Do not modify source data, include geojson from file and compute bbox consistently

docs/example/input-types.md

@@ -88,6 +88,61 @@ tp.Topology(gdf, prequantize=False).to_json()
 </div>
 </div>
 
+* * *
+
+## GeoJSON data from file
+A GeoJSON file can be parsed as json dictionary.
+
+<div class="code-example mx-1 bg-example">
+<div class="example-label" markdown="1">
+Example 🔧
+{: .label .label-blue-000 }
+</div>
+<div class="example-text" markdown="1">
+
+```python
+import topojson as tp
+import json
+
+with open("tests/files_topojson/example_data_africa.geojson", 'r') as f:
+    data = json.load(f)
+
+assert data['type']) == 'FeatureCollection'
+topo = topojson.Topology(data)
+
+# to visualise we use the (optional!) package Altair.
+topo.toposimplify(4).to_alt()
+```
+<div id="embed_output_mesh_altair_from_geojson"></div>
+</div>
+</div>
+
+* * * 
+
+## TopoJSON data from file
+A TopoJSON file can be postprocessed using json dictionary.
+
+<div class="code-example mx-1 bg-example">
+<div class="example-label" markdown="1">
+Example 🔧
+{: .label .label-blue-000 }
+</div>
+<div class="example-text" markdown="1">
+
+```python
+import topojson as tp
+import json
+
+with open("tests/files_topojson/naturalearth_lowres_africa.topojson", 'r') as f:
+    data = json.load(f)
+# parse topojson file using `object_name`
+topo = topojson.Topology(data, object_name="data")
+topo.toposimplify(4).to_svg()
+```
+<img src="../images/africa_toposimp.svg">
+</div>
+</div>
+
 * * * 
 
 
@@ -311,28 +366,19 @@ tp.Topology(dict_in, prequantize=False).to_json()
 </div>
 </div>
 
-* * * 
-
-## TopoJSON file loaded as json-dict
-A TopoJSON file can be postprocessed.
-
-<div class="code-example mx-1 bg-example">
-<div class="example-label" markdown="1">
-Example 🔧
-{: .label .label-blue-000 }
-</div>
-<div class="example-text" markdown="1">
-
-```python
-import topojson as tp
-import json
-
-with open("tests/files_topojson/naturalearth_lowres_africa.topojson", 'r') as f:
-    data = json.load(f)
-# parse topojson file using `object_name`
-topo = topojson.Topology(data, object_name="data")
-topo.toposimplify(4).to_svg()
-```
-<img src="../images/africa_toposimp.svg">
-</div>
-</div>
+<script>
+window.addEventListener("DOMContentLoaded", event => {
+    var opt = {
+        mode: "vega-lite",
+        renderer: "svg",
+        actions: false
+    };
+
+    var spec_mesh_altair = "{{site.baseurl}}/json/example_mesh.vl.json";
+    vegaEmbed("#embed_output_mesh_altair_from_geojson", spec_mesh_altair, opt).catch(console.err); 
+
+});
+</script>
+<script type="text/javascript" src="https://cdn.jsdelivr.net/npm/vega@5"></script>
+<script type="text/javascript" src="https://cdn.jsdelivr.net/npm/vega-lite@4"></script>
+<script type="text/javascript" src="https://cdn.jsdelivr.net/npm/vega-embed@6"></script>

tests/test_extract.py

@@ -3,6 +3,7 @@
 from shapely import geometry
 import geopandas
 import geojson
+from geojson import Feature, Polygon, FeatureCollection
 import fiona
 
 
@@ -361,18 +362,20 @@ def test_extract_fiona_file_gpkg():
 
     assert len(topo["bookkeeping_geoms"]) == 4
 
+
 # test to check if original data is not modified
 def test_extract_dict_org_data_untouched():
     data = {
         "foo": {"type": "LineString", "coordinates": [[0, 0], [1, 0], [2, 0]]},
         "bar": {"type": "LineString", "coordinates": [[0, 0], [1, 0], [2, 0]]},
     }
     topo = Extract(data).to_dict()
-    topo_foo = topo['objects']['foo']
+    topo_foo = topo["objects"]["foo"]
     data_foo = data["foo"]
 
-    assert 'arcs' in topo_foo.keys()
-    assert 'arcs' not in data_foo.keys()
+    assert "arcs" in topo_foo.keys()
+    assert "arcs" not in data_foo.keys()
+
 
 # test to check if original data is not modified
 def test_extract_list_org_data_untouched():
@@ -381,41 +384,76 @@ def test_extract_list_org_data_untouched():
         geometry.Polygon([[1, 0], [2, 0], [2, 1], [1, 1], [1, 0]]),
     ]
     topo = Extract(data).to_dict()
-    topo_0 = topo['objects'][0]
+    topo_0 = topo["objects"][0]
     data_0 = data[0]
 
-    assert 'arcs' in topo_0.keys()
-    assert data_0.type == 'Polygon'
+    assert "arcs" in topo_0.keys()
+    assert data_0.type == "Polygon"
+
 
