This release sees a complete reorganisation of the code around parsing and indexing place names. The code is now encapsulated in so-called tokenizer modules. The algorithms from the 3.x.x versions have been moved to the legacy tokenizer module. The release also introduces a new ICU tokenizer module. It uses a complete different approach to handling place names with
- normalisation and transliteration based on libICU
- configurable abbreviation handling at import time
- and configurable pre-processing of place names
The former PHP utils have been dropped in favour of the nominatim CLI tool. The format of some external data sources (postcodes, TIGER data) has changed to CSV. You need to download new versions of them.
A more complete list of changes can be found in the Changelog.
Updating from 3.6.x or 3.7.x to the new release is straightforward. Please follow the instructions from the migration guide. Updating from older versions is possible but not recommended. A complete reimport is required for using the ICU tokenizer.
Please note that this release drops support for Python 3.5 and PostgreSQL < 9.5.