]> git.openstreetmap.org Git - nominatim.git/log
nominatim.git
2 years agoMerge pull request #2685 from lonvia/show-inherited-housenumber
Sarah Hoffmann [Fri, 29 Apr 2022 10:15:33 +0000 (12:15 +0200)]
Merge pull request #2685 from lonvia/show-inherited-housenumber

Keep address parts inherited from surrounding buildings after indexing

2 years agoMerge pull request #2684 from lonvia/translit-keep-spacing-marks
Sarah Hoffmann [Fri, 29 Apr 2022 08:38:28 +0000 (10:38 +0200)]
Merge pull request #2684 from lonvia/translit-keep-spacing-marks

ICU: better letter identification in normalization

2 years agokeep inherited address parts after indexing
Sarah Hoffmann [Thu, 28 Apr 2022 19:38:00 +0000 (21:38 +0200)]
keep inherited address parts after indexing

The inherited housenumber is needed for display output. We can't
take the one from the housenumber field because it is already
normalized. Remove the inherited address only when reindexing.

Fixes #2683.

2 years agoICU: better letter identification in normalization
Sarah Hoffmann [Thu, 28 Apr 2022 15:20:56 +0000 (17:20 +0200)]
ICU: better letter identification in normalization

The Letter class does not include non-spacing marks that can also
have a consonant or vowel meaning, especially in Indian languages.
Use the alnum propoerty instead which includes them all. Also
include the vowel-canceling Virama, which is not a letter by itself
but changes the transliteration.

2 years agoMerge pull request #2678 from lonvia/address-part-order
Sarah Hoffmann [Fri, 22 Apr 2022 18:32:10 +0000 (20:32 +0200)]
Merge pull request #2678 from lonvia/address-part-order

Change selection of primary address part for ways that cross boundaries

2 years agofurther tweaking of address distance
Sarah Hoffmann [Fri, 22 Apr 2022 12:32:19 +0000 (14:32 +0200)]
further tweaking of address distance

For point features, keep using the distance to centroid.
For area features, add a tie breaker for the case where the
center point falls on the boundary.

2 years agochange distance computation between place and address part
Sarah Hoffmann [Thu, 21 Apr 2022 19:56:59 +0000 (21:56 +0200)]
change distance computation between place and address part

Instead of computing the distance to the centroid of the area
compute the distance of the area to the centroid of the feature.
This means we give preference to the area that covers the centroid.
It's still a heuristics but one that is a bit less random.

2 years agoMerge pull request #2666 from lonvia/admin-command-for-forced-indexing
Sarah Hoffmann [Thu, 14 Apr 2022 19:44:08 +0000 (21:44 +0200)]
Merge pull request #2666 from lonvia/admin-command-for-forced-indexing

Admin command for forced indexing

2 years agoadd documentation for new refresh command
Sarah Hoffmann [Thu, 14 Apr 2022 13:10:24 +0000 (15:10 +0200)]
add documentation for new refresh command

2 years agoadd tests for new data invalidation functions
Sarah Hoffmann [Thu, 14 Apr 2022 12:52:13 +0000 (14:52 +0200)]
add tests for new data invalidation functions

2 years agoadd new commands for forced invalidation before indexing
Sarah Hoffmann [Wed, 13 Apr 2022 20:00:37 +0000 (22:00 +0200)]
add new commands for forced invalidation before indexing

2 years agorecommend PostgreSQL 13+
Sarah Hoffmann [Thu, 14 Apr 2022 07:21:25 +0000 (09:21 +0200)]
recommend PostgreSQL 13+

See https://github.com/osm-search/Nominatim/discussions/2659.

2 years agodocs: different default for format on osm.org
Sarah Hoffmann [Fri, 8 Apr 2022 15:13:42 +0000 (17:13 +0200)]
docs: different default for format on osm.org

Add a note that the format parameter is needed for
nominatim.openstreetmap.org for historical reasons.