 # test to check if original data is not modified
 def test_extract_gdf_org_data_untouched():
     data = geopandas.read_file(
         "tests/files_geojson/naturalearth_alb_grc.geojson", driver="GeoJSON"
     )
     topo = Extract(data).to_dict()
-    topo_0 = topo['objects'][0]
+    topo_0 = topo["objects"][0]
     data_0 = data.iloc[0]
 
-    assert 'arcs' in topo_0.keys()
-    assert data_0.geometry.type == 'Polygon'
+    assert "arcs" in topo_0.keys()
+    assert data_0.geometry.type == "Polygon"
+
 
 # test to check if original data is not modified
 def test_extract_shapely_org_data_untouched():
     data = geometry.LineString([[0, 0], [1, 0], [1, 1], [0, 1]])
-    topo = Extract(data).to_dict()  
-    topo_0 = topo['objects'][0]
+    topo = Extract(data).to_dict()
+    topo_0 = topo["objects"][0]
+
+    assert "arcs" in topo_0.keys()
+    assert data.type == "LineString"
 
-    assert 'arcs' in topo_0.keys()
-    assert data.type == 'LineString'    
 
 # test to check if original data is not modified
 def test_extract_shapefile_org_data_untouched():
     import shapefile
 
     data = shapefile.Reader("tests/files_shapefile/southamerica.shp")
     topo = Extract(data).to_dict()
-    topo_0 = topo['objects']['feature_00']['geometries'][0]
-    data_0 = data.__geo_interface__['features'][0]['geometry']
+    topo_0 = topo["objects"]["feature_00"]["geometries"][0]
+    data_0 = data.__geo_interface__["features"][0]["geometry"]
+
+    assert "arcs" in topo_0.keys()
+    assert "arcs" not in data_0.keys()
+
+
+# issue 137 do not modify source data
+def test_extract_source_data_modify():
+    # prepare data
+    feat_1 = Feature(
+        geometry=Polygon([[[0, 0], [1, 0], [1, 1], [0, 1], [0, 0]]]),
+        properties={"name": "abc"},
+    )
+    feat_2 = Feature(
+        geometry=Polygon([[[1, 0], [2, 0], [2, 1], [1, 1], [1, 0]]]),
+        properties={"name": "def"},
+    )
+    data = FeatureCollection([feat_1, feat_2])
+
+    # before Topology()
+    assert "geometry" in data["features"][0].keys()
+
+    # apply Topology()
+    topo = Extract(data)
+
+    # after Topology()
+    assert "geometry" in data["features"][0].keys()
+
+
+# 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:
+        data = json.load(f)
+    topo = Extract(data).to_dict()
 
-    assert 'arcs' in topo_0.keys()
-    assert 'arcs' not in data_0.keys()  
+    assert len(topo["linestrings"]) == 287

tests/test_topology.py

@@ -423,58 +423,65 @@ def test_topology_geodataframe_valid():
 
 def test_topology_geojson_duplicates():
 
-    p0 = wkt.loads('POLYGON ((0 0, 0 1, 1 1, 2 1, 2 0, 1 0, 0 0))')
-    p1 = wkt.loads('POLYGON ((0 1, 0 2, 1 2, 1 1, 0 1))')
-    p2 = wkt.loads('POLYGON ((1 0, 2 0, 2 -1, 1 -1, 1 0))')
-    data = geopandas.GeoDataFrame({"name": ["abc", "def", "ghi"], "geometry": [p0, p1, p2]})
+    p0 = wkt.loads("POLYGON ((0 0, 0 1, 1 1, 2 1, 2 0, 1 0, 0 0))")
+    p1 = wkt.loads("POLYGON ((0 1, 0 2, 1 2, 1 1, 0 1))")
+    p2 = wkt.loads("POLYGON ((1 0, 2 0, 2 -1, 1 -1, 1 0))")
+    data = geopandas.GeoDataFrame(
+        {"name": ["abc", "def", "ghi"], "geometry": [p0, p1, p2]}
+    )
     topo = topojson.Topology(data, prequantize=False)
     p0_wkt = topo.to_gdf().geometry[0].wkt
 
-    assert p0_wkt == 'POLYGON ((0 1, 0 0, 1 0, 2 0, 2 1, 1 1, 0 1))'
+    assert p0_wkt == "POLYGON ((0 1, 0 0, 1 0, 2 0, 2 1, 1 1, 0 1))"
+
 
 # test for https://github.com/mattijn/topojson/issues/110
 def test_topology_topoquantization_dups():
     gdf = geopandas.read_file(geopandas.datasets.get_path("naturalearth_lowres"))
-    gdf = gdf[gdf.name.isin(['France', 'Belgium', 'Netherlands'])]
+    gdf = gdf[gdf.name.isin(["France", "Belgium", "Netherlands"])]
     topo = topojson.Topology(data=gdf, prequantize=False).toposimplify(4)
     topo = topo.topoquantize(50).to_dict()
 
