tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[osm] build OSM training data for intersections using the JSON output from intersections.py rather having to compute each time

Browse Source
This commit is contained in:
Al
2016-08-17 18:11:55 -04:00
parent a3ae1eb330
commit 145af9331e
2 changed files with 6 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
scripts/geodata/osm/formatter.py
View File

@@ -1027,7 +1027,7 @@ class OSMAddressFormatter(object):
if i % 1000 == 0 and i > 0:
print('did {} formatted places'.format(i))
def build_intersections_training_data(self, infile, out_dir, way_db_dir, tag_components=True):
def build_intersections_training_data(self, infile, out_dir, tag_components=True):
'''
Intersection addresses like "4th & Main Street" are represented in OSM
by ways that share at least one node.
@@ -1053,8 +1053,7 @@ class OSMAddressFormatter(object):
replace_with_base_name_prob = float(nested_get(self.config, ('intersections', 'replace_with_base_name_probability'), default=0.0))
reader = OSMIntersectionReader(infile, way_db_dir)
for node_id, latitude, longitude, ways in reader.intersections():
for node_id, latitude, longitude, ways in OSMIntersectionReader.read_intersections(infile):
if not ways or len(ways) < 2:
continue

12
scripts/geodata/osm/osm_address_training_data.py
View File

@@ -26,6 +26,9 @@ python osm_address_training_data.py -a $(OSM_DIR)/planet-addresses.osm -f --lang
Formatted addresses (untagged):
python osm_address_training_data.py -a $(OSM_DIR)/planet-addresses.osm -f -u --language-rtree-dir=$(LANG_RTREE_DIR) --neighborhoods-rtree-dir=$(NEIGHBORHOODS_RTREE_DIR) --rtree-dir=$(RTREE_DIR) --quattroshapes-rtree-dir=$(QS_TREE_DIR) --geonames-db=$(GEONAMES_DB_PATH) -o $(OUT_DIR)
Intersections (after running intersections.py to create the JSON file):
python osm_address_training_data -x $(OSM_DIR)/intersections.json -f --language-rtree-dir=$(LANG_RTREE_DIR) --neighborhoods-rtree-dir=$(NEIGHBORHOODS_RTREE_DIR) --rtree-dir=$(RTREE_DIR) --quattroshapes-rtree-dir=$(QS_TREE_DIR) --geonames-db=$(GEONAMES_DB_PATH) -o $(OUT_DIR)
Toponyms:
python osm_address_training_data.py -b $(OSM_DIR)/planet-borders.osm --language-rtree-dir=$(LANG_RTREE_DIR) -o $(OUT_DIR)
'''
@@ -452,9 +455,6 @@ if __name__ == '__main__':
parser.add_argument('-x', '--intersections-file',
help='Path to planet-ways-latlons.osm')
parser.add_argument('--ways-db-dir',
help='Path to temporary ways db')
parser.add_argument('--language-rtree-dir',
required=True,
help='Language RTree directory')
@@ -556,10 +556,6 @@ if __name__ == '__main__':
osm_formatter.build_place_training_data(args.place_nodes_file, args.out_dir, tag_components=not args.untagged)
if args.intersections_file and args.format:
if args.ways_db_dir is None:
parser.error('--ways-db-dir required for intersections')
logging.basicConfig(level=logging.INFO)
components = AddressComponents(osm_rtree, language_rtree, neighborhoods_rtree, quattroshapes_rtree, geonames, metro_stations_index)
osm_formatter = OSMAddressFormatter(components, subdivisions_rtree, buildings_rtree, metro_stations_index)
osm_formatter.build_intersections_training_data(args.intersections_file, args.out_dir, args.ways_db_dir, tag_components=not args.untagged)
osm_formatter.build_intersections_training_data(args.intersections_file, args.out_dir, tag_components=not args.untagged)