add address counts to tokens

[nominatim.git] / ChangeLog

4.4.0
 * add export to SQLite database and SQLite support for the frontend
 * switch to Python frontend as the default frontend
 * update to osm2pgsql 1.11.0
 * add support for new osm2pgsql middle table format
 * simplify geometry for large polygon objects not used in addresses
 * various performance tweaks for search in Python frontend
 * fix regression in search with categories where it was confused with near
   search
 * partially roll back use of SQLAlchemy lambda statements due to bugs
   in SQLAchemy
 * fix handling of timezones for timestamps from the database
 * fix handling of full address searches in connection with a viewbox
 * fix postcode computation of highway areas
 * fix handling of timeout errors for Python <= 3.10
 * fix address computation for postcode areas
 * fix variable shadowing in osm2pgsql flex script, causing bugs with LuaJIT
 * make sure extratags are always null when empty
 * reduce importance of places without wikipedia reference
 * improve performance of word count computations
 * drop support for wikipedia tags with full URLs
 * replace get_addressdata() SQL implementation with a Python function
 * improve display name for non-address features
 * fix postcode validation for postcodes with country code
   (thanks @pawel-wroniszewski)
 * add possibility to run imports without superuser database rights
   (thanks @robbe-haesendonck)
 * new CLI command for cleaning deleted relations (thanks @lujoh)
 * add check for database version in the CLI check command
 * updates to import styles ignoring more unused objects
 * various typo fixes (thanks @kumarUjjawal)

4.3.2
 * fix potential SQL injection issue for 'nominatim admin --collect-os-info'
 * PHP frontend: fix on-the-fly lookup of postcode areas near boundaries
 * Python frontend: improve handling of viewbox
 * Python frontend: correct deployment instructions

4.3.1
 * reintroduce result rematching
 * improve search of multi-part names
 * fix accidentally switched meaning of --reverse-only and --search-only in
   warm command

4.3.0
 * fix failing importance recalculation command
 * fix merging of linked names into unnamed boundaries
 * fix a number of corner cases with interpolation splitting resulting in
   invalid geometries
 * fix failure in website generation when password contains curly brackets
 * fix broken use of ST_Project in PostGIS 3.4
 * new NOMINATIM_SEARCH_WITHIN_COUNTRIES setting to restrict reverse lookups
   to known countries (thanks @alfmarcua)
 * allow negative OSM IDs (thanks @alfmarcua)
 * disallow import of Tiger data in a frozen DB
 * avoid UPDATE to change settings to be compatible with r/o DBs (thanks @t-tomek)
 * update bundled osm2pgsql to 1.9.2
 * reorganise osm2pgsql flex style and make it the default
 * exclude names ending in :wikipedia from indexing
 * no longer accept comma as a list separator in name tags
 * process forward dependencies on update to catch updates in geometries
   of ways and relations
 * fix handling of isolated silent letters during transliteration
 * no longer assign postcodes to large linear features like rivers
 * introduce nominatim.paths module for finding data and libraries
 * documentation layout changed to material theme
 * new documentation section for library
 * various smaller fixes to existing documentation
   (thanks @woodpeck, @bloom256, @biswajit-k)
 * updates to vagrant install scripts, drop support for Ubuntu 18
   (thanks @n-timofeev)
 * removed obsolete configuration variables from env.defaults
 * add script for generating a taginfo description (thanks @biswajit-k)
 * modernize Python code around BDD test and add testing of Python frontend
 * lots of new BDD tests for API output

4.2.3

 * fix deletion handling for 'nominatim add-data'
 * adapt place_force_delete() to new deletion handling
 * flex style: avoid dropping of postcode areas
 * fix update errors on address interpolation handling

4.2.2

 * extend flex-style library to fully support all default styles
 * fix handling of Hebrew aleph
 * do not assign postcodes to rivers
 * fix string matching in PHP code
 * update osm2pgsql (various updates to flex)
 * fix slow query when deleting places on update
 * fix CLI details query
 * fix recalculation of importance values
 * fix polygon simplification in reverse results
 * add class/type information to reverse geocodejson result
 * minor improvements to default tokenizer configuration
 * various smaller fixes to documentation

