(self, uri, sql_expr=None, table=None,
         if sql_expr:
             self.sql_expr = sql_expr
         elif table:
-            self.sql_expr = f"SELECT * FROM {table}"
+            self.sql_expr = f'SELECT * FROM {table}'
         else:
-            raise ValueError("Must provide either a sql_expr or a table")
+            raise ValueError('Must provide either a sql_expr or a table')
 
         self._geopandas_kwargs = geopandas_kwargs or {}
         self._dataframe = None
@@ -179,12 +182,13 @@ def __init__(self, uri, sql_expr=None, table=None,
 
     def _open_dataset(self):
         self._dataframe = geopandas.read_postgis(
-            self.sql_expr, self.uri, **self._geopandas_kwargs)
+            self.sql_expr, self.uri, **self._geopandas_kwargs
+        )
 
 
 class PostGISSource(GeoPandasSQLSource):
-    name = "postgis"
+    name = 'postgis'
 
 
 class SpatiaLiteSource(GeoPandasSQLSource):
-    name = "spatialite"
+    name = 'spatialite'

intake_geopandas/regionmask.py

@@ -0,0 +1,78 @@
+from intake.source.base import Schema
+
+from .geopandas import GeoPandasFileSource
+
+
+class RegionmaskSource(GeoPandasFileSource):
+    name = 'regionmask'
+
+    def __init__(
+        self,
+        urlpath,
+        use_fsspec=False,
+        storage_options=None,
+        bbox=None,
+        geopandas_kwargs=None,
+        metadata=None,
+        regionmask_kwargs=None,
+    ):
+        """
+        urlpath : str or iterable, location of data
+            Either the absolute or relative path to the file or URL to be
+            opened. Some examples:
+            - ``{{ CATALOG_DIR }}data/states.shp``
+            - ``http://some.domain.com/data/states.geo.json``
+        use_fsspec: bool
+            Whether to use fsspec to open `urlpath`. By default, `urlpath` is passed
+            directly to GeoPandas, which opens the file using `fiona`. However, for some
+            use cases it may be beneficial to read the file using `fsspec` before
+            passing the resulting bytes to GeoPandas (e.g., when using `fsspec`
+            caching).
+            Note that fiona/GDAL and `fsspec` have mutually-incompatible URL chaining
+            syntaxes, so the URLs passed to each may be significantly different.
+        storage_options: dict
+            Storage options to pass to fsspec when opening. Only used when
+            `use_fsspec=True`.
+        bbox : tuple | GeoDataFrame or GeoSeries, default None
+            Filter features by given bounding box, GeoSeries, or GeoDataFrame.
+            CRS mis-matches are resolved if given a GeoSeries or GeoDataFrame.
+        geopandas_kwargs : dict
+            Any further arguments to pass to geopandas's read_file function.
+        regionmask_kwargs : dict
+            Any further arguments to pass to regionmask.from_geopandas.
+        """
+        self._regionmask_kwargs = regionmask_kwargs or {}
+
+        super().__init__(
+            urlpath=urlpath,
+            metadata=metadata,
+            use_fsspec=use_fsspec,
+            storage_options=storage_options,
+            geopandas_kwargs=geopandas_kwargs,
+            bbox=bbox,
+        )
+
+    def _open_dataset(self):
+        try:
+            import regionmask
+        except ImportError:
+            raise ImportError('please install regionmask')
+        super()._open_dataset()
+        self._dtypes = self._dataframe.dtypes.to_dict()
+        self._dtypes = {n: str(t) for (n, t) in self._dtypes.items()}
+        self._dataframe = regionmask.from_geopandas(
+            self._dataframe, **self._regionmask_kwargs
+        )
+
+    def _get_schema(self):
+
+        if self._dataframe is None:
+            self._open_dataset()
+
+        return Schema(
+            datashape=None,
+            dtype=self._dtypes,
+            shape=(None, len(self._dtypes)),
+            npartitions=1,
+            extra_metadata={},
+        )

setup.py

@@ -1,6 +1,7 @@
 # -*- coding: utf-8 -*-
 
-from setuptools import setup, find_packages
+from setuptools import find_packages, setup
+
 import versioneer
 
 requires = open('requirements.txt').read().strip().split('\n')
@@ -10,7 +11,7 @@
     version=versioneer.get_version(),
     cmdclass=versioneer.get_cmdclass(),
     description='Geopandas plugin for Intake',
-    url='https://github.com/informatics-lab/intake_geopandas',
+    url='https://github.com/intake/intake_geopandas',
     maintainer='Jacob Tomlinson',
     maintainer_email='jacob.tomlinson@informaticslab.co.uk',
     license='BSD',
@@ -23,10 +24,13 @@
             'postgis = intake_geopandas.geopandas:PostGISSource',
             'shapefile = intake_geopandas.geopandas:ShapefileSource',
             'spatialite = intake_geopandas.geopandas:SpatiaLiteSource',
+            'regionmask = intake_geopandas.regionmask:RegionmaskSource',
         ]
     },
     include_package_data=True,
     install_requires=requires,
+    extras_require={'':['regionmask']},
     long_description_content_type='text/markdown',
     long_description=open('README.md').read(),
-    zip_safe=False, )
+    zip_safe=False,
+)

tests/data/shape.catalog.yaml

@@ -9,10 +9,39 @@ sources:
       urlpath: '{CATALOG_PATH}/stations/stations.shp'
 
   MEOW:
-      description: MEOW
-      driver: intake_geopandas.geopandas.ShapefileSource
-      args:
-        urlpath: http://maps.tnc.org/files/shp/MEOW-TNC.zip
+    description: MEOW
+    driver: intake_geopandas.geopandas.ShapefileSource
+    args:
+      urlpath: http://maps.tnc.org/files/shp/MEOW-TNC.zip
+
+  MEOW_regionmask:
+    description: MEOW for regionmask
+    driver: intake_geopandas.regionmask.RegionmaskSource
+    args:
+      urlpath: http://maps.tnc.org/files/shp/MEOW-TNC.zip
+      regionmask_kwargs:
+        names: ECOREGION
+        abbrevs: _from_name
+        source: http://maps.tnc.org
+        numbers: ECO_CODE_X
+        name: MEOW
+
+  MEOW_regionmask_cache:
+    description: MEOW for regionmask and cache
+    driver: intake_geopandas.regionmask.RegionmaskSource
+    args:
+      urlpath: simplecache::http://maps.tnc.org/files/shp/MEOW-TNC.zip
+      use_fsspec: true
+      storage_options:
+        simplecache:
+          same_names: true
+          cache_storage: tmp/intake_geopandas
+      regionmask_kwargs:
+        names: ECOREGION
+        abbrevs: _from_name
+        source: http://maps.tnc.org
+        numbers: ECO_CODE_X
+        name: MEOW
 
 
   MEOW_simplecache:
@@ -32,6 +61,7 @@ sources:
     driver: intake_geopandas.geopandas.ShapefileSource
     args:
       urlpath: simplecache::zip://gadm36_ALA_0*::https://biogeo.ucdavis.edu/data/gadm3.6/shp/gadm36_ALA_shp.zip
+      use_fsspec: true
       storage_options:
         simplecache:
           same_names: true