2 years agoMerge pull request #2660 from lonvia/pyosmium-contextmanager
Sarah Hoffmann [Fri, 8 Apr 2022 15:07:33 +0000 (17:07 +0200)]
Merge pull request #2660 from lonvia/pyosmium-contextmanager

Support using ReplicationServer as contextmanager

2 years agosupport new ReplicationServer as contextmanager
Sarah Hoffmann [Thu, 7 Apr 2022 15:58:04 +0000 (17:58 +0200)]
support new ReplicationServer as contextmanager

2 years agoMerge pull request #2466 from I70l0teN4ik/state-code
Sarah Hoffmann [Thu, 7 Apr 2022 14:39:50 +0000 (16:39 +0200)]
Merge pull request #2466 from I70l0teN4ik/state-code

add ISO3166-2-lvl<admin_level> field to response address details

2 years agofix bdd tests and docs
Artem Ziablytskyi [Thu, 7 Apr 2022 12:09:08 +0000 (14:09 +0200)]
fix bdd tests and docs

2 years agouse ISO3166-2-lvl<admin_level> instead of typeLabel prefix
Artem Ziablytskyi [Wed, 6 Apr 2022 17:24:39 +0000 (19:24 +0200)]
use ISO3166-2-lvl<admin_level> instead of typeLabel prefix

2 years agoproper instruction to import data
Artem Ziablytskyi [Wed, 23 Mar 2022 13:32:03 +0000 (14:32 +0100)]
proper instruction to import data

2 years agoFix API docs and Vagrant instructions to import data
Artem Ziablytskyi [Thu, 10 Mar 2022 09:57:26 +0000 (10:57 +0100)]
Fix API docs and Vagrant instructions to import data

2 years agoFix API docs and Vagrant instructions to import data
Artem Ziablytskyi [Thu, 10 Mar 2022 09:57:26 +0000 (10:57 +0100)]
Fix API docs and Vagrant instructions to import data

2 years agoChange the key to `<addresspart_type>-ISO3166-2` to support xml response correctly
Artem Ziablytskyi [Wed, 26 Jan 2022 16:40:39 +0000 (17:40 +0100)]
Change the key to `<addresspart_type>-ISO3166-2` to support xml response correctly

2 years agoadd `<addresspart_type>:ISO3166-2` field to response address details
Artem Ziablytskyi [Mon, 17 Jan 2022 16:20:06 +0000 (17:20 +0100)]
add `<addresspart_type>:ISO3166-2` field to response address details

2 years agoadd `state_code` field to response address details
Artem Ziablytskyi [Tue, 12 Oct 2021 15:35:15 +0000 (17:35 +0200)]
add `state_code` field to response address details

2 years agoMerge pull request #2629 from tareqpi/country-names-yaml-configuration
Sarah Hoffmann [Mon, 4 Apr 2022 07:04:25 +0000 (09:04 +0200)]
Merge pull request #2629 from tareqpi/country-names-yaml-configuration

Move default country names into yaml configuration

2 years agofix linting error
Tareq Al-Ahdal [Fri, 1 Apr 2022 16:14:18 +0000 (00:14 +0800)]
fix linting error

2 years agoremove the conversion to json for name
Tareq Al-Ahdal [Fri, 1 Apr 2022 14:54:14 +0000 (22:54 +0800)]
remove the conversion to json for name

2 years ago'read_config' is no longer a fixture
Tareq Al-Ahdal [Fri, 1 Apr 2022 14:52:17 +0000 (22:52 +0800)]
'read_config' is no longer a fixture

add 'read_config' to test cases that need it

2 years agotest for loading special characters from country_settings.yaml
Tareq Al-Ahdal [Fri, 1 Apr 2022 13:58:57 +0000 (21:58 +0800)]
test for loading special characters from country_settings.yaml

2 years agofix reset country info before loading custom data
Tareq Al-Ahdal [Fri, 1 Apr 2022 13:55:34 +0000 (21:55 +0800)]
fix reset country info before loading custom data

2 years agocustom country config loads correctly
Tareq Al-Ahdal [Fri, 1 Apr 2022 13:46:56 +0000 (21:46 +0800)]
custom country config loads correctly

2 years agoMerge pull request #2655 from lonvia/migration-internal-country-name
Sarah Hoffmann [Thu, 31 Mar 2022 16:04:18 +0000 (18:04 +0200)]
Merge pull request #2655 from lonvia/migration-internal-country-name

