Geojson ids option (#220)

tests/test_extract.py

@@ -1,4 +1,5 @@
 import json
+import pytest
 from topojson.core.extract import Extract
 from shapely import geometry
 import geopandas
@@ -469,6 +470,36 @@ def test_extract_keep_properties():
     assert topo["objects"]["feature_1"]["properties"]["name"]["def"] == "ghi"
 
 
+def test_extract_geojson_keep_index():
+    feat_1 = Feature(
+        id="custom_index",
+        geometry=Polygon([[[0, 0], [1, 0], [1, 1], [0, 1], [0, 0]]]),
+    )
+    feat_2 = Feature(
+        geometry=Polygon([[[1, 0], [2, 0], [2, 1], [1, 1], [1, 0]]]),
+    )
+    data = FeatureCollection([feat_1, feat_2])
+    topo = Extract(data).to_dict()
+    objects = topo["objects"]
+
+    assert bool(objects.get("custom_index")) == True
+    assert bool(objects.get("feature_1")) == True
+
+
+def test_extract_geojson_keep_index_duplicates():
+    feat_1 = Feature(
+        id="duplicate_id",
+        geometry=Polygon([[[0, 0], [1, 0], [1, 1], [0, 1], [0, 0]]]),
+    )
+    feat_2 = Feature(
+        id="duplicate_id",
+        geometry=Polygon([[[1, 0], [2, 0], [2, 1], [1, 1], [1, 0]]]),
+    )
+    data = FeatureCollection([feat_1,feat_2])
+    with pytest.raises(IndexError):
+        Extract(data)
+
+
 # why cannot load geojson file using json module?
 def test_extract_read_geojson_from_json_dict():
     with open("tests/files_geojson/naturalearth_lowres.geojson") as f:

tests/test_topology.py

@@ -96,7 +96,7 @@ def test_topology_winding_order_TopoOptions():
     topo = topojson.Topology(data, winding_order="CW_CCW").to_dict(options=True)
 
     assert len(topo["objects"]) == 1
-    assert len(topo["options"]) == 11
+    assert len(topo["options"]) == 12
 
 
 # test winding order using kwarg variables
@@ -106,7 +106,7 @@ def test_topology_winding_order_kwarg_vars():
     topo = topojson.Topology(data, winding_order="CW_CCW").to_dict(options=True)
 
     assert len(topo["objects"]) == 1
-    assert len(topo["options"]) == 11
+    assert len(topo["options"]) == 12
 
 
 def test_topology_computing_topology():
@@ -699,3 +699,23 @@ def test_topology_write_multiple_object_json_dict():
     topo_dict = topo.to_dict()
 
     assert len(topo_dict["objects"]) == 2
+
+def test_topology_ignore_index_true_geojson():
+
+    from geojson import Feature, FeatureCollection, Polygon
+    feat_1 = Feature(
+        id="duplicate_id",
+        geometry=Polygon([[[0, 0], [1, 0], [1, 1], [0, 1], [0, 0]]]),
+    )
+    feat_2 = Feature(
+        id="duplicate_id",
+        geometry=Polygon([[[1, 0], [2, 0], [2, 1], [1, 1], [1, 0]]]),
+    )
+    fc = FeatureCollection([feat_1,feat_2])
+
+    # Using ignore_index to use default feature ids.
+    topo = topojson.Topology(fc, ignore_index=True).to_dict(options=True)
+    geom = topo["objects"]["data"]["geometries"]
+
+    index = [obj["id"] for obj in geom]
+    assert index == ["feature_0","feature_1"]

topojson/core/extract.py

@@ -460,9 +460,19 @@ def _extract_featurecollection(self, geom):
 
             if feature["type"] == "GeometryCollection":
                 feature_dict["geometries"] = feature["geometry"]["geometries"]
-            data[
-                "feature_{}".format(str(idx).zfill(zfill_value))
-            ] = feature_dict  # feature
+
+            if self.options.ignore_index or not feature.get("id"):
+                data[
+                    "feature_{}".format(str(idx).zfill(zfill_value))
+                ] = feature_dict
+            else:
+                data[feature.get("id")] = feature_dict  # feature
+
+        # check for overwritten duplicate keys
+        if len(data) < len(obj["features"]):
+            msg = "index in data duplicated, use `ignore_index=True` to overwrite index"
+            raise IndexError(msg)
+
         # new data dictionary is created, throw the geometries back to main()
         self._is_single = False
         self._extractor(data)

topojson/core/topology.py

@@ -98,6 +98,11 @@ class Topology(Hashmap):
         case it is required to provide a list of the referenced `object_name` in
         combination with an equal length list of `data` objects.
         Default is a single object named `data`.
+    ignore_index : bool
+        If set to true existing ids/indexes of geojson FeatureCollections will be
+        ignored and overwritten. Otherwise features with ids will use their existing one.
+        If indexes are not ignored and a duplicate id exists an exception will be raised.
+        Default is false.
     """
 
     def __init__(
@@ -114,6 +119,7 @@ def __init__(
         simplify_algorithm="dp",
         winding_order="CW_CCW",
         object_name="data",
+        ignore_index=False,
     ):
         options = TopoOptions(locals())
 

topojson/utils.py

@@ -27,6 +27,7 @@ def __init__(
         simplify_algorithm="dp",
         winding_order=None,
         object_name="data",
+        ignore_index=False,
     ):
         # get all arguments
         arguments = locals()
@@ -91,6 +92,11 @@ def __init__(
         else:
             self.object_name = ["data"]
 
+        if "ignore_index" in arguments:
+            self.ignore_index = arguments["ignore_index"]
+        else:
+            self.ignore_index = False
+
     def __repr__(self):
         return "TopoOptions(\n  {}\n)".format(pprint.pformat(self.__dict__))