Newer
Older
The full imagefusion framework, including libimagefusion, utils and python interface, can be installed by `pip install .` using pip from a conda environment or `pip3 install .` using pip from your distribution. scikit-build is responsible for building and pip installs it in your home directory where the other python packages live. For the link dependencies, CMake is set up to prefer the python packages over the distribution packages.
Alternatively, you can install it also without this script by calling `sudo make install`, which will install it to /usr/local by default. This requires the dependencies to be installed from your distribution.
from skbuild import setup
if __name__ == "__main__":
pkg_name = os.getenv('PYTHON_PACKAGE_NAME', 'imagefusion') # name for the egg directory
more_kw_args = {}
if os.getenv('IMAGEFUSION_BUILD_PYTHON_INTERFACE', 'YES') == 'YES':
# this will include the __init__.py file
more_kw_args['packages'] = ['imagefusion'] # package python files like __init__.py
more_kw_args['package_dir'] = {'imagefusion': 'python'} # location of package python files like python/__init__.py
setup(
install_requires=['numpy', 'gdal'],
version='0.0.1',
author='Christof Kaufmann',
author_email='christofkaufmann.dev@gmail.com',
description='Framework to fuse satellite images of different resolutions and dates e. g. with STARFM',
long_description='This framework provides everything required to use imagefusion algorithms like STARFM, ESTARFM, SPSTFM and more. These compute a high-resolution image at the date of an low-resolution image with help of more low- and high-resolution images. This basically improves the resolution for satellite images like MODIS (with help of Landsat images) or Sentinel-3 (with help of Sentinel-2)',
**more_kw_args