split up large osmid index on placex

[nominatim.git] / settings / env.defaults

diff --git a/settings/env.defaults b/settings/env.defaults
index 9fefc622bd9fa5666d5cedb1829b83f724ec3f1d..3115f4382aacf582c5a1054e78c03130bde9f00f 100644 (file)
--- a/settings/env.defaults
+++ b/settings/env.defaults
@@ -5,7 +5,7 @@
 # Database connection string.
 # Add host, port, user etc through additional semicolon-separated attributes.
 # e.g. ;host=...;port=...;user=...;password=...
 # Database connection string.
 # Add host, port, user etc through additional semicolon-separated attributes.
 # e.g. ;host=...;port=...;user=...;password=...

-# Changing this variable requires to run 'setup.php --setup-website'.
+# Changing this variable requires to run 'nominatim refresh --website'.

 NOMINATIM_DATABASE_DSN="pgsql:dbname=nominatim"
 
 # Database web user.
 NOMINATIM_DATABASE_DSN="pgsql:dbname=nominatim"
 
 # Database web user.


@@ -15,10 +15,16 @@ NOMINATIM_DATABASE_WEBUSER="www-data"
 # Directory where to find the PostgreSQL server module.
 # When empty the module is expected to be located in the 'module' subdirectory
 # in the project directory.
 # Directory where to find the PostgreSQL server module.
 # When empty the module is expected to be located in the 'module' subdirectory
 # in the project directory.

-# Changing this value requires to run ./utils/setup --create-functions.
+# Changing this value requires to run 'nominatim refresh --functions'.

 NOMINATIM_DATABASE_MODULE_PATH=
 
 NOMINATIM_DATABASE_MODULE_PATH=
 

-# Number of occurances of a word before it is considered frequent.
+# Tokenizer used for normalizing and parsing queries and names.
+# The tokenizer is set up during import and cannot be changed afterwards
+# without a reimport.
+# Currently available tokenizers: icu, legacy
+NOMINATIM_TOKENIZER="icu"
+
+# Number of occurrences of a word before it is considered frequent.

 # Similar to the concept of stop words. Frequent partial words get ignored
 # or handled differently during search.
 # Changing this value requires a reimport.
 # Similar to the concept of stop words. Frequent partial words get ignored
 # or handled differently during search.
 # Changing this value requires a reimport.


@@ -40,12 +46,18 @@ NOMINATIM_LANGUAGES=
 # Changing this value requires a reimport.
 NOMINATIM_TERM_NORMALIZATION=":: NFD (); [[:Nonspacing Mark:] [:Cf:]] >;  :: lower (); [[:Punctuation:][:Space:]]+ > ' '; :: NFC ();"
 
 # Changing this value requires a reimport.
 NOMINATIM_TERM_NORMALIZATION=":: NFD (); [[:Nonspacing Mark:] [:Cf:]] >;  :: lower (); [[:Punctuation:][:Space:]]+ > ' '; :: NFC ();"
 

+# Configuration file for the tokenizer.
+# The content depends on the tokenizer used. If left empty the default settings
+# for the chosen tokenizer will be used. The configuration can only be set
+# on import and not be changed afterwards.
+NOMINATIM_TOKENIZER_CONFIG=
+

 # Search in the Tiger house number data for the US.
 # Note: The tables must already exist or queries will throw errors.
 # Changing this value requires to run ./utils/setup --create-functions --setup-website.
 NOMINATIM_USE_US_TIGER_DATA=no
 
 # Search in the Tiger house number data for the US.
 # Note: The tables must already exist or queries will throw errors.
 # Changing this value requires to run ./utils/setup --create-functions --setup-website.
 NOMINATIM_USE_US_TIGER_DATA=no
 

-# Search in the auxilary housenumber table.
+# Search in the auxiliary housenumber table.

 # Changing this value requires to run ./utils/setup --create-functions --setup-website.
 NOMINATIM_USE_AUX_LOCATION_DATA=no
 
 # Changing this value requires to run ./utils/setup --create-functions --setup-website.
 NOMINATIM_USE_AUX_LOCATION_DATA=no
 


@@ -53,7 +65,7 @@ NOMINATIM_USE_AUX_LOCATION_DATA=no
 # The following settings allow to set a proxy to use when remotely downloading
 # data. Host and port are required. Login and password are optional.
 NOMINATIM_HTTP_PROXY=no
 # The following settings allow to set a proxy to use when remotely downloading
 # data. Host and port are required. Login and password are optional.
 NOMINATIM_HTTP_PROXY=no

-NOMINATIM_HTTP_PROXY_HOST=proxy.mydomain.com
+NOMINATIM_HTTP_PROXY_HOST=

 NOMINATIM_HTTP_PROXY_PORT=3128
 NOMINATIM_HTTP_PROXY_LOGIN=
 NOMINATIM_HTTP_PROXY_PASSWORD=
 NOMINATIM_HTTP_PROXY_PORT=3128
 NOMINATIM_HTTP_PROXY_LOGIN=
 NOMINATIM_HTTP_PROXY_PASSWORD=


@@ -62,34 +74,25 @@ NOMINATIM_HTTP_PROXY_PASSWORD=
 # HTTPS_PROXY="http://user:pass@10.10.1.10:1080"
 
 # Location of the osm2pgsql binary.
 # HTTPS_PROXY="http://user:pass@10.10.1.10:1080"
 
 # Location of the osm2pgsql binary.

-# When empty, osm2pgsql is expected to reside in the osm2pgsql directory in
-# the project directory.

