Release 0.7.2

This is a bug fix release fixing an inconsistency between JSON dumper and JSON reader, see #919. If you want to read self-created dumps or the experimental JSON dumps at https://download1.graphhopper.com/public/raw/ with any extra tags included, please update to this release.

Release 0.7.1

This is a bug fix release that fixes the following regressions:

• do not overwrite street of street addresses when an addr:place is present
• return only one result for reverse queries per default
• sort reverse results by distance per default

The release also updates dependencies to their latest version.

Release 0.7.0

This release brings the OpenSearch version of photon out of its experimental state and makes it the recommended version to use. The ElasticSearch version is now deprecated and will be removed with the next release. If you are using the pre-computed database dumps, please use the dumps at https://download1.graphhopper.com/public/experimental for the time being when using photon-opensearch. The download server is up for a refresh and reorganisation in the coming weeks. Once this is done, we'll provide the usual stable release dumps for photon-opensearch as well. Look out for announcements over at https://nominatim.org/blog/.

The other major new feature in this release is the ability to write the export from Nominatim into a json file and reimport the file later. The export filtering options for countries, languages and extratags work with json export and import. The export format is not completely finalised yet and may still change slightly depending on the feedback we get from users. So, please try it out and let us know.

The Nominatim exports has been reorganised and reads the data now country by country, caching address information along the way. This saves millions of SQL queries and brings down the time for a planet export to half a day. The export needs a special country index on the placex table, to be able to extract the information efficiently. It will create the index automatically if it is not yet there. If you have a user with read-only access, you can manually create the index before exporting.

The new '-import-geometry-column' option allows to import not only centroids for places but their full geometry. Be aware that this comes at the price of more than doubling the database size. Many thanks to @red-fenix for this feature!

Other changes include:

• rework Nominatim export to cache
• fix missing OSM IDs and importance values in photon-opensearch output
• fix startup error when synonym file was wrong
• remove most plufins from bundled opensearch
• correctly handle precedence between postcode boundaries and postcode address tags
• move to javlin webframework to replace unmaintained spark
• drop json dependency in favour of jackson
• add ISO3166-2 state codes to recognised address parts

The release also features a lot of code cleanup, ever increasing test coverage
and a completely overhauled README.

Release 0.6.2

This is a bug fix release which fixes the following issue:

• avoid sending too much data to the database during updates

Release 0.6.1

This release fixes some smaller issues in the new OpenSearch version of Photon:

• correctly set timeout for reverse endpoint
• correctly translate 64bit numbers and floating point numbers when reading from OpenSearch database
• dump raw result content when running with debug mode (also moves score and importance into debug output)
• make sure to reopen the index when installation of synonym fails
• reduce installed and loaded plugins to the necessary minimum

The release also increases the queue size for the importer, which results in a nice speed-up of around 20%.

Release 0.6.0

This release adds a port of Photon to the latest version of OpenSearch. The version is not database compatible with the ElasticSearch version and makes some changes to the way search works, creating simpler and faster search queries for some especially complex cases. This version of Photon also adds a new structured search endpoint (thanks @tobiass-sdl). The OpenSearch version is still experimental at this stage. In particular, expect the database schema to change some more as we are starting to take advantage of the improvements that happend since the currently used ElasticSearch 5 version. The long-term plan is to make the OpenSearch version the main and only Photon version with the major 1.0 release.

This release also sees a switch from Maven to Gradle for the build system. Gradle offers better support for building multiple flavours of Photon and gives better control over dependencies.

Other changes in this release are:

• new configuration parameters for setting the maximum number of returned results (thanks @karussell)
• fix importing addresses of objects that are not part of an address themselves (e.g. lakes, mountains)
• allow multiple CORS origins on the command line (thanks @burleight)

The ElasticSearch version of Photon is compatible with databases from version 0.5.0. To update, simply download and start the newest Photon version.

The OpenSearch version of Photon needs a different database format. We currently provide dumps in the experimental section of the download server. Please note that experimental dumps are usually made for the master version of Photon, not the 0.6.0 release specifically. The dumps are not prepared for structured searches. You need to create your own database to make use of this new feature.

Release 0.5.0

This release brings back the ability to update a Photon database from a Nominatim source. The new mechanism is decoupled from the Nominatim update process, making it easier to handle. A long-standing issue around UIDs of house numbers documents has been fixed so that all data should now correctly be updated. This version also introduces a new API endpoint /nominatim-update/status which allows scripts to check if an update is already in progress.

This version also introduces a /status endpoint which returns information about the age of the database. (Many thanks to @ybert.)

If you do not use the update facility, then the release remains compatible with version 0.4 database dumps including the ones available on https://download1.graphhopper.com/public. If you want to run updates, you need to start with a fresh import using this release. Please be aware that updates now require an additional preparation step. Consult the README for more information.

Further changes:

• add command-line parameter to set the timeout for ElasticSearch queries
• fix issue where Photon couldn't be run from directories with spaces in the name
• fix off-by-one error when expanding interpolations
• remove lombok and replace uses with explicit code
• typo and grammar fixes to log messages and documentation
• update dependencies to latest versions

Release 0.4.4

This is a minor release which mainly updates the dependent libraries, including a security critical update of the json library.

The release drops support for Java 8 and enables support for Java 21.

Thanks to @TKYK93 and @sebix for giving the README a refresh.

Release 0.4.3

This release adds the tag filter functionality to reverse geocoding. The osm-tag parameter can now be used with reverse. Many thanks to @ybert for this new feature.

Other changes include some improvements to the build process (thanks @otbutz) and updates to versions of the dependencies.

This will be the last version to support Java 8.

Release 0.4.2

This release adapts the import process to some recent changes in the name handling of linked places in Nominatim. Names of linked places now appear in an artificial _place_name tag which Photon needs to read in addition to the usual name tags.

On the search side, two changes attempt to improve fuzzy matching of search results:

• Skipping of words is no longer allowed for short two-word queries. ES otherwise prefers to sometimes skip one of two words instead of matching against a wrongly spelled word.
• Fuzzy matching is now attempted against all languages variants, not only against the chosen output language. There still is a restriction that multi-lingual queries don't work with fuzzy matching.