Add migration for new country name handling in ICU tokenizer

2 years agoadd test for new table_has_column function
Sarah Hoffmann [Thu, 31 Mar 2022 13:50:02 +0000 (15:50 +0200)]
add test for new table_has_column function

2 years agoadd migration to mark internal country names
Sarah Hoffmann [Thu, 31 Mar 2022 13:43:46 +0000 (15:43 +0200)]
add migration to mark internal country names

2 years agoseparate unit test function into three functions
Tareq Al-Ahdal [Wed, 30 Mar 2022 14:06:59 +0000 (22:06 +0800)]
separate unit test function into three functions

2 years agoremove temporary file
Sarah Hoffmann [Wed, 30 Mar 2022 13:13:31 +0000 (15:13 +0200)]
remove temporary file

2 years agoMerge pull request #2650 from mtmail/update-lookup-examples
Sarah Hoffmann [Mon, 28 Mar 2022 18:21:45 +0000 (20:21 +0200)]
Merge pull request #2650 from mtmail/update-lookup-examples

documentation: update example output of lookup endpoint

2 years agodocumentation: update example output of lookup endpoint
Marc Tobias [Mon, 28 Mar 2022 14:41:10 +0000 (16:41 +0200)]
documentation: update example output of lookup endpoint

2 years agofix edge case handling when 'names' is not there
Tareq Al-Ahdal [Fri, 25 Mar 2022 14:25:55 +0000 (22:25 +0800)]
fix edge case handling when 'names' is not there

2 years agoAdded unit tests for loading country info from yaml file
Tareq Al-Ahdal [Fri, 25 Mar 2022 14:22:44 +0000 (22:22 +0800)]
Added unit tests for loading country info from yaml file

2 years agofix linting error
Tareq Al-Ahdal [Thu, 24 Mar 2022 05:27:24 +0000 (13:27 +0800)]
fix linting error

2 years agofix storing of escape sequences in database
Tareq Al-Ahdal [Thu, 24 Mar 2022 05:18:44 +0000 (13:18 +0800)]
fix storing of escape sequences in database

2 years agoclean up
Tareq Al-Ahdal [Thu, 24 Mar 2022 05:16:59 +0000 (13:16 +0800)]
clean up

2 years agomerge of insert query + modularity enhancements
Tareq Al-Ahdal [Thu, 24 Mar 2022 05:13:38 +0000 (13:13 +0800)]
merge of insert query + modularity enhancements

2 years agoMerge pull request #2641 from lonvia/reinit-tokenizer-dir
Sarah Hoffmann [Sun, 20 Mar 2022 20:46:07 +0000 (21:46 +0100)]
Merge pull request #2641 from lonvia/reinit-tokenizer-dir

Transparantly reinitialize tokenizer directory when necessary

2 years agoreinit the tokenizer directory on website refresh
Sarah Hoffmann [Sun, 20 Mar 2022 16:49:22 +0000 (17:49 +0100)]
reinit the tokenizer directory on website refresh

This means the project directory is usable again, once refresh --website
was run.

2 years agodocs: copying project dir no longer necessary
Sarah Hoffmann [Sun, 20 Mar 2022 15:01:27 +0000 (16:01 +0100)]
docs: copying project dir no longer necessary

2 years agorestore the tokenizer directory when missing
Sarah Hoffmann [Sun, 20 Mar 2022 10:31:42 +0000 (11:31 +0100)]
restore the tokenizer directory when missing

Automatically repopulate the tokenizer/ directory with the PHP stub
and the postgresql module, when the directory is missing. This allows
to switch working directories and in particular run the service
from a different maschine then where it was installed.
Users still need to make sure that .env files are set up correctly
or they will shoot themselves in the foot.

See #2515.

2 years agocache loaded configuration
Sarah Hoffmann [Sun, 20 Mar 2022 10:30:03 +0000 (11:30 +0100)]
cache loaded configuration

Reading the YAML files is fairly expensive and slows down the BDD tests
significantly. Therefore cache the results from reading the file.

