Any possibility to get the link to the sample tiles.db used here if available publicly?

Its just an extract from the overture-maps building polygons dataset stored in a duckdb database. It's about 270mb on my machine so a bit too big to share easily, but it just contains a single geometry column with a r-tree index on. You should be able to reproduce something similar to it easily.

@Maxxen, do you think it's possible to make it work by just reading the parquet file instead of using the persistent database?

UPD. It seems like I was having a data source issue. When I tried to read it directly, the app was failing with a UserWarning: resource_tracker: There appear to be 1 leaked semaphore objects to clean up at shutdown error at some point, but the other parquet file works just fine.

Yeah sure, but its probably going to be significantly slower instead of using a persistent DB with an rtree index, as you need to scan a lot more of the parquet file for each tile, depending on the row-group size/spatial partitioning/sorting of the file.

Any possibility to get the link to the sample tiles.db used here if available publicly?

Here is duckdb SQL to generate a tiles.db file (~170mb) you can use. It holds New York City buildings using data from Overture Maps Buildings.

install spatial;
load spatial;

attach 'tiles.db';
use tiles;

create or replace table t1 as (
  select
    -- transform into Web Mercator which the flask server expects
    st_transform(geometry, 'EPSG:4326', 'EPSG:3857', always_xy := true) as geometry,
    subtype,
    class,
    height,
  from
    read_parquet(
       -- You may need to change to whatever release is available at
       -- https://docs.overturemaps.org/guides/buildings/
      's3://overturemaps-us-west-2/release/2025-11-19.0/theme=buildings/type=building/*',
      filename = true,
      hive_partitioning = 1
    )
  where
    -- extent of new york city
    bbox.xmin between -74.2 and -73.6 and bbox.ymin between 40.5 and 40.9
    and bbox.xmax between -74.2 and -73.6
    and bbox.ymax between 40.5 and 40.9
    -- arbitrary filters to reduce size
    and subtype is not null
    and class is not null
    and height is not null
);
create index my_idx on t1 using rtree (geometry);

use memory.main;
detach tiles;