tappa is the Python version of the R terra package.

Note that this is an alpha (very early) version!

Please try it, and report any bugs you find or suggestions you may have.

Requirements

Windows installation

Option A — conda / mamba (recommended)

conda from miniforge or Anaconda provides pre-built GDAL, GEOS, and PROJ binaries with matching CMake packages, which is by far the easiest route on Windows.

1. Install Visual Studio Build Tools (free) with the Desktop development with C++ workload. From PowerShell:

   winget install Microsoft.VisualStudio.2022.BuildTools --override "--add Microsoft.VisualStudio.Workload.VCTools --includeRecommended --passive"

   Or download the installer from https://visualstudio.microsoft.com/visual-cpp-build-tools/ and select Desktop development with C++.

2. Create a conda environment and install the geospatial dependencies:

   conda create -n tappa python=3.12
   conda activate tappa
   conda install -c conda-forge gdal geos proj pybind11 cmake ninja

3. Install Python build and runtime dependencies:

   pip install scikit-build-core numpy pandas matplotlib

4. Activate the MSVC compiler in your Miniforge / conda prompt. This step is needed every time you open a new terminal:

   "C:\Program Files (x86)\Microsoft Visual Studio\2022\BuildTools\VC\Auxiliary\Build\vcvarsall.bat" x64

5. Build and install from the tappa repository root. The SKBUILD_CMAKE_ARGS variable tells scikit-build-core to use Ninja instead of Unix Makefiles on Windows:

   cd C:\path\to\tappa
   set SKBUILD_CMAKE_ARGS=-GNinja
   pip install -e .

   scikit-build-core runs CMake automatically and places the compiled _terra.pyd extension inside src/tappa/.

   Note: if you see a CMake error about mismatched generators, delete the build/ directory first: rmdir /s /q build

6. Verify:

   import tappa as pt
   r = pt.rast()
   print(r)

Tip: if CMake cannot find the conda libraries, pass the prefix explicitly:

pip install -e . -C cmake.define.CMAKE_PREFIX_PATH="%CONDA_PREFIX%\Library"

Option B — vcpkg

vcpkg is Microsoft's C++ package manager and is pre-installed on GitHub Actions Windows runners.

1. Install Visual Studio Build Tools (see step 1 above).

2. Install vcpkg and the geospatial libraries:

   git clone https://github.com/microsoft/vcpkg C:\vcpkg
   C:\vcpkg\bootstrap-vcpkg.bat
   C:\vcpkg\vcpkg install gdal geos proj --triplet x64-windows

3. Install Python build and runtime dependencies:

   pip install scikit-build-core pybind11 numpy pandas matplotlib

4. Build the package, pointing CMake to the vcpkg toolchain:

   cd C:\path\to\tappa
   pip install -e . `
     -C cmake.define.CMAKE_TOOLCHAIN_FILE=C:/vcpkg/scripts/buildsystems/vcpkg.cmake `
     -C cmake.define.VCPKG_TARGET_TRIPLET=x64-windows

DLL search path: when using vcpkg you must add the vcpkg DLL directory to PATH before importing tappa so that gdal.dll, geos_c.dll, and proj.dll are found:

$env:PATH = "C:\vcpkg\installed\x64-windows\bin;" + $env:PATH

This is not needed with conda, where the environment's Library\bin is already on PATH.

Linux installation

# Debian / Ubuntu
sudo apt-get install -y libgdal-dev libgeos-dev libproj-dev cmake
# Fedora / RHEL
sudo dnf install -y gdal-devel geos-devel proj-devel cmake
cd /path/to/tappa
python3 -m venv .venv
source .venv/bin/activate
pip install scikit-build-core pybind11 numpy pandas matplotlib
pip install -e .

macOS installation

brew install gdal geos proj cmake
cd /path/to/tappa
python3 -m venv .venv
source .venv/bin/activate
pip install scikit-build-core pybind11 numpy pandas matplotlib
pip install -e .

Usage

import tappa
import matplotlib.pyplot as plt
# Create / read a raster
f = "https://github.com/rspatial/terra/raw/refs/heads/master/inst/ex/elev.tif"
r = tappa.rast(f)
e = tappa.ext(6, 6.4, 49.5, 50)
x = r.crop(e)
# Inspect
print(r)
# Plot
tappa.plot(r)
plt.show()
# Work with extents
e = tappa.ext(-180, 180, -90, 90)
print(e)
# Arithmetic operators
r2 = r * 2
r3 = r + r2
mask = r > 500
# Vector
ff = "https://github.com/rspatial/terra/raw/refs/heads/master/inst/ex/lux.shp"
v = tappa.vect(ff)
v

License

GPL-3.0-or-later