Available now
The geospatial core is here and ready to use.
Spatial filtering, metric grid, variogram, IDW, ordinary kriging, cokriging and cross-validation, with native Fortran/OpenMP kernels for high performance. The same core in Python and Rust.
Python
Prebuilt wheels for Linux, macOS and Windows, with nothing to compile.
pip install apbaseRust
Fortran kernels ship precompiled inside the crate, for Linux, macOS and Windows, using the system GEOS and PROJ.
cargo add apbase- Requires
- Rust 1.85+
- Platforms
- Linux, macOS, Windows
- License
- Apache 2.0
- Core
- Fortran + OpenMP
- Code
- GitHub ↗
- Package
- crates.io ↗
Roadmap
What’s coming in upcoming releases.
First, apbase gets more complete: more methods, more ways to view the map, and visual diagnostics. Then it scales, with batch processing and a dashboard.
Core
One-call pipelineAvailable
Filters outliers, builds the grid and chooses between IDW and ordinary kriging.
result = apbase.create_map(x, y, z)KNN and indicator krigingNext release
Two new interpolators. Indicator kriging estimates the probability of exceeding a threshold, such as a nutrient’s critical level.
result.method # can now be "knn" # or "indicatriz"Export and visualizePlanned
From the result straight to an interactive map: an HTML file to share or a leafmap in your notebook.
result.to_html("field_07.html") result.to_leafmap()Diagnostic plotsPlanned
Experimental semivariogram with the fitted model, plus cross-validation plots, to audit the map.
result.plot_variogram() result.plot_cross_validation()
Scale
Batch processingPlanned
Many fields or seasons at once, across the machine’s cores or on workers.
apbase.create_maps(fields)DashboardPlanned
A dashboard to track the maps, layers and diagnostics you’ve generated, without opening a notebook.
Planned items may change. Available today: apbase 0.2.0 for Python and 0.2.0 for Rust.