-
Notifications
You must be signed in to change notification settings - Fork 1
/
jenny.py
324 lines (262 loc) · 9.96 KB
/
jenny.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
import sqlite3
import xml.etree.ElementTree as ET
import csv
from shapely.geometry import Point, LineString, Polygon
from shapely.wkb import loads, dumps
tree = ET.parse("sfsubset.osm")
root = tree.getroot()
test_csv_f = open("layers/pointtest.csv", "wb")
test_csv = csv.writer(test_csv_f)
test_csv.writerow([ "name", "wkt" ])
nodes = {}
ways = {}
relations = {}
def tags_from_child_elems(elem):
tags = {}
for child_elem in elem:
if child_elem.tag == "tag":
tags[child_elem.attrib["k"]] = child_elem.attrib["v"]
return tags
class Node(object):
def __init__(self, osm_id, lon, lat, tags):
self.id = osm_id
self.lon = lon
self.lat = lat
self.tags = tags
self.point = Point(lon, lat)
@classmethod
def from_xml_elem(cls, elem):
osm_id = elem.attrib["id"]
lat = float(elem.attrib["lat"])
lon = float(elem.attrib["lon"])
tags = tags_from_child_elems(elem)
return cls(osm_id, lon, lat, tags)
class Way(object):
def __init__(self, osm_id, nodes, tags):
self.id = osm_id
self.nodes = nodes
self.tags = tags
coords = [(n.lon, n.lat) for n in nodes]
if len(coords) == 0:
# Invalid, but we'll clean up this mess in the caller.
return
if len(coords) < 2:
# If we find a way with only one node then just double-up
# that node to keep the data model sensible even if the
# data itself isn;t.
coords = coords + coords
self.line_string = LineString(coords)
if len(coords) > 3:
self.polygon = Polygon(list(coords[:-1]))
else:
# Make a fake polygon just so there's always something
# at self.polygon and we don't blow up when we encounter
# malformed data. This will make a pretty dumb polygon, though.
self.polygon = Polygon(coords * 3)
@classmethod
def from_xml_elem(cls, elem):
osm_id = elem.attrib["id"]
tags = tags_from_child_elems(elem)
new_nodes = []
for child_elem in elem:
if child_elem.tag == "nd":
node_id = child_elem.attrib["ref"]
node = nodes[node_id]
new_nodes.append(node)
return cls(osm_id, new_nodes, tags)
class RelationMember(object):
def __init__(self, obj, role):
self.object = obj
self.role = role
@classmethod
def from_xml_elem(cls, elem):
member_type = elem.attrib["type"]
ref = elem.attrib["ref"]
role = elem.attrib["role"]
if member_type == "way":
obj = ways.get(ref, None)
elif member_type == "node":
obj = nodes.get(ref, None)
elif member_type == "relation":
# Our target relation might not have been loaded yet,
# so we'll stash its id for now and fix it up later.
obj = ref
else:
# Unknown target type
obj = None
return cls(obj, role)
class Relation(object):
def __init__(self, osm_id, members, tags):
self.id = osm_id
self.members = members
self.tags = tags
@classmethod
def from_xml_elem(cls, elem):
osm_id = elem.attrib["id"]
tags = tags_from_child_elems(elem)
members = []
for child_elem in elem:
if child_elem.tag == "member":
member = RelationMember.from_xml_elem(child_elem)
if member is not None:
members.append(member)
return cls(osm_id, members, tags)
def members_by_type(self, member_type):
return (m for m in self.members if type(m.object) is member_type)
class DataFile(object):
def __init__(self, csv, tag_keys):
self.csv = csv
self.tag_keys = tag_keys
self.csv.writerow(["wkt"] + [key.replace(":", "__") for key in tag_keys])
@classmethod
def open_new(cls, fn, tag_keys):
full_name = "layers/%s.csv" % fn
f = open(full_name, 'wb')
csv_w = csv.writer(f)
return cls(csv_w, tag_keys)
def add_point(self, obj):
wkt = obj.point.wkt
others = [obj.tags.get(k, "").encode("utf-8") for k in self.tag_keys]
self.csv.writerow([wkt] + others)
def add_way(self, obj):
wkt = obj.line_string.wkt
others = [obj.tags.get(k, "").encode("utf-8") for k in self.tag_keys]
self.csv.writerow([wkt] + others)
def add_polygon(self, obj):
wkt = obj.polygon.wkt
others = [obj.tags.get(k, "").encode("utf-8") for k in self.tag_keys]
self.csv.writerow([wkt] + others)
for elem in root:
if elem.tag == "node":
node = Node.from_xml_elem(elem)
nodes[node.id] = node
#point = node.point
#name = node.tags.get("name", None)
#if name is not None:
# print repr([name, point.wkt])
# test_csv.writerow([name.encode("utf-8"), point.wkt])
elif elem.tag == "way":
way = Way.from_xml_elem(elem)
if getattr(way, "polygon", None) is not None:
ways[way.id] = way
elif elem.tag == "relation":
relation = Relation.from_xml_elem(elem)
relations[relation.id] = relation
# Because the relations are a graph we save just relation id strings as
# members during loading and need to now fix them up to be real relation
# object references.
for relation_id, relation in relations.iteritems():
for member in relation.members:
if type(member.object) is str:
member.object = relations.get(member.object, None)
