Fix OSM→OSW zone _w_id references missing from nodes

sujata-m · sujata-m · commit 45e70c0052ee · 2026-05-19T23:39:01.000+05:30
## Dev Board Ticket - https://dev.azure.com/TDEI-UW/TDEI/_workitems/edit/3665 - https://dev.azure.com/TDEI-UW/TDEI/_workitems/edit/3191/ ## Changes - **Zone boundary protection in `construct_geometries()`**: collect zone boundary IDs from `_w_id` lists and filter them out of `internal_nodes` before `self.G.remove_nodes_from(...)`. This prevents zone boundary nodes from being deleted by simplification edge cases (e.g., circular-way neighbor combination), so they remain in `node_id_map` for `_w_id` → sequential `_id` remapping in `to_geojson()`. - **Regression fixture**: added [tests/unit_tests/test_files/zone_boundary.xml](tests/unit_tests/test_files/zone_boundary.xml) — a minimal `highway=pedestrian` plaza plus connecting footways with simplifiable intermediate nodes. - **Regression test**: `test_osm2osw_zone_boundary_is_osw_compliant` in [test_osm_compliance.py](tests/unit_tests/test_osm_compliance/test_osm_compliance.py) runs `Formatter.osm2osw()`, zips the output, and asserts `OSWValidation` reports zero issues. - **README fixes**: - Corrected invalid Python import `from osm-osw-reformatter` → `from osm_osw_reformatter`. - Aligned conda command (`python==3.10.3` → `python=3.10`) with the documented 3.10.x requirement. - Updated sample snippet to `return` the result and access `results.generated_files`. - **Version bump**: `0.3.4` → `0.3.5` in [version.py](src/osm_osw_reformatter/version.py) with matching [CHANGELOG.md](CHANGELOG.md) entry. ## Testing - Ran the full compliance test suite — all 3 pass, including the new zone-boundary regression test: ``` pytest tests/unit_tests/test_osm_compliance/ ===== 3 passed in 4.81s ===== ``` - Validated end-to-end via the `example.py` flow (`osm2osw` → zip → `python-osw-validation`) on the fixture: zero validation issues reported.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,10 @@
 # Change log
 
+### 0.3.5
+- [BUG-3665](https://dev.azure.com/TDEI-UW/TDEI/_workitems/edit/3665) - Fix OSM→OSW export so zone boundary nodes are retained in `nodes.geojson` and zone `_w_id` references resolve to remapped sequential `_id`s, restoring OSW validation compliance for pedestrian-area geometries.
+- Add regression coverage that converts a `highway=pedestrian` plaza fixture and asserts `python-osw-validation` reports zero issues.
+- [ISSUE-3191](https://dev.azure.com/TDEI-UW/TDEI/_workitems/edit/3191/) - Fixed Documentation  
+
 ### 0.3.4
 - Fix OSM→OSW conversion when OSM Way contains consecutive duplicate nodes (bug 3286). Instead of generating an invalid 0 length geometry, the segment is ignored.
 
diff --git a/README.md b/README.md
@@ -30,7 +30,7 @@ The conversion of OSW data to OSM is beneficial for incorporating detailed pedes
 Running the code base requires a proper Python environment set up. The following lines of code helps one establish such env named `tdei-osw`. replace `tdei-osw` with the name of your choice.
 
 ```
-conda create -n tdei-osw python==3.10.3 gdal
+conda create -n tdei-osw python=3.10 gdal
 conda activate tdei-osw
 pip install -r requirements.txt
 ```
@@ -92,24 +92,25 @@ To install the GDAL library (Geospatial Data Abstraction Library) on your system
 
 ```python  
 import asyncio
-from osm-osw-reformatter import Formatter  
+from osm_osw_reformatter import Formatter  
   
 async def osm_convert():
     f = Formatter(workdir=<OUTPUT_DIR>, file_path=<OSM_INPUT_FILE>)
-    await f.osm2osw()
+    return await f.osm2osw()
     # Uncomment below line to clean up the generated files
     # f.cleanup()
 
 
 def osw_convert():
     f = Formatter(workdir=<OUTPUT_DIR>, file_path=<OSW_INPUT_FILE>)
-    f.osw2osm()
+    return f.osw2osm()
     # Uncomment below line to clean up the generated files
     # f.cleanup()
 
 
 if __name__ == '__main__':
-    asyncio.run(osm_convert())
+    results = asyncio.run(osm_convert())
+    print(results.generated_files)
     osw_convert()  
 ```  
   
diff --git a/requirements.txt b/requirements.txt
@@ -5,4 +5,4 @@ shapely~=2.0.2
 pyproj~=3.6.1
 coverage~=7.5.1
 ogr2osm==1.2.0
-python-osw-validation==0.3.5
+python-osw-validation==0.4.0
diff --git a/setup.py b/setup.py
@@ -31,6 +31,7 @@
         'asyncio~=3.4.3',
         'networkx~=3.2',
         'shapely~=2.0.2',
+        'pyproj~=3.6.1',
         'ogr2osm==1.2.0'
     ],
     packages=find_packages(where='src'),
diff --git a/src/osm_osw_reformatter/serializer/osm/osm_graph.py b/src/osm_osw_reformatter/serializer/osm/osm_graph.py
@@ -588,6 +588,21 @@ def construct_geometries(self, progressbar: Optional[callable] = None) -> None:
                 if progressbar:
                     progressbar.update(1)
                 