2 years agoMerge pull request #2639 from lonvia/remove-operator
Sarah Hoffmann [Fri, 18 Mar 2022 15:42:18 +0000 (16:42 +0100)]
Merge pull request #2639 from lonvia/remove-operator

No longer use operator tag as a name

2 years agofix linting error
Tareq Al-Ahdal [Fri, 18 Mar 2022 13:05:47 +0000 (21:05 +0800)]
fix linting error

2 years agoremove special case for operator names
Sarah Hoffmann [Fri, 18 Mar 2022 09:48:53 +0000 (10:48 +0100)]
remove special case for operator names

The OSM data has been sufficiently cleaned up by now that
the operator no longer needs to be considered a name tag.
Use 'brand' as the searchable alternative.

2 years agomodify logic of _include_key function
Tareq Al-Ahdal [Thu, 17 Mar 2022 22:52:16 +0000 (06:52 +0800)]
modify logic of _include_key function

2 years agoadding prefix to keys is now more configurable
Tareq Al-Ahdal [Thu, 17 Mar 2022 22:20:00 +0000 (06:20 +0800)]
adding prefix to keys is now more configurable

2 years agoRevert the removal of new line at the end of the file
Tareq Al-Ahdal [Thu, 17 Mar 2022 22:07:48 +0000 (06:07 +0800)]
Revert the removal of new line at the end of the file

2 years agomodify logic to prepend 'name:' to keys'
Tareq Al-Ahdal [Thu, 17 Mar 2022 22:01:25 +0000 (06:01 +0800)]
modify logic to prepend 'name:' to keys'

2 years agoreattach 'name:' prefix to keys
Tareq Al-Ahdal [Thu, 17 Mar 2022 21:46:23 +0000 (05:46 +0800)]
reattach 'name:' prefix to keys

2 years agoinitialize an empty dictionary for nested name key
Tareq Al-Ahdal [Thu, 17 Mar 2022 18:50:33 +0000 (02:50 +0800)]
initialize an empty dictionary for nested name key

2 years agoremove duplicate values
Tareq Al-Ahdal [Thu, 17 Mar 2022 18:43:42 +0000 (02:43 +0800)]
remove duplicate values

2 years agofix comment
Tareq Al-Ahdal [Thu, 17 Mar 2022 18:38:04 +0000 (02:38 +0800)]
fix comment

2 years agoMerge branch 'master' into country-names-yaml-configuration
Tareq Al-Ahdal [Thu, 17 Mar 2022 18:36:12 +0000 (02:36 +0800)]
Merge branch 'master' into country-names-yaml-configuration

2 years agoReformatting of country keys
Tareq Al-Ahdal [Thu, 17 Mar 2022 18:23:11 +0000 (02:23 +0800)]
Reformatting of country keys

2 years agoMerge pull request #2637 from lonvia/keep-linked-place-names
Sarah Hoffmann [Thu, 17 Mar 2022 15:39:30 +0000 (16:39 +0100)]
Merge pull request #2637 from lonvia/keep-linked-place-names

Introduce separation of names from linked places

2 years agoadapt ParameterParser tests to new key list
Sarah Hoffmann [Thu, 17 Mar 2022 10:45:05 +0000 (11:45 +0100)]
adapt ParameterParser tests to new key list

2 years agofix linting
Sarah Hoffmann [Thu, 17 Mar 2022 10:05:32 +0000 (11:05 +0100)]
fix linting

2 years agomerge linked names correctly into namedetails
Sarah Hoffmann [Thu, 17 Mar 2022 10:02:02 +0000 (11:02 +0100)]
merge linked names correctly into namedetails

Convert the '_place_*' entries back to normal entries before
returning them in the 'namedetails' section. If the name field is
duplicated, kept the '_place_*' notation. This preserves the previous
behaviour before _place_ names were introduces but adds the additional
names from the linked place for reference.

2 years agomake sure outputs take into account linked place names
Sarah Hoffmann [Wed, 16 Mar 2022 20:44:52 +0000 (21:44 +0100)]
make sure outputs take into account linked place names

2 years agofix return code for PHP exceptions
Sarah Hoffmann [Wed, 16 Mar 2022 20:44:02 +0000 (21:44 +0100)]
fix return code for PHP exceptions

These have returned a 0 until now.

