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

[fix] Use local language with script e.g. ja_rm in place training data

Browse Source
This commit is contained in:
Al
2016-07-26 18:00:38 -04:00
parent 72c3723b43
commit 0c1b12b65c
1 changed files with 9 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
scripts/geodata/osm/formatter.py
View File

@@ -392,6 +392,15 @@ class OSMAddressFormatter(object):
all_local_languages = set([l for l, d in local_languages]) all_local_languages = set([l for l, d in local_languages])
random_languages = set(INTERNET_LANGUAGE_DISTRIBUTION) random_languages = set(INTERNET_LANGUAGE_DISTRIBUTION)
language_defaults = OrderedDict(local_languages)
for tag in tags:
if ':' in tag:
tag, lang = tag.rsplt(':', 1)
if lang.lower() not in all_local_languages and lang.lower().split('_', 1)[0] in all_local_languages:
local_languages.append((lang, language_defaults[lang.split('_')[0]]))
all_local_languages.add(lang)
more_than_one_official_language = len([lang for lang, default in local_languages if default]) > 1 more_than_one_official_language = len([lang for lang, default in local_languages if default]) > 1
containing_ids = [(b['type'], b['id']) for b in osm_components] containing_ids = [(b['type'], b['id']) for b in osm_components]