-    assert topo['arcs'][6] == [[44, 47], [0, 0]]
-
+    assert topo["arcs"][6] == [[44, 47], [0, 0]]
+
+
 # parse topojson from file
 def test_topology_topojson_from_file():
-    with open("tests/files_topojson/naturalearth.topojson", 'r') as f:
+    with open("tests/files_topojson/naturalearth.topojson", "r") as f:
         data = json.load(f)
 
     topo = topojson.Topology(data).to_dict()
 
-    assert len(topo["objects"]) == 1   
+    assert len(topo["objects"]) == 1
+
 
 # parse topojson file and plot with altair
 def test_topology_topojson_to_alt():
     # load topojson file into dict
-    with open("tests/files_topojson/naturalearth_lowres_africa.topojson", 'r') as f:
+    with open("tests/files_topojson/naturalearth_lowres_africa.topojson", "r") as f:
         data = json.load(f)
-        
+
     # parse topojson file using `object_name`
     topo = topojson.Topology(data, object_name="data")
     # apply toposimplify and serialize to altair
-    chart = topo.toposimplify(1).to_alt()    
+    chart = topo.toposimplify(1).to_alt()
 
     assert len(chart.__dict__.keys()) == 2
 
+
 # Object of type int64 is not JSON serializable
 def test_topology_topojson_to_alt_int64():
     # load topojson file into dict
-    with open("tests/files_topojson/mesh2d.topojson", 'r') as f:
-        data = json.load(f)    
+    with open("tests/files_topojson/mesh2d.topojson", "r") as f:
+        data = json.load(f)
 
     # parse topojson file using `object_name`
     topo = topojson.Topology(data, object_name="mesh2d_flowelem_bl")
     # apply toposimplify and serialize to altair
-    chart = topo.toposimplify(1).to_alt()  
+    chart = topo.toposimplify(1).to_alt()
+
+    assert len(chart.__dict__.keys()) == 2
 
-    assert len(chart.__dict__.keys()) == 2     
 
 def test_topology_nested_list_properties():
     from geojson import Feature, Polygon, FeatureCollection
@@ -516,12 +523,24 @@ def test_topology_nested_list_properties():
 
     assert len(topo) == 4
 
+
 def test_topology_update_bbox_topoquantize_toposimplify():
     # load example data representing continental Africa
-    data = topojson.utils.example_data_africa()  
+    data = topojson.utils.example_data_africa()
     # compute the topology
-    topo = topojson.Topology(data)  
+    topo = topojson.Topology(data)
     # apply simplification on the topology and render as SVG
-    bbox = topo.topoquantize(10).to_dict()['bbox'] 
+    bbox = topo.topoquantize(10).to_dict()["bbox"]
+
+    assert round(bbox[0], 1) == -17.6
+
+
+def test_topology_bbox_no_delta_transform():
+    data = {
+        "foo": {"type": "LineString", "coordinates": [[0, 0], [1, 0], [2, 0]]},
+        "bar": {"type": "LineString", "coordinates": [[0, 0], [1, 0], [2, 0]]},
+    }
+    topo_1 = topojson.Topology(data, object_name="topo_1").to_dict()
+    topo_2 = topojson.Topology(topo_1, object_name="topo_1").to_dict()
 
-    assert bbox == (0, 0, 9, 9)    
+    assert topo_1["bbox"] == topo_2["bbox"]

topojson/core/dedup.py

@@ -298,7 +298,7 @@ def _pop_merged_arcs(self, bk_dups, linestring_list, array_bk):
 
         # collect new indices of shared arcs
         u, c = np.unique(arr_new, return_counts=True)
-        arr_bk_sarcs = u[c > 1] 
+        arr_bk_sarcs = u[c > 1]
         arr_bk_sarcs = arr_bk_sarcs[~np.isnan(arr_bk_sarcs)]
-        
+
         return arr_new, arr_bk_sarcs

topojson/core/extract.py

@@ -447,10 +447,12 @@ def _extract_featurecollection(self, geom):
 
         # convert FeatureCollection into a dict of features
         if not hasattr(self, "_obj"):
+            geom = copy.deepcopy(geom)
             obj = geom
             self._obj = geom
         else:
             obj = self._obj
+
         data = {}
         zfill_value = len(str(len(obj["features"])))
 
@@ -460,7 +462,9 @@ def _extract_featurecollection(self, geom):
             feature["type"] = "GeometryCollection"
             feature["geometries"] = [feature["geometry"]]
             feature.pop("geometry", None)
-            data["feature_{}".format(str(idx).zfill(zfill_value))] = feature
+            data["feature_{}".format(str(idx).zfill(zfill_value))] = geometry.shape(
+                feature
+            )  # feature
 
         # new data dictionary is created, throw the geometries back to main()
         self._is_single = False
@@ -599,6 +603,11 @@ def _extract_dictionary(self, geom):
         """
 
         self._is_single = False
+        if (
+            "type" in geom.keys()
+            and geom["type"].casefold() == "FeatureCollection".casefold()
+        ):
+            return self._extract_featurecollection(geom)
         self._data = copy.deepcopy(self._data)
 
         # iterate over the input dictionary or geographical object