From 9dbf8ba3afc47515a6db2b43ca522351167b6dc8 Mon Sep 17 00:00:00 2001
From: Gareth Simons <garethsimons@me.com>
Date: Tue, 16 Jul 2024 21:14:41 +0200
Subject: [PATCH] cleans up OSM  tag handling

---
 pyproject.toml                 |  2 +-
 pysrc/cityseer/tools/graphs.py |  2 +-
 pysrc/cityseer/tools/io.py     | 14 ++++++++++++++
 pysrc/cityseer/tools/util.py   | 12 ++++++------
 tests/tools/test_io.py         | 24 +++++++++++++++++++++++-
 5 files changed, 45 insertions(+), 9 deletions(-)

diff --git a/pyproject.toml b/pyproject.toml
index 9c6809ba..45cafed5 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,6 +1,6 @@
 [project]
 name = "cityseer"
-version = '4.15.0'
+version = '4.15.1'
 description = "Computational tools for network-based pedestrian-scale urban analysis"
 readme = "README.md"
 requires-python = ">=3.10, <3.13"
diff --git a/pysrc/cityseer/tools/graphs.py b/pysrc/cityseer/tools/graphs.py
index 235a769e..7d421c9d 100644
--- a/pysrc/cityseer/tools/graphs.py
+++ b/pysrc/cityseer/tools/graphs.py
@@ -304,7 +304,7 @@ def _extract_tags_to_set(
     if tags_list is not None:
         if not isinstance(tags_list, (list, set, tuple)):
             raise ValueError(f"Tags should be provided as a `list` of `str` instead of {type(tags_list)}.")
-        tags_list = [t.lower() for t in tags_list]
+        tags_list = [t.strip().lower() for t in tags_list if t not in ["", " ", None]]
         tags.update(tags_list)
     return tags
 
diff --git a/pysrc/cityseer/tools/io.py b/pysrc/cityseer/tools/io.py
index 47e7f50b..c076965a 100644
--- a/pysrc/cityseer/tools/io.py
+++ b/pysrc/cityseer/tools/io.py
@@ -1166,6 +1166,20 @@ def _node_key(node_coords):
         for k in ["geom", "geometry"]:
             if k in props:
                 del props[k]
+        # names, routes, highways
+        for k in ["names", "routes", "highways"]:
+            if not k in props:
+                props[k] = []  # type: ignore
+            else:
+                prop = props[k]
+                if isinstance(prop, str):
+                    prop = prop.strip("()[] ,'\"")
+                    prop = prop.split(",")
+                    prop = [p.strip("\" '") for p in prop]
+                    prop = [p for p in prop if p not in ["", " ", None]]
+                if not isinstance(prop, (tuple, list)):
+                    raise TypeError(f"Expected key {k} to be a list type to retain compatibility with OSM workflows.")
+                props[k] = prop  # type: ignore
         g_multi.add_edge(node_key_a, node_key_b, src_edge_idx=edge_idx, geom=edge_geom, **props)
 
     # deduplicate
diff --git a/pysrc/cityseer/tools/util.py b/pysrc/cityseer/tools/util.py
index 28ec6642..573a78d0 100644
--- a/pysrc/cityseer/tools/util.py
+++ b/pysrc/cityseer/tools/util.py
@@ -357,18 +357,18 @@ class EdgeInfo:
 
     @property
     def names(self):
-        """Returns a set of street names."""
-        return tuple(set(self._names))
+        """Returns a list of street names."""
+        return list(set(self._names))
 
     @property
     def routes(self):
-        """Returns a set of routes - e.g. route numbers."""
-        return tuple(set(self._refs))
+        """Returns a list of routes - e.g. route numbers."""
+        return list(set(self._refs))
 
     @property
     def highways(self):
-        """Returns a set of highway types - e.g. footway."""
-        return tuple(set(self._highways))
+        """Returns a list of highway types - e.g. footway."""
+        return list(set(self._highways))
 
     def __init__(self):
         """Initialises a network information structure."""
diff --git a/tests/tools/test_io.py b/tests/tools/test_io.py
index d7137934..9b5e053c 100644
--- a/tests/tools/test_io.py
+++ b/tests/tools/test_io.py
@@ -1,7 +1,7 @@
 import networkx as nx
 import numpy as np
 import pytest
-from pyproj import CRS, Transformer
+from pyproj import Transformer
 from shapely import geometry, wkt
 
 from cityseer import config
@@ -782,3 +782,25 @@ def test_nx_from_generic_geopandas(primal_graph):
     for s, e, d in nx_from_generic.edges(data=True):
         total_lens_generic += d["geom"].length
     assert total_lens_input - total_lens_generic < config.ATOL
+    # test OSM keys
+    out_gpd = io.geopandas_from_nx(primal_graph, crs=3395)
+    out_gpd["names"] = ""
+    out_gpd["names"].iloc[0] = "(boo,)"
+    out_gpd["routes"] = ""
+    out_gpd["routes"].iloc[0] = "(boo,)"
+    out_gpd["highways"] = ""
+    out_gpd["highways"].iloc[0] = "boo"
+    out_gpd["highways"].iloc[1] = "boo,"
+    out_gpd["highways"].iloc[2] = "(boo)"
+    out_gpd["highways"].iloc[3] = "(boo,)"
+    out_gpd["highways"].iloc[4] = "(boo),"
+    out_gpd["highways"].iloc[5] = "(boo, baa)"
+    in_G = io.nx_from_generic_geopandas(out_gpd)
+    G = graphs.nx_remove_filler_nodes(in_G)
+    for s, e in G.edges():
+        hwy = G[s][e][0]["highways"]
+        assert isinstance(hwy, list)
+        if len(hwy) == 2:
+            assert hwy == ["boo", "baa"]
+        elif len(hwy) == 1:
+            assert hwy == ["boo"]