|
Hey HN! We're releasing OxiGDAL v0.1.0, a production-grade geospatial data abstraction library written in 100% pure Rust -- no C/C++/Fortran toolchain required.
If you've ever fought with gdal-sys linking errors, bloated Docker images just to read a GeoTIFF, or GDAL's data-race bugs in multithreaded code, this is for you. What it covers:
- 11 format drivers (GeoTIFF/COG, GeoJSON, GeoParquet, Zarr, FlatGeobuf, Shapefile, NetCDF, HDF5, GRIB, JPEG2000, VRT)
- Full CRS transformations (pure Rust PROJ with 1000+ EPSG codes)
- SIMD-accelerated raster/vector algorithms (AVX2, AVX-512, NEON auto-detected)
- Cloud-native I/O (HTTP range reads for COGs, async throughout)
- Cross-platform bindings: Python (numpy/arrow-native), Node.js, WASM, iOS, Android
- ~500K SLoC across 68 workspace crates Just cargo add oxigdal -- no system libs, no pkg-config, no cmake. Compiles to a single static binary. Works in WASM and Lambda out of the box.
GitHub: https://github.com/cool-japan/oxigdal
We'd love feedback from the geo community. Happy to answer any questions! |