+        # Protect zone boundary nodes: even if they ended up in internal_nodes
+        # due to circular-way simplification edge cases, they must not be removed
+        # because zones' _w_id still references them and to_geojson() needs to
+        # remap those IDs to sequential _id values.
+        zone_boundary_ids = set()
+        for _n, _d in self.G.nodes(data=True):
+            w_ids = _d.get("_w_id")
+            if isinstance(w_ids, list):
+                for ref in w_ids:
+                    try:
+                        zone_boundary_ids.add(int(ref))
+                    except (TypeError, ValueError):
+                        pass
+        if zone_boundary_ids:
+            internal_nodes = [n for n in internal_nodes if n not in zone_boundary_ids]
         self.G.remove_nodes_from(internal_nodes)
 
     def to_undirected(self):
@@ -675,15 +690,30 @@ def _remap_node_ref(ref, node_id_map):
         zone_features = []
         polygon_features = []
         node_id_map = {}
+        zone_node_refs = set()
+        for _, d in self.G.nodes(data=True):
+            if OSWZoneNormalizer.osw_zone_filter(d):
+                w_ids = d.get("_w_id", d.get("ndref", []))
+                if not isinstance(w_ids, list):
+                    w_ids = [w_ids]
+                for ref in w_ids:
+                    zone_node_refs.add(ref)
+                    zone_node_refs.add(str(ref))
+                    try:
+                        zone_node_refs.add(int(ref))
+                    except (TypeError, ValueError):
+                        pass
+
         for n, d in self.G.nodes(data=True):
             d_copy = {**d}
             source_id = _source_id(n)
             geometry_obj = d_copy.pop("geometry")
             geometry = mapping(geometry_obj)
             geometry_type = geometry_obj.geom_type
             is_topology_node = geometry_type == "Point" and self.G.degree(n) > 0
+            is_zone_node = geometry_type == "Point" and n in zone_node_refs
 
-            if is_topology_node:
+            if is_topology_node or is_zone_node:
                 _assign_ids(d_copy, node_id_counter, source_id)
                 node_id_map[n] = d_copy["_id"]
                 node_id_counter += 1
diff --git a/src/osm_osw_reformatter/version.py b/src/osm_osw_reformatter/version.py
@@ -1 +1 @@
-__version__ = '0.3.4'
+__version__ = '0.3.5'
diff --git a/tests/unit_tests/test_files/zone_boundary.xml b/tests/unit_tests/test_files/zone_boundary.xml
@@ -0,0 +1,44 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<osm version="0.6" generator="test" upload="false">
+  <!-- Pedestrian plaza corners -->
+  <node id="1" visible="true" version="1" lat="47.6151489" lon="-122.3206583"/>
+  <node id="2" visible="true" version="1" lat="47.6149868" lon="-122.3206509"/>
+  <node id="3" visible="true" version="1" lat="47.6149898" lon="-122.3203263"/>
+  <node id="4" visible="true" version="1" lat="47.6151507" lon="-122.3203266"/>
+
+  <!-- External footway nodes (one intermediate node that should get simplified) -->
+  <node id="5" visible="true" version="1" lat="47.6153000" lon="-122.3206583"/>
+  <node id="6" visible="true" version="1" lat="47.6152500" lon="-122.3206583"/>
+  <node id="7" visible="true" version="1" lat="47.6149898" lon="-122.3201000"/>
+  <node id="8" visible="true" version="1" lat="47.6149898" lon="-122.3202000"/>
+
+  <!-- Pedestrian plaza: closed way forming a zone -->
+  <way id="100" visible="true" version="1">
+    <nd ref="1"/>
+    <nd ref="2"/>
+    <nd ref="3"/>
+    <nd ref="4"/>
+    <nd ref="1"/>
+    <tag k="highway" v="pedestrian"/>
+    <tag k="area" v="yes"/>
+    <tag k="name" v="Test Plaza"/>
+  </way>
+
+  <!-- Footway connecting external node 5 to plaza corner 1 via intermediate node 6 -->
+  <way id="200" visible="true" version="1">
+    <nd ref="5"/>
+    <nd ref="6"/>
+    <nd ref="1"/>
+    <tag k="highway" v="footway"/>
+    <tag k="footway" v="sidewalk"/>
+  </way>
+
+  <!-- Footway connecting plaza corner 3 to external node 7 via intermediate node 8 -->
+  <way id="201" visible="true" version="1">
+    <nd ref="3"/>
+    <nd ref="8"/>
+    <nd ref="7"/>
+    <tag k="highway" v="footway"/>
+    <tag k="footway" v="sidewalk"/>
+  </way>
+</osm>
diff --git a/tests/unit_tests/test_osm_compliance/test_osm_compliance.py b/tests/unit_tests/test_osm_compliance/test_osm_compliance.py
@@ -10,6 +10,7 @@
 ROOT_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 OUTPUT_DIR = os.path.join(os.path.dirname(os.path.dirname(ROOT_DIR)), 'output')
 TEST_DATA_WITH_INCLINE_ZIP_FILE = os.path.join(ROOT_DIR, 'test_files/dataset_with_incline.zip')
+TEST_ZONE_BOUNDARY_FILE = os.path.join(ROOT_DIR, 'test_files/zone_boundary.xml')
 
