Merge pull request #1224 from CartoDB/alasarr/grid_polyfill

Grid polyfill

cartoframes/analysis/grid.py

@@ -0,0 +1,29 @@
+import mercantile
+from shapely.geometry import box
+import pandas as pd
+from cartoframes import CartoDataFrame
+
+
+class QuadGrid():
+
+    def polyfill(self, input_gdf, zoom_level):
+
+        dfs = []
+
+        for _, row in input_gdf.iterrows():
+            bounds = row['geometry'].bounds
+            tiles = mercantile.tiles(bounds[0], bounds[1], bounds[2], bounds[3], zoom_level)
+            resp = []
+            for t in tiles:
+                r = row.copy()
+                geometry = box(*mercantile.bounds(t))
+                if geometry.intersects(row['geometry']):
+                    r['geometry'] = geometry
+                    r['quadkey'] = mercantile.quadkey(t)
+                    resp.append(r)
+
+            dfs.append(pd.DataFrame(resp))
+
+        cdf = CartoDataFrame(pd.concat(dfs).reset_index(drop=True))
+        cdf.crs = 'epsg:4326'
+        return cdf

cartoframes/data/observatory/catalog/repository/repo_client.py

@@ -27,7 +27,7 @@ def get_categories(self, filters=None):
 
     def get_categories_joined_datasets(self, filters=None):
         query = 'SELECT DISTINCT c.* FROM categories_public c, datasets_public t'
-        return self._run_query(query,  filters, ['c.id = t.category_id'])
+        return self._run_query(query, filters, ['c.id = t.category_id'])
 
     def get_providers(self, filters=None):
         query = 'SELECT t.* FROM providers_public t'

setup.py

@@ -32,6 +32,7 @@ def get_version():
     'unidecode>=1.1.0,<2.0',
     'pyarrow>=0.14.1,<1.0',
     'google-cloud-bigquery>=1.19.0,<2.0',
+    'mercantile>=1.1.2,<2.0'
     # 'Rtree>=0.8.3,<1.0'
 ]
 

tests/unit/analysis/test_grid.py

@@ -0,0 +1,95 @@
+# """Unit tests for cartoframes.analysis.grid"""
+
+import pandas as pd
+import numpy as np
+import os
+
+from shapely.geometry import box, shape
+from shapely import wkt
+from cartoframes.analysis.grid import QuadGrid
+from cartoframes import CartoDataFrame
+
+import geopandas as gpd
+from geopandas.testing import assert_geodataframe_equal
+
+# DATA FRAME SRC BBOX
+pol_1 = box(1, 1, 2, 2)
+pol_2 = box(3, 3, 4, 4)
+GDF_BOX = gpd.GeoDataFrame({'id': [1, 2], 'geometry': [pol_1, pol_2]}, columns=['id', 'geometry'])
+
+pol_geojson = {
+    "type": "Polygon",
+    "coordinates": [
+        [
+            [
+                -5.899658203125,
+                38.436379603
+            ],
+            [
+                -6.690673828125,
+                37.67512527892127
+            ],
+            [
+                -6.15234375,
+                37.43997405227057
+            ],
+            [
+                -5.8447265625,
+                37.70120736474139
+            ],
+            [
+                -6.13037109375,
+                37.82280243352756
+            ],
+            [
+                -5.877685546874999,
+                38.02213147353745
+            ],
+            [
+                -6.009521484375,
+                38.12591462924157
+            ],
+            [
+                -5.5810546875,
+                38.1777509666256
+            ],
+            [
+                -5.899658203125,
+                38.436379603
+            ]
+        ]
+    ]
+}
+
+GDF_IRREGULAR = gpd.GeoDataFrame({'id': [1], 'geometry': [shape(pol_geojson)]}, columns=['id', 'geometry'])
+
+BASE_FOLDER = os.path.dirname(os.path.abspath(__file__))
+
+
+class TestGrid(object):
+
+    def _load_test_gdf(self, fname):
+        fname = os.path.join(BASE_FOLDER, fname)
+        df = pd.read_csv(fname, dtype={'id': np.int64, 'geometry': object, 'quadkey': object})
+        df['geometry'] = df['geometry'].apply(wkt.loads)
+        cdf_test = CartoDataFrame(df)
+        cdf_test.crs = 'epsg:4326'
+        return cdf_test
+
+    def test_quadgrid_polyfill_box(self, mocker):
+        """cartoframes.analysis.grid.QuadGrid.polyfill"""
+        cdf = QuadGrid().polyfill(GDF_BOX, 12)
+        assert isinstance(cdf, CartoDataFrame)
+
+        # Check both dataframes are equals
+        cdf_test = self._load_test_gdf('grid_quadkey_bbox.csv')
+        assert_geodataframe_equal(cdf, cdf_test, check_less_precise=True)
+
+    def test_quadgrid_polyfill_pol(self, mocker):
+        """cartoframes.analysis.grid.QuadGrid.polyfill"""
+        cdf = QuadGrid().polyfill(GDF_IRREGULAR, 12)
+        assert isinstance(cdf, CartoDataFrame)
+
+        # Check both dataframes are equals
+        cdf_test = self._load_test_gdf('grid_quadkey_pol.csv')
+        assert_geodataframe_equal(cdf, cdf_test, check_less_precise=True)