Fix OSM→OSW conversion when OSM Way contains consecutive duplicate nodes

https://dev.azure.com/TDEI-UW/TDEI/_workitems/edit/3286

**PR Title**

Fix OSM→OSW conversion when OSM Way contains consecutive duplicate nodes

**PR Summary**

- 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.

- Tested on file attached to dev board bug: osm.b8106236-b919-4e50-b956-99399635c0bb.xml
- Added a new unit test

Author: wisamuw

CHANGELOG.md

@@ -1,5 +1,8 @@
 # Change log
 
+### 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.
+
 ### 0.3.3
 - Fix OSM→OSW export classification so canonical OSM tags are used for semantic recognition and `ext:*` tags are preserved as extensions instead of being treated as feature-defining tags.
 - Fix closed ext-only ways such as `ext:demolished:building=yes` to emit polygon output without falling through to point geometry construction.

src/osm_osw_reformatter/serializer/osm/osm_graph.py

@@ -34,6 +34,7 @@ def way(self, w) -> None:
 
         d2 = {**d, **OSWWayNormalizer(tags).normalize()}
 
+        segment_n = 0
         for i in range(len(w.nodes) - 1):
             u = w.nodes[i]
             v = w.nodes[i + 1]
@@ -49,12 +50,19 @@ def way(self, w) -> None:
             v_lon = float(v.lon)
             v_lat = float(v.lat)
 
+            # Skip consecutive duplicate nodes. They create zero-length segments.
+            if u_ref == v_ref:
+                del u
+                del v
+                continue
+
             d3 = {**d2}
-            d3['segment'] = i
+            d3['segment'] = segment_n
             d3['ndref'] = [u_ref, v_ref]
             self.G.add_edges_from([(u_ref, v_ref, d3)])
             self.G.add_node(u_ref, lon=u_lon, lat=u_lat)
             self.G.add_node(v_ref, lon=v_lon, lat=v_lat)
+            segment_n += 1
             del u
             del v
 

src/osm_osw_reformatter/version.py

@@ -1 +1 @@
-__version__ = '0.3.3'
+__version__ = '0.3.4'

tests/unit_tests/test_files/bug_3286.xml

@@ -0,0 +1,10 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<osm version="0.6" generator="TDEI exporter" upload="false">
+<node visible="true" version="1" id="8223" lat="47.9217092" lon="-122.2499566"><tag k="ext:osm_version" v="2"/></node>
+<node visible="true" version="1" id="8224" lat="47.9217137" lon="-122.2499283"><tag k="ext:osm_version" v="2"/></node>
+<node visible="true" version="1" id="8232" lat="47.9217122" lon="-122.2499433"><tag k="ext:osm_version" v="2"/></node>
+<node visible="true" version="1" id="8237" lat="47.9217066" lon="-122.2499667"><tag k="ext:osm_version" v="2"/></node>
+<node visible="true" version="1" id="8241" lat="47.9217132" lon="-122.2499356"><tag k="ext:osm_version" v="2"/></node>
+<node visible="true" version="1" id="16255" lat="47.9217166" lon="-122.2495229"><tag k="ext:osm_version" v="2"/></node>
+<way visible="true"  version="1" id="19002"><nd ref="8237"/><nd ref="8237"/><nd ref="8223"/><nd ref="8232"/><nd ref="8241"/><nd ref="8224"/><nd ref="16255"/><tag k="width" v="2.0"/><tag k="footway" v="sidewalk"/><tag k="highway" v="footway"/><tag k="incline" v="0.009"/><tag k="surface" v="concrete"/><tag k="ext:osm_version" v="3"/><tag k="ext:width_confidence" v="1.0"/><tag k="ext:sidewalk_corrected" v="yes"/></way>
+</osm>

tests/unit_tests/test_osm2osw/test_osm2osw.py

@@ -15,6 +15,7 @@
 TEST_INVALID_NODE_TAGS_FILE = os.path.join(ROOT_DIR, 'test_files/node_with_invalid_tags.xml')
 TEST_TREE_FILE = os.path.join(ROOT_DIR, 'test_files/tree-test.xml')
 TEST_BUG_3477_FILE = os.path.join(ROOT_DIR, 'test_files/bug_3477.xml')
+TEST_BUG_3286_FILE = os.path.join(ROOT_DIR, 'test_files/bug_3286.xml')
 
 
 def is_valid_float(value):
@@ -350,6 +351,26 @@ async def run_test():
 
         asyncio.run(run_test())
 
+    def test_bug_3286_consecutive_duplicate_nodes(self):
+        osm_file_path = TEST_BUG_3286_FILE
+
+        async def run_test():
+            osm2osw = OSM2OSW(osm_file=osm_file_path, workdir=OUTPUT_DIR, prefix='test')
+            result = await osm2osw.convert()
+            self.assertTrue(result.status)
+            self.assertEqual(len(result.generated_files), 2)
+
+            for file_path in result.generated_files:
+                if file_path.endswith('edges.geojson'):
+                    with open(file_path) as f:
+                        geojson = json.load(f)
+                        self.assertEqual(len(geojson.get("features", [])), 1)
+
+            for file_path in result.generated_files:
+                os.remove(file_path)
+
+        asyncio.run(run_test())
+
 
 if __name__ == '__main__':
     unittest.main()