4.2.1

 * fix XSS vulnerability in debug view

4.2.0

 * add experimental support for osm2pgsql flex style
 * introduce secondary importance value to be retrieved from a raster data file
   (currently still unused, to replace address importance, thanks to @tareqpi)
 * add new report tool `nominatim admin --collect-os-info`
   (thanks @micahcochran, @tareqpi)
 * reorganise index to improve lookup performance and size
 * run index creation after import in parallel
 * run ANALYZE more selectively to speed up continuation of indexing
 * fix crash on update when addr:interpolation receives an illegal value
 * fix minimum number of retrieved results to be at least 10
 * fix search for combinations of special term + name (e.g Hotel Bellevue)
 * do not return interpolations without a parent street on reverse search
 * improve invalidation of linked places on updates
 * fix address parsing for interpolation lines
 * make sure socket timeouts are respected during replication
   (working around a bug in some versions of pyosmium)
 * update bundled osm2pgsql to 1.7.1
 * add support for PostgreSQL 15
 * typing fixes to work with latest type annotations from typeshed
 * smaller improvements to documentation (thanks to @mausch)

4.1.1

 * fix XSS vulnerability in debug view

4.1.0

 * switch to ICU tokenizer as default
 * add housenumber normalization and support optional spaces during search
 * add postcode format checking and support optional spaces during search
 * add function for cleaning housenumbers in word table
 * add updates/deletion of country names imported from OSM
 * linked places no longer overwrite names from a place permanently
 * move default country name configuration into yaml file (thanks @tareqpi)
 * more compact layout for interpolation and TIGER tables
 * introduce mutations to ICU tokenizer (used for German umlauts)
 * support reinitializing a full project directory with refresh --website
 * fix various issues with linked places on updates
 * add support for external sanitizers and token analyzers
 * add CLI commands for forced indexing
 * add CLI command for version report
 * add offline import mode
 * change geocodejson to return a feature class in the 'type' field
 * add ISO3166-2 to address output (thanks @I70l0teN4ik)
 * improve parsing and matching of addr: tags
 * support relations as street members of associatedStreet
 * better ranking for address results from TIGER data
 * adapt rank classification to changed tag usage in OSM
 * update bundled osm2pgsql to 1.6.0
 * add typing information to Python code
 * improve unit test coverage
 * reorganise and speed up code for BDD tests, drop support for scenes
 * move PHP unit tests to PHP 9.5
 * extensive typo fixes in documentation (thanks @woodpeck,@StephanGeorg,
   @amandasaurus, @nslxndr, @stefkiourk, @Luflosi, @kianmeng)
 * drop official support for installation on CentOS
 * add installation instructions for Ubuntu 22.04
 * add support for PHP8
 * add setup instructions for updates and systemd
 * drop support for PostgreSQL 9.5

4.0.2

 * fix XSS vulnerability in debug view

4.0.1

 * fix initialisation error in replication script
 * ICU tokenizer: avoid any special characters in word tokens
 * better error message when API php script does not exist
 * fix quoting of house numbers in SQL queries
 * small fixes and improvements in search query parsing
 * add documentation for moving the database to a different machine

4.0.0

 * refactor name token computation and introduce ICU tokenizer
   * name processing now happens in the indexer outside the DB
   * reorganizes abbreviation handling and moves it to the indexing phases
   * adds preprocessing of names
 * add country-specific ranking for Spain, Slovakia
 * partially switch to using SP-GIST indexes
 * better updating of dependent addresses for name changes in streets
 * remove unused/broken tables for external housenumbers
 * move external postcodes to CSV format and no longer save them in tables
   (adds support for postcodes for arbitrary countries)
 * remove postcode helper entries from placex (thanks @AntoJvlt)
 * change required format for TIGER data to CSV
 * move configuration of default languages from wiki into config file
 * expect customized configuration files in project directory by default
 * disable search API for reverse-only import (thanks @darkshredder)
 * port most of maintenance/import code to Python and remove PHP utils
 * add catch-up mode for replication
 * add updating of special phrases (thanks @AntoJvlt)
 * add support for special phrases in CSV files (thanks @AntoJvlt)
 * switch to case-independent matching between place and boundary names
 * remove disabling of reverse query parsing
 * minor tweaks to search algorithm to avoid more false positives
 * major overhaul of the administrator and developer documentation
 * add security disclosure policy
 * add testing of installation scripts via CI
 * drop support for Python < 3.6 and Postgresql < 9.5

