osm-search/Nominatim
1
1
Fork 0
mirror of https://github.com/osm-search/Nominatim.git synced 2024-11-24 06:22:08 +03:00
Code Issues Packages Projects Releases Wiki Activity
2,126 Commits 28 Branches 41 Tags 164 MiB
d643ca8dee
Commit Graph

367 Commits

Author SHA1 Message Date
Sarah Hoffmann
d643ca8dee move address line computation in its own function 2020-02-21 16:38:14 +01:00
Sarah Hoffmann
6189e0c79b improve parenting for large areas with rank 30 
Instead of unconditionally parenting them to a street, the
larger areas get a parent area that contains them. To keep
things computationally light-weight, only use the centroid and
bbox to determine if an area is contained.

Requires renaming of parenting functions because renaming
a parameter of the function causes issues when updating the
function (it requires a manual delete, which I'd like to
avoid).
 2020-02-19 08:43:53 +01:00
Sarah Hoffmann
92c5d3b720 make sure that linked places are within a boundary 
This is a regression from previous code refactoring.

Fixes #1684.
 2020-02-18 22:46:32 +01:00
Sarah Hoffmann
2a6e8ad68e add bbox whereclause to make postgis 3.0 happy 
Normally ST_Covers() should include a bbox index use,
so adding a bbox where clause is not really necessary.
However, the query planner messes up and uses a parallel
index search with a second index instead of exclusively
running on the geometry index, when the bbox part is
missing.
 2020-02-16 14:10:22 +01:00
Sarah Hoffmann
b9171dd10b clean up intermediate tables earlier with --drop 
When --drop is given, we can remove all node geometry information
already after the import with osm2pgsql. Also drop all unnecessary
tables before creating the final indices.
 2020-02-12 11:03:20 +01:00
Sarah Hoffmann
5ec25122f6 rename functions where return parameter changed 
Postgresql cannot cleanly reimport these functions when
upgrading, so simply rename to avoid errors.
 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
9371b1aeb9 forgot new trigger sql 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
6f6d116451 adapt index for changes name lookup 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
3ff6eccfd7 move trigger creation later in setup 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
5d1fa597ea clean up get_word_id function 
Replaced by addr_ids_from_name() which also normalises the
string.
 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
3b6c2c9155 getNearestNamed*Feature functions better return values 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
f863040b38 factor out parent search from addr:street/addr:place 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
1033f8bce7 factor out searching for parent road for pois 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
cf4dbbd681 remove unused function 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
6dccc693d0 factor out computation of default names 
Also moves the computation down the line so that we never
have to do it twice.
 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
c3dc66ce9c factor out place linking sql 2020-01-23 22:28:43 +01:00
Sarah Hoffmann
ddaf1b79d4 remove special handling of rail 
Skip railway=rail in the style, so that installations can remove
it if they wish.
 2020-01-08 23:53:23 +01:00
Sarah Hoffmann
7a194789bc remove remaining sql functions into function/ directory 2020-01-08 11:45:51 +01:00
Sarah Hoffmann
827d7a9a62 move postcode table triggers to own file 2020-01-08 11:22:23 +01:00
Sarah Hoffmann
dae2761137 move placex triggers into own file 2020-01-08 11:18:42 +01:00
Sarah Hoffmann
4304c1a7bb move place triggers into own file 2020-01-07 23:55:38 +01:00
Sarah Hoffmann
c537ea18a4 move functions for interpolation table in own file 2020-01-05 16:36:46 +01:00
Sarah Hoffmann
28fa7be75a move functions for address lookup into own file 2020-01-05 16:16:21 +01:00
Sarah Hoffmann
f1a5862f3d move creation function for aux data into own file 
This function is currently unused, so don't even load it.
 2020-01-05 16:04:04 +01:00
Sarah Hoffmann
4088e4e371 move importance/wikipedia functions into separate file 2020-01-05 15:55:39 +01:00
Sarah Hoffmann
0ef6425847 move SQL functions for normalisation in separate file 2020-01-05 15:38:20 +01:00
Sarah Hoffmann
12f830fbbb fix loading of data without wikipedia files 
Also removes unused place_boundingbox table.
 2019-11-17 23:58:43 +01:00
Sarah Hoffmann
6d764a05b0 prefer English wikipedia for wikidata matches 2019-11-17 19:53:42 +01:00
Sarah Hoffmann
e4555a208d fix wikidata import 
The loop was not skipping empty results of get_wikipedia_match().
 2019-11-17 10:06:34 +01:00
Sarah Hoffmann
d1a9dc0f24 use wikidata links for importance as well 2019-11-16 16:52:23 +01:00
Sarah Hoffmann
5fb850982a move importance computation into its own function 2019-11-16 16:52:23 +01:00
Sarah Hoffmann
e3e9f69654 fix rank of postcode results 
Fixes #1487.
 2019-10-19 17:57:57 +02:00
Sarah Hoffmann
34a4a9b08f create/drop indexes concurrently 
Fixes #1507.
 2019-10-19 17:13:05 +02:00
Sarah Hoffmann
511204c158
Merge pull request #1443 from lonvia/reorganize-search-name-partition-tables 
Reorganize search name partition tables
 2019-07-28 15:18:12 +02:00
Sarah Hoffmann
79b81d39d8 streamline indexes of search_name partition tables 
Remove index on name_vector. We always do near search where the
geometry index is sufficient. Also split centroid index in low
and high rank indexes. Reduces index size by about 25%.
 2019-07-28 13:29:35 +02:00
Sarah Hoffmann
2bbe5017d4 use bbox of geometry when searching for attached streets 
As we are doing a distance search, this improves results for
large places like airports.