 
 class TestOSMCompliance(unittest.IsolatedAsyncioTestCase):
@@ -37,6 +38,25 @@ async def test_output_is_osm_compliant(self):
         os.remove(zip_path)
         formatter.cleanup()
 
+    async def test_osm2osw_zone_boundary_is_osw_compliant(self):
+        formatter = Formatter(workdir=OUTPUT_DIR, file_path=TEST_ZONE_BOUNDARY_FILE, prefix='zone_boundary')
+        res = await formatter.osm2osw()
+        osw_files = res.generated_files
+
+        zip_path = os.path.join(OUTPUT_DIR, 'zone_boundary_osw.zip')
+        with zipfile.ZipFile(zip_path, 'w', zipfile.ZIP_DEFLATED) as zipf:
+            for f in osw_files:
+                zipf.write(f, os.path.basename(f))
+
+        validator = OSWValidation(zipfile_path=zip_path)
+        result = validator.validate()
+        self.assertEqual(len(result.issues), 0, f'OSW Validation issues: {json.dumps(result.issues)}')
+
+        for f in osw_files:
+            os.remove(f)
+        os.remove(zip_path)
+        formatter.cleanup()
+
     async def test_incline_tag_preserved(self):
         osw2osm = OSW2OSM(
             zip_file_path=TEST_DATA_WITH_INCLINE_ZIP_FILE,
diff --git a/tests/unit_tests/test_serializer/test_osm_graph.py b/tests/unit_tests/test_serializer/test_osm_graph.py
@@ -1058,6 +1058,58 @@ def test_to_geojson_zones_reference_remapped_node_ids_in_w_id(self):
             zone_w_ids = zone_data["features"][0]["properties"]["_w_id"]
             self.assertTrue(all(wid in node_ids for wid in zone_w_ids))
 
+    def test_to_geojson_zone_boundary_custom_points_stay_in_nodes_file(self):
+        graph = nx.MultiDiGraph()
+        graph.add_node(
+            10,
+            geometry=Point(0, 0),
+            lon=0.0,
+            lat=0.0,
+            **{"ext:entrance": "yes"},
+        )
+        graph.add_node(20, geometry=Point(1, 0), lon=1.0, lat=0.0)
+        graph.add_node(
+            "z100",
+            geometry=Polygon([(0, 0), (1, 0), (0, 1), (0, 0)]),
+            highway="pedestrian",
+            _w_id=["10", "20"],
+        )
+
+        osm_graph = OSMGraph(G=graph)
+
+        with TemporaryDirectory() as tmpdir:
+            nodes_path = os.path.join(tmpdir, 'nodes.geojson')
+            edges_path = os.path.join(tmpdir, 'edges.geojson')
+            points_path = os.path.join(tmpdir, 'points.geojson')
+            lines_path = os.path.join(tmpdir, 'lines.geojson')
+            zones_path = os.path.join(tmpdir, 'zones.geojson')
+            polygons_path = os.path.join(tmpdir, 'polygons.geojson')
+
+            osm_graph.to_geojson(
+                nodes_path,
+                edges_path,
+                points_path,
+                lines_path,
+                zones_path,
+                polygons_path,
+            )
+
+            with open(nodes_path) as f:
+                node_data = json.load(f)
+            with open(zones_path) as f:
+                zone_data = json.load(f)
+
+            node_ids = {feat["properties"]["_id"] for feat in node_data["features"]}
+            node_ext_ids = {
+                feat["properties"].get("ext:osm_id")
+                for feat in node_data["features"]
+            }
+            zone_w_ids = zone_data["features"][0]["properties"]["_w_id"]
+
+            self.assertIn("10", node_ext_ids)
+            self.assertFalse(os.path.exists(points_path))
+            self.assertTrue(all(wid in node_ids for wid in zone_w_ids))
+
     def test_to_undirected_on_simple_graph(self):
         g = nx.Graph()
         g.add_edge(1, 2)

Original file line number	Diff line number	Diff line change
`@@ -1 +1 @@`
`1`		`-__version__ = '0.3.4'`
	`1`	`+__version__ = '0.3.5'`