 # EXPERT ONLY. You should usually use the supplied osm2pgsql.
 NOMINATIM_OSM2PGSQL_BINARY=
 
 # EXPERT ONLY. You should usually use the supplied osm2pgsql.
 NOMINATIM_OSM2PGSQL_BINARY=
 

-# Location of pyosmium-get-changes.
-# Only needed when running updates.
-NOMINATIM_PYOSMIUM_BINARY=
-

 # Directory where to find US Tiger data files to import.
 # Directory where to find US Tiger data files to import.

-# Used with setup.php --import-tiger-data. When unset, the data is expected
-# to be located under 'data/tiger' in the source tree.
+# OBSOLETE: use `nominatim add-data --tiger-data <dir>` to explicitly state
+#           the directory on import

 NOMINATIM_TIGER_DATA_PATH=
 
 # Directory where to find pre-computed Wikipedia importance files.
 NOMINATIM_TIGER_DATA_PATH=
 
 # Directory where to find pre-computed Wikipedia importance files.

-# When unset, the data is expected to be located in the 'data' directory
-# in the source tree.
+# When unset, the data is expected to be located in the project directory.

 NOMINATIM_WIKIPEDIA_DATA_PATH=
 
 # Configuration file for special phrase import.
 NOMINATIM_WIKIPEDIA_DATA_PATH=
 
 # Configuration file for special phrase import.

-# When unset, the internal default settings from 'settings/phrase_settings.php'
-# are used.
+# OBSOLETE: use `nominatim special-phrases --config <file>` or simply put
+#           a custom phrase-settings.json into your project directory.

 NOMINATIM_PHRASE_CONFIG=
 
 # Configuration file for rank assignments.
 NOMINATIM_PHRASE_CONFIG=
 
 # Configuration file for rank assignments.

-# When unset, the internal default settings from 'settings/address-levels.json'
-# are used.
-NOMINATIM_ADDRESS_LEVEL_CONFIG=
+NOMINATIM_ADDRESS_LEVEL_CONFIG=address-levels.json

 
 # Configuration file for OSM data import.
 # This may either be the name of one of an internal style or point
 
 # Configuration file for OSM data import.
 # This may either be the name of one of an internal style or point


@@ -148,28 +151,28 @@ NOMINATIM_TABLESPACE_AUX_INDEX=
 # These are used to keep the database up to date. Per default it points to
 # the minutely updates for the main OSM database. There are other services
 # geared towards larger update intervals or data extracts.
 # These are used to keep the database up to date. Per default it points to
 # the minutely updates for the main OSM database. There are other services
 # geared towards larger update intervals or data extracts.

-# Changing this value requires to rerun 'update/php --init-updates'.
+# Changing this value requires to rerun 'nominatim replication --init'.

 NOMINATIM_REPLICATION_URL="https://planet.openstreetmap.org/replication/minute"
 
 # Maximum amount of data to download per batch.
 # Size is in MB.
 NOMINATIM_REPLICATION_MAX_DIFF=50
 
 NOMINATIM_REPLICATION_URL="https://planet.openstreetmap.org/replication/minute"
 
 # Maximum amount of data to download per batch.
 # Size is in MB.
 NOMINATIM_REPLICATION_MAX_DIFF=50
 

-# Publication interval of the replication service.
+# Publication interval of the replication service (in seconds).

 # Determines when Nominatim will attempt again to download again a new
 # update. The time is computed from the publication date of the last diff
 # downloaded. Setting this to a slightly higher value than the actual
 # publication interval avoids unnecessary rechecks.
 NOMINATIM_REPLICATION_UPDATE_INTERVAL=75
 
 # Determines when Nominatim will attempt again to download again a new
 # update. The time is computed from the publication date of the last diff
 # downloaded. Setting this to a slightly higher value than the actual
 # publication interval avoids unnecessary rechecks.
 NOMINATIM_REPLICATION_UPDATE_INTERVAL=75
 

-# Wait time to recheck for a pending update.
+# Wait time to recheck for a pending update (in seconds).

 # Time to wait after an expected update was not available on the server.
 NOMINATIM_REPLICATION_RECHECK_INTERVAL=60
 
 ### API settings
 #
 # The following settings configure the API responses. You must rerun
 # Time to wait after an expected update was not available on the server.
 NOMINATIM_REPLICATION_RECHECK_INTERVAL=60
 
 ### API settings
 #
 # The following settings configure the API responses. You must rerun

-# setup.php --setup-website after changing any of them.
+# 'nominatim refresh --website' after changing any of them.

 
 # Send permissive CORS access headers.
 # When enabled, send CORS headers to allow access to everybody.
 
 # Send permissive CORS access headers.
 # When enabled, send CORS headers to allow access to everybody.


@@ -185,7 +188,7 @@ NOMINATIM_MAPICON_URL=
 NOMINATIM_DEFAULT_LANGUAGE=
 
 # Enable a special batch query mode.
 NOMINATIM_DEFAULT_LANGUAGE=
 
 # Enable a special batch query mode.

-# This features is currently undocumented and potentially broken.
+# This feature is currently undocumented and potentially broken.

 NOMINATIM_SEARCH_BATCH_MODE=no
 
 # Threshold for searches by name only.
 NOMINATIM_SEARCH_BATCH_MODE=no
 
 # Threshold for searches by name only.


@@ -204,7 +207,7 @@ NOMINATIM_POLYGON_OUTPUT_MAX_TYPES=1
 ### Log settings
 #
 # The following options allow to enable logging of API requests.
 ### Log settings
 #
 # The following options allow to enable logging of API requests.

-# You must rerun setup.php --setup-website after changing any of them.
+# You must rerun 'nominatim refresh --website' after changing any of them.

 #
 # Enable logging of requests into the DB.
 # The request will be logged into the new_query_log table.
 #
 # Enable logging of requests into the DB.
 # The request will be logged into the new_query_log table.