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

[abbreviations] Adding toponyms gazetteer for probabilistically abbreviating things like Mount=>Mt, Saint=>St, Fort=>Ft in place names

Browse Source
This commit is contained in:
Al
2016-08-24 18:52:00 -04:00
parent dfa5c8e0a6
commit 2e7f8f1ae7
1 changed files with 8 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
scripts/geodata/address_expansions/gazetteers.py
View File

@@ -157,6 +157,13 @@ HOUSE_NUMBER_DICTIONARIES = ('house_number', 'no_number')
POSTCODE_DICTIONARIES = ('postcode',)
TOPONYM_ABBREVIATION_DICTIONARIES = STREET_TYPES_DICTIONARIES + ('qualifiers',
'personal_titles',
'synonyms',
'toponyms',
)
UNIT_ABBREVIATION_DICTIONARIES = ('level_types_basement',
'level_types_mezzanine',
'level_types_numbered',
@@ -191,4 +198,5 @@ chains_gazetteer = create_gazetteer(CHAIN_DICTIONARY)
unit_types_gazetteer = create_gazetteer(*UNIT_ABBREVIATION_DICTIONARIES)
street_and_synonyms_gazetteer = create_gazetteer(*(STREET_TYPES_DICTIONARIES + (SYNONYM_DICTIONARY, )))
abbreviations_gazetteer = create_gazetteer(*ALL_ABBREVIATION_DICTIONARIES)
toponym_gazetteer = create_gazetteer(*TOPONYM_ABBREVIATION_DICTIONARIES)
given_name_gazetteer = create_gazetteer(GIVEN_NAME_DICTIONARY)