New release is out: 1.12.0

In this release, long waited postal code support was introduced. As a result, geocoder database format has been extended and you have to update the datasets. Press "Check for updates" and proceed with the download. If datasets were up to date, only geocoder-nlp database will be updated.

Postal codes are imported from OSM and, for UK, from https://www.freemaptools.com/. Extra postal codes were reversed geocoded on import to figure out admin hierarchy of them (which part of UK they come from).

When searching with postal codes, you could either specify the code alone or together with some part of hierarchy. When postal code is specified together with address parts, its searching for records that have both - postal codes AND address - components in the record.

Postal codes should be detected by libpostal automatically. For some UK tests that I did, I could see that it was not working. This could be due to older training dataset or its incompleteness from the beginning. Unfortunately, libpostal development has been slow these days, but maybe, together with other users, we can fix few bugs and figure out what to do with the training data.

Primitive parser supports entry of postal codes using prefix "post:" . NB! IT'S CaSe SenSITIVE. So, for example:

post: AA BB CC , city

would be parsed with city as a part of admin and AA BB CC as a postal code. For primitive parser, postal codes can be entered anywhere in the string, just keep that part separated by commas. Documentation at https://rinigus.github.io/osmscout-server/en/ covers this as well.

In addition to postal codes, geocoder now returns phone number and website, if these data were at OSM. Pure Maps will be returning it in the next release, modRana has to be patched. All API is backward compatible, so it works already, just these components are not shown to the users.

I have optimized a bit further loading of libpostal for users of primitive parser only. Would be interesting to see, if @pichlo or others on J1 would get search results faster.

Translations have been updated - great work by all translators, thank you!