Allow application developers to directly download the output tilesets to facilitate easier self-hosting

[antonmarsden]

User Story

In order to build a high traffic map-based application (web/mobile) that uses LINZ tiles, as an application developer I want to be able to self-host the Web Mercator output tilesets on my own infrastructure. (instead of trying to figure out how to build the outputs from this GitHub repo).

I'm sure there could be use cases where developers may want to customise the output, but I suspect most will be happy to consume what you're currently producing.

Acceptance Criteria

• [ ] Can conveniently get access to the download the output tileset as a single large file, e.g., a pmtiles file (the pmtiles format can bundle both vector and raster tiles). For reference, see this publically available repository containing OSM vector data in pmtiles format
• [ ] If not pmtiles format, would be useful to have documentation regarding how to do format shifting.
• [ ] ...

Add an Assignee, Milestone, Release and any relevant Labels.

[blacha]

@antonmarsden are you after the aerial imagery (raster) or vector datasets?

All the source raster imagery is stored in a public S3 bucket s3://nz-imagery for more information on how to access and work with the imagery https://github.com/linz/imagery, It is approx 20-30TB of imagery.

For the vector datasets as our schema isn't finalised so we haven't publically published it, it is stored as a ~3GB mbtiles file, its been on my todo list for a while to make the bucket where its stored more public and to just document that the schema may change.

Would you like copy of the latest mbtiles file to see if it would meet your needs?

[antonmarsden]

@blacha The vector dataset is the most useful I think. It's a nice alternative to the raster Topo50. An mbtiles file would do the trick.

[blacha]

Here is a presigned URL that expires in a week with the current topographic mbtiles

topographic.mbtiles

We dont have many (any?) public documents on the structure or how the file was made, you can have a look at basemaps-config's topoographic style json for this file to get a rough idea as how we style it

The general gist is we take data from the LINZ data service and then simplify and rename/remap the attributes.

Here is a STAC collection.json of datasets from the LDS that we use: collection.json