Replies: 2 comments 1 reply
-
Regardless, you should ensure you're using the latest version of geoarrow-rust. At this point, geometry simplification is not in scope of Lonboard, and you should do any geometry simplification you need before passing the data to Lonboard |
Beta Was this translation helpful? Give feedback.
0 replies
-
Sorry for that. geoarrow.rust.io 0.3.0 read_parquet works fine. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Hi. I downloaded building footprint data (polygons) from 'https://ftp.maps.canada.ca/pub/nrcan_rncan/extraction/auto_building/shp/, combined all shapefiles there into one big geodataframe and saved the result as geoparquet with
geometry_encoding='geoarrow'
. This gives a table of 11M lines.I tried opening the parquet file with
geoarrow.rust.core
read_parquet
and use it in a PolygonLayer :'geoarrow.rust.core._rust.GeoTable' object [table_parquet_geoarrow_rust] has no attribute 'select'
So I ended up opening the parquet file with
pyarrow.parquet
read_table
and now I can get a PolygonLayer. I have a GPU-less computer and got to see parts of the map but display is awfully slow. I would think 11M polygons is just too much to draw.I'm trying to mimick this map, which shows rough perimeter of areas with data and more and more detail as one zooms in, only with the data in lonboard (and eventually Polars/DuckDB for analysis). Have there been discussions on ways to adapt the number of vertices in polygons as a function of zoom level e.g. take a bbox or some rough perimeter of individual polygons ? Or get some rough perimeter of groups of polygons, a kind of "geometry aggregation", as a function of zoom level ?
Beta Was this translation helpful? Give feedback.
All reactions