Frédéric Rodrigo

The proof of concept for keeping osm_id is on this extra branch : https://github.com/frodrigo/openmaptiles/commits/building_cluster_with_osm_id It not deals with update yet. Waiting for update on switching to osm_id for feature_id or...

I updated with the support of the osm_id of the building. While building are clustered, I group osm_id in array. On read, I extract ST_Collection and array on parallel.

I rebased on the last master and removed the Draft/WIP status. I used this code now for years ! But since I wrote it, the generalized z13 building levels was...

Yes. Imposm output only one level zoom tiles on update. You can use this script to compute the others https://github.com/makina-maps/makina-maps/blob/master/nginx/tile_multiplier.py

> @frodrigo, > > From a legal perspective, can we get permission to copy and integrate the code from https://github.com/makina-maps/makina-maps/blob/master/nginx/tile_multiplier.py into OMT or OMT-tools? > > If so, I will...

It is how OpenMapTiles works for now. On each update you have to regenerate materialized views for the whole planet. No matter the size of your diff.

Overview of parts to improve for quicker diff update Simple Update on full table: - [x] ~update_peak_point.sql #896~ - [x] ~update_country_point.sql #944~ - [x] ~update_state_point.sql #944~ - [x] ~update_island_polygon.sql #944~...

As you can see by the long list of commit here, it was largely improved in the last release, but not fully solved yet.

> @frodrigo actually is it possible that this will be very different based on whether `import-sql` was run before `update-osm`? I think that maybe once the indexes etc are added,...

Yes, also for me the isochrone are the big missing thing on OSRM. I already need to write myself a wrapper on to of OSRM using a one-to-may matrix, or...