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

[fix] adding house to ignore keys rather than aliasing it

Browse Source
This commit is contained in:
Al
2015-09-04 12:40:08 -04:00
parent d64a27bc57
commit 17fcfa8b59
1 changed files with 15 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
scripts/geodata/osm/osm_address_training_data.py
View File

@@ -141,8 +141,6 @@ class AddressFormatter(object):
aliases = OrderedDict([ aliases = OrderedDict([
('name', 'house'), ('name', 'house'),
# in OSM this is the type of house, not a real tag
('house', 'house_type'),
('addr:housename', 'house'), ('addr:housename', 'house'),
('addr:housenumber', 'house_number'), ('addr:housenumber', 'house_number'),
('addr:street', 'road'), ('addr:street', 'road'),
@@ -482,6 +480,15 @@ def build_ways_training_data(language_rtree, infile, out_dir):
i += 1 i += 1
f.close() f.close()
OSM_IGNORE_KEYS = (
'house',
)
def strip_keys(value, ignore_keys):
for key in ignore_keys:
value.pop(key, None)
def build_address_format_training_data(language_rtree, infile, out_dir): def build_address_format_training_data(language_rtree, infile, out_dir):
i = 0 i = 0
@@ -494,6 +501,8 @@ def build_address_format_training_data(language_rtree, infile, out_dir):
formatted_tagged_file = open(os.path.join(out_dir, ADDRESS_FORMAT_DATA_TAGGED_FILENAME), 'w') formatted_tagged_file = open(os.path.join(out_dir, ADDRESS_FORMAT_DATA_TAGGED_FILENAME), 'w')
formatted_tagged_writer = csv.writer(formatted_tagged_file, 'tsv_no_quote') formatted_tagged_writer = csv.writer(formatted_tagged_file, 'tsv_no_quote')
remove_keys = OSM_IGNORE_KEYS
for key, value in parse_osm(infile): for key, value in parse_osm(infile):
try: try:
latitude, longitude = latlon_to_floats(value['lat'], value['lon']) latitude, longitude = latlon_to_floats(value['lat'], value['lon'])
@@ -504,6 +513,9 @@ def build_address_format_training_data(language_rtree, infile, out_dir):
if not (country and default_languages): if not (country and default_languages):
continue continue
for key in remove_keys:
_ = value.pop(key, None)
formatted_address_tagged = formatter.format_address(country, value) formatted_address_tagged = formatter.format_address(country, value)
formatted_address_untagged = formatter.format_address(country, value, tag_components=False) formatted_address_untagged = formatter.format_address(country, value, tag_components=False)
if formatted_address_tagged is not None: if formatted_address_tagged is not None:
@@ -545,7 +557,7 @@ def build_address_format_training_data_limited(language_rtree, infile, out_dir):
f = open(os.path.join(out_dir, ADDRESS_FORMAT_DATA_LANGUAGE_FILENAME), 'w') f = open(os.path.join(out_dir, ADDRESS_FORMAT_DATA_LANGUAGE_FILENAME), 'w')
writer = csv.writer(f, 'tsv_no_quote') writer = csv.writer(f, 'tsv_no_quote')
remove_keys = NAME_KEYS + COUNTRY_KEYS + POSTAL_KEYS remove_keys = NAME_KEYS + COUNTRY_KEYS + POSTAL_KEYS + OSM_IGNORE_KEYS
for key, value in parse_osm(infile): for key, value in parse_osm(infile):
try: try: