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

[osm] Treating components that are all punctuation as blank in address parsing (e.g. a single comma)

Browse Source
This commit is contained in:
Al
2015-09-03 17:46:57 -04:00
parent e1e5c16637
commit 294101ad80
1 changed files with 8 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
scripts/geodata/osm/osm_address_training_data.py
View File

@@ -244,6 +244,12 @@ class AddressFormatter(object):
text = re.sub(regex, replacement, text) text = re.sub(regex, replacement, text)
return text return text
def tokenize_component(self, v):
tokens = tokenize(value)
if sum((1 for c, t in tokens if c.value < token_types.PERIOD.value)) > 0:
return [t for c, t in tokenize(v)]
return []
def format_address(self, country, components, minimal_only=True, tag_components=True): def format_address(self, country, components, minimal_only=True, tag_components=True):
template = self.config.get(country.upper()) template = self.config.get(country.upper())
if not template: if not template:
@@ -263,10 +269,10 @@ class AddressFormatter(object):
if tag_components: if tag_components:
components = {k: u' '.join([u'{}/{}'.format(t, k.replace(' ', '_')) components = {k: u' '.join([u'{}/{}'.format(t, k.replace(' ', '_'))
for c, t in tokenize(v)]) for t in self.tokenize_component(v)])
for k, v in components.iteritems()} for k, v in components.iteritems()}
else: else:
components = {k: u' '.join([t for c, t in tokenize(v)]) components = {k: u' '.join(self.tokenize_component(v))
for k, v in components.iteritems()} for k, v in components.iteritems()}
text = self.render_template(template_text, **components) text = self.render_template(template_text, **components)