3.7.3

 * fix XSS vulnerability in debug view

3.7.2

 * fix database check for reverse-only imports
 * do not error out in status API result when import date is missing
 * add array_key_last function for PHP < 7.3 (thanks to @woodpeck)
 * fix more url when server name is unknown (thanks to @mogita)
 * commit changes to replication log table

3.7.1

 * fix smaller issues with special phrases import (thanks @AntoJvlt)
 * add index to speed up continued indexing during import
 * fix index on location_property_tiger(parent_place_id) (thanks @changpingc)
 * make sure Python code is backward-compatible with Python 3.5
 * various documentation fixes

3.7.0

 * switch to dotenv for configuration file
 * introduce 'make install' (reorganising most of the code)
 * introduce nominatim tool as replacement for various php scripts
 * introduce project directories and allow multiple installations from same build
 * clean up BDD tests: drop nose, reorganise step code
 * simplify test database for API BDD tests and autoinstall database
 * port most of the code for command-line tools to Python
   (thanks to @darkshredder and @AntoJvlt)
 * add tests for all tooling
 * replace pyosmium-get-changes with custom internal implementation using
   pyosmium
 * improve search for queries with housenumber and partial terms
 * add database versioning
 * use jinja2 for preprocessing SQL files
 * introduce automatic migrations
 * reverse fix preference of interpolations over housenumbers
 * parallelize indexing of postcodes
 * add non-key indexes to speed up housenumber + street searches
 * switch housenumber field in placex to save transliterated names

3.6.0

 * add full support for searching by and displaying of addr:* tags
 * improve address output for large-area objects
 * better use of country names from OSM data for search and display
 * better debug output for reverse call
 * add support for addr:place links without an place equivalent in OSM
 * improve finding postcodes with normalisation artefacts
 * batch object to index for rank 30, avoiding a wrap-around of transaction
   IDs in PostgreSQL
 * introduce dynamic address rank computation for administrative boundaries
   depending on linked objects and their place in the admin level hierarchy
 * add country-specific address ranking for Indonesia, Russia, Belgium and
   the Netherlands (thanks @hendrikmoree)
 * make sure wikidata/wikipedia tags are imported for all styles
 * make POIs searchable by name and housenumber (thanks @joy-yyd)
 * reverse geocoding now ignores places without an address rank (rivers etc.)
 * installation of a webserver is no longer mandatory, for development
   use the php internal webserver via 'make serve
 * reduce the influence of place nodes in addresses
 * drop support for the unspecific is_in tag
 * various minor tweaks to supplied styles
 * move HTML web frontend into its own project
 * move scripts for processing external data sources into separate directories
 * introduce separate configuration for website (thanks @krahulreddy)
 * update documentation, in particular, clean up development docs
 * update osm2pgsql to 1.4.0

3.5.2

 * ensure that wikipedia tags are imported for all styles
 * reinstate verbosity for indexing during updates
 * make house number reappear in display name on named POIs
 * introduce batch processing in indexer to avoid transaction ID overrun
 * increase splitting for large geometries to improve indexing speed
 * remove deprecated get_magic_quotes_gpc() function
 * make sure that all postcodes have an entry in word and are thus searchable
 * remove use of ST_Covers in conjunction with ST_Intersects,
   causes bad query planning and slow updates in Postgis3
 * update osm2pgsql

3.5.1

 * disable jit and parallel processing in PostgreSQL for osm2pgsql
 * update libosmium to 2.15.6 (fixes an issue with processing hanging
   on large multipolygons)