2 years agosave differing linked polace names in extra fields
Sarah Hoffmann [Wed, 16 Mar 2022 15:38:52 +0000 (16:38 +0100)]
save differing linked polace names in extra fields

This keeps the names tracable and ensures that all names are searchable
when they differ. Do not keep names when they are exactly the same
to save some space. Linked names are cleaned out before relinking.

2 years agoMerge pull request #2633 from lonvia/fix-reverse-single-interpolation-point
Sarah Hoffmann [Wed, 16 Mar 2022 13:22:59 +0000 (14:22 +0100)]
Merge pull request #2633 from lonvia/fix-reverse-single-interpolation-point

Correctly handle single-point interpolations in reverse

2 years agouse https protocol for cloning from github
Sarah Hoffmann [Wed, 16 Mar 2022 11:05:58 +0000 (12:05 +0100)]
use https protocol for cloning from github

Does not need authentication.

2 years agocorrectly handle single-point interpolations in reverse
Sarah Hoffmann [Wed, 16 Mar 2022 10:19:09 +0000 (11:19 +0100)]
correctly handle single-point interpolations in reverse

Lookup in location_property_osmline needs to be special cased
for startnumber = endnumber. Also adds tests for the case.

Fixes #2680.

2 years agofix linting error
Tareq Al-Ahdal [Tue, 15 Mar 2022 11:14:04 +0000 (19:14 +0800)]
fix linting error

2 years agoMerge pull request #2632 from nslxndr/fix-log-typo
Sarah Hoffmann [Tue, 15 Mar 2022 10:01:57 +0000 (11:01 +0100)]
Merge pull request #2632 from nslxndr/fix-log-typo

Fix typo in log message on replication initialisation

2 years agoFix typo in log message on replication initialisation
Sandor Nagy [Tue, 15 Mar 2022 06:50:47 +0000 (07:50 +0100)]
Fix typo in log message on replication initialisation

2 years agoreintroduce 'name:' prefix to country name keys
Tareq Al-Ahdal [Sun, 13 Mar 2022 10:58:27 +0000 (18:58 +0800)]
reintroduce 'name:' prefix to country name keys

2 years agoRemove country.sql from CMakeLists.txt
Tareq Al-Ahdal [Sun, 13 Mar 2022 10:56:19 +0000 (18:56 +0800)]
Remove country.sql from CMakeLists.txt

2 years agomodify data import logic to load country names from yaml
Tareq Al-Ahdal [Sat, 12 Mar 2022 07:20:57 +0000 (15:20 +0800)]
modify data import logic to load country names from yaml

2 years agomove default country names into yaml configuration
Tareq Al-Ahdal [Sat, 12 Mar 2022 07:17:01 +0000 (15:17 +0800)]
move default country names into yaml configuration

2 years agoMerge pull request #2627 from mtmail/location-of-osm2pgsql
Sarah Hoffmann [Thu, 10 Mar 2022 14:39:10 +0000 (15:39 +0100)]
Merge pull request #2627 from mtmail/location-of-osm2pgsql

documentation: clarify osm2pgsql isnt in project directory by default

2 years agodocumentation: clarify osm2pgsql isnt in project directory by default
Marc Tobias [Thu, 10 Mar 2022 13:16:12 +0000 (14:16 +0100)]
documentation: clarify osm2pgsql isnt in project directory by default

2 years agoMerge pull request #2621 from lonvia/housenumber-analyzer
Sarah Hoffmann [Tue, 1 Mar 2022 14:19:07 +0000 (15:19 +0100)]
Merge pull request #2621 from lonvia/housenumber-analyzer

Introduce optional token analysis for housenumbers

2 years agodo not expand records in select list
Sarah Hoffmann [Tue, 1 Mar 2022 07:54:15 +0000 (08:54 +0100)]
do not expand records in select list

An expression of the form 'SELECT (func()).*' will be expanded
by Postgresql _before_ execution with the result that the function
will be called as many times as there are fields in the record.
This is not what we want. The function call needs to go into
the FROM clause instead.