Fixes #1442.
 2019-07-28 13:28:27 +02:00
marc tobias
765a932561 make US postcode data to an optional download 2019-07-24 01:13:57 +02:00
marc tobias
0e44659033 during TIGER import skip records with negative house number range 2019-07-21 21:41:12 +02:00
Sarah Hoffmann
be47cd2549 remove country from place_addressline 
The country information can be determined sufficiently from
the country code. We only loose the specific OSM object
behind the address.

Also streamlines the location_area_country table.
 2019-07-10 21:29:47 +02:00
Sarah Hoffmann
4c1793b4e3 recreate interpolations when one of their support nodes changes 
A simple update is not enough because the interpolation splits
might change as well as the housenumbers.

Fixes #1360.
 2019-07-03 23:15:54 +02:00
Sarah Hoffmann
d1ca73f813 Reset housenumber on every place update 
As it is a computed field, it needs to be computed from scratch
to take into account any surrounding changes.

Fixes #1395.
 2019-07-03 20:56:35 +02:00
Sarah Hoffmann
b0e6fb73c6 generally remove all country and state places from address 
Gets rid of the hard-coded expection for place nodes and sets
the address rank generally via the address level config instead.
That means only administrative boundaries are now used at that
level in addresses.
 2019-06-30 22:27:44 +02:00
Sarah Hoffmann
e05e413cc4 use real centroid when looking for near roads 
The point-on-surface may be at the corner in large objects, so
that roads are too far away.

Fixes #1389.
 2019-06-10 18:23:12 +02:00
marc tobias
10fbda702b exclude postcode ranges separated by colon from centre point calculation 2019-05-25 20:43:38 +02:00
Sarah Hoffmann
403ee260f6 Ensure that postcode relations are used in addresses 
Postcode nodes are normally thrown away as they only play
a role for computing artifical postcodes. However, if we
have a postcode area this still should take part of the
address.

Fixes #1330.
 2019-04-14 11:20:03 +02:00
Sarah Hoffmann
189da9afb3 add osm_id index for osmline table 
Needed when deleting address interpolation.
 2019-02-21 23:26:31 +01:00
marc tobias
853b536394 Remove get_addressrank_label. Move get_searchrank_label to PHP 2019-02-09 20:38:36 +01:00
Sarah Hoffmann
3811d916b9 remove self-reference from place-addressline 2019-02-09 15:26:10 +01:00
Sarah Hoffmann
c822012aad ignore admin boundary ways for countries and states 
Countries and states are mapped world-wide as relations by now.

Fixes #543 and #1291.
 2019-01-26 13:37:10 +01:00
Sarah Hoffmann
cc17aa8d6b Remove postcodes also from word table when they no longer exist 
Also adds tests for postcode updates.

Fixes #1273.
 2019-01-04 23:11:47 +01:00