3.5.0

 * structured select on HTML search page
 * new PHP Nominatim\Shell class to wrap shell escaping
 * remove polygon parameter from all API calls
 * improve handling of postcode areas
 * reorganise place linking algorithm, now using wikidata tag as well
 * remove linkees from search_name and larger_area tables
 * introduce country-specific address ranks
 * reorganise rank address computation
 * cleanup of partition function
 * improve parenting for large POIs
 * add support for Postgresql 12 and Postgis 3
 * add earlier cleanup when --drop is given, to reduce memory usage
 * remove use of place_id in URLs
 * replace C nominatim indexer with a simpler Python implementation
 * split up the huge sql/functions.sql file
 * move osm2pgsql tests to osm2pgsql
 * add new extratags style which imports all tags from OSM
 * add new script for checking the import after completion
 * update osm2pgsql, reducing memory usage
 * use new wikipedia importance and add processing of wikidata tags
 * add search form for details page
 * use ExtraDataPath for country_grid table
 * remove short_name from list of names to be displayed
 * split up CMakeFile, so that all parts can be built separately
 * update installation instructions for CentOS and Ubuntu
 * add script for importing/updating multiple country extracts
 * various documentation improvements

3.4.2

 * fix security bug in /details endpoint where user input was not
   properly sanitized

3.4.1

 * update osm2pgsql to fix hans during updates and lost address numbers
   during updates

3.4.0

 * increase required version for PostgreSQL(9.3), PostGIS(2.2) and PHP(7.0)
 * better error reporting for out-of-memory errors
 * exclude postcode ranges separated by colon from centre point calculation
 * update osm2pgsql, better handling of imports without flatnode file
 * switch to more efficient algorithm for word set computation
 * use only boundaries for country and state parts of addresses
 * improve updates of addresses with housenumbers and interpolations
 * remove country from place_addressline table and use country_code instead
 * optimise indexes on search_name partition tables
 * improve searching of attached streets for large objects like airports
 * drop support for python 2
 * new scripts for importing Wikidata for importance
 * create and drop indexes concurrently to not clash with auto vacuum
 * various documentation improvements


3.3.0

 * zoom 17 in reverse now zooms in on minor streets
 * fix use of postcode relations in address
 * support for housenumber 0 on interpolations
 * replace database abstraction DB with PDO and switch to using exceptions
 * exclude line features at rank 30 from reverse geocoding
 * remove self-reference and country from place_addressline
 * make json output more readable (less escaping)
 * update conversion scripts for postcodes
 * scripts in utils/ are no longer executable (always use scripts in build dir)
 * remove Natural Earth country fallback (OSM is complete enough)
 * make rank assignments configurable
 * allow accept languages with underscore
 * new reverse-only import mode (without search index table)
 * rely on boundaries only for states and countries
 * update osm2pgsql, now using a configurable style
 * provide multiple import styles
 * improve search when house number and postcodes are dropped
 * overhaul of setup code
 * add support for PHPUnit 6
 * update test database
 * various documentation improvements

3.2.0

 * complete rewrite of reverse search algorithm
 * add new geojson and geocodejson output formats
 * add simple export script to export addresses to CSV
 * remove is_in terms from address computation
 * remove unused search_name_country tables
 * various smaller fixes to query parsing
 * convert Tokens and token types to class types
 * correctly handle update when boundary object type is changed
 * improve debug output for /search endpoint
 * update to latest osm2pgsql and leaflet.js
 * overhaul of /details endpoint:
     * new class parameter when using osmtype/osmid parameters
     * permalink to instance-independent osmtype/osmid parameter format
     * new json output format
 * update CentOS vagrant machine to use SELinux
 * add vagrant scripts for Ubuntu 18.04
 * fix build process for BSD
 * enable running the database on a different host than the setup scripts
 * allow to configure use of custom PHP binaries (PHP_BIN)
 * extensive coding style improvements to PHP code
 * more PHP unit tests for new classes
 * increase coverage for API tests
 * add documentation for API

3.1.0

 * rework postcode handling and introduce location_postcode table
 * make setup less verbose and print a summary in the end
 * setup: error out when web site user does not exist
 * add more API tests to complete code coverage
 * reinstate key-value amenity search (special term [key=value])
 * fix detection of coordinates in query
 * various smaller tweaks to ranking of search interpretations
 * complete overhaul of PHP frontend code using OOP
 * add address rank to details page
 * update Tiger scripts for 2017 data and clean up unused code
 * various bug fixes and improvements to UI
 * improve reverse geocoding performance close to coasts
 * more PHP style cleanup (quoting)
 * allow unnamed road in reverse geocoding to avoid too far off results
 * add function to recalculate counts for full-word search term
 * add function to check if new updates are available
 * update documentation and switch to mkdocs for generating HTML