2 years agofix linting issue
Sarah Hoffmann [Fri, 18 Feb 2022 22:24:50 +0000 (23:24 +0100)]
fix linting issue

2 years agoadd tests for new analyzed housenumbers
Sarah Hoffmann [Fri, 18 Feb 2022 20:51:19 +0000 (21:51 +0100)]
add tests for new analyzed housenumbers

2 years agoadapt housenumber cleanup to new word table structure
Sarah Hoffmann [Fri, 18 Feb 2022 15:43:17 +0000 (16:43 +0100)]
adapt housenumber cleanup to new word table structure

2 years agobdd: disable some housenumber tests for legacy
Sarah Hoffmann [Fri, 18 Feb 2022 14:58:38 +0000 (15:58 +0100)]
bdd: disable some housenumber tests for legacy

Optional spaces in housenumbers are not supported by legacy tokenizer,
so disable those tests.

2 years agoadd documentation for housenumber analyzer
Sarah Hoffmann [Fri, 18 Feb 2022 14:55:10 +0000 (15:55 +0100)]
add documentation for housenumber analyzer

2 years agohousenumber analyzer: avoid creating too many variants
Sarah Hoffmann [Wed, 16 Feb 2022 19:36:30 +0000 (20:36 +0100)]
housenumber analyzer: avoid creating too many variants

Housenumber fields with lots of text are likely bad data. So is
data with many changes from letter to digit. Exclude them from adding
optional spaces.

2 years agoadd new analyser for houenumbers
Sarah Hoffmann [Wed, 16 Feb 2022 16:18:23 +0000 (17:18 +0100)]
add new analyser for houenumbers

This analyser makes spaces optional.

2 years agoadd framework for analysing housenumbers
Sarah Hoffmann [Wed, 16 Feb 2022 10:15:43 +0000 (11:15 +0100)]
add framework for analysing housenumbers

This lays the groundwork for adding variants for housenumbers.
When analysis is enabled, then the 'word' field in the word table
is used as usual, so that variants can be created. There will be
only one analyser allowed which must have the fixed name
'@housenumber'.

2 years agoicu: move token deduplication into TokenInfo
Sarah Hoffmann [Tue, 15 Feb 2022 20:51:49 +0000 (21:51 +0100)]
icu: move token deduplication into TokenInfo

Puts collection into one common place.

2 years agoicu: move housenumber token computation out of TokenInfo
Sarah Hoffmann [Tue, 15 Feb 2022 20:20:47 +0000 (21:20 +0100)]
icu: move housenumber token computation out of TokenInfo

This was the last function to use the cache. There is a more clean
separation of responsibility now.

2 years agohandle unknown analyzer
Sarah Hoffmann [Tue, 15 Feb 2022 13:38:03 +0000 (14:38 +0100)]
handle unknown analyzer

When changing something in the default configuration of the sanatizers
that refers to an analyzer that is not yet loaded, there shouldn't be
any errors.

2 years agomove generation of normalized token form to analyzer
Sarah Hoffmann [Tue, 15 Feb 2022 11:15:18 +0000 (12:15 +0100)]
move generation of normalized token form to analyzer

This gives the analyzer more flexibility in choosing the normalized
form. In particular, an analyzer creating different variants can choose
the variant that will be used as the canonical form.

2 years agoMerge pull request #2614 from lonvia/reorganise-country-names
Sarah Hoffmann [Fri, 25 Feb 2022 08:46:20 +0000 (09:46 +0100)]
Merge pull request #2614 from lonvia/reorganise-country-names

Reorganise handling of country names imported from OSM

2 years agoadd migration to add new derived_names column
Sarah Hoffmann [Thu, 24 Feb 2022 19:49:39 +0000 (20:49 +0100)]
add migration to add new derived_names column

2 years agoadd tests for country updates
Sarah Hoffmann [Thu, 24 Feb 2022 15:18:49 +0000 (16:18 +0100)]
add tests for country updates

2 years agobdd: run full import on tests
Sarah Hoffmann [Thu, 24 Feb 2022 10:35:21 +0000 (11:35 +0100)]
bdd: run full import on tests

This uncovered a couple of outdated/wrong tests which have been
fixed, too.

2 years agocountry_name: use separate columns for names from OSM
Sarah Hoffmann [Tue, 22 Feb 2022 22:35:48 +0000 (23:35 +0100)]
country_name: use separate columns for names from OSM

This allows us to distinguish between base names and imported ones
and consiquently removing imported ones if necessary.