[osm] splitting streets (way names) on semicolons

This commit is contained in:
Al
2016-01-24 00:42:25 -05:00
parent 3485738c2b
commit bab7a0f961

View File

@@ -498,12 +498,13 @@ def build_ways_training_data(language_rtree, infile, out_dir):
continue continue
for lang, val in name_language.iteritems(): for lang, val in name_language.iteritems():
for s in val: for v in val:
if lang in languages: for s in v.split(';'):
writer.writerow((lang, country, tsv_string(s))) if lang in languages:
abbrev = osm_abbreviate(street_types_gazetteer, s, lang, abbreviate_prob=1.0, separate_prob=0.5) writer.writerow((lang, country, tsv_string(s)))
if abbrev != s: abbrev = osm_abbreviate(street_types_gazetteer, s, lang, abbreviate_prob=1.0, separate_prob=0.5)
writer.writerow((lang, country, tsv_string(abbrev))) if abbrev != s:
writer.writerow((lang, country, tsv_string(abbrev)))
if i % 1000 == 0 and i > 0: if i % 1000 == 0 and i > 0:
print('did {} ways'.format(i)) print('did {} ways'.format(i))
i += 1 i += 1