3.0.1

 * fix bug in geometry building algorithm in osm2pgsql
 * fix typos in build instructions

3.0.0

 * move to cmake build system
 * various fixes to HTML output
 * reverse endpoint now can return geometries
 * add support for PHP7
 * move to on-the-fly computation of interpolations
 * improve handling of linked places (updating)
 * improve test framework:
    * replace lettuce with behave
    * use smaller database for API tests
 * drop support for postgres < 9.1, postgis < 2.0 and PHP < 5.4
 * make external data use optional (useful for imports without US)
 * detect postgres and postgis versions automatically
 * clean up query logging and remove unused tables
 * move installation documentation into this repo
 * add self-documenting vagrant scripts
 * remove --create-website, recommend to use website directory in build
 * add accessor functions for URL parameters and improve error checking
 * remove IP blocking and rate-limiting code
 * enable CI via travis
 * reformatting for more consistent coding style
 * make country search term creation part of setup
 * update country names and country grid
 * handle roads that cross boundaries better
 * keep full information on address tags
 * update to refactored osm2pgsql which use libosmium based types
 * switch from osmosis to pyosmium for updates
 * be more strict when matching against special search terms
 * handle postcode entries with multiple values correctly

2.5

 * reverse geocoding includes looking up housenumbers from Tiger data
 * added parameter to return simplified geometries
 * new lookup call for getting address information for OSM objects
 * new namedetails and extratags parameters that expose the name and extratags
   fields of the placex table
 * mobile website
 * reverse web view

2.4

 * drop support for postgres 8.4
 * rewrite address interpolation
 * switch to C++ version of osm2pgsql and rewrite tag filtering
 * support for bridge:name and tunnel:name, man_made, junction
 * drop way-node index (reduces database size by about 15%)
 * add support for configuring tablespaces and webserver user
 * better evaluation of search queries in right-to-left notation
 * improve house number search for streets with many duplicate entries
 * code cleanup (remove unused functions and tables)

2.3

 * further improve ordering of results
 * support for more lat/lon formats in search-as-reverse
 * fix handling of GB postcodes
 * new functional test suite
 * support for waterway relations
 * inherit postcodes from street to poi
 * fix housenumber normalisation to find non-latin house numbers
 * take viewbox into account for ordering of results
 * pois may now inherit address tags from surrounding buildings
 * improve what objects may participate in an address
 * clean up handled class/type combinations to current OSM usage
 * lots of bug fixes

2.2

 * correct database rights for www-data
 * add timestamps for update output
 * load postgis via extension for postgis >= 2.0
 * remove non-admin boundaries from addresses
 * further improve ordering of results with same importance
 * merge addr:postcode tags into object addresses
 * include rank and importance in reverse geocode output
 * replace ST_Line_Interpolate_Point with ST_LineInterpolatePoint
  (for postgis >= 2.1)
 * update osm2pgsql to latest version
 * properly detect changes of admin_level
 * remove landuses when name is removed
 * smaller fixes

2.1

 * update to openlayers 2.12 (now custom built)
 * update fallback OSM boundaries
 * add support for postgresql 9.2/9.3 and postgis 2.x
 * add structured queries
 * add options for polygon output in various formats
   (geojson, svg, kml, postgis text)
 * maintenance functions for deleting objects and updating regions
   (plcae_force_update/place_force_delete)
 * web view for polygons that need deleting
 * rate limiting using memcache
 * improve layout of details page
 * add support for boundary:postal_code
 * full CORS support
 * improve parenting of POIs
 * support for extract daily diffs from Geofabrik
 * support for addresses without a street
   (addr:place and conscription number house numbers)
 * improve layout of word and search_name_* tables
 * support for US ZIP+4 codes
 * refactoring of front-end PHP code
 * lots of smaller bug fixes

2.0.1

 * delete outdated entries from location_area_country
 * remove remaining uses of INTEGER, to allow node ids larger than 2^31