csvjson: Support types other than Point

Parses and removes 'geojson' from the properties,
putting the coordinate into the GeoJSON feature geometry.

Fixes #866

csvkit/utilities/csvjson.py

@@ -92,11 +92,40 @@ def dump_json(data, newline=False):
             )
 
             features = []
+
+            global min_lon, min_lat, max_lon, max_lat
+
             min_lon = None
             min_lat = None
             max_lon = None
             max_lat = None
 
+            def update_boundary_lat(lat):
+                global min_lat, max_lat
+
+                if min_lat is None or lat < min_lat:
+                    min_lat = lat
+                if max_lat is None or lat > max_lat:
+                    max_lat = lat
+
+            def update_boundary_lon(lon):
+                global min_lon, max_lon
+
+                if min_lon is None or lon < min_lon:
+                    min_lon = lon
+                if max_lon is None or lon > max_lon:
+                    max_lon = lon
+
+            def update_boundary_coords(coords):
+                if (isinstance(coords, list) and len(coords) == 2 and
+                        isinstance(coords[0], (float, int)) and
+                        isinstance(coords[1], (float, int))):
+                    update_boundary_lon(coords[0])
+                    update_boundary_lat(coords[1])
+                else:
+                    for coord in coords:
+                        update_boundary_coords(coord)
+
             lat_column = match_column_identifier(table.column_names, self.args.lat, self.args.zero_based)
             lon_column = match_column_identifier(table.column_names, self.args.lon, self.args.zero_based)
 
@@ -110,39 +139,48 @@ def dump_json(data, newline=False):
                 feature['type'] = 'Feature'
                 properties = OrderedDict()
                 geoid = None
+                geo_type = 'Point'
                 lat = None
                 lon = None
+                coords = None
 
                 for i, c in enumerate(row):
                     if i == lat_column:
+                        if not c:
+                            continue
                         try:
                             lat = float(c)
                         except ValueError:
                             lat = None
-                        if min_lat is None or lat < min_lat:
-                            min_lat = lat
-                        if max_lat is None or lat > max_lat:
-                            max_lat = lat
+                        update_boundary_lat(lat)
                     elif i == lon_column:
+                        if not c:
+                            continue
                         try:
                             lon = float(c)
                         except ValueError:
                             lon = None
-                        if min_lon is None or lon < min_lon:
-                            min_lon = lon
-                        if max_lon is None or lon > max_lon:
-                            max_lon = lon
+                        update_boundary_lon(lon)
                     elif i == id_column:
                         geoid = c
+                    elif table.column_names[i] == 'type':
+                        geo_type = c
+                    elif table.column_names[i] == 'geojson':
+                        geojson = json.loads(c)
+                        coords = geojson['coordinates']
+                        update_boundary_coords(coords)
                     elif c is not None:
                         properties[table.column_names[i]] = c
 
                 if id_column is not None:
                     feature['id'] = geoid
 
+                if lon and lat:
+                    coords = [lon, lat]
+
                 feature['geometry'] = OrderedDict([
-                    ('type', 'Point'),
-                    ('coordinates', [lon, lat])
+                    ('type', geo_type),
+                    ('coordinates', coords)
                 ])
                 feature['properties'] = properties
                 features.append(feature)

tests/test_utilities/test_csvjson.py

@@ -62,7 +62,7 @@ def test_duplicate_keys(self):
         self.assertRaisesRegex(ValueError, 'Value True is not unique in the key column.', utility.run)
         output_file.close()
 
-    def test_geojson(self):
+    def test_geojson_point(self):
         geojson = json.loads(self.get_output(['--lat', 'latitude', '--lon', 'longitude', 'examples/test_geo.csv']))
 
         self.assertEqual(geojson['type'], 'FeatureCollection')
@@ -83,6 +83,36 @@ def test_geojson(self):
             self.assertTrue(isinstance(geometry['coordinates'][0], float))
             self.assertTrue(isinstance(geometry['coordinates'][1], float))
 
+    def test_geojson_shape(self):
+        geojson = json.loads(self.get_output(['--lat', 'latitude', '--lon', 'longitude', 'examples/test_geojson.csv']))
+
+        self.assertEqual(geojson['type'], 'FeatureCollection')
+        self.assertFalse('crs' in geojson)
+        self.assertEqual(geojson['bbox'], [100.0, 0.0, 105.0, 1.0])
+        self.assertEqual(len(geojson['features']), 3)
+
+        for feature in geojson['features']:
+            self.assertEqual(feature['type'], 'Feature')
+            self.assertFalse('id' in feature)
+            self.assertIn('properties', feature)
+            self.assertIsInstance(feature['properties'], dict)
+            self.assertIn('prop0', feature['properties'].keys())
+
+            geometry = feature['geometry']
+
+            self.assertIn('coordinates', geometry)
+            self.assertIsNotNone(geometry['coordinates'])
+
+        self.assertEqual(geojson['features'][0]['geometry']['type'], 'Point')
+        self.assertEqual(geojson['features'][1]['geometry']['type'], 'LineString')
+        self.assertEqual(geojson['features'][2]['geometry']['type'], 'Polygon')
+
+        self.assertEqual(geojson['features'][0]['geometry']['coordinates'], [102.0, 0.5])
+        self.assertEqual(geojson['features'][1]['geometry']['coordinates'],
+                         [[102.0, 0.0], [103.0, 1.0], [104.0, 0.0], [105.0, 1.0]])
+        self.assertEqual(geojson['features'][2]['geometry']['coordinates'],
+                         [[[100.0, 0.0], [101.0, 0.0], [101.0, 1.0], [100.0, 1.0], [100.0, 0.0]]])
+
     def test_geojson_with_id(self):
         geojson = json.loads(self.get_output(['--lat', 'latitude', '--lon', 'longitude', '-k', 'slug', 'examples/test_geo.csv']))