Puy de Dôme, France

Venosc, Isère

Generate 1:1 Minecraft worlds (Java Edition 1.21.10+) from DEM, landuse and OSM data.

1. Install the Rust toolchain and collect GeoTIFFs that share the same projected coordinate system (e.g., LAMB93). DEM data from IGN is recommended: Download here. LiDAR point clouds (COPC/LAZ) for buildings are available at IGN LiDAR HD.
2. Run francegen [options] <tif-folder> <output-world> to build terrain, use francegen locate <world> <real-x> <real-z> to translate coordinates, and francegen info <world> for printing general information about a generated world.
3. Use a JSON config file provided in examples/ or create your own
4. Dive into User Getting Started and the User Configuration Reference for detailed CLI flags, metadata, and overlay configuration tips.

1. Install the latest stable Rust via rustup.
2. Use cargo fmt --all, and cargo build before sending changes.
3. Get DEM data from cartes.gouv.fr; LiDAR point clouds are at IGN LiDAR HD
4. Iterate with cargo run -- --config examples/french_alps.json <tif-folder> <output-world> to validate code against sample data.

A good CPU and GPU are recommended, and large RAM allocation is a must. Worlds were generated and tested on an Ryzen 9 3900x, RTX 3070, with 40GB RAM allocated.

We recommend the use of Voxy and C2ME for rendering/chunk loading, and Chunky for distant generation. Distant Horizons should also work.

• User Getting Started
• User Configuration Reference