From 1b09b7f2e5ba5968632eae69942d820209f96bab Mon Sep 17 00:00:00 2001 From: Al Date: Thu, 28 Jul 2016 16:18:32 -0400 Subject: [PATCH] [fix] Adding country_region to address_parser_train --- src/address_parser.h | 3 ++- src/address_parser_train.c | 4 ++++ 2 files changed, 6 insertions(+), 1 deletion(-) diff --git a/src/address_parser.h b/src/address_parser.h index d813077c..f8d5fa86 100644 --- a/src/address_parser.h +++ b/src/address_parser.h @@ -82,7 +82,8 @@ with the general error-driven averaged perceptron. #define ADDRESS_COMPONENT_STATE_DISTRICT 1 << 11 #define ADDRESS_COMPONENT_STATE 1 << 12 #define ADDRESS_COMPONENT_POSTAL_CODE 1 << 13 -#define ADDRESS_COMPONENT_COUNTRY 1 << 14 +#define ADDRESS_COMPONENT_COUNTRY_REGION 1 << 14 +#define ADDRESS_COMPONENT_COUNTRY 1 << 15 typedef enum { ADDRESS_PARSER_HOUSE, diff --git a/src/address_parser_train.c b/src/address_parser_train.c index 220c5d66..622e1a7d 100644 --- a/src/address_parser_train.c +++ b/src/address_parser_train.c @@ -34,6 +34,7 @@ static inline bool is_phrase_component(char *label) { string_equals(label, ADDRESS_PARSER_LABEL_ISLAND) || string_equals(label, ADDRESS_PARSER_LABEL_STATE) || string_equals(label, ADDRESS_PARSER_LABEL_POSTAL_CODE) || + string_equals(label, ADDRESS_PARSER_LABEL_COUNTRY_REGION) || string_equals(label, ADDRESS_PARSER_LABEL_COUNTRY)); } @@ -182,6 +183,9 @@ address_parser_t *address_parser_init(char *filename) { } else if (string_equals(prev_label, ADDRESS_PARSER_LABEL_COUNTRY)) { class_id = ADDRESS_PARSER_COUNTRY; component = ADDRESS_COMPONENT_COUNTRY; + } else if (string_equals(prev_label, ADDRESS_PARSER_LABEL_COUNTRY_REGION)) { + class_id = ADDRESS_PARSER_COUNTRY_REGION; + component = ADDRESS_COMPONENT_COUNTRY; } else if (string_equals(prev_label, ADDRESS_PARSER_LABEL_STATE_DISTRICT)) { class_id = ADDRESS_PARSER_STATE_DISTRICT; component = ADDRESS_COMPONENT_STATE_DISTRICT;