### HERE COMES THE PART WHERE WE ACTUALLY DO THE DATA MANIPULATION!
# First, we process the relations and add extra annotations to the
# members of them so that this data is visible to the renderer, because the
# renderer doesn't "see" relations, only nodes and ways.
for relation_id, relation in relations.iteritems():
relation_type = relation.tags.get("type")
if relation_type == "route":
route_type = relation.tags.get("route")
if route_type in ("bus", "tram", "light_rail"):
for member in relation.members_by_type(Way):
member.object.tags["on_bus_route"] = "yes"
# Now we can generate the geometry CSV files
highways = DataFile.open_new("highways", [
"name",
"highway",
"sidewalk_left",
"sidewalk_right",
"on_bus_route",
"interesting_to_peds",
"steps",
"foot",
"tunnel",
])
landuse = DataFile.open_new("landuse", [
"landuse",
])
buildings = DataFile.open_new("buildings", [
"name",
"building:levels",
"shop",
])
pois = DataFile.open_new("pois", [
"name",
"amenity",
"shop",
"office",
"food",
"cuisine",
])
barrierlines = DataFile.open_new("barrierlines", [
"barrier",
])
rails = DataFile.open_new("rails", [
"railway",
"tracks",
])
for way_id, way in ways.iteritems():
if "highway" in way.tags:
if way.tags.get("area", None) != "yes":
# Just turn paths into footways for ease of common rendering
if way.tags["highway"] in ("path", "pedestrian"):
way.tags["highway"] = "footway"
if way.tags["highway"] == "steps":
way.tags["highway"] = "footway"
way.tags["steps"] = "yes"
sidewalk = way.tags.get("sidewalk", None)
sidewalk_left = False
sidewalk_right = False
if sidewalk == "both":
sidewalk_left = True
sidewalk_right = True
elif sidewalk == "left":
sidewalk_left = True
elif sidewalk == "right":
sidewalk_right = True
elif sidewalk in ("none", "separate"):
pass
else:
if way.tags.get("highway") not in ("motorway", "cycleway", "motorway_link", "service") and way.tags.get("foot") != "no":
sidewalk_left = True
sidewalk_right = True
way.tags["sidewalk_left"] = "yes" if sidewalk_left else "no"
way.tags["sidewalk_right"] = "yes" if sidewalk_right else "no"
interesting_to_peds = (sidewalk_left or sidewalk_right)
foot = False
if way.tags.get("foot") in ("yes", "permissive") and not (sidewalk_left or sidewalk_right):
interesting_to_peds = True
foot = True
way.tags["foot"] = "yes" if foot else "no"
way.tags["interesting_to_peds"] = "yes" if interesting_to_peds else "no"
highways.add_way(way)
# Turn some area stuff into landuse for rendering simplicity
if way.tags.get("amenity") == "parking" and "landuse" not in way.tags:
way.tags["landuse"] = "parking"
if way.tags.get("leisure") == "park" and "landuse" not in way.tags:
way.tags["landuse"] = "park"
if way.tags.get("leisure") == "common" and "landuse" not in way.tags:
way.tags["landuse"] = "common"
if way.tags.get("highway") == "pedestrian" and way.tags.get("area") == "yes" and "landuse" not in way.tags:
way.tags["landuse"] = "pedestrian"
if "landuse" in way.tags:
landuse.add_polygon(way)
if "barrier" in way.tags and way.tags.get("area") != "yes":
barrierlines.add_way(way)
if "building" in way.tags:
if way.tags.get("location") not in ("underground", "underwater"):
buildings.add_polygon(way)
if way.tags.get("railway") in ("tram", "light_rail", "rail"):
if way.tags.get("tunnel") == "yes":
# Don't care about underground railways.
continue
if way.tags["railway"] == "light_rail":
# normalize for easy rendering
way.tags["railway"] = "tram"
rails.add_way(way)
for node_id, node in nodes.iteritems():
if node.tags.get("amenity") in ("fuel","parking","bicycle_parking","telephone","taxi","car_sharing","bench"):
continue
for poi_key in ("amenity", "shop", "food", "cuisine"):
if poi_key in node.tags:
pois.add_point(node)