Commit Graph

3814 Commits

Author SHA1 Message Date
Al
c9c1d912b0 [dictionaries] removing spaces in the Japanese dictionaries 2016-09-12 22:21:27 -04:00
Al
14c20091f4 [fix] abbreviations in hyphenated phrases like Saint-Germaine. Hyphenation should use the phrase length not the token length 2016-09-12 22:20:25 -04:00
Al
0f8e7cd9dc [boundaries] adding exception for the direct-controlled municipalities in China (Shanghai, Beijing, Tianjin and Chongqing) to be treated as city instead of state. These will use the admin_centre properties 60% of the time, relation propreties 40% of the time 2016-09-12 18:46:08 -04:00
Al
5bd2fac514 [openaddresses] CITY => city_district in distrito-federal (Mexico City), add OSM boundaries to pick up Ciudad de México 2016-09-12 18:24:40 -04:00
Al
62a5f6002d [geonames] adding query method to GeoNamesDB 2016-09-12 16:42:42 -04:00
Al
e8408d39fd [fix] unzip_file checks status code 2016-09-12 16:42:02 -04:00
Al
23c9fbe3fb [wof] don't need to crawl for admin data now that whosonfirst-data is separate from venues, etc. 2016-09-12 13:52:49 -04:00
Al
2057536bd9 [requirements] adding boto3 and gevent to requirements.txt for geodata package 2016-09-12 04:09:15 -04:00
Al
6fd43a1f20 [wof] script to download all the WoF postal code repos and their dependencies 2016-09-12 04:03:41 -04:00
Al
d41905aa18 [wof] geven-based hierarchy crawler for WoF 2016-09-12 04:03:02 -04:00
Al
45338b43b5 [wof] adding basic WhosOnFirst client for interaction with S3, downloading by ID, etc. 2016-09-12 03:56:27 -04:00
Al
eaacf55f42 [boundaries] exceptions for city_district and barrios in Buenos Aires and Rosario in Argentina 2016-09-12 02:57:49 -04:00
Al
55e9ab1978 [places] adding world_region tag and adding the phrase West Indies with small random probability for English-speaking Caribbean nations. Ref: #113 2016-09-11 21:54:56 -04:00
Al
069e4c348c [openaddresses] adding a few US counties from the latest PR 2016-09-11 16:19:39 -04:00
Al
917493834d [dictionaries] removing R as an abbreviation for river, can re-add it in a street name only dictionary if needed, but don't want to abbreviate towns with River in the name simply to R 2016-09-10 16:19:31 -04:00
Al
551cce8cb1 [fix] making a separate gazetteer for toponym abbreviations 2016-09-10 01:08:58 -04:00
Al
bcde9e2fe7 [fix] toponym abbreviations after country name, may want to use it 2016-09-10 00:49:31 -04:00
Al
bbc5131cb6 [fix] toponym abbreviations 2016-09-10 00:48:31 -04:00
Al
19a044f7f3 [fix] imports 2016-09-10 00:09:11 -04:00
Al
ae02b0769d [openaddresses] abbreviating boundary components for OpenAddresses 2016-09-10 00:04:11 -04:00
Al
604e898d65 [fix] using toponym_gazetteer in OSM boundary abbreviations 2016-09-10 00:02:59 -04:00
Al
8e86b5caf7 [addresses] reverting AUS probabilties 2016-09-09 23:39:07 -04:00
Al
543d2335b7 [addresses] alternatives 2016-09-09 23:17:15 -04:00
Al
5d26ab41e7 [openaddresses] removing OpenAddresses hacks now that upstream changes are merged 2016-09-09 09:40:45 -04:00
Al
9cdcd7f21a [fix] indentation 2016-09-09 08:59:50 -04:00
Al
a14202fc7a [fix] default value 2016-09-09 01:46:03 -04:00
Al
85ad3bf0f4 [formatting] allowing a non-default option for components that can be inserted between road and house number 2016-09-09 01:38:39 -04:00
Al
4c6bcda3b2 [fix] config 2016-09-08 15:21:19 -04:00
Al
d1e3c6a24a [openaddresses] adding Italy countrywide to a pre_release_downloads set so it can be used in libpostal without having been merged yet 2016-09-08 15:16:35 -04:00
Al
0edbe5a593 [formatting] don't allow insertions between house number and road name 2016-09-08 15:00:36 -04:00
Al
a329503155 [formatting] adding structural exceptions for house_number before road in continental Europe (and house_number after road in France). Ref: https://github.com/pelias/api/pull/655 2016-09-08 12:10:57 -04:00
Al
170e8d74d8 [fix] checking for components 2016-09-08 03:19:10 -04:00
Al
769a65b808 [openaddresses] adding place-only and place+postcode probability to OpenAddresses to capture more place names not in OSM as standalone queries 2016-09-08 03:17:21 -04:00
Al
6ffd697d7e [places] using probability 1.0 for cities in the places config. Can do full dropouts of place components separately 2016-09-07 17:22:08 -04:00
Al
6543afaaa3 [utils] checking for 200 from download_file 2016-09-07 14:13:42 -04:00
Al
317e4caca9 [fix] only percent quote the filename 2016-09-07 13:55:34 -04:00
Al
f061d4239b [fix] quote, not urlencode 2016-09-07 13:51:08 -04:00
Al
5c915855c9 [fix] urlencode 2016-09-07 13:48:39 -04:00
Al
62c8fa9048 [fix] encoding III 2016-09-07 13:39:38 -04:00
Al
6e5908385a [fix] encoding again 2016-09-07 13:34:04 -04:00
Al
3bc122f7a0 [fix] encoding 2016-09-07 13:28:18 -04:00
Al
cb3fe5273a [components] using gnis:class=Populated Place to map to city for the US when admin_level is not specified and the place key is not specified/not mapped 2016-09-07 11:49:44 -04:00
Al
88be23c85b [boundaries] adding borough mapping to Mexico as it's used for Mexico City 2016-09-07 11:25:06 -04:00
Al
ec24c4b6ac [boundaries] removing the place=borough mapping because it's used on the east coast and PA to mean city 2016-09-07 11:24:37 -04:00
Al
978c436071 [boundaries] added Kingston, Jamaica as an admin_centre in OSM for Kingston and St Andrew Corporation in Jamaica. Overriding with a 0.9 probability so the full name of the combined parish gets used (as "state") some of the time, but mostly just stands in as a city boundary for Kingston 2016-09-06 15:27:09 -04:00
Al
6ff643f5a9 [addresses] adding "Lot" with more probability in the default/residential zone for English. Increasing sampling probabilities for AU/NZ 2016-09-06 01:21:51 -04:00
Al
7bcf29cb69 [openaddresses] adding Cape Town for city since it's not quite in OSM as such 2016-09-03 12:37:33 -04:00
Al
7e7ee7462a [fix] dutch house number formatting, strip spaces 2016-09-02 14:47:52 -04:00
Al
95384e5a2c [openaddresses adding hack for Honolulu until join function can handle null in OpenAddresses 2016-09-02 14:29:40 -04:00
Al
eab8765ac8 [openaddresses] adding Honolulu 2016-09-02 14:28:53 -04:00