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

[countries] use ISO 3166 country name 5% of the time for general addresses, 10% of the time for OpenAddresses. Gives the parser examples of names like "Korea, Republic of" in #168

Browse Source
This commit is contained in:
Al
2017-03-25 19:41:59 -04:00
parent ecfa6855e7
commit 81c59e116a
5 changed files with 21 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
resources/parser/data_sets/openaddresses.yaml
View File

@@ -1,8 +1,9 @@
global:
cldr_country_probability: 0.5
localized_name_probability: 0.7
localized_name_probability: 0.6
iso_alpha_2_code_probability: 0.2
iso_alpha_3_code_probability: 0.1
iso_3166_name_probability: 0.1
abbreviate_street_probability: 0.3
separate_street_probability: 0.2

3
resources/parser/default.yaml
View File

@@ -139,6 +139,7 @@ country:
# When the user-specified country is an ISO code, remove it from the components with this probability (fall back on geocoded components)
remove_iso_code_probability: 0.1
cldr:
localized_name_probability: 0.97
localized_name_probability: 0.92
iso_alpha_2_code_probability: 0.02
iso_alpha_3_code_probability: 0.01
iso_3166